Approach to the adult with dyspepsia

Authors:
George F Longstreth, MD
Brian E Lacy, MD, PhD
Section Editor:
Nicholas J Talley, MD, PhD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Dec 17, 2021.
INTRODUCTIONDyspepsia is a common symptom with an extensive differential
diagnosis and a heterogeneous pathophysiology. It occurs in up to 20 percent of the
population, although prevalence rates are lower using different iterations of the Rome
criteria [1,2]. Most affected people do not seek medical evaluation for their symptoms
[1,2]. Although dyspepsia does not affect survival, it is responsible for substantial health
care costs and significantly affects quality of life [3-6].
This topic will review the definition, etiology, and general approach to the evaluation and
management of the patient with dyspepsia. The evaluation and recommendations are
largely consistent with the American College of Gastroenterology and American
Gastroenterological Association guidelines for the evaluation of dyspepsia [7,8].
ETIOLOGYApproximately 20 to 25 percent of patients with dyspepsia have an
underlying organic cause (table 1). However, up to 75 to 80 percent of patients have
functional (idiopathic or nonulcer) dyspepsia with no underlying cause on diagnostic
evaluation [9]. (See 'Diagnostic strategies and initial management' below.)
Dyspepsia secondary to organic disease — Although there are several organic
causes for dyspepsia, the main causes are peptic ulcer disease, Helicobacter pylori,
gastroesophageal reflux, medications (nonsteroidal antiinflammatory agents being the
most common offender), and gastric malignancy (table 1). Although gastric cancer as a
cause of dyspepsia is a concern for both health care providers and patients, it is
uncommon in North America [8].
●Peptic ulcer disease – Upper abdominal pain or discomfort is the most
prominent symptom in patients with peptic ulcers. Although discomfort from ulcers
is usually centered in the epigastrium, it may occasionally localize to the right or
left upper quadrants [10]. While classic symptoms of duodenal ulcer occur when
acid is secreted in the absence of a food buffer (ie, two to five hours after meals or
on an empty stomach), peptic ulcers can be associated with food-provoked
symptoms, and thus, the utility of using symptoms related to food ingestion to
predict the presence of an ulcer is unreliable. Peptic ulcers can also be associated
with postprandial belching, epigastric fullness, early satiation, fatty food
intolerance, nausea, and occasional vomiting. (See "Peptic ulcer disease: Clinical
manifestations and diagnosis".)
●Gastroesophageal malignancy – Gastroesophageal malignancy is an
uncommon cause of chronic dyspepsia in the Western hemisphere, but the
incidence is higher in Asian Americans, Hispanic Americans, and Afro-Caribbean
Americans. The incidence of gastroesophageal malignancy increases with age.
When present, abdominal pain tends to be epigastric, vague, and mild early in the
1
 disease but more severe and constant as the disease progresses. In addition,
other symptoms and signs typically evolve with disease progression (eg, anemia,
fatigue, weight loss). (See "Epidemiology and pathobiology of esophageal
cancer", section on 'Epidemiology' and "Epidemiology of gastric
cancer" and "Clinical features, diagnosis, and staging of gastric cancer", section
on 'Clinical features'.)
●Biliary pain – Classic biliary pain is characterized by episodic intense dull pain
located in the right upper quadrant, epigastrium, or (less often) substernal area
that may radiate to the back (particularly the right shoulder blade). The pain is
often associated with diaphoresis, nausea, and vomiting. The pain is constant and
not colicky [11]. It is not exacerbated or reproduced by movement and is not
relieved by squatting, belching, bowel movements, or passage of flatus. The pain
typically lasts at least 30 minutes, plateauing within an hour. The pain then starts
to subside, with an entire attack usually lasting less than six hours.
(See "Approach to the management of gallstones" and "Overview of gallstone
disease in adults".)
●Drug-induced dyspepsia – NSAIDs and COX-2 selective inhibitors can cause
dyspepsia even in the absence of peptic ulcer disease. Other drugs that have
been implicated in drug-induced dyspepsia include calcium channel blockers,
methylxanthines, alendronate, orlistat, potassium
supplements, acarbose, dabigatran, iron, vitamin D, selective serotonin reuptake
inhibitors, sildenafil, sulfonylureas, and certain antibiotics,
including erythromycin [12,13]. (See "Nonselective NSAIDs: Overview of adverse
effects", section on 'Gastrointestinal effects'.)
●Other causes – Celiac disease and chronic pancreatitis may rarely present with
dyspeptic symptoms alone. Other rare causes for dyspepsia include infiltrative
diseases of the stomach (eg, eosinophilic gastroenteritis [14], Crohn disease,
sarcoidosis [15], lymphoma [16], and amyloidosis [17,18]), diabetic radiculopathy
[19], metabolic disturbances (eg, hypercalcemia, heavy metal toxicity), hepatoma,
steatohepatitis, celiac artery compression syndrome, superior mesenteric artery
syndrome, abdominal wall pain [20], and intestinal angina (table 1).
(See "Granulomatous gastritis", section on 'Crohn disease' and "Extrapulmonary
manifestations of sarcoidosis", section on 'Gastrointestinal' and "Celiac artery
compression syndrome" and "Superior mesenteric artery syndrome" and "Chronic
mesenteric ischemia", section on 'Clinical features' and "Epidemiology,
pathogenesis, and clinical manifestations of celiac disease in adults", section on
'Gastrointestinal manifestations'.)
Functional dyspepsia — The diagnosis of functional (idiopathic or nonulcer) dyspepsia
requires exclusion of other organic causes of dyspepsia [21]. It is defined by the
presence of one or more of the following: postprandial fullness, early satiation,
epigastric pain, or burning, and no evidence of structural disease to explain the
symptoms [22]. The pathophysiology, diagnosis, and management of functional
dyspepsia are discussed in detail, separately. (See 'Diagnostic strategies and initial
management' below and "Functional dyspepsia in adults".)

2
INITIAL EVALUATIONA history, physical examination, and laboratory evaluation
are the first steps in the evaluation of a patient with new onset of dyspeptic symptoms.
(See 'History' below and 'Physical examination' below and 'Laboratory tests' below.)
The goal of the initial evaluation is to identify alarm features for gastroesophageal
malignancy (table 2), which will direct the diagnostic approach. (See 'Diagnostic
strategies and initial management' below.)
History — A detailed history is necessary to determine the underlying cause and to
identify patients with alarm features (table 2).

As examples:

●A dominant history of heartburn or regurgitation, is suggestive of

gastroesophageal reflux disease (GERD), recognizing that some patients have
overlapping GERD and functional dyspepsia [23,24]. (See "Clinical manifestations
and diagnosis of gastroesophageal reflux in adults", section on 'Clinical
manifestations'.)
●Aspirin and other NSAID use raises the possibility of NSAID dyspepsia and
peptic ulcer disease. Radiation of pain to the back or personal or family history of
pancreatitis may be indicative of underlying chronic pancreatitis.
(See "Nonselective NSAIDs: Overview of adverse effects", section on
'Gastrointestinal effects'.)
●Significant weight loss, anorexia, anemia, vomiting, dysphagia, odynophagia, and
a family history of gastrointestinal cancers suggest the presence of an underlying
gastroesophageal malignancy. (See "Clinical features, diagnosis, and staging of
gastric cancer", section on 'Clinical features'.)
●The presence of severe episodic epigastric or right upper quadrant abdominal
pain, usually in association with nausea or vomiting, lasting at least 30 minutes is
suggestive of symptomatic cholelithiasis [25]. (See "Acute calculous cholecystitis:
Clinical features and diagnosis", section on 'Clinical manifestations'.)
●Nausea and vomiting, with or without weight loss, occurring with recurrent or
persistent upper abdominal pain raises the possibility of gastroparesis, especially
in patients with risk factors. However, both the pathophysiology and symptom
expression of functional dyspepsia and gastroparesis are quite similar [26]. An
analysis of patients in a multicenter gastroparesis registry recognized the
significant overlap in functional dyspepsia and gastroparesis symptoms [27].
Physical examination — The physical examination in patients with dyspepsia is
usually normal, except for epigastric tenderness. The presence of epigastric tenderness
cannot accurately distinguish organic dyspepsia from functional dyspepsia. Abdominal
tenderness on palpation should be evaluated for the presence of Carnett's sign to
determine if it is due to pain arising from the abdominal wall rather than due to
inflammation of the underlying viscera. The presence of increased local tenderness
during muscle tensing (positive Carnett's sign) suggests the presence of abdominal wall
pain. However, if the pain is decreased (negative Carnett's sign), the origin of pain is not
from the abdominal wall and likely from an intra-abdominal organ, as the tensed
abdominal wall muscles protect the viscera. (See "Anterior cutaneous nerve entrapment
syndrome", section on 'Diagnostic approach'.)

3
Other informative findings on physical examination include a palpable abdominal mass
(eg, hepatoma) or lymphadenopathy (eg, left supraclavicular or periumbilical in gastric
cancer), jaundice (eg, secondary to liver metastasis), a bruit (celiac artery compression
syndrome), a succussion splash (gastric outlet obstruction), or pallor secondary to
anemia. Ascites may indicate the presence of peritoneal carcinomatosis. Patients with
an underlying malignancy may have evidence of muscle wasting, loss of subcutaneous
fat, and peripheral edema due to weight loss.

Laboratory tests — Routine blood counts and blood chemistry including liver function
tests, serum lipase, and amylase, should be performed to identify patients with alarm
features (eg, iron deficiency anemia) and underlying metabolic diseases that can cause
dyspepsia (eg, diabetes, hypercalcemia) (table 2). (See "Clinical manifestations of
hypercalcemia", section on 'Gastrointestinal abnormalities' and "Diabetic autonomic
neuropathy of the gastrointestinal tract".)
DIAGNOSTIC STRATEGIES AND INITIAL MANAGEMENT The approach
to, and extent of, diagnostic evaluation of a patient with dyspepsia is based on the
clinical presentation, the patient's age, and the presence of alarm features (table 2). An
approach to the evaluation of a patient with dyspepsia is outlined in the algorithm
(algorithm 1) [28]. Our approach is largely consistent with the American College of
Gastroenterology and Canadian Association of Gastroenterology guidelines [8].
The optimal age cut-off for endoscopic evaluation in patients with dyspepsia is
controversial and is supported by limited evidence that suggests that the risk of
malignancy in most United States populations below the age of 60 years is low.
Guidelines also suggest that the age cutoff may vary between countries, depending
upon the prevalence of gastric cancer. The American Gastroenterological Association
guidelines suggest that it may be reasonable in some resource-rich countries to
consider the age of 60 or 65 years as the threshold age at which endoscopy should be
offered to all new dyspeptic patients, while an age cutoff of 45 or 50 years may be more
appropriate for Asian Americans, Hispanic Americans, and Afro-Caribbean Americans
due to an increased risk for gastric cancer, or in populations with a high incidence of
gastric cancer in young individuals [11]. A European consensus statement recommends
endoscopy in adults older than 45 years old who present with persistent dyspepsia [29].
These recommendations highlight the fact that diagnostic evaluation of the patient with
dyspepsia need to be individualized based on symptoms, age, ethnic background,
family history, nationality, and regional incidence of gastric cancer. (See "Clinical
features, diagnosis, and staging of gastric cancer".)
Patient age ≥60 years
Upper endoscopy — We perform an upper endoscopy to evaluate dyspepsia in
patients age ≥60 years [8]. Biopsies of the stomach should be obtained to rule out H.
pylori. Patients with H. pylori should receive eradication therapy in addition to treatment
based on the underlying diagnosis (eg, peptic ulcer disease). After H. pylori treatment,
eradication should be assessed. (See "Medical management of gastroesophageal reflux
disease in adults" and "Peptic ulcer disease: Treatment and secondary prevention",
section on 'Initial management' and "Treatment regimens for Helicobacter pylori in
adults".)

4
Multiple studies have evaluated the yield of upper endoscopy in patients with dyspepsia
[30-33]. A meta-analysis of nine studies with 5389 patients found that the most
prevalent findings in patients with dyspepsia were erosive esophagitis and peptic ulcer
disease (pooled prevalence 6 and 8 percent, respectively) [34]. The diagnostic yield of
upper endoscopy increases with age [30,32]. In the absence of warning signs, upper
endoscopy in younger patients is unlikely to find a worrisome cause.
Additional evaluation and management — Most patients with a normal upper
endoscopy and routine laboratory tests have functional dyspepsia. However, additional
evaluation may be required based on symptoms. (See 'Laboratory tests' above
and 'Evaluation of persistent symptoms' below and "Functional dyspepsia in adults".)
Patient age <60 years — Patients <60 years of age should be tested and treated for H.
pylori, and upper endoscopy should be performed selectively. Patients who are H.
pylori negative or who continue to have symptoms after successful eradication of H.
pylori should be treated with antisecretory therapy with a proton pump inhibitor (PPI)
(algorithm 1) [11,35]. In patients whose symptoms do not improve after eight weeks of
PPI therapy, we initiate a therapeutic trial with a tricyclic antidepressant. Given the side
effects associated with prokinetics and the limited evidence of efficacy, we reserve the
use of prokinetics for patients who fail tricyclic antidepressants. (See 'Evaluation of
persistent symptoms' below.)
Upper endoscopy in selected patients — Endoscopic evaluation of patients <60
years is reserved for patients with any one of the following:
●Clinically significant weight loss (>5 percent usual body weight over 6 to 12
months).
●Overt gastrointestinal bleeding.
●>1 other alarm feature (table 2).
●Rapidly progressive alarm features.

Alarm features include:

●Unintentional weight loss

●Dysphagia  
●Odynophagia
●Unexplained iron deficiency anemia
●Persistent vomiting
●Palpable mass or lymphadenopathy
●Family history of upper gastrointestinal cancer
In such patients, upper endoscopy should be performed early, preferably within two to
four weeks. Biopsies of the stomach should be obtained to rule out H. pylori and
patients with evidence of infection should be treated with eradication therapy. (See 'Test
and treat for Helicobacter pylori' below and "Treatment regimens for Helicobacter pylori
in adults", section on 'Initial antibiotic therapy'.)
While prior guidelines have recommended upper endoscopy for all patients with alarm
features regardless of age, the recommendation to selectively perform upper endoscopy
in patients <60 years is based on evidence that individual alarm features have low
positive predictive value for an upper gastrointestinal tract malignancy [8,36].

5
Test and treat for Helicobacter pylori — The rationale for H. pylori testing in patients
with dyspepsia is based upon the recognition of H. pylori as an etiologic factor in peptic
ulcer disease.
Mucosal biopsies for H. pylori should be obtained use the Sydney protocol, which
includes specimens from the lesser and greater curve of the antrum within 2 to 3 cm of
the pylorus, from the lesser curvature of the corpus (4 cm proximal to the angularis),
from the middle portion of the greater curvature of the corpus (8 cm from the cardia),
and one from the incisura angularis [37].
In patients who do not require an upper endoscopy or in whom biopsies were not
performed at the time of upper endoscopy, testing for H. pylori should be performed with
a test for active infection (eg, urea breath test or stool antigen assay). Serologic testing
should not be used due to their low positive predictive value [38].
Patients who test positive for an infection with H. pylori should undergo treatment with
eradication therapy. Most dyspeptic patients who are H. pylori positive and who are
treated with appropriate antibiotic therapy persist with dyspeptic symptoms; the number
needed to successfully relieve dyspeptic symptoms is estimated at one in seven. In a
meta-analysis of two randomized trials in which 563 H. pylori- infected dyspepsia
patients were assigned to eradication therapy or placebo, H. pylori eradication therapy
resulted in a significant reduction in dyspepsia (RR remaining dyspeptic 0.81; 95% CI
0.70–0.94) with a number needed to treat of seven [8,39]. (See "Treatment regimens for
Helicobacter pylori in adults", section on 'Initial antibiotic therapy'.)
Antisecretory therapy — Antisecretory therapy with a PPI can relieve dyspepsia
symptoms [11]. This was illustrated in a meta-analysis of six randomized controlled
trials that included 2709 patients with dyspepsia who were assigned to PPI therapy or
control (placebo or antacid therapy) [8]. Dyspepsia symptoms were present in a
significantly lower proportion of the PPI group as compared with controls (50 versus 73
percent). PPIs are also likely to be more effective at relieving symptoms of dyspepsia as
compared with H2 receptor antagonists (H2RA) [40,41]. Although a meta-analysis of
seven randomized trials that included 2456 patients with dyspepsia did not demonstrate
a statistically significant difference between PPIs and H2RAs in providing symptom
relief, the meta-analysis was limited by significant heterogeneity between studies [8].
Studies have shown that a twice-daily PPI is not more effective than a once-daily PPI at
relieving dyspeptic symptoms [42]. (See "Antiulcer medications: Mechanism of action,
pharmacology, and side effects", section on 'Antisecretory agents'.)
Other therapies
Tricyclic antidepressants and azapirones — Patients who test negative for H.
pylori or remain symptomatic after its eradication and have inadequate response to a
PPI can be considered for therapy with a tricyclic agent (algorithm 1). For patients with
persistent symptoms, the tetracyclic antidepressant mirtazapine may prove beneficial
[43]. Alternatively, low-dose buspirone (an azapirone) before meals may improve post-
prandial symptoms [44].
Prokinetics — Treatment with prokinetic agents (most evaluated agents are
unavailable in North America) is discussed in detail separately [45]. (See "Functional
dyspepsia in adults", section on 'Prokinetic agents' and "Functional dyspepsia in adults",
section on 'Antidepressants'.)

6
Evaluation of persistent symptoms — Despite the approaches described above,
some patients continue to have symptoms of dyspepsia. Patients with continued
symptoms of dyspepsia fall into the following categories: patients with persistent H.
pylori infection, patients with an alternate diagnosis, and patients with functional
dyspepsia. (See "Functional dyspepsia in adults".)

Patients with continued symptoms of dyspepsia should be carefully reassessed, paying

specific attention to the type of ongoing symptoms, the degree to which symptoms have
improved or worsened, and compliance with medications.

An upper endoscopy should be performed in patients with persistent dyspepsia

(algorithm 1). During upper endoscopy, biopsies for H. pylori should be performed in
patients who have not previously been tested, while culture and sensitivity testing
should be performed in patients previously treated for H. pylori. Biopsies of the
duodenum should also be performed to rule out celiac disease or inflammatory
conditions.
In patients with a normal upper endoscopy, further evaluation should be performed
selectively based on the type of ongoing symptoms (eg, abdominal imaging with an
ultrasound or computed tomography scan in patients with concurrent jaundice or pain
suggestive of a biliary/pancreatic source, a four-hour solid-phase scintigraphic gastric
emptying scan for those with persistent nausea and vomiting). Approximately 75
percent of patients with dyspepsia have functional (idiopathic or nonulcer) dyspepsia
with no underlying cause on diagnostic evaluation. (See "Functional dyspepsia in
adults".)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Dyspepsia".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topic (see "Patient education: Stomach ache and stomach upset (The
Basics)")
●Beyond the Basics topic (see "Patient education: Upset stomach (functional
dyspepsia) in adults (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
7
●Dyspepsia is defined as one or more of the following symptoms: postprandial
fullness, early satiation, epigastric pain, or burning. Approximately 25 percent of
patients with dyspepsia are found to have an underlying organic disease on
diagnostic evaluation (table 1). However, approximately 75 percent of patients
have functional (idiopathic or nonulcer) dyspepsia.
●A detailed history, physical examination, and laboratory studies are necessary to
determine the underlying etiology and identify alarm features that may warrant
additional evaluation (table 2).
●The approach to and extent of diagnostic evaluation of a patient with dyspepsia is
based on the presence or absence of alarm features, patient age, and the
prevalence of H. pylori infection (table 2 and algorithm 1).
•Patients ≥60 years of age with dyspepsia should undergo an upper
endoscopy. Most patients with a normal upper endoscopy and routine
laboratory tests have functional dyspepsia. However, additional evaluation may
be required based on symptoms (eg, abdominal imaging with an ultrasound or
computed tomography scan in patients with concurrent jaundice or pain
suggestive of a biliary/pancreatic source).
•In patients <60 years indications for upper endoscopy include:
-Clinically significant weight loss (>5 percent usual body weight over 6 to
12 months).
-Overt gastrointestinal bleeding.
->1 alarm feature (table 2).
-Rapidly progressive alarm features.
•Patients <60 years of age should be tested and treated for H. pylori. Patients
who test positive for an infection with H. pylori should undergo treatment with
eradication therapy.
●In patients who test negative for H. pylori and for those with persistent symptoms
after H. pylori eradication, we suggest treatment with antisecretory therapy with a
proton pump inhibitor for four to eight weeks (Grade 2A).
●Patients with continued symptoms of dyspepsia despite a trial of a tricyclic
antidepressant and prokinetic, should undergo endoscopic evaluation with an
upper endoscopy and biopsies, if not previously performed (algorithm 1). Further
evaluation for an alternate diagnosis should be performed selectively based on the
patient’s symptoms. Patients with continued symptoms of dyspepsia for three
months with symptom onset at least six months before diagnosis and no evidence
of structural disease to explain the symptoms should be diagnosed and treated as
functional dyspepsia. (See 'Evaluation of persistent symptoms' above
and "Functional dyspepsia in adults".)

8
Functional dyspepsia in adults
Authors:
George F Longstreth, MD
Brian E Lacy, MD, PhD
Section Editor:
Nicholas J Talley, MD, PhD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jan 12, 2022.
INTRODUCTIONDyspepsia is a common symptom with an extensive differential
diagnosis and a heterogeneous pathophysiology [1]. The prevalence of uninvestigated
dyspepsia worldwide is up to 20 percent, especially in females, smokers, and people
taking nonsteroidal antiinflammatory agents. The prevalence varies according to the
definition used for dyspepsia [2,3]. Dyspepsia can significantly impair quality of life [4].
The proportion of affected people who seek medical care ranges from 14 to 66 percent
in various countries and ethnic groups [5]. The majority of patients (75 to 80 percent)
with symptoms of dyspepsia are eventually categorized as having functional (idiopathic,
nonulcer) dyspepsia. Many authorities regard it as a "disorder of gut-brain interaction"
along with irritable bowel syndrome and other symptom-based gastrointestinal disorders
[6], but the term "functional" is still commonly used. This topic review will provide an
overview of the pathophysiology and treatment of functional dyspepsia.
Our recommendations for the diagnosis and management of functional dyspepsia are
largely consistent with the American College of Gastroenterology, Canadian
Gastroenterological Association, and American Gastroenterological Association
guidelines [7,8], and a European expert consensus [9]. The etiology, general approach
to the evaluation, and management of the patient with uninvestigated dyspepsia are
presented separately. (See "Approach to the adult with dyspepsia".)
EPIDEMIOLOGY AND PATHOPHYSIOLOGYThe prevalence of functional
dyspepsia ranges from 5 to 11 percent worldwide [2,3]. The pathophysiology of
functional dyspepsia is not well understood. However, several potential mechanisms
have been suggested. These mechanisms may differ between subtypes of functional
dyspepsia (postprandial distress syndrome and epigastric pain syndrome) [10].
(See 'Diagnostic criteria' below.)
●Gastric emptying, accommodation, and vagal function – Functional
dyspepsia has been associated with several motility disorders. These include mild
delays in gastric emptying, rapid gastric emptying, antral hypomotility, gastric
dysrhythmias, impaired gastric accommodation in response to a meal, and
abdominal vagal dysfunction [10-13]. However, these findings vary among
patients. (See "Gastroparesis: Etiology, clinical manifestations, and diagnosis".)
●Visceral hypersensitivity – Visceral hypersensitivity is characterized by a
lowered threshold for induction of pain in the presence of normal gastric
compliance. Visceral hypersensitivity independent of delayed gastric emptying has
been demonstrated in patients with functional dyspepsia [14]. Both
9
mechanoreceptor dysfunction and aberrant processing of afferent input in the
spinal cord or brain may play a role in the pathophysiology of visceral
hypersensitivity [15]. In one study that included 270 patients with functional
dyspepsia, 37 percent had hypersensitivity to gastric distension, and
hypersensitive patients reported higher cumulative symptom scores as compared
with normosensitive patients [10]. Intraduodenal infusion of dextrose and lipids
was also associated with symptoms in patients with functional dyspepsia, and an
association of hypersensitivity with higher plasma concentrations of enteral
hormones suggested hormonal mediation of the sensitivity [16].
●Helicobacter pylori infection – The role of H. pylori infection in the
pathogenesis of functional dyspepsia remains unclear (see "Pathophysiology of
and immune response to Helicobacter pylori infection"). Furthermore, only a small
minority of patients with functional dyspepsia report symptom improvement after
eradication therapy for H. pylori [17].
●Altered gut microbiome – Alterations in the upper gastrointestinal tract
microbiome may result in the development of dyspepsia. This hypothesis is
supported by the observation that dyspeptic symptoms are more likely to occur
after an episode of gastroenteritis [18-21]. In one study, risk factors for the
persistence of dyspepsia eight years after exposure to a waterborne outbreak of
bacterial dysentery were female sex, smoking, premorbid irritable bowel syndrome
(IBS), anxiety, depression, and >7 days of diarrhea or abdominal cramps during
the acute illness [19]. It has been hypothesized that the efficacy of H.
pylori therapy in improving symptoms of functional dyspepsia in some patients is
due to the impact on the gut microbiome rather than the eradication of H.
pylori alone [22,23].
●Duodenal inflammation and immune activation – Increased eosinophils and
mast cells and altered lymphocyte populations, including "gut-homing"
lymphocytes, have been reported in the duodenum of patients with functional
dyspepsia [24-28]. Structural and functional differences in duodenal submucosal
ganglia between patients with functional dyspepsia and controls have also been
described [29]. Confocal endomicroscopy has revealed increased epithelial gaps
in patients with functional dyspepsia plus reduced transepithelial electrical
resistance that was inversely correlated with symptom severity [30]. The proton
pump inhibitor pantoprazole reduced duodenal eosinophils, mast cells, mucosal
permeability, and symptoms in patients with functional dyspepsia [31]. These
findings support the suggestion that luminal factors, such as acid and bile acids,
cause low-grade inflammation that impairs mucosal integrity, resulting in abnormal
gastrointestinal neuroregulation and symptoms [32].
●The hypothalamic-pituitary-adrenal (HPA) axis and stress – Patients with
functional gastrointestinal disorders who experience mental stress can have
increased activation of the amygdala and dysregulation of the HPA axis [33]. In
healthy subjects, acute stress increases salivary cortisol levels and intestinal
permeability [34]. In patients with functional dyspepsia, especially those with the
epigastric pain subtype, magnetic resonance imaging of the brain revealed
functional abnormalities in areas that process afferent signals [35].

10
 ●Psychosocial dysfunction – Functional dyspepsia has been associated with
generalized anxiety disorder, somatization, and major depression [36-38]. There is
also a higher prevalence of functional gastrointestinal disorders in patients with a
self-reported history of childhood abuse [39-41]. In patients with the prototypical
functional gastrointestinal disorder, IBS, a history of adverse early life events was
related to HPA hyper-responsiveness to a visceral stressor [42]. A long-term
follow-up study revealed initial trait anxiety was independently correlated with
symptoms and quality of life; depression, anxiety, and comorbid functional somatic
symptoms (IBS and chronic fatigue) were cross-sectionally associated with
dyspepsia symptoms at follow-up [43]. The importance of psychosocial factors is
further evident from an analysis of a large number of patients with functional
dyspepsia that assessed gastrointestinal symptoms, psychological comorbidity,
and extra-intestinal symptoms. The subjects clustered into four groups, two of
which were characterized by high psychological burden [44].
CLINICAL MANIFESTATIONSPatients with functional dyspepsia usually
describe postprandial fullness, early satiety, bloating and/or epigastric pain/burning.
Postprandial fullness is the most intense symptom in patients with meal-induced
symptoms [45]. Symptoms may be severe enough to limit usual activities. Some
patients may have nausea, vomiting, or heartburn, however, these symptoms are
usually infrequent.
DIAGNOSIS
Overview of diagnostic approach — Functional dyspepsia is suspected in patients
with a clinical history of postprandial fullness, early satiety, or epigastric pain/burning. A
clinical diagnosis of functional dyspepsia requires the fulfillment of symptom-based
diagnostic criteria and an evaluation to exclude other causes of dyspepsia. This
evaluation consists of a history (eg, dietary, medical, surgical, family, and
medications/supplements), physical examination, laboratory studies, and endoscopic
evaluation to exclude organic/structural disease to explain the symptoms (algorithm 1).
An approach to the evaluation of a patient with dyspepsia is discussed in detail
separately. (See "Approach to the adult with dyspepsia", section on 'Initial
evaluation' and "Approach to the adult with dyspepsia", section on 'Diagnostic strategies
and initial management'.)
Diagnostic criteria — Symptom-based criteria have been proposed to standardize the
diagnosis of functional dyspepsia.
●Rome IV criteria for functional dyspepsia – According to the Rome IV criteria,
functional dyspepsia is defined as the presence of one or more of the following
symptoms: postprandial fullness, early satiation, epigastric pain or epigastric
burning, and no evidence of structural disease (including at upper endoscopy) to
explain the symptoms (table 1) [46].
While patients with these symptoms and a negative diagnostic evaluation likely
have functional dyspepsia, according to the Rome IV guidelines, the criteria
should be fulfilled for the last three months with symptom onset at least six months
before diagnosis. Criteria for symptom frequency and duration are particularly
useful in defining patient eligibility for research, but clinician judgement may allow
diagnosis in practice without rigid adherence to them.

11
 ●Functional dyspepsia subtypes – Two subtypes of functional dyspepsia are
recognized based on the predominant symptoms. However, overlap between
these subtypes is common [10,47].
•Postprandial distress syndrome is characterized by bothersome postprandial
fullness and/or early satiation (table 1).
•Epigastric pain syndrome is characterized by bothersome epigastric pain or
burning that is not exclusively postprandial (table 1).
DIFFERENTIAL DIAGNOSISThe differential diagnosis of functional dyspepsia
includes other organic causes of dyspepsia (table 2). Although there are several organic
causes for dyspepsia, the main causes are peptic ulcer disease, gastritis,
gastroesophageal reflux, and medications (eg, nonsteroidal anti-inflammatory drugs
[NSAIDs]). An underlying gastric malignancy is a rare cause of dyspepsia in North
America. Functional dyspepsia is differentiated from these by clinical assessment,
laboratory testing, and upper endoscopy. (See "Approach to the adult with dyspepsia",
section on 'Dyspepsia secondary to organic disease'.)
●Gastroesophageal reflux disease – The most common symptoms of
gastroesophageal reflux disease are retrosternal burning pain and regurgitation.
Dyspepsia symptoms can coexist with heartburn, but in patients with functional
dyspepsia, epigastric pain, and fullness are the predominant symptoms [48].
●Gastroparesis – Gastroparesis is less prevalent than functional dyspepsia, but
overlaps with it, as gastric emptying may be slow and symptoms of dyspepsia
occur in both disorders [49,50]. Patients with functional dyspepsia may have
nausea. However, in patients with gastroparesis, vomiting, rather than abdominal
pain or epigastric fullness, is generally the predominant symptom.
●Irritable bowel syndrome – More than 60 percent of patients with functional
dyspepsia may have overlapping IBS symptoms [8,46,51], and the overlap may be
more likely when the symptoms of either disorder are severe [52]. Rather than
epigastric pain associated with functional dyspepsia, IBS is characterized by
abdominal pain or discomfort associated with a change in stool form or frequency.
As compared with patients with IBS alone, patients with both IBS and functional
dyspepsia had increased bloating and abdominal pain after a lactulose-nutrient
challenge test [53].  
MANAGEMENTThe management of patients with functional dyspepsia is
controversial and alleviates symptoms in only a small proportion of patients (algorithm
1) [46,54].
Initial approach — Patients with functional dyspepsia should be tested and treated
for Helicobacter pylori. We treat patients with functional dyspepsia who test negative
for H. pylori and those with persistent symptoms four weeks after eradication of H.
pylori with a proton pump inhibitor (PPI). A Cochrane systematic review revealed 14
patients needed to be treated to cure one case of dyspepsia [17]. (See 'Helicobacter
pylori test and treat' below and 'Proton pump inhibitors' below and 'Subsequent
approach' below.)
Helicobacter pylori test and treat — The diagnosis of H. pylori should be made with a
test for active infection (stool antigen assay or urea breath test) if testing was not
performed at the time of upper endoscopy performed for evaluation of dyspepsia
(algorithm 1). Serologic testing should not be performed due to the low positive
12
predictive value. (See "Indications and diagnostic tests for Helicobacter pylori infection
in adults", section on 'Approach to diagnostic testing' and "Approach to the adult with
dyspepsia", section on 'Diagnostic strategies and initial management'.)
H. pylori eradication may improve dyspeptic symptoms by altering acid secretion or
modification of intestinal microbiota [55,56]. It also has the benefit of preventing
unrecognized peptic ulcers associated with H. pylori. (See "Treatment regimens for
Helicobacter pylori in adults" and 'Epidemiology and pathophysiology' above.)
Proton pump inhibitors — PPIs appear to be moderately effective in the treatment of
some patients with functional dyspepsia. In a systematic review of 18 randomized
controlled trials, PPIs were more effective than placebo in relieving overall dyspepsia
symptoms (RR 0.88, 95% CI 0.82-0.94) with a NNT of 11 [57]. Low- and standard-dose
PPIs had similar effectiveness. Twice daily PPI use has not been shown to be better
than once daily use. PPIs may exert their beneficial effect by reducing duodenal
eosinophils, mast cells, and mucosal permeability [31]. In patients with functional
dyspepsia who respond to PPI therapy, attempts should be made to discontinue PPIs
every 6 to 12 months to minimize long-term risk of therapy.
H2-receptor antagonists — In a meta-analysis of 12 trials with a total of 2183 patients,
H2RAs were associated with a 23 percent reduction in symptoms compared with
placebo (relative risk reduction [RRR] 23 percent, 95% CI 8-35 percent) with an NNT of
7 [58]. In the systematic review of PPI therapy for functional dyspepsia, only two studies
compared the efficacy of PPIs with H2AAs, and no difference was evident [57].
However, the quality of most trials included was poor, and there was significant
heterogeneity among studies. Another limitation of these studies is that patients with
gastroesophageal reflux disease may have been misclassified as having functional
dyspepsia.
Subsequent approach
Antidepressants — In patients with functional dyspepsia whose symptoms do not
improve after eight weeks of PPI therapy, we initiate a therapeutic trial with a tricyclic
antidepressant (TCA). For patients with a partial clinical response to a PPI, a TCA can
be initiated as combination therapy. For patients who fail to improve on a PPI, the PPI
should be stopped and a TCA initiated.
We begin with a TCA at a low dose (eg, amitriptyline 10 mg or desipramine 25 mg at
night). The dose may be increased at two-week intervals. A dose of 20 to 30 mg is
adequate in many patients, and we do not exceed a dose of 75 mg in most patients.
Higher doses may not be more effective than lower doses and may be associated with
daytime sedation and other anticholinergic side effects. We usually continue the TCA for
8 to 12 weeks before stopping, if it is ineffective. If the patient responds, we usually
continue the drug for appropriately six months and then consider slowly tapering the
medication off. The TCA can be resumed if dyspepsia recurs. While some clinicians
prefer to use trazodone rather than amitriptyline or desipramine, there are few clinical
data to support its use. Low dose TCAs may also improve associated symptoms of
insomnia and fibromyalgia in patients with functional dyspepsia [59].
The TCA amitriptyline is beneficial for some patients. In a large trial, adequate relief was
obtained by 53 percent of patients taking amitriptyline 50 mg once daily versus 40
percent of patients taking placebo [60]. A network meta-analysis revealed TCAs ranked
second for efficacy (relative risk [RR] of remaining symptomatic 0.71, 95% CI 0.58-0.87)

13
and first when only low risk of bias trials were included. Most of the trials that assessed
TCAs studied patients who were refractory to other drugs analyzed in the network [61].
The drug did not delay gastric emptying and improved accommodation and postprandial
bloating, but the exact mechanism of action remains unknown [62].
Mirtazapine has also demonstrated benefit in patients with functional dyspepsia and
unintentional weight loss that may be due to a central mechanism of action [63-65]. In
34 patients with >10 percent weight loss randomized to mirtazapine 15 mg per day or
placebo for eight weeks, mirtazapine reduced global symptoms of dyspepsia, early
satiation, and gastrointestinal-specific anxiety. In addition, patients treated with
mirtazapine reported improved nutrient tolerance and quality of life and weight gain [63].
In another study in which 60 patients with functional dyspepsia and >5 percent weight
loss randomized to mirtazapine 30 mg per day, paroxetine 20 mg per day or
conventional (acid inhibitor or prokinetic) therapy for eight weeks, mirtazapine reduced
dyspepsia and depression and resulted in weight gain compared with the other two
groups [64]. Mirtazapine-associated drowsiness was common in both studies.
Prokinetic agents — We reserve the use of prokinetics (eg, metoclopramide 5 to 10
mg three times daily one-half an hour before meals and at night for four weeks), to
patients in whom other therapies have failed and limit their duration to four weeks
before discontinuing treatment [8]. If symptoms recur, we repeat a course of therapy,
recognizing that up to 30 percent of patients may have side effects, most of which are
generally mild and resolve with cessation of therapy [8].
In a systematic review and meta-analysis that included 29 trials of six individual agents
(cisapride, acotiamide, itopride, tegaserod, mosapride, and ABT-229) in patients with
functional dyspepsia, overall, global symptom improvement was greater with individual
agents than placebo (40 versus 26 percent, respectively) [66]. However, the quality of
evidence, according to the GRADE system, was very low. Comparisons
of domperidone versus newer prokinetics revealed no superiority, also with a very low
quality of evidence. No improvement in quality of life was documented. All agents were
well tolerated over a short term except cisapride. Few studies related efficacy to gastric
emptying data, providing no support for assessing gastric emptying to direct therapy.
Notably, no eligible studies assessed metoclopramide or domperidone, the only agents
available in North America, in functional dyspepsia. Methodologic difficulties, including
the frequent overlap of functional dyspepsia with gastroesophageal reflux disease, lack
of validated endpoints, and presence of delayed gastric emptying in only a minority of
patients [13] complicate designing studies of prokinetics [67].
If postprandial nausea is a predominant symptom, trials of other antiemetic agents can
be employed (eg, promethazine, prochlorperazine, meclizine), although data from
clinical trials are lacking.
THERAPIES WITH LIMITED OR UNCLEAR ROLE
●Psychotherapy – We reserve psychotherapy for motivated patients who
associate symptoms with stressors and in patients who fail initial empiric medical
therapy. The psychological factors associated with functional dyspepsia suggest
that psychological therapy could help some patients. A systematic review of four
trials (relaxation therapy and hypnosis, psychodrama, psychotherapy, and
cognitive-behavioral therapy) suggested therapy could be beneficial for one year,
but methodological deficiencies led to uncertainty about the results [68].
14
 (See "Overview of psychotherapies", section on 'Cognitive and behavioral
therapies'.)
●Fundic relaxant drugs – There is limited evidence that relaxing the gastric
fundus may improve early satiation and postprandial fullness. In a small
randomized trial, buspirone (10 mg, three times daily for four weeks) as compared
with placebo increased gastric accommodation and reduced the overall severity of
symptoms of dyspepsia, despite slowing gastric emptying of liquids [69]. In
another study that included 32 patients with functional dyspepsia refractory to
PPIs and/or domperidone, buspirone improved gastric emptying measured by 13C-
octanoic breath test and reduced early satiation [70].
●Antinociceptive agents – It is hypothesized that antinociceptive agents
(eg, carbamazepine, tramadol, or pregabalin) may impact the central processing
of pain, thereby decreasing visceral hypersensitivity that has been associated with
functional dyspepsia. A post hoc analysis of data obtained from six randomized
controlled trials in patients with generalized anxiety disorder and prominent
gastrointestinal symptoms showed that pregabalin was significantly more effective
than placebo in treating both anxiety and gastrointestinal symptoms [71]. In a
placebo-controlled trial of pregabalin in patients with irritable bowel syndrome
reduced pain, but adequate relief did not differ from the placebo group [72].
However, results of trials in functional dyspepsia are needed. We do not routinely
use of tramadol for the treatment of functional gastrointestinal disorders due to the
potential for addiction.
●Complementary and alternative medicine – Several complementary and
alternative medicine approaches to functional dyspepsia have been described.
However, further studies are needed before they can be recommended [73,74]. A
systematic review of several low quality studies involving herbal and natural
products, acupuncture, and homeopathy suggested a benefit from peppermint oil
and STW5, a European multiherbal preparation that includes peppermint and
caraway [75]. STW5 may improve symptoms of functional dyspepsia by
stimulating gastric fundic relaxation and antral motility [76]. An eight-week
placebo-controlled trial of STW5 found statistically significant but marginal
symptomatic improvement [77]. A study in 95 patients with FD (Rome III criteria)
found that a combination of L-menthol and caraway oil improved postprandial FD
symptoms in some patients [78].
●Dietary modification – The postprandial timing of symptoms suggests
improvement could be obtained through dietary modification. However, a
population case-control study failed to find an association between various foods
and functional gastrointestinal disorders [79]. The multifactorial nature of
functional dyspepsia and difficulty of conducting controlled dietary studies has
markedly limited information useful to practice. Possibly, dieticians could help
some patients through individualized approaches [80].
ADDITIONAL EVALUATION OF PERSISTENT SYMPTOMS Patients with
functional dyspepsia who have persistent symptoms of dyspepsia may require
additional testing for an alternate diagnosis. We perform a gastric emptying study to
evaluate for gastroparesis in selected patients with refractory functional dyspepsia who
have persistent nausea and vomiting or risk factors for delayed gastric emptying (eg,
15
diabetes mellitus). However, it is important that a significant overlap exists between
dyspepsia and gastroparesis [49,50]; and treatment directed at accelerating delayed
gastric emptying in these patients may not necessarily improve symptoms.
(See "Gastroparesis: Etiology, clinical manifestations, and diagnosis", section on
'Evaluation' and "Treatment of gastroparesis".)
PROGNOSISFunctional dyspepsia has a chronic disease course with symptoms that
vary in severity over time [51]. Patients may be asymptomatic for periods of time and
have periodic symptomatic relapses. In two population-based studies of the natural
history of functional dyspepsia, over a follow-up of 10 to 12 years approximately 15 to
20 percent of individuals had persistent symptoms and 40 to 52 percent had symptom
resolution [81]. In 30 to 35 percent of patients symptoms fluctuated over time and
patients met criteria for another functional gastrointestinal disorder.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Dyspepsia".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Stomach ache and stomach upset (The
Basics)")
●Beyond the Basics topics (see "Patient education: Upset stomach (functional
dyspepsia) in adults (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●The prevalence of functional dyspepsia is up to 20 percent worldwide. The
pathophysiology of functional dyspepsia is not well understood. Several potential
mechanisms have been suggested. (See 'Epidemiology and
pathophysiology' above.)
●Patients with functional dyspepsia describe postprandial fullness, early satiety,
and/or epigastric pain/ burning. Nausea, vomiting, or heartburn are less frequent.
(See 'Epidemiology and pathophysiology' above and 'Clinical
manifestations' above.)
●A clinical diagnosis of functional dyspepsia requires the fulfillment of symptom-
based diagnostic criteria and an evaluation to exclude other causes of dyspepsia
(algorithm 1 and table 1). This evaluation consists of a history, physical
16
examination, laboratory studies, and endoscopic evaluation to exclude
organic/structural disease underlying the symptoms. The evaluation of a patient
with dyspepsia to establish the cause is discussed in detail, separately.
(See "Approach to the adult with dyspepsia".)
●Non-invasive testing for active Helicobacter pylori (H. pylori) infection should be
performed in patients with functional dyspepsia if gastric biopsies were not
obtained for H. pylori on upper endoscopy (algorithm 1). We suggest treatment
for H. pylori in patients with functional dyspepsia who test positive for an infection
(Grade 2A). (See "Treatment regimens for Helicobacter pylori in adults".)
●We suggest a four- to eight-week trial of a once daily proton pump inhibitor (PPI)
in patients with functional dyspepsia and no evidence of H. pylori and patients with
persistent symptoms after eradication of H. pylori (Grade 2A). (See 'Proton pump
inhibitors' above.)
●We suggest a tricyclic antidepressant drug for patients with persistent symptoms
after an eight-week trial of a PPI (Grade 2C). We start with a low dose
(eg, amitriptyline 10 mg at bedtime or desipramine 25 mg at bedtime) and
gradually increase the dose as tolerated. (See 'Antidepressants' above.)
●We suggest the use of prokinetics in patients in whom eradication of H.
pylori and a trial of proton pump inhibitor and tricyclic antidepressant has failed
(Grade 2C). In such patients, we generally limit a trial of metoclopramide to 5 to
10 mg three times daily one-half an hour before meals and at night for about four
weeks. The risk of side effects, including tardive dyskinesia, increase with the
cumulative dose and duration of treatment. We refer motivated patients who fail
medical therapy and patients who associate symptoms with stressors for
psychotherapy. (See 'Prokinetic agents' above.)
●Functional dyspepsia has a chronic disease course with symptoms that vary in
severity over time. Patients may be asymptomatic for periods of time followed by
symptomatic relapses. (See 'Prognosis' above.)

17
Acute hemorrhagic erosive gastropathy and reactive
gastropathy
Authors:
Pamela J Jensen, MD
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Mar 02, 2020.
INTRODUCTIONInjury to the gastric mucosa leads to epithelial cell damage and
regeneration. Epithelial cell damage and reactive regeneration without associated
inflammation is defined as "gastropathy" [1,2]. In contrast, the term gastritis is used to
denote inflammation associated with gastric mucosal injury. A practical etiology-based
framework for the classification of gastropathy and gastritis is presented [1].
(See "Gastritis: Etiology and diagnosis".)
This topic review will discuss acute hemorrhagic and erosive gastropathy and reactive
gastropathy. The classification and diagnosis of gastritis, and the causes of acute and
chronic gastritis are presented separately. (See "Gastritis: Etiology and
diagnosis" and "Acute and chronic gastritis due to Helicobacter pylori" and "Metaplastic
(chronic) atrophic gastritis".)
ACUTE HEMORRHAGIC EROSIVE GASTROPATHY
Etiology — Acute hemorrhagic erosive gastropathy is characterized by hemorrhagic
and erosive lesions that develop shortly after exposure of the gastric mucosa to
injurious substances or after a substantial reduction in mucosal blood flow (ischemic
gastropathy).
Gastric and duodenal ulceroinflammatory lesions occurring during severe damage to
the central nervous system (Cushing's ulcers) are often considered in this group,
although some authorities consider them to be clinically and pathogenetically distinct
[3,4]. (See "Stress ulcers in the intensive care unit: Diagnosis, management, and
prevention".)
Drugs/toxins — The most frequent injurious agents include:
●Alcohol – Alcohol exerts a direct toxic effect on gastric epithelium. In addition, it
impairs gastric motility, leading to delayed gastric emptying and consequent
prolonged mucosal contact time, further exacerbating the toxic injury [5]. Injury is
dose-dependent and usually requires a gastric alcohol concentration of >10
percent [6].
●Medications
•Nonsteroidal anti-inflammatory drugs (NSAIDs) – Damage can include
acute hemorrhagic gastropathy, erosions, reactive gastropathy, ulcers, and
perforation. Proton pump inhibitors (PPIs) have been used to mitigate
gastrointestinal toxicity; however, there have been reports suggesting that
PPIs can exacerbate NSAID-induced enteropathy, particularly within the small
18
 intestine [7]. (See "NSAIDs (including aspirin): Treatment of gastroduodenal
toxicity".)
•Iron pills – The consumption of large amounts of oral iron may lead to gastric
mucosal necrosis, erosions, and ulcers. Brown pigment may be seen overlying
the damaged epithelium [8].
•Oral sodium phosphate preparations – Oral sodium phosphate preparations,
used for colonoscopy bowel preparation, are associated with acute
hemorrhagic gastropathy [9]. (See "Bowel preparation before colonoscopy in
adults", section on 'Sodium phosphate preparations'.)
•Cancer chemotherapy – Damage to gastric mucosa in patients undergoing
cancer chemotherapy is multifactorial and includes direct mucosal injury by the
medication(s), vomiting-induced mucosal injury, and cancer-mediated or
chemotherapy-induced immunosuppression (with opportunistic gastric
infections). Hepatic arterial infusion of chemotherapy agents can produce
alarming regenerative cellular atypia, easily confused with carcinoma [10].
Ischemia — Cocaine can cause ischemic gastropathy by intense vasoconstriction with
subsequent ischemia due to activation of alpha-adrenergic receptors in the mesentery
[11]. Other causes of ischemic injury include hypovolemia, sepsis, trauma, and mucosal
prolapse.
Pathogenesis — As a result of the acute mucosal injury, the normal protective barrier
(which includes secreted mucins, bicarbonate, and the epithelium itself) is disrupted.
This permits acid and other luminal substances (eg, pepsin and bile acids) to penetrate
into the lamina propria, where they cause injury to the vasculature, stimulate nerves,
and cause the release of histamine and other inflammatory mediators [12,13]. The
mucosal injury increases significantly during the reperfusion that follows ischemia due to
the production of toxic oxygen-free radicals by infiltrating neutrophils [14].
An additional pathogenetic factor in NSAID-induced acute hemorrhagic and erosive
gastropathy is the inhibition of mucosal prostaglandin production. Prostaglandins protect
against acute mucosal injury due to injurious substances by several mechanisms,
including stimulation of mucus and bicarbonate secretion and increased vascular
permeability, resulting in perivascular edema that dilutes and delays toxic agents
reaching subepithelial capillaries [15,16]. (See "NSAIDs (including aspirin):
Pathogenesis of gastroduodenal toxicity".)
Clinical features — Acute gastropathies often present with abdominal discomfort/pain,
heartburn, nausea, vomiting, and hematemesis. Bleeding may be delayed until three to
seven days following the stressful event and can range from occult blood to massive
hemorrhage. Gastropathy (often referred to as gastritis) is commonly identified at the
time of endoscopy, but such gastropathies rarely lead to significant upper
gastrointestinal bleeding in the absence of other factors such as anticoagulation,
coagulopathy or concurrent gastric/duodenal ulcers. (See "Causes of upper
gastrointestinal bleeding in adults", section on 'Gastritis/gastropathy and
duodenitis/duodenopathy' and "Approach to acute upper gastrointestinal bleeding in
adults".)
Endoscopy and pathology
Gastric erosions — Acute hemorrhagic erosive gastropathy appears as multiple
petechial hemorrhages and small red or black erosions on endoscopy [17,18]. Stress-

19
related lesions (including Curling's ulcers) usually appear in the fundus near the
gastroesophageal junction and spread distally, but remain confined to the fundus and
body [18]. In contrast, gastropathy due to NSAIDs and alcohol ingestion involves the
entire stomach from the start, although it may be most evident in the antrum [18]. The
erosions in NSAID and alcohol-induced hemorrhagic gastropathy are usually smaller
and more rapidly reversible than those seen in ischemic gastropathy [18].
Erosions may have only subtle histological features because mucosal restitution and
repair are rapid. In many cases, an attenuated, regenerative-appearing epithelium
provides the only evidence for their occurrence [4]. Inflammation is usually slight
(consisting of no more than a few neutrophils) or absent.
Gastric/duodenal ulcers — Acute hemorrhagic erosive gastropathy may be associated
with the development of acute gastric and duodenal ulcers. Acute stress ulceration is
most likely to occur in relation to shock-induced hemodynamic instability. Ulcers are
usually multiple, shallow, 0.5 to 2.0 cm in diameter, and almost exclusively located in
the corpus and fundus [3,4,18,19].
The pathologic appearance of the ulcers depends upon their age. Ulcers initially appear
necrotic, while older ulcers often demonstrate inflammation, granulation tissue, and
epithelial cell regeneration. Ulcers of various ages can usually be found in the same
patient [19].
Gastric (or duodenal) ulcers developing in the setting of acute hemorrhagic gastropathy
may be associated with significant bleeding if they extend below the muscularis
mucosae into deeper layers where large vessels are located [3,4,19]. Bleeding is more
common from gastric than duodenal ulcers.
Management and prevention — Management of acute hemorrhagic erosive
gastropathy is based on the cause and the clinical presentation. Treatment includes
removing the causative agent and employing a limited course of acid suppression with a
PPI. Bleeding is typically self-limited. Additional measures that may be required include
withholding anticoagulants when they may be contributing (if possible), and, if bleeding
is severe, endoscopic therapy with modalities such as argon plasma coagulation.
(See "Causes of upper gastrointestinal bleeding in adults", section on
'Gastritis/gastropathy and duodenitis/duodenopathy'.)
A variety of measures may also be effective for prevention of acute hemorrhagic
gastropathy in high-risk patients. Examples include antisecretory agents (PPIs or H2
receptor antagonists) to prevent stress ulceration in the intensive care unit and for those
undergoing cancer chemotherapy. (See "Stress ulcers in the intensive care unit:
Diagnosis, management, and prevention".)
REACTIVE GASTROPATHYReactive gastropathy is usually caused by long-term
exposure of the gastric mucosa to substances capable of injuring the gastric mucosa
(medications, ethanol, bile reflux), or by mucosal ischemia or vascular congestion.
Reactive gastropathy is characterized by a paucity of inflammation, variable degrees of
foveolar hyperplasia, mucosal edema, proliferation of smooth muscle fibers in the
lamina propria, and vascular dilation and congestion.
Etiology — These changes were first recognized in severe bile reflux gastropathy [20],
and have subsequently been observed in biopsy specimens from patients with chronic
nonsteroidal anti-inflammatory drug (NSAID)-related mucosal injury [21]. Alendronate,
which can cause acute gastropathy [22], can also cause reactive gastropathy.
20
Chronic NSAID use — The pathogenetic connection between the acute hemorrhagic
erosive gastropathy discussed above and the chronic reactive gastropathy associated
with NSAIDs is not well understood. Continued, regular administration of NSAIDs in
humans and experimental animals can lead to a reduction in visible hemorrhages and
erosions as a result of mucosal adaptation [23-27]. However, many patients who take
these drugs for prolonged periods continue to have ulceroinflammatory lesions with
mucosal hemorrhages and erosions or, in approximately 15 percent of patients, gastric
or duodenal ulceration [28,29]. This outcome has been referred to as "failed adaptation"
[30]. H. pylori infection is known to exacerbate severe gastric mucosal injury in chronic
NSAID users, possibly due to impairment of this adaptation process [31]. The chronic,
poorly healing erosions and ulcers that often accompany NSAID-induced reactive
gastropathy are usually associated with suppressed epithelial cell regeneration, which
can be detected in biopsies obtained from the edge of ulcers [32,33]. The presence of
scarring in the muscularis mucosae and submucosa in gastric resection specimens from
NSAID users is also consistent with repeated episodes of erosion, ulceration, and
repair. (See "NSAIDs (including aspirin): Pathogenesis of gastroduodenal toxicity".)
Bile reflux gastropathy — Bile reflux gastropathy typically results from the entry of bile
into the stomach because of an operative stoma, an incompetent pyloric sphincter, or
abnormal duodenal motility [34]. The degree of gastric injury reflects the amount of
exposure to bile salts. Patients who have a gastroenterostomy, for example,
demonstrate the greatest degree of foveolar hyperplasia near the stoma, with
decreasing changes proximally [20,35,36]. Bile salt concentration in gastric juice
correlates with the severity of foveolar hyperplasia in these patients.
The effect of bile salts on the gastric mucosa is comparable to that seen after chronic
NSAID use [34]. Bile salts also appear to be a risk factor for the development of
metaplastic (chronic) atrophic gastritis in the antrum [37]. Animal studies have
demonstrated that bile salts and the phospholipid lysolecithin cause gastric mucosal
injury [38,39]. Their degree of toxicity is influenced by a number of variables, including
the type of bile salt, the presence or absence of deconjugation of bile salts by bacteria
in the intestine or stomach, the gastric pH, and the presence of pancreatic enzymes.
Pancreatic enzymes in particular enhance injury from bile salts, in part because they
can catalyze formation of lysolecithin from lecithin. (See "Metaplastic (chronic) atrophic
gastritis".)
Lysolecithin and bile salts break down the gastric mucosal barrier [40], leading to back-
diffusion of hydrogen ions and mucosal injury [38,39]. Concomitant infection by H.
pylori can increase inflammation [41,42]. Cytoprotection from bile-induced injury by
prostaglandins has been demonstrated in animal studies [43].
Other causes — Other oral agents known to be injurious to gastric mucosa and likely
causing reactive gastropathy with chronic use include potassium and iron supplements
[44-47].
Convincing evidence that smoking or alcohol causes reactive gastropathy are lacking
[5,21,23,48-50]. Exposure of the gastric mucosa to alcohol is associated with acute
hemorrhagic and erosive gastropathy (see 'Drugs/toxins' above) and chronic alcoholism
correlates with presence of hemorrhagic and erosive lesions, [51-53]. Investigations
directed at correlating alcohol consumption and chronic reactive gastropathy are difficult
to conduct because of numerous confounding variables, including the presence of other

21
potential causes of gastric injury such as NSAID use, H. pylori infection, and perhaps
smoking. The types of alcoholic beverages consumed also may be important because
of variations in alcohol concentration as well as important differences in content
between distilled and fermented alcohols [5].
Clinical features — Symptoms in patients with chronic bile acid reflux gastropathy and
other forms of reactive gastropathy are variable, ranging from histologic changes only
without symptoms to abdominal pain, bilious vomiting, and weight loss [54].
Diagnosis — The diagnosis of reactive gastropathy is usually established on biopsy
specimens from patients undergoing upper endoscopy for evaluation of upper
abdominal symptoms.
Endoscopy and histopathology — Although there are no endoscopic findings that are
diagnostic, mucosal changes are usually limited to hyperemia, and are confined to the
antrum or gastroenterostomy sites.
Reactive gastropathy is characterized by a paucity of inflammation, variable degrees of
foveolar hyperplasia, mucosal edema, proliferation of smooth muscle fibers in the
lamina propria, and vascular dilation and congestion [20]. Foveolar hyperplasia is
characterized by lengthening of the foveolae, with individual cells having reduced height
and depletion of intracellular mucin. The glands may take on a tortuous, corkscrew
appearance in severe cases, as is commonly seen in biopsy specimens taken near
gastroenterostomy stomas (in which gastropathy is usually due to bile reflux). Foveolar
hyperplasia is part of a spectrum of hyperplastic changes that are found near
gastroenterostomies; others include discrete polypoid lesions (ie, polypoid hypertrophic
gastritis or gastritis cystica polyposa) [35]. Foveolar hyperplasia can regress after bile is
diverted from the gastroenterostomies [55]. These changes were first recognized in
severe bile reflux gastropathy [20], and have subsequently been observed in biopsy
specimens from patients with chronic NSAID-related mucosal injury [21]. Alendronate,
which can cause acute gastropathy [22], can also cause reactive gastropathy.
Determining the etiology — Whereas a potential cause of reactive gastropathy may
be elicited based on initial history, it is often difficult to establish with certainty that a
particular agent has caused gastropathy [23,56]. In patients with surgically altered
anatomy, reactive gastropathy can be attributed to bile reflux gastropathy in the
absence of other causes. In contrast, gastropathy in patients with non-operated
stomachs should not be attributed to exposure to bile without evidence of
duodenogastric reflux. Duodenogastric reflux can be demonstrated by visualization of
bile in the stomach during endoscopy, bile salt analysis in gastric juice, or radionuclide
scanning [34,42,57,58].
Management — Treatment includes removing the causative agent. In patients with
symptomatic chronic bile acid reflux gastropathy, surgery to divert bile from refluxing
into the stomach (usually a Roux-en-Y revision), has been associated with improvement
in symptoms in 50 to 90 percent of patients in several case series [54,59-61].
Improvement is less likely in patients with delayed gastric emptying [59].

While a number of medical treatments have been evaluated, there are limited data to
support their use:

●Ursodeoxycholic acid (ursodiol) – In one small placebo-controlled crossover trial

that included 12 patients with symptomatic bile reflux gastropathy,
22
 ursodeoxycholic acid for 1 month decreased pain and improved nausea and
vomiting, but did not improve endoscopic appearance or histology [62]. A
subsequent non-randomized trial suggested that ursodiol is superior to a proton
pump inhibitor in these patients [63].
●Sucralfate – Sucralfate improved histologic features but not symptoms in a
randomized controlled trial that included 23 patients with symptoms of bile reflux
gastropathy following Billroth I, Billroth II, or vagotomy and pyloroplasty [64].
●Other – Prostaglandin E2 was ineffective in improving endoscopic features or
symptoms in a double-blind crossover trial of bile reflux gastropathy
[65]. Cholestyramine combined with alginates (to improve contact time in the
gastric remnant) was also ineffective on symptoms or histology in a placebo-
controlled trial that included 32 patients [66].
SUMMARY AND RECOMMENDATIONS
●Injury to the gastric mucosa leads to epithelial cell damage and regeneration.
Epithelial cell damage and regeneration without associated inflammation is
defined as "gastropathy." (See 'Introduction' above.)
●Acute hemorrhagic erosive gastropathy is characterized by hemorrhagic and
erosive lesions that develop shortly after exposure of the gastric mucosa to
injurious substances or after a substantial reduction in mucosal blood flow. The
most frequent injurious agents include drugs (eg, nonsteroidal anti-inflammatory
drugs and alcohol), hypovolemia, and chronic congestion. (See 'Etiology' above.)
●Acute hemorrhagic and erosive gastropathy appears as multiple petechial
hemorrhages and small red or black erosions on endoscopy. Erosions may have
only subtle histological features because mucosal restitution and repair are rapid.
In many cases, an attenuated regenerative appearing epithelium provides the only
evidence for their occurrence. Inflammation is usually minimal or absent.
(See 'Endoscopy and pathology' above.)
●Patients with acute hemorrhagic erosive gastropathy may present as abdominal
discomfort, pain, heartburn, nausea, vomiting, and hematemesis. Bleeding may be
seen three to seven days following the stressful event and can range from occult
blood to massive hemorrhage. Gastropathy (often referred to as gastritis) is
commonly identified at the time of endoscopy, but rarely lead to significant upper
gastrointestinal bleeding in the absence of other factors such as anticoagulation or
a coagulopathy or concurrent gastric/duodenal ulcers. (See 'Clinical
features' above.)
●Management of acute hemorrhagic erosive gastropathy is based on the cause
and the clinical presentation. Treatment includes removing the causative agent
and a limited course of acid suppression with a proton pump inhibitor. Bleeding
from gastropathy is typically self-limited. Additional measures that may be required
include withholding anticoagulants when they may be contributing (if possible),
and, if bleeding is severe, endoscopic therapy with modalities such as argon
plasma coagulation. (See 'Management and prevention' above.)
●Reactive gastropathy is usually caused by long-term exposure to substances
capable of injuring the gastric mucosa (medications, ethanol, bile), or by mucosal
ischemia or vascular congestion. (See 'Etiology' above.)

23
 ●Symptoms in patients with chronic bile reflux gastropathy and other forms of
reactive gastropathy are variable. Patients may be asymptomatic or have
symptoms of abdominal pain, bilious vomiting, and weight loss. (See 'Clinical
features' above.)
●The diagnosis of reactive gastropathy is usually established on biopsy specimens
from patients undergoing upper endoscopy for evaluation of upper abdominal
symptoms. Reactive gastropathy is characterized by a paucity of inflammation,
variable degrees of foveolar hyperplasia, mucosal edema, proliferation of smooth
muscle fibers in the lamina propria, and vascular dilation and congestion.
(See 'Endoscopy and histopathology' above.)
●The presumed offending agent should be eliminated in patients with reactive
gastropathy, if possible. In symptomatic patients with bile reflux gastropathy,
surgery diverting bile from refluxing into the stomach is the definitive treatment.
(See 'Management' above.)

24
Acute and chronic gastritis due to Helicobacter
pylori
Authors:
Pamela J Jensen, MD
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Sep 29, 2021.
INTRODUCTIONGastritis denotes inflammation associated with gastric mucosal
injury. Epithelial cell damage and regeneration without associated inflammation is
referred to as "gastropathy" [1,2]. Gastritis is usually caused by infectious agents
(eg, Helicobacter pylori) or is immune mediated, although in many cases the cause of
the gastritis is unknown.
This topic will review acute (active) and chronic gastritis due to H. pylori [3-5]. The other
forms of gastritis and gastropathy and other issues related to H. pylori are discussed
separately. (See "Gastritis: Etiology and diagnosis" and "Indications and diagnostic tests
for Helicobacter pylori infection in adults" and "Treatment regimens for Helicobacter
pylori in adults" and "Association between Helicobacter pylori infection and
gastrointestinal malignancy" and "Association between Helicobacter pylori infection and
duodenal ulcer" and "Helicobacter pylori and gastroesophageal reflux
disease" and "Pathophysiology of and immune response to Helicobacter pylori
infection".)
HELICOBACTER PYLORI GASTRITISH. pylori gastritis affects two-thirds of
the world's population and is one of the most common chronic inflammatory disorders
[6]. Most patients with H. pylori infection will show features of both acute and chronic
gastritis (chronic active gastritis).
Pathophysiology — H. pylori resides primarily in the unstirred layer of gastric mucus,
adjacent to epithelial cells at the mucosal surface and in gastric pits (picture 1A-B) [7].
Gastric glands are usually not involved. The epithelial localization reflects the affinity
of H. pylori for gastric mucous cells [8,9]. H. pylori does not attach to small intestinal or
other gastric epithelial cell types. The organisms are uncommonly found in the lamina
propria. (See "Pathophysiology of and immune response to Helicobacter pylori
infection".)
The usual natural history of H. pylori gastritis is of an antral predominant early stage of
infection with only minimal corpus involvement. This stage is associated with
exaggerated gastrin release and reduced somatostatin release, often precipitating an
increase in acid secretion, enough to cause duodenal ulcers in some patients [10].
With continued inflammation, gastrin producing (G) cells and acid producing parietal
cells are gradually lost, precipitating a fall in acid secretion and the development of
atrophy with intestinal metaplasia [11]. These changes facilitate the proximal migration
of the bacteria, leading to corpus gastritis [12]. Thus, the natural history of H.
25
pylori gastritis is of diffuse antral inflammation spreading to the corpus, resulting in an
atrophic front of advancing corpus injury with concomitant reduction in acid secretion.
This scenario is accelerated by hypochlorhydria such as that caused by chronic therapy
with proton pump inhibitors (PPIs). However, this orad evolution is not inevitable since it
can be modified by antibiotic treatment.
Patients in whom H. pylori colonization is heaviest in the gastric body may differ from
those with antral predominant infection. Duodenal ulcers are typically associated with
antral predominant gastritis, little or no atrophy, and high-normal or increased acid
secretion. By contrast, gastric ulcers and gastric cancer are typically associated with
more extensive gastritis, widespread intestinal metaplasia, and low-normal or reduced
gastric acid secretion [13].
Acute gastritis
Clinical manifestations — Patients with acute H. pylori are asymptomatic or develop
mild self-limited dyspeptic symptoms. Few examples of spontaneous acute infection
have been recognized since the majority of patients who develop nonspecific dyspeptic
complaints (which may signal acute infection) may not seek medical attention and thus
are not immediately investigated [14-16].
The ability of H. pylori to cause acute gastritis was first demonstrated after healthy
volunteers ingested the organisms and developed a mild illness (consisting of epigastric
pain, nausea, and vomiting without fever) associated with acute inflammatory changes
on gastric biopsy [17,18]. Acute infection was also demonstrated in volunteers
undergoing gastric secretory studies who were inadvertently infected by contaminated
equipment [19-21]. These cases also demonstrated that acute infection is associated
with the development of transient hypochlorhydria, a phenomenon that was suspected
to be caused by an infectious agent and was referred to as "epidemic hypochlorhydria"
[21]. (See "Pathophysiology of and immune response to Helicobacter pylori infection".)
Acute gastritis almost always evolves into active chronic gastritis unless the patient is
treated with appropriate antibiotics. (See 'Chronic gastritis' below.)
Endoscopic and histopathologic features — The endoscopic appearance of
acute H. pylori gastritis is variable and, in severe cases, can resemble lymphoma or
gastric carcinoma [22]. In early infection, H. pylori gastritis preferentially involves the
gastric antrum.
Histologic changes of acute H. pylori gastritis include intense neutrophilic infiltration of
the mucous neck region and lamina propria. When severe, pit abscesses occur, along
with mucin loss, erosion of the juxtaluminal cytoplasm, and desquamation of surface
foveolar cells. Both neutrophils and bacteria are responsible for destruction of the
epithelium. Lymphoid follicles appear within one week after the onset of acute H.
pylori infection, and are uncommon in non-H. pylori-infected gastric mucosa [20,23]. In
general, lymphoid follicles represent an immune response to the organism, and are
composed of aggregates of lymphocytes and other lymphoid cells associated with a
central germinal center made up of larger, paler mononuclear cells. The number of
lymphoid follicles present correlates with the titer of serum IgG anti-H. pylori antibodies
[24]. Lymphoid follicles accompanying H. pylori gastritis are involved in the genesis of
primary gastric B cell lymphoma [25,26]. The pathogenesis may involve stimulation of B
cells by activated T cells within the follicles. (See "Association between Helicobacter
pylori infection and gastrointestinal malignancy".)

26
Chronic gastritis
Clinical presentation
Gastroduodenal manifestations — It is unclear if chronic H. pylori infection causes
abdominal pain in the absence of peptic ulcer disease. However, it has been associated
with functional dyspepsia. Patients with chronic H. pylori gastritis can present with
complications of peptic ulcer disease or gastroduodenal complications of chronic
infection including gastric atrophy, intestinal metaplasia, gastric cancer, and mucosa-
associated lymphoid tissue (MALT) lymphoma. (See "Functional dyspepsia in adults",
section on 'Epidemiology and pathophysiology' and "Peptic ulcer disease: Clinical
manifestations and diagnosis", section on 'Clinical manifestations' and "Clinical features,
diagnosis, and staging of gastric cancer", section on 'Clinical features' and "Clinical
manifestations, pathologic features, and diagnosis of extranodal marginal zone
lymphoma of mucosa associated lymphoid tissue (MALT)", section on 'Clinical
features'.)
Extragastrointestinal manifestations — In some studies, H. pylori infection has been
associated with other extragastrointestinal diseases including:
●Iron deficiency anemia – H. pylori gastritis has been associated with iron
deficiency anemia [27,28]. The most plausible mechanism is decreased iron
absorption due to H. pylori-associated gastric atrophy and hypochlorhydria [28].
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults",
section on 'Indications for testing'.)
●Idiopathic thrombocytopenic purpura (ITP) – In some adults with ITP who
were infected with H. pylori, platelet counts increase following eradication therapy.
A proposed mechanism is molecular mimicry with cross-reactive antibodies.
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults",
section on 'Indications for testing'.)
●Vitamin B12 deficiency – H. pylori infection can lead to chronic (metaplastic)
atrophic gastritis resulting in hypochlorhydria and vitamin B12 malabsorption.
Also, the bacteria may elicit production of antibodies that cross-react with the
gastric parietal H+ K+ ATPase. Eradication of chronic H. pylori infection has been
associated with increases in vitamin B12 levels [28,29]. (See "Causes and
pathophysiology of vitamin B12 and folate deficiencies", section on 'H. pylori
infection' and "Metaplastic (chronic) atrophic gastritis", section on 'Helicobacter
pylori'.)
Associations have been noted between H. pylori infection and a large number of other
conditions, such as coronary artery disease, neurologic disorders, metabolic syndrome,
cerebrovascular disease, nonalcoholic fatty liver disease, and diabetes mellitus.
However, the data are controversial and insufficient to establish a causal link [28,30-32].
Endoscopic features — The endoscopic appearance is normal in as many as 50
percent of patients with chronic H. pylori gastritis [33,34]. Other patients may have
nonspecific endoscopic features including mucosal erythema, friable gastric mucosa,
and diffuse antral nodularity.
Histopathology — Histopathology plays a major role in diagnosing H. pylori infection,
establishing the presence, severity, and extent of gastritis, and detecting associated
complications of H. pylori infection (eg, intestinal metaplasia, MALT lymphoma, and

27
gastric carcinoma). H. pylori is almost always accompanied by gastritis and the
diagnosis of H. pylori should be suspect in its absence.
●Identification of H. pylori – A definitive histopathologic diagnosis of H.
pylori infection depends upon the demonstration of the typical spiral shaped bacilli
on a biopsy specimen. During antimicrobial treatment, H. pylori bacteria may lose
their typical spiral shape and assume new forms, including U-shaped, rod-like,
and coccoid forms. The coccoid forms appear as round basophilic dots, 0.4 to 1.2
micrometer in diameter.
The organisms can be detected in both the antrum and the body of the stomach in
the majority (80 percent) of chronically infected patients (image 1) [35]. H. pylori is
localized to the antrum alone or the body alone in 8 and 10 percent of patients,
respectively. Localization of H. pylori to the body alone is usually due to
concurrent PPI use or marked gastric atrophy and intestinal metaplasia.
Given the variable distribution of H. pylori in the stomach, and the attenuated
growth observed during treatment with PPIs, diagnostic accuracy can be
increased when biopsies are taken from multiple sites in the stomach. The
updated Sydney system [1] recommends that biopsy specimens be taken from
five different sites for optimal assessment of both gastritis and H. pylori status:
lesser and greater curvature of the antrum, lesser and greater curvature of the
corpus, and the incisura angularis. (See "Indications and diagnostic tests for
Helicobacter pylori infection in adults", section on 'Patient undergoing upper
endoscopy' and "Indications and diagnostic tests for Helicobacter pylori infection
in adults", section on 'Histology'.)
It is frequently possible to identify H. pylori in standard hematoxylin and eosin
(H&E) preparations. However, when a low density of H. pylori and atrophic
mucosal changes are both present, visualization of the organism becomes
unreliable on H&E alone [36,37]. Most pathologists use H&E plus a second stain
for H. pylori visualization. A variety of stains are available and can be divided into
non-silver-based stains, silver-based stains, and immunohistochemical stains [38]:
•Immunohistochemical stains – Immunostaining techniques are highly
sensitive, specific, and reliable (picture 2). They have a particular advantage in
patients partially treated for H. pylori gastritis, a setting that can result in
atypical (including coccoid) forms, which may mimic bacteria or cell debris on
hematoxylin and eosin preparations. When PPIs and other hypochlorhydric
states facilitate survival and overgrowth of non-H. pylori bacteria,
immunohistochemical stains can confirm the absence of H. pylori [39].
•Non-silver-based stains – The quick Giemsa method (eg, Diff-Quik) is easy to
use, inexpensive, and gives consistent results [40]. It is the preferred method
in many laboratories, particularly for screening, and when
immunohistochemistry is not readily available [41]. However, since it is a
morphologic stain, it is not as specific as immunohistochemistry.
•Silver-based stains – Silver stains (such Warthin-Starry and Genta methods),
which were crucial to the original demonstration of H. pylori, are expensive and
technologically complex. The results are not always reliable due to abundant
background artifact [40]. This stain has largely been replaced by the simpler
(Giemsa) and more specific (immunohistochemistry) stains.

28
 ●Associated gastritis – Acute and chronic inflammatory cells are concentrated in
the upper part of the mucosa, beginning just below the surface epithelium and
giving the appearance of superficial gastritis. This pattern is so characteristic
that H. pylori gastritis may be suspected even at the lowest magnifications.
The chronic inflammatory elements in H. pylori gastritis primarily consist of
lymphocytes and plasma cells, scattered macrophages, and often eosinophils
[42]. Lymphoid follicles are frequently present and represent an immune response
to the bacteria; their presence provides a useful marker for H. pylori infection.
Similarly, a prominence of plasma cells is a valuable clue to H. pylori infection.
The acute (active) inflammatory component consists of neutrophilic infiltration of
the surface and foveolar epithelium and the lamina propria, usually in scattered
foci, often with small pit abscesses. The intensity of the inflammation varies
among patients and sometimes from specimen to specimen in the same patient.
Active inflammation is somewhat more common in antral than in oxyntic H.
pylori infection [43]. Although casual observation reveals no obvious relation
between the numbers of organisms and the severity of the acute or chronic
inflammation, a correlation with bacterial density and active gastritis has been
described [38]. Inflammation associated with chronic H. pylori gastritis improves
dramatically after eradication of the organisms with appropriate antibiotics.
●Neutrophils disappear rapidly; the persistence of even small numbers of
neutrophils may be predictive of relapse [44].
●Lymphocytes and eosinophils decrease more slowly, and some chronic
inflammation can still be seen after one year. Lymphoid follicles are the slowest to
disappear, and usually persist for more than one year [44].
●Studies have shown that intestinal metaplasia and atrophy (if present) usually do
not resolve by one year but that significant improvement has been recognized
after 10 years [45,46]. Organism eradication may also help prevent the
development of further gastric atrophy and intestinal metaplasia [47].
●Fibrosis and architectural distortion, including foveolar hyperplasia, may persist
long after H. pylori infection is eliminated and often resembles chemical
gastropathy. (See "Acute hemorrhagic erosive gastropathy and reactive
gastropathy".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

29
 ●Basics topics (see "Patient education: H. pylori infection (The
Basics)" and "Patient education: Gastritis (The Basics)" and "Patient education:
Upper endoscopy (The Basics)")
●Beyond the Basics topics (see "Patient education: Helicobacter pylori infection
and treatment (Beyond the Basics)" and "Patient education: Upper endoscopy
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Gastritis denotes inflammation associated with gastric mucosal injury. Epithelial
cell damage and regeneration without associated inflammation is referred to as
"gastropathy." Gastritis is usually caused by infectious agents (eg, H. pylori) or is
immune mediated, although in many cases the cause of the gastritis is unknown.
(See 'Pathophysiology' above.)
●Patients with acute H. pylori are asymptomatic or develop mild self-limited
dyspeptic symptoms. Acute gastritis almost always evolves into active chronic
gastritis unless the patient is treated with appropriate antibiotics. (See 'Acute
gastritis' above.)
●It is unclear if chronic H. pylori infection causes abdominal pain in the absence of
peptic ulcer disease. However, it has been associated with functional dyspepsia.
Patients with chronic H. pylori gastritis can present with complications of peptic
ulcer disease or gastroduodenal complications of chronic infection including
gastric atrophy, intestinal metaplasia, gastric cancer, and mucosa-associated
lymphoid tissue lymphoma. The endoscopic appearance is normal in as many as
50 percent of patients with chronic H. pylori gastritis. Other patients may have
nonspecific endoscopic features including mucosal erythema, friable gastric
mucosa, and diffuse antral nodularity. (See 'Chronic gastritis' above.)
●H. pylori gastritis typically begins as a diffuse antral gastritis, which subsequently
spreads to the gastric corpus if untreated. Changes of chronic active gastritis may
be associated with or result in intestinal metaplasia (atrophy). Chronic use of
proton pump inhibitors (PPIs) may facilitate proximal migration of the organisms
leading to corpus gastritis. Given the variable distribution of H. pylori in the
stomach, and the attenuated growth observed during treatment with PPIs, we
obtain one to two biopsies from five different sites within the stomach (greater and
lesser curvature of antrum, greater and lesser curvature of the corpus, and the
incisura angularis). (See 'Chronic gastritis' above.)
●A definitive histopathologic diagnosis of H. pylori infection depends upon the
demonstration of the typical spiral shaped bacilli on a biopsy specimen.
Immunohistochemistry may be necessary for the detection of H. pylori organisms
in patients on an antibiotic, chronic PPI therapy, or with other hypochlorhydric
states that predispose to gastric bacterial overgrowth. Acute inflammation
disappears rapidly with treatment, but the chronic inflammation, including
lymphoid follicles, can persist for years. (See 'Endoscopic and histopathologic
features' above and 'Histopathology' above.)

30
Approach to the patient with large gastric folds
Authors:
Mark Feldman, MD, MACP, AGAF, FACG
Pamela J Jensen, MD
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jun 24, 2021.
INTRODUCTIONEnlarged or giant mucosal folds in the stomach (diffuse mucosal
hypertrophy) are attributable to a variety of proliferative, inflammatory, and infiltrative
conditions of the stomach. These conditions may primarily involve the mucosa,
submucosa, or both but have a similar endoscopic appearance, and the underlying
etiology cannot be determined without additional evaluation. This topic reviews the
epidemiology, etiology, and evaluation of large gastric folds.
HISTOPATHOLOGYThe normal gastric mucosa is composed of two
compartments: the superficial mucosa, which includes the surface mucous cells and
foveolar cells, and the deep compartment, which is composed of the specialized
epithelium [1]. Surface cells and foveolar cells line the upper portion of the gastric
glands. Slightly deeper in the gland is the isthmus, which contains the mucous neck
cells. The base of the oxyntic gland contains the specialized epithelium, which includes
the enterochromaffin cell (serotonin), the enterochromaffin-like cell (histamine), the chief
cell (pepsinogen), the D cell (somatostatin), and the parietal cell (acid and intrinsic
factor).

Enlarged or giant gastric folds are due to diffuse mucosal hypertrophy, which may be
due to a hyperplastic or nonhyperplastic gastropathy.

●Hyperplastic gastropathy – The gastric epithelial cells that compose the oxyntic
glands become hyperplastic and give rise to giant mucosal folds. Hyperplastic
gastropathies include Ménétrier's disease and its rare variant, hyperplastic
hypersecretory gastropathy, and Zollinger-Ellison syndrome. (See 'Hyperplastic
gastropathies' below.)
●Nonhyperplastic gastropathy – The gastric mucosa may contain other (non-
intrinsic) cell types that result in enlargement of the gastric folds. Nonhyperplastic
gastropathy may be due to Helicobacter pylori gastritis, infiltrative diseases,
infections, and malignancy. (See 'Nonhyperplastic gastropathies' below.)
In one series, endoscopic snare biopsy was performed in 52 patients with large gastric
folds (greater than 1.0 cm in width and persisting after air insufflation) [2]. The
diagnoses included the following:
●Chronic gastritis/lymphoid hyperplasia – 40 percent
●Benign tumors – 16 percent
●Gastric malignancy (adenocarcinoma, B cell lymphoma) – 12 percent

31
 ●Zollinger-Ellisonsyndrome – 10 percent
●Ménétrier's disease – 8 percent
ETIOLOGY OF ENLARGED GASTRIC FOLDS
Hyperplastic gastropathies
Ménétrier's disease — Ménétrier's disease (also called protein-losing hypertrophic
gastropathy) is a rare acquired condition of the stomach that is characterized by the
following features:
●Giant mucosal folds in the gastric fundus and body (picture 1)
●Diminished acid secretory capacity
●A protein-losing state resulting in hypoalbuminemia
Epidemiology — Ménétrier's disease is most common in middle-aged males and cases
have been reported in both adults and in children. In contrast to children, Ménétrier's
disease is usually progressive in adults. There are rare reports of spontaneous
regression in adults, and virtually none in adults with symptoms longer than six months'
duration [3]. An association with ulcerative colitis has been reported [4].
Pathogenesis — The pathogenesis of Ménétrier's disease is not well understood. It is
thought to involve increased signaling of the epidermal growth factor receptor (EGFR),
which results in proliferation of epithelial cells of the surface mucous cell compartment
[1,5-7]. The enhanced signaling of EGFR in Ménétrier's disease is driven by
overproduction of transforming growth factor-alpha (TGF-alpha), a ligand that can bind
and activate EGFR. TGF-alpha is produced by the oxyntic mucosa, predominantly by
the parietal cells; however, the molecular defect that results in the localized upregulation
of TGF-alpha is not known [1,6,7]. TGF-alpha is known to stimulate growth of epithelial
cells and to inhibit gastric acid secretion. The decreased acid production may be due to
both direct effects on the parietal cells and indirectly via stimulation of somatostatin
release.
Gastric infection with cytomegalovirus (CMV) is postulated to be an important cause of
Ménétrier's disease in children. One-third of the children with Ménétrier's reportedly
have CMV-associated Ménétrier's disease, which usually spontaneously resolves in two
to four weeks. It has been difficult to prove that an actual infection with CMV is present
in these pediatric cases due to the high prevalence of CMV infection in children. CMV
may infect epithelial cells by interacting with EGFR, thereby activating a tyrosine kinase,
a requirement for viral particle entry [8,9]. TGF-alpha was found to be increased by
immunohistochemistry in one childhood and one adult case of CMV-associated
Ménétrier's disease [8,10]. As CMV is cleared from the stomach, the stimulus to the
EGFR is decreased, thus resulting in less TGF-alpha production, lessened cell
proliferation, and a return to normal anatomy.
Clinical manifestations
●Signs and symptoms – Patients with Ménétrier's disease can present with a
variety of clinical features including epigastric pain, nausea, vomiting, diarrhea,
weight loss, and more rarely, gastrointestinal bleeding [11-14]. In addition, patients
with Ménétrier's disease also have manifestations of protein-losing
gastroenteropathy. These include peripheral edema, ascites, pleural and
pericardial effusions. Patients with the hypertrophic, hypersecretory variant of
Ménétrier's disease may have peptic ulcers due to high rates of gastric acid

32
 secretion. Otherwise, the clinical features are similar to the more common form of
Ménétrier's disease [13,15-19]. (See "Protein-losing gastroenteropathy".)
In one series of 40 patients, the main complaints were epigastric pain (65
percent), weakness (60 percent), anorexia (45 percent), weight loss (45 percent),
edema (38 percent), and vomiting (38 percent) [14]. In addition, approximately 80
percent of patients had hypoalbuminemia and increased enteric protein loss.
●Laboratory findings – Hypoalbuminemia is present in 20 to 100 percent of
patients with Ménétrier's disease [11-13,20,21]. Other laboratory abnormalities
include lymphopenia, hypochlorhydria, and reduced serum levels of globulins,
cholesterol, alpha-1 antitrypsin, fibrinogen, and ceruloplasmin; these reductions
are similar to patients with other forms of protein-losing gastroenteropathy.
Small to moderate elevation of serum gastrin levels may be present [22].
Intragastric pH is usually high (median pH of 6). Basal and stimulated gastric acid
secretion is usually low or normal in Ménétrier's disease, but varies depending
upon the point in the disease course when such measurements are made
[11,13,21,23]. Patients with the rare hypertrophic, hypersecretory variant of
Ménétrier's disease have elevated gastric acid secretion without significant
hypergastrinemia.
Cancer risk — There may be an increased risk of gastric neoplasia (dysplasia or
adenocarcinoma) in patients with Ménétrier's disease. However, there are few studies
that have evaluated the risk of gastric cancer in Ménétrier's disease [24]. Risk estimates
vary from 2 to 15 percent [11,13,21,25-27]. In many studies that have reported an
increased risk of gastric cancer in patients with Ménétrier's disease, adequate histologic
confirmation of the two conditions is lacking.  
Zollinger-Ellison syndrome — Zollinger-Ellison syndrome (gastric acid hypersecretion
due to hypergastrinemia from a gastrinoma, either sporadic or as a component of
multiple endocrine neoplasia type 1 [MEN1]) is a hyperplastic gastropathy involving the
body and fundus of the stomach. One hyperplastic cell type in Zollinger-Ellison
syndrome is the parietal cell that results in the hypersecretion of gastric acid. The high
levels of circulating gastrin seen in patients with gastrinoma lead to parietal cell
hyperplasia via increased expression of parathyroid hormone-related peptide (PTHrP),
also known as parathyroid hormone-like hormone (PTHLH), expression by parietal cells
[28]. There is also an increase in histamine-secreting enterochromaffin-like cells, as
gastrin is trophic for these cells. (See "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis".)
Nonhyperplastic gastropathies
Helicobacter pylori gastritis — Acute gastritis due to H. pylori can be associated with
striking acute inflammation and enlarged gastric folds, giving the impression of a gastric
malignancy [29]. Chronic H. pylori gastritis may also be associated with enlarged gastric
folds [30,31]. In a report of 48 patients biopsied with isolated thickened gastric folds,
chronic active gastritis was identified most frequently on biopsy (81 percent), with
evidence of H. pylori in 92 percent of those biopsied [32]. (See "Acute and chronic
gastritis due to Helicobacter pylori".)
Other observations have suggested an association between H. pylori gastritis and
enlarged gastric folds [15-17,33-36]. In one such series, 18 of 32 patients with thickened

33
gastric folds were infected with H. pylori [34]. H. pylori eradication therapy led to
normalization of gastric fold size.
The pathogenesis of diffuse mucosal hypertrophy in H. pylori gastritis is uncertain. One
case report suggested that production of an enteroendocrine mucosal growth factor,
glucagon-like peptide (GLP)-2, is upregulated in H. pylori infection. It is hypothesized
that the increase in this trophic hormone may result in the formation of giant folds, which
then recede along with the serum GLP-2 levels with resolution of infection [37].
Another study demonstrated different levels of cytokine production associated with
different strains of H. pylori [38]. They found that the H. pylori isolated from a patient
with Ménétrier's disease induced the production of high levels of hepatocyte growth
factor (HGF) mRNA expression. HGF has been reported to alter the tight junctions of
gastric epithelial cells, which would lead to protein loss.
Gastric neoplasia — Neoplasms, particularly adenocarcinoma and lymphoma or even
lymphoblastic leukemia [39], adenocarcinoma, and carcinoid tumors, should be
excluded in patients with enlarged gastric folds [2]. Endoscopic ultrasonography may be
helpful in this setting [40,41]; however, a biopsy is needed to make the diagnosis.
(See "Clinical presentation and diagnosis of primary gastrointestinal
lymphomas" and "Clinical features, diagnosis, and staging of gastric
cancer" and "Clinical characteristics of well-differentiated neuroendocrine (carcinoid)
tumors arising in the gastrointestinal and genitourinary tracts".)
Other causes — Other diseases that may be associated with large gastric folds include
lymphocytic gastritis (picture 2) (which may be related in some cases to H.
pylori [33,42]), sarcoidosis, eosinophilic gastroenteritis, HIV-associated hypertrophic
gastritis, and very rarely, Cronkhite-Canada syndrome [43]. (See "Gastrointestinal,
hepatic, pancreatic, and peritoneal sarcoidosis" and "Eosinophilic gastrointestinal
diseases" and "Overview of colon polyps".)
EVALUATION
History and examination — The patient's medical history and physical examination
may provide valuable clues as to the possible etiology of enlarged gastric folds. For
example, patients with pronounced constitutional symptoms (eg, persistent fevers, night
sweats, weight loss, fatigue) should be evaluated for systemic diseases such as gastric
lymphoma. Patients with Cronkhite-Canada syndrome have alopecia, cutaneous
hyperpigmentation, gastrointestinal polyposis, onychodystrophy, diarrhea, weight loss,
and abdominal pain. Patients with Zollinger-Ellison syndrome have a history of multiple
or refractory peptic ulcers; ulcers distal to the duodenum; peptic ulcer disease and
diarrhea; or multiple endocrine neoplasia type 1 (MEN1). (See "Overview of colon
polyps", section on 'Cronkhite-Canada syndrome'.)
In some cases, the presence of an underlying condition is already known (eg, H. pylori,
sarcoidosis, or eosinophilic gastroenteritis).
Upper endoscopy — The underlying etiology cannot be determined without additional
evaluation (eg, gastric biopsy). In patients with Ménétrier's disease, the enlarged folds
are usually confined to the oxyntic mucosa (body and fundus), with sparing of the
antrum (picture 1), and may reach several centimeters in thickness. Although gastric
folds are usually enlarged symmetrically, the appearance may be asymmetric and
polypoid. Gastric rugal folds may fill the lumen, and often have surface erosions, and

34
are accompanied by copious amounts of thick mucus that may form bridges across the
gastric lumen and obscure visualization.
Patients with Zollinger-Ellison syndrome and eosinophilic gastroenteritis may have
concurrent ulcers in the stomach or small intestine. The endoscopic appearance of
polyposis syndromes (familial adenomatous polyposis, Peutz-Jeghers syndrome) and,
rarely, proton pump inhibitor-induced fundal gland polyposis can mimic Ménétrier's
disease when polyposis is profuse [44,45]. However, these are distinguished by
histology. (See "Gastric polyps" and "Clinical manifestations and diagnosis of familial
adenomatous polyposis" and "Peutz-Jeghers syndrome: Clinical manifestations,
diagnosis, and management".)
Rarely, gastric varices in the fundus can give the appearance of enlarged folds on
imaging or on upper endoscopy. If gastric varices are suspected, gastric biopsies
should be avoided, and endoscopic ultrasonography should be performed to distinguish
between the two [46]. Varices are round anechoic structures arising from the lamina
propria or submucosa. Benign causes of giant gastric folds (eg, gastritis) tend to thicken
the second and third of the five layers of the stomach, whereas malignancies tend to
involve the fourth layer. Gastric lymphoma may be hypoechoic, whereas gastric
adenocarcinoma can show thick noncompressible folds [41]. (See "Endoscopic
ultrasound for the characterization of subepithelial lesions of the upper gastrointestinal
tract", section on 'Varices'.)
Gastric biopsy — Determining the etiology of large gastric folds is often possible based
on histopathology. A full-thickness gastric biopsy (or endoscopic snare or suction
biopsy) is usually required to evaluate for the loss of the deep glandular component
[21,47]. The diagnosis of Ménétrier's disease is established on biopsy by the
demonstration of extreme foveolar hyperplasia with glandular atrophy, involving loss of
parietal and chief cells. Ménétrier's disease usually is associated with a marked
reduction in parietal cell mass. The diagnosis of gastric sarcoidosis rests upon histologic
evidence of non-necrotizing granulomas in the biopsy specimens, exclusion of other
causes of granulomatous inflammation (eg, some infections), and clinical and pathologic
evidence of sarcoidosis affecting other organ systems. Additional stains for H. pylori and
cytomegalovirus should be performed to rule out concurrent infection.
Laboratory tests — In patients with Zollinger-Ellison syndrome, a serum gastrin value
greater than 10 times the upper limit of normal (1000 pg/mL) in the presence of a gastric
pH below 2 is diagnostic. In patients with Ménétrier's disease, intragastric pH is usually
high and small to moderate elevation of fasting serum gastrin levels may be present
[22].
We also obtain a complete blood count to evaluate for anemia and total protein and
serum albumin levels to screen for a protein-losing enteropathy. In patients with edema
and hypoalbuminemia, additional evaluation to confirm the presence of a protein-losing
gastroenteropathy is discussed in detail separately. (See "Protein-losing
gastroenteropathy", section on 'Laboratory studies'.)
MANAGEMENT
Treatment of underlying diseases — The mainstay of management is to treat the
underlying disease. The treatments used for specific disorders are described in detail
separately.

35
 ●Zollinger-Ellison (See "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis".)
●Sarcoidosis (See "Gastrointestinal, hepatic, pancreatic, and peritoneal
sarcoidosis", section on 'Gastric'.)
●Eosinophilic gastroenteritis (See "Eosinophilic gastrointestinal diseases", section
on 'Clinical manifestations'.)
●Cronkhite-Canada syndrome (See "Overview of colon polyps", section on
'Cronkhite-Canada syndrome'.)
●Lymphocytic gastritis (See "Gastritis: Etiology and diagnosis", section on 'Etiology
and classification'.)
●H. pylori (See "Treatment regimens for Helicobacter pylori in adults".)
Ménétrier's disease
Initial management — The treatment of Ménétrier's disease is challenging and the best
approach to treatment is unclear. Randomized controlled trials are lacking and data on
treatment efficacy are limited to small case series and case reports.
●General measures – Supportive care in patients with a protein-losing
enteropathy is with a low-fat, high-protein, high medium-chain triglyceride diet to
reduce protein loss. The diet should be low in long-chain fatty acids, generally
present as long-chain triglycerides. (See "Protein-losing gastroenteropathy",
section on 'Dietary therapy'.)
Patients with evidence of cytomegalovirus or H. pylori infections on gastric biopsy
should undergo treatment. Treatment of these infections has been associated with
resolution of Ménétrier's disease in case reports [33-36,42,48-50].
(See 'Helicobacter pylori gastritis' above and "Epidemiology, clinical
manifestations, and treatment of cytomegalovirus infection in immunocompetent
adults", section on 'Organ-specific complications'.)
The management of gastrointestinal bleeding in patients with Ménétrier's disease is
similar to other causes of GI bleeding and is discussed in detail separately.
(See "Approach to acute upper gastrointestinal bleeding in adults".)
●Gastric antisecretory agents – In patients with persistent symptoms despite
supportive measures, we perform a one- to three-month trial of an antisecretory
agent (histamine 2 receptor antagonist or proton pump inhibitor) in order to reduce
gastric acid secretion and gastric protein loss while monitoring the patient for a
reduction in symptoms and signs (eg, edema) and improvement in serum protein
levels. However, studies supporting the effectiveness of these agents are lacking.
Persistent symptoms — There are limited data to guide treatment in patients with
persistent symptoms despite antisecretory therapy, and the choice of treatment is based
on the nature and severity of symptoms, ease of administration, individual preference,
and insurance considerations.
●Octreotide – Multiple case reports of the use of octreotide, a somatostatin
analogue, for 3 to 12 months in patients with Ménétrier's disease suggest that its
use is associated with improvement of serum protein levels and in clinical
symptoms [51-54]. Molecular evidence suggests that octreotide may reduce
epidermal growth factor receptor (EGFR) signaling downstream [7]. However,
octreotide does not appear to consistently reverse the underlying gastric rugal
hypertrophy, and long-term outcome data are lacking [53-55]. In addition,

36
 octreotide and its analogues require subcutaneous administration. Although an
oral preparation is now available, it has not been evaluated in patients with
Ménétrier's disease.  
●Cetuximab – Cetuximab is a monoclonal antibody to EGFR that has been used to
treat Ménétrier's disease. In one case series that included nine patients with
Ménétrier's disease, seven patients completed treatment with cetuximab for one
month [4]. All seven patients demonstrated statistically significant improvement in
quality-of-life indices as well as in biochemical measures (gastric acidity). Four
patients also had almost complete reversal of foveolar hyperplasia. However, four
of the seven eventually required total gastrectomy [4]. In another case series of
two patients who were treated for four months and one year, respectively, both
had resolution of symptoms while on treatment with cetuximab [56]. In the United
States, the Food and Drug Administration (FDA) has allowed cetuximab to be
used on a compassionate need basis in some patients with Ménétrier's disease.
●Gastrectomy – Gastrectomy is reserved for those patients with intractable
symptoms and signs refractory to medical therapy. In rare cases with massive or
unrelenting bleeding that is uncontrollable endoscopically, gastrectomy is
required. Even though Ménétrier's disease is primarily a disease of the gastric
fundus and body, a total gastrectomy is the surgery of choice due to fewer surgical
complications and better quality of life as compared with partial proximal
gastrectomy.
Gastric cancer screening — In patients with Ménétrier's disease who have not
undergone gastrectomy, we perform endoscopic surveillance every one to two years for
early detection of dysplasia or carcinoma [8,57].
SUMMARY AND RECOMMENDATIONS
●Enlarged or giant mucosal folds in the stomach (diffuse mucosal hypertrophy) are
attributable to a variety of proliferative, inflammatory, and infiltrative conditions of
the stomach, which can cause a hyperplastic or nonhyperplastic gastropathy. In
hyperplastic gastropathies, gastric epithelial cells that compose the oxyntic glands
become hyperplastic and give rise to giant mucosal folds. Hyperplastic
gastropathies include Ménétrier's disease and its rare variant, hyperplastic
hypersecretory gastropathy, and Zollinger-Ellison syndrome. Large gastric folds
are associated with other non-hyperplastic gastropathies including infections
(eg, H. pylori, cytomegalovirus [CMV]), infiltrative diseases, and neoplasms.
(See 'Histopathology' above and 'Etiology of enlarged gastric folds' above.)
●Ménétrier's disease is a rare acquired condition of the stomach. Patients with
Ménétrier's disease can present with a variety of clinical features including
epigastric pain, weight loss, nausea, vomiting, diarrhea, peripheral edema due to
protein-losing gastroenteropathy, and gastrointestinal bleeding. (See 'Ménétrier's
disease' above.)
●The underlying etiology of large gastric folds cannot be determined based on
endoscopic imaging alone. Although the patient's medical history and physical
examination may provide valuable clues as to the underlying etiology of enlarged
gastric folds, a full-thickness gastric biopsy (or endoscopic snare or suction
biopsy) is usually required. The diagnosis of Ménétrier's disease is established by
a deep gastric biopsy showing extreme foveolar hyperplasia with glandular
37
atrophy in a patient with striking enlargement of gastric folds or rugae seen on
endoscopy or abdominal imaging. (See 'Upper endoscopy' above.)
●In patients with large gastric folds, we also obtain a complete blood count to
evaluate for anemia and total protein and serum albumin levels to screen for a
protein-losing enteropathy. (See 'Laboratory tests' above.)
●The mainstay of management is to treat the underlying disease. The treatment of
Ménétrier's disease is challenging and the best approach to treatment is unclear.
(See 'Ménétrier's disease' above.)
•In addition to a low-fat, high-protein, high medium-chain triglyceride diet to
reduce protein loss, initial management includes treatment of H. pylori or CMV
infection, if present, and a one- to three-month trial of acid suppressive
therapy.
•For those with persistent symptoms in spite of antisecretory therapy, options
include octreotide or cetuximab, an epidermal growth factor receptor inhibitor.
However, data on the efficacy of both options are limited in this population.
•Gastrectomy is reserved for those patients with intractable symptoms and
signs refractory to medical therapy. In rare cases with massive or unrelenting
bleeding that is uncontrollable endoscopically, gastrectomy is required.
(See 'Ménétrier's disease' above.)
●In patients with Ménétrier's disease who have not had total gastrectomy, we
perform endoscopic surveillance every one to two years because of the possible
increase in risk of gastric cancer. (See 'Cancer risk' above.)

38
Gastritis: Etiology and diagnosis
Authors:
Mark Feldman, MD, MACP, AGAF, FACG
Pamela J Jensen, MD
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Mar 02, 2020.
INTRODUCTIONGastric inflammatory disease can be broadly categorized into
gastritis and gastropathy based on the presence of associated mucosal inflammation
due to gastric injury. Gastritis is predominantly an inflammatory process, while the term
gastropathy denotes a gastric mucosal disorder with minimal to no inflammation.
Although the term "gastritis" is often used to describe endoscopic or radiologic
characteristics of abnormal-appearing gastric mucosa, a diagnosis of gastritis requires
histopathologic evidence of inflammation.
This topic will review the etiology, classification, and diagnosis of gastritis. Specific
causes of acute and chronic gastritis and gastropathy are presented in detail separately.
(See "Acute hemorrhagic erosive gastropathy and reactive gastropathy" and "Acute and
chronic gastritis due to Helicobacter pylori" and "Metaplastic (chronic) atrophic
gastritis" and "Granulomatous gastritis".)
DEFINITIONS
Gastropathy — Epithelial cell damage and regeneration with minimal or no associated
inflammation is termed "gastropathy."
Gastritis — The term gastritis is used to denote inflammation associated with gastric
mucosal injury.
ETIOLOGY AND CLASSIFICATIONGastritis is usually caused by infectious
agents (eg, H. pylori) or is immune-mediated, although in many cases the cause of the
gastritis is unknown.
Gastritis has been classified by time course (acute versus chronic), proposed
pathophysiology and histologic features, and etiology (table 1) [1]. Although several
classifications have been proposed, there is no universally accepted classification of
gastritis [1-12]. Classification of gastritis remains controversial because of gaps in
knowledge of etiology and pathogenesis, variation in nomenclature, and the frequent
coexistence of more than one type in an individual patient. Furthermore, the entities
overlap morphologically, and so some conditions are classified by etiology and others
by morphologic pattern. Most classification systems distinguish acute, short-term
disease from chronic, long-term disease. The terms acute and chronic are also used to
describe the type of inflammatory cell infiltrate. Acute inflammation is represented by
neutrophilic infiltration, while chronic inflammation is characterized by a mixture of
mononuclear cells, chiefly lymphocytes, plasma cells, and macrophages.
The most comprehensive and widely used classification system, the updated Sydney
system, classifies gastritis on the basis of the morphology, topography, and possible
39
etiology into acute, chronic, and special (or distinctive) [5]. Special forms of gastritis
include those without a clear etiology and gastropathies. Chronic gastritis is categorized
as nonatrophic or atrophic. However, the updated Sydney system does not yield
actionable prognostic information concerning the risk of gastric cancer in patients with
chronic atrophic gastritis and intestinal metaplasia. (See "Metaplastic (chronic) atrophic
gastritis".)
In patients with chronic atrophic gastritis and intestinal metaplasia, the Operative Link
for Gastritis Assessment [OLGA] and the Operative Link on Gastric Intestinal
Metaplasia Assessment (OLGIM) system for staging aid in prediction of cancer risk
(table 2 and table 3). Some experts suggest using a combination of both to stage
chronic atrophic gastritis and intestinal metaplasia. The OLGA/OLGIM staging systems,
incorporates histologic phenotypes with extent of disease to aid in prediction of cancer
risks [8-10]. High-stage disease (OLGA/OLGIM III/IV) are associated with a high risk of
gastric cancer. The OLGIM staging system shows less inter-observer variability as
compared to OLGA but may be less sensitive in identifying high-risk gastritis [8,9,11,12].
(See "Metaplastic (chronic) atrophic gastritis", section on 'Histologic staging of severity'.)
DIAGNOSISThe diagnosis of gastritis is usually established on biopsy specimens
from patients undergoing upper endoscopy for evaluation of often unrelated upper
abdominal symptoms.
Endoscopy and gastric mucosal biopsy — Endoscopic information should include a
brief narrative and/or endoscopic pictures of focal lesions (eg, thickened folds, polyps,
masses, erosions, or ulcers) and abnormal-looking areas. Communication between the
endoscopist and pathologist is necessary to optimize the interpretation of the biopsy
specimens.
●Endoscopic features – Endoscopic features of gastritis include erythema,
mucosal erosions, the absence of rugal folds, and the presence of visible vessels.
However, these features have a low sensitivity and there is significant inter-
observer variability for some features [13]. An illustrative study included 488 adults
who were randomly selected from the general population and were screened by
gastroscopy and biopsy [14]. The authors evaluated the sensitivity of macroscopic
features (including erythema, mucosal erosions, the absence of rugal folds, and
the presence of visible vessels) compared with histologic findings. None of these
endoscopic findings had sensitivity greater than 57 percent for determining the
presence of gastritis or H. pylori infection.
●Gastric biopsy protocol and histology – Histologic findings of gastritis can
vary over a wide spectrum ranging from epithelial hyperplasia with minimal
inflammation to extensive epithelial cell damage with infiltration by inflammatory
cells. Accurate histologic assessment of gastritis and gastropathy depends upon
optimizing the site and number of biopsy specimens. Magnification endoscopy
may help in the identification of areas to biopsy. (See "Magnification endoscopy".)
Biopsy site preferences and number vary in clinical practice based on the
incidence and risk factors for gastric cancer. However, there is general consensus
among experts on the following biopsy approach [1].
•Multiple biopsies of both the corpus and the antrum should be obtained when
attempting to evaluate gastritis and establish the etiology (eg, H. pylori or
autoimmune gastritis) (figure 1). For optimal assessment, one to two biopsies
40
 from five biopsy sites are taken (greater and lesser curvature of antrum,
greater and lesser curvature of the corpus, and the incisura angularis). This
approach is endorsed in the updated Sydney Classification [5].
•In first-degree relatives of patients with early onset gastric carcinoma (age
<45 years), endoscopic evaluation with at least four biopsies from the gastric
antrum and corpus is suggested. Multiple biopsies with adequate sampling are
necessary for scoring and staging of premalignant lesions in this subgroup of
patients [9].
•All visible abnormalities should be biopsied and submitted in separate
containers.
•Normal-appearing mucosa adjacent to the lesional tissue should also be
biopsied.
The pathologist should include a morphologic classification of the gastritis or
gastropathy in the biopsy report (table 4) [15].
Determining the etiology — The etiology of gastritis may be clear based on the clinical
history and histopathologic examination of the stomach [16]. However, in other cases,
additional testing may be required.
History — Key elements of the history include relevant predisposing factors/associated
conditions (eg, Crohn disease, sarcoidosis, celiac disease).
Duodenal histology if obtained — Biopsies of the duodenum may also be helpful for
diagnosing some forms of chronic gastritis. As examples, duodenal biopsies may show
evidence of Crohn disease in patients with granulomatous gastritis or celiac disease in
patients with lymphocytic gastritis. (See "Granulomatous gastritis", section on 'Etiology'.)
Additional tests in selected patients
●Testing for H. pylori – In patients with chronic atrophic gastritis without a clear
underlying etiology, hematoxylin and eosin (H&E) plus a second stain should be
used for H. pylori visualization (non-silver-based stains, silver-based stains, or
immunohistochemical stains). It is frequently possible to identify H. pylori in
standard H&E preparations. However, the distribution of H. pylori in the stomach is
variable, and the growth may be attenuated during treatment with proton pump
inhibitors. When a low density of H. pylori and atrophic mucosal changes are
combined, visualization of the organism becomes unreliable on H&E alone.
(See "Acute and chronic gastritis due to Helicobacter pylori", section on
'Histopathology'.)
Noninvasive testing for H. pylori with urea breath testing and stool antigen testing
has high sensitivity and specificity for active H. pylori infection. (See "Indications
and diagnostic tests for Helicobacter pylori infection in adults", section on
'Diagnostic tests'.)
●Immunologic markers – In patients with evidence of early/evolving histologic
features of autoimmune chronic atrophic gastritis, the presence of antibodies to
intrinsic factor or parietal cells, in conjunction with an elevated fasting serum
gastrin level, can support the diagnosis. (See "Metaplastic (chronic) atrophic
gastritis", section on 'Histologic features of AMAG and EMAG'.)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
41
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Gastritis (The Basics)" and "Patient

education: Upper endoscopy (The Basics)")
●Beyond the Basics topics (see "Patient education: Upper endoscopy (Beyond the
Basics)")
SUMMARY AND RECOMMENDATIONS
●Gastric inflammatory disease can be broadly categorized into gastritis and
gastropathy. Gastritis is predominantly an inflammatory process while gastropathy
denotes gastric mucosal pathology with minimal to no inflammation.
(See 'Introduction' above.)
●Gastritis is usually caused by infectious agents (most commonly, H. pylori) or is
immune-mediated, although in many cases the cause of the gastritis is unknown.
There is no universally accepted classification of gastritis/gastropathy, although
several classifications of gastritis and gastropathy have been proposed (table 1).
(See 'Etiology and classification' above.)
●The diagnosis of gastritis is usually established on biopsy specimens from
patients undergoing upper endoscopy for evaluation of often unrelated upper
abdominal symptoms. Endoscopic features of gastritis include erythema, mucosal
erosions, the absence of rugal folds, and the presence of visible vessels.
Histologic findings of gastritis can vary over a wide spectrum, ranging from
epithelial hyperplasia with minimal inflammation to extensive epithelial cell
damage with infiltration by inflammatory cells. Accurate histologic assessment of
gastritis and gastropathy depends upon optimizing the site and number of biopsy
specimens. (See 'Diagnosis' above.)
●The etiology of gastritis may be clear based on the clinical history and
histopathologic examination. However, in other cases (often in mild gastritis),
additional testing for H. pylori and autoimmune gastritis may be required.
(See 'Determining the etiology' above.)

42
Granulomatous gastritis
Authors:
Mark Feldman, MD, MACP, AGAF, FACG
Pamela J Jensen, MD
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jan 11, 2022.
INTRODUCTIONA granuloma is an organized aggregation of combined histiocytic,
lymphocytic, and plasmacytic infiltrates (granulomatous inflammation). When this
organized collection of cells is identified in the stomach, it is referred to as
granulomatous gastritis.
This topic review will discuss the epidemiology, etiology, histopathology, and diagnosis
of granulomatous gastritis. The classification of gastritis and causes of acute and
chronic gastritis are presented separately. (See "Gastritis: Etiology and
diagnosis" and "Acute and chronic gastritis due to Helicobacter pylori".)
EPIDEMIOLOGYGranulomatous gastritis is rare, with an incidence in gastric
biopsies of approximately 0.35 percent [1-5]. In one large series, only 56 cases were
diagnosed after review of 15,947 biopsies/resections [5].
ETIOLOGYGranulomatous gastritis is a type of chronic gastritis and can be
subclassified based on the etiology into infectious, noninfectious, and idiopathic
granulomatous gastritis [5].
A specific cause can be identified in most patients found to have granulomatous
gastritis [2,6]. The most likely causes vary by geography and ethnicity. Most cases of
granulomatous gastritis in developed countries are noninfectious with the most common
causes being Crohn disease and sarcoidosis [2,3,7]. In contrast, in resource-limited
countries, infections, particularly tuberculosis, are the most common cause of
granulomatous gastritis [8].
In a Belgian cohort in which biopsies from 71 patients with granulomatous gastritis were
reviewed, the most common causes of granulomatous gastritis were [2]:
●Crohn disease (55 percent) or possible Crohn disease (2 percent)
●Sarcoidosis (21 percent)
●Distal esophageal adenocarcinoma with chronic active gastritis (5 percent)
●Mucosa-associated lymphoid tissue (MALT) lymphoma (5 percent)
●Peptic ulcer complications (5 percent)
●Hypertrophic gastropathy and chronic active gastritis (2 percent)
In a study from Korea, 18 of 22,643 patients (0.08 percent) biopsied were diagnosed
with granulomatous gastritis over a six-year period (1997 to 2002) [4]. Of those, 14 had
chronic Helicobacter pylori gastritis, two had gastric adenocarcinoma, one had chronic
gastritis without H. pylori, and one had Crohn disease.
A retrospective review of 23 children and 23 adults with granulomatous gastritis in a
southern United States medical center demonstrated age- and race-related differences
43
in the etiologies of granulomatous gastritis [9]. In the children (average age 12 years),
the most common diagnosis was Crohn disease. In the adults (average age 49 years),
who were most often African American, the leading diagnoses were sarcoidosis and
idiopathic granulomatous gastritis.
Noninfectious causes
Crohn disease — Although granulomatous gastritis is rare in Crohn disease, this is the
most common etiology of granulomatous gastritis in developed countries. While
clinically apparent gastritis is seen in only 2 to 7 percent of patients with Crohn disease,
microscopic involvement may be detected in up to 75 percent of patients with Crohn
disease and endoscopically normal mucosa [9-12].
Individuals with Crohn disease of the stomach may have focally enhanced gastritis
(FEG), a focal inflammatory lesion composed of lymphocytes, histiocytes, and admixed
granulocytes (which involve either one or a few adjacent foveolae) in a background of
normal gastric mucosa. FEG is more frequent in the gastric antrum than in the gastric
body [13]. Lymphocytic gastritis may also be present [6]. Approximately 10 to 33
percent of these patients also have H. pylori infection with chronic active gastritis [6].
Granulomatous gastritis is present in approximately one-third of individuals with Crohn
disease and microscopic gastric involvement. The rate may be higher with increased
tissue sampling [14]. These granulomas are mucosal but are sharply separated from the
foveolar epithelium and often poorly formed (picture 1) [4].
Granulomatous gastritis may be present in patients with known Crohn colitis or ileitis or
may be the first manifestation of Crohn disease, particularly in children. In patients with
the histopathologic findings of FEG or granulomatous gastritis, or with symptoms
suggestive of inflammatory bowel disease, a search for intestinal Crohn disease should
be undertaken [2,3]. (See "Clinical manifestations, diagnosis, and prognosis of Crohn
disease in adults".)
Sarcoidosis — Within the gastrointestinal (GI) tract, the stomach is the most commonly
affected site in sarcoidosis [15]. Gastric granulomas found in patients with systemic
sarcoidosis are usually asymptomatic. Subclinical GI tract involvement has been
identified in 5 to 10 percent of patients with systemic sarcoidosis, while symptomatic
granulomatous gastritis has been reported in less than 1 percent [16]. Symptomatic
forms are dependent on the site of involvement in the stomach and range from
ulcerative lesions to lesions that are polypoid with gastric outlet obstruction. Infiltrative
and hyperplastic patterns of involvement may mimic Ménétrier's disease or linitis
plastica carcinoma [17]. (See "Gastrointestinal, hepatic, pancreatic, and peritoneal
sarcoidosis".)
Other causes — Rare noninfectious causes of granulomatous gastritis include tumors
(upper GI tract adenocarcinoma and MALT lymphoma), vasculitides (eg,
granulomatosis with polyangiitis [GPA] and eosinophilic granulomatosis with polyangiitis
[EGPA]), drugs (eg, cocaine, carbimazole, and interferon), xanthogranulomatous
gastritis, foreign body reactions to stomach contents within an ulcer repair zone or to
surgical material (eg, suture), crystalline drugs (eg, antacids containing magnesium,
aluminium, and silicon), reaction to endogenous materials (mucin, lipid, or crystalline
material), Langerhans cell histiocytosis (which is more common in childhood), and
chronic granulomatous disease (which is also more common in childhood and
adolescence) [18]. (See "Granulomatosis with polyangiitis and microscopic polyangiitis:

44
Clinical manifestations and diagnosis" and "Clinical manifestations, pathologic features,
and diagnosis of Langerhans cell histiocytosis" and "Chronic granulomatous disease:
Pathogenesis, clinical manifestations, and diagnosis".)
Infectious causes — An association with granulomatous gastritis and H. pylori has
been described [3,4,19,20], although evidence does not support H. pylori as an etiologic
agent in most cases of granulomatous gastritis [18,21]. In a review of 18,000 gastric
biopsies, H. pylori was only identified in 6 of 16 granulomatous biopsies [22]. In a
subsequent series of 71 patients, 92 percent of biopsies contained H. pylori; however,
many diverse etiologies for the associated granulomatous gastritis were also found [2].
Isolated granulomatous gastritis associated with H. pylori has reportedly shown
improvement after eradication therapy for the H. pylori, although persistence of
granulomas after 12 to 17 months was noted [23].
While H. pylori infection with related chronic active gastritis may be responsible for rare
cases of granulomatous gastritis, a search for other causes of granulomata must be
undertaken when H. pylori is present [1,22]. It is currently not known whether
granulomatous gastritis responds to H. pylori eradication [24].
Other infections that can cause granulomatous gastritis include tuberculosis, syphilis,
Whipple disease, a variety of fungal infections including cryptococcosis, as well as
anisakidosis, taeniasis, and infestations with other parasitic worms and schistosomes.
(See "Clinical manifestations, diagnosis, and treatment of miliary
tuberculosis" and "Syphilis: Treatment and monitoring" and "Whipple's
disease" and "Tapeworm infections" and "Schistosomiasis: Epidemiology and clinical
manifestations" and "Schistosomiasis: Diagnosis".)
Idiopathic granulomatous gastritis — In up to 25 percent of cases a cause of
granulomatous gastritis cannot be identified [2]. These patients are considered to have
idiopathic or isolated granulomatous gastritis [18,25]. The clinical importance of
idiopathic granulomatous gastritis and its existence as a distinctive condition are
uncertain [25,26]. It is possible that "idiopathic" granulomatous gastritis represents a
nonspecific inflammatory reaction to an as yet unknown or undiscovered agent [2,4].
Some of these patients may also eventually be diagnosed with a known cause (eg,
sarcoidosis or Crohn disease). Therefore the use of the descriptive terminology
granulomatous gastritis of uncertain etiology rather than the term idiopathic or isolated
granulomatous gastritis as a specific diagnosis has been suggested [21].
HISTOPATHOLOGIC FEATURESGranulomatous gastritis encompasses a
broad spectrum of histopathologic features ranging from vague, ill-defined collections of
mucosal/submucosal macrophages to fully formed granulomas with giant cells and
palisading histiocytes.
Granuloma type — Granulomas may be non-caseating, caseating, or necrotizing.
Caseation is usually associated with an infectious etiology that may be bacterial or
fungal (eg, Mycobacterium tuberculosis, atypical mycobacteria, histoplasmosis). Well-
defined, non-caseating epithelioid granulomas with a circumscribed solid appearance
are commonly seen in sarcoidosis. Non-caseating granulomas are also seen in Crohn
disease and occasionally in association with H. pylori. Necrotizing granulomas,
especially if associated with vasculitis, suggest eosinophilic granulomatosis with
polyangiitis (EGPA) or granulomatosis with polyangiitis (GPA).

45
Whereas the morphologic features of the granulomas may yield a clue to the etiology, in
many cases the granulomas are morphologically nonspecific. Special stains are needed
to rule out infectious etiologies (acid-fast stains and silver impregnation stains, or
specific immunohistochemical stains for designated microorganisms).

Granuloma environment — The tissue surrounding the granulomas may also contain

a clue to the etiology of the granulomatous gastritis. In Crohn disease, the gastric
mucosa may demonstrate focally enhanced gastritis, and in granulomatous disease
associated with H. pylori, a chronic active gastritis or chronic atrophic gastritis may be
present [21]. An infiltrate rich in eosinophils is suggestive of a parasitic infection, drug-
induced mucosal injury, or EGPA. The presence of isolated compact granulomas
sometimes surrounded by lymphocytes in otherwise normal-appearing mucosa is
suggestive of sarcoidosis [21]. Gastric involvement in syphilis is characterized by a
dense lymphoplasmacytic and granulomatous infiltrate often in a perivascular
distribution [27-29].
Other features — Diseases with granulomatous infiltrates, instead of having well-
formed granulomas, may have specific morphologic features, such as Langerhans cell
histiocytosis, in which the macrophages are reniform and are CD1a-positive. The
presence of necrotizing granulomatous vasculitis is consistent with GPA. EGPA (Churg-
Strauss syndrome) is characterized by extravascular granulomatous inflammation with
eosinophils comprising the predominant cell type in the inflammatory infiltrate.
Xanthogranulomatous gastritis (XGG), a specific subtype of granulomatous gastritis,
can present as a mass lesion, mimicking a gastrointestinal stromal tumor or other
gastric tumor when it involves the muscularis propria of the gastric wall [30-32]. A case
of XGG associated with gastric actinomycosis has been described [33].
CLINICAL FEATURESPatients may be asymptomatic or present with epigastric
abdominal pain. Nausea, vomiting, and weight loss may result from gastric outlet
obstruction. Patients with concomitant intestinal involvement (eg, from Crohn disease)
may have diarrhea. Upper gastrointestinal bleeding, although rare, has been reported in
patients with sarcoidosis and may occasionally be massive and fatal [34,35]. Other
clinical features vary based on the underlying etiology. (See "Gastrointestinal, hepatic,
pancreatic, and peritoneal sarcoidosis" and "Sarcoid myopathy" and "Clinical
manifestations, diagnosis, and prognosis of Crohn disease in adults" and "Whipple's
disease".)
DIAGNOSISThe diagnosis of granulomatous gastritis is based on the histopathologic
evaluation of gastric tissue. Granulomatous gastritis is considered to be present when
there is a nodular, fairly circumscribed collection of epithelioid histiocytes, often mixed
with a lymphocytic and plasmacytic infiltrate in the stomach. (See 'Histopathologic
features' above.)
An approach to the diagnosis of granulomatous gastritis is outlined in the algorithm
(algorithm 1) [17].
Evaluation to determine the underlying etiology — The following approach may be
used to determine the underlying cause of granulomatous gastritis:
●A clinical history should include a history of residence in or travel to areas
endemic for certain infectious diseases (eg, Mycobacterium tuberculosis, atypical
mycobacteria, histoplasmosis, H. pylori, schistosomiasis, teniasis), high-risk
46
 behaviors (eg, cocaine, syphilis), medications (eg, antacids), pulmonary or
systemic symptoms (eg, sarcoidosis, eosinophilic granulomatosis with polyangiitis,
Whipple's disease, mucosa-associated lymphoid tissue lymphoma), and/or
evidence of lower gastrointestinal symptoms (eg, Crohn disease, Whipple's
disease).
●A careful endoscopic inspection of the stomach with biopsies to map the extent of
the disease process should be undertaken. The endoscopic appearance of
granulomatous gastritis may reveal erythema, erosions, ulcers, rugal hypertrophy
or luminal obstruction simulating malignancy or be completely normal [22,36].
Note should be made of the presence of suture material. Multiple biopsies of the
gastric mucosa should be obtained to increase diagnostic yield, as granulomas
may be present in endoscopically normal mucosa.
●Histopathological examination should include hematoxylin and eosin stains to
evaluate the number of granulomas, location of the granulomas, and the nature of
the inflammatory background of surrounding gastric mucosa. Certain
histopathologic features may be suggestive of the underlying etiology. For
example, the presence of caseating necrosis may be due to mycobacterial or
fungal infections (Mycobacterium tuberculosis, atypical mycobacteria,
histoplasmosis). The presence of well-defined epithelioid granulomas surrounded
by normal-appearing mucosa is suggestive of sarcoidosis. Non-caseating
granulomas are seen with Crohn disease and occasionally in association with H.
pylori. In Crohn disease, the gastric mucosa may demonstrate focally enhanced
gastritis. Necrotizing granulomas, especially if associated with vasculitis, suggest
granulomatosis with polyangiitis. Special stains should be performed for specific
bacteria/fungi (acid-fast stains and silver impregnation stains, or specific
immunohistochemical stains for designated microorganisms). Polarization
microscopy should be performed to look for birefringent material (foreign body).
●Laboratory testing, colonoscopy, and imaging studies should be based on clinical
suspicion (eg, chest radiography for tuberculosis and sarcoidosis, angiotensin-
converting enzyme level for sarcoidosis, serum antineutrophil cytoplasmic
antibodies for GPA, and serologic tests for syphilis). (See "Clinical manifestations,
diagnosis, and treatment of miliary tuberculosis" and "Clinical manifestations and
diagnosis of pulmonary sarcoidosis" and "Granulomatosis with polyangiitis and
microscopic polyangiitis: Clinical manifestations and diagnosis" and "Syphilis:
Screening and diagnostic testing".)
SUMMARY AND RECOMMENDATIONS
●A granuloma is an organized aggregation of combined histiocytic, lymphocytic,
and plasma cell infiltrates (granulomatous inflammation). When this organized
collection of cells is identified in the stomach, it is referred to as granulomatous
gastritis. (See 'Introduction' above.)
●Granulomatous gastritis encompasses a broad spectrum of histopathologic
features ranging from vague, ill-defined collections of mucosal/submucosal
macrophages to fully formed granulomas with giant cells and palisades of
histiocytes. (See 'Histopathologic features' above.)
●Granulomatous gastritis is a type of chronic gastritis that may be infectious,
noninfectious, or idiopathic in etiology. Granulomatous gastritis is rare, with an
47
incidence in gastric biopsies of approximately 0.35 percent.
(See 'Epidemiology' above.)
●In developed countries, the most common causes of granulomatous gastritis are
Crohn disease and sarcoidosis, while in developing countries, infectious causes
are more common.
●The cause of granulomatous gastritis cannot be identified in up to 25 percent of
patients. These individuals are considered to have "granulomatous gastritis of
uncertain etiology." Some of these patients will eventually present with an
identifiable etiology for the granulomatous gastritis. (See 'Etiology' above.)
●Determining the etiology of granulomatous gastritis requires a detailed clinical
history and endoscopic evaluation including biopsies for histopathological
examination that includes hematoxylin and eosin stains. Special stains should be
performed for designated organisms/fungi (acid-fast stains and silver impregnation
stains, or specific immunohistochemical stains) and polarization.
(See 'Histopathologic features' above.)
Additional laboratory testing, colonoscopy, and imaging studies should be
performed to establish the etiology of granulomatous gastritis, based on the
clinical setting. (See 'Diagnosis' above.)

48
Metaplastic (chronic) atrophic gastritis
Authors:
Pamela J Jensen, MD
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Sep 02, 2020.
INTRODUCTIONGastritis usually has an infectious or autoimmune etiology. This
topic review discusses the two types of metaplastic (chronic) atrophic gastritis [1,2].
Other forms of gastritis and gastropathy are presented separately. (See "Gastritis:
Etiology and diagnosis" and "Acute and chronic gastritis due to Helicobacter
pylori" and "Acute hemorrhagic erosive gastropathy and reactive
gastropathy" and "Granulomatous gastritis" and "Approach to the patient with large
gastric folds".)
DEFINITIONThe term metaplastic (chronic) atrophic gastritis, also referred to as
gastric atrophy, is used to describe a form of chronic gastritis that, in addition to
inflammation, is associated with mucosal thinning, loss of specialized cells in gastric
glands, and changes in epithelial cell types (ie, metaplasia).
SUBTYPESMetaplastic (chronic) atrophic gastritis includes two main subtypes,
autoimmune and environmental metaplastic atrophic gastritis (AMAG and EMAG).
Although the AMAG and EMAG may be pathogenetically and clinically distinct, they
often share histologic features and may overlap clinically. (See 'Autoimmune
metaplastic atrophic gastritis' below and 'Environmental metaplastic atrophic
gastritis' below.)
Autoimmune metaplastic atrophic gastritis — AMAG is a form of metaplastic
(chronic) atrophic gastritis that results in the replacement of the normal oxyntic mucosa
in the gastric corpus by atrophic and metaplastic mucosa, leading to a corpus
predominant atrophic gastritis, reduced or absent acid and pepsin production, and loss
of intrinsic factor, which may progress to a severe form of vitamin B12-deficiency
anemia known as pernicious anemia (PA).
Epidemiology — AMAG and PA have prevalences of 2 percent and 0.15 to 1 percent,
respectively [3]. The prevalence of AMAG increases with age and is higher in women as
compared with men. In the United States, AMAG has a similar prevalence among
White, Hispanic, and African American populations, and occurs over a wide age range
[4]. There is an association of AMAG with other autoimmune diseases [5,6]. Up to one-
third of patients with autoimmune thyroid disease and 6 to 10 percent of patients with
type 1 diabetes mellitus (DM) have concurrent AMAG [7].
Etiopathogenesis — AMAG is associated with a T-cell mediated destruction of the
oxyntic mucosa and production of autoantibodies directed against parietal cell antigens
and intrinsic factor. In a mouse model of autoimmune gastritis, antibodies were directed
against the hydrogen-potassium ATPase enzyme in parietal cells [8]. The chronic
49
inflammation, glandular atrophy, and epithelial metaplasia of AMAG are closely
paralleled by elevated serum antibodies to parietal cell antigens and to intrinsic factor,
reflecting its autoimmune origin. Serum pepsinogen I levels, produced in the oxyntic
mucosa, decrease also, as does gastric pepsinogen secretion. (See 'Laboratory
features' below and 'Laboratory testing' below.)
Both genetic and environmental factors may play a role in the pathogenesis of AMAG.
Autoimmune gastritis susceptibility genes (Gasa 1, 2, 3, and 4) have been discovered
on chromosomes 4 and 6 and the H2 gene complex in murine models. Some of these
genes are located on the same locus as mouse DM susceptibility genes, which may
account for the strong association between AMAG and type 1 DM in humans [9].
Asymptomatic relatives of patients with AMAG are sometimes found to have established
AMAG and even PA [5].
There is evidence that H. pylori might serve as a trigger of AMAG and PA and that H.
pylori infection at a younger age triggers the development of autoimmune gastritis,
which progresses clinically over time from iron deficiency anemia (perhaps due to iron
malabsorption because of hypochlorhydria/achlorhydria) to cobalamin deficiency with
macrocytosis [10]. H. pylori infection has also been associated with iron deficiency
anemia in patients without evidence of gastrointestinal blood loss or metaplastic
atrophic gastritis by mechanisms that are unclear. The association between AMAG
and H. pylori infection in many patients has led to speculation that H. pylori may, in
some cases, induce AMAG via antigenic mimicry or cross-reactivity. However, this
hypothesis is controversial and the subject of active research [11]. Patients with AMAG
are less likely to be infected by H. pylori than age-matched controls [12]. Two possible
explanations are that the metaplastic intestinal epithelium becomes unsuitable (lacks
host receptors) for H. pylori colonization, and that the associated hypochlorhydria
encourages overgrowth of the stomach by other bacterial species [13].
Despite their inverse relationship, chronic H. pylori gastritis and AMAG share some
clinical and pathologic features. Both can be associated with an atrophic form of body
gastritis and hypochlorhydria. In addition, some patients with H. pylori infection have
circulating parietal cell autoantibodies [14]. Serologic testing for past H. pylori infection
may be helpful in such patients. (See "Indications and diagnostic tests for Helicobacter
pylori infection in adults", section on 'Serology'.)
Clinical features — Patients with AMAG may be asymptomatic, but many of them have
dyspepsia with postprandial distress [15]. Patients with AMAG may be symptomatic
from vitamin B12 malabsorption and PA [2,5,15]. The presence of symptoms
attributable to anemia depends on the rate that deficiency has developed, the severity
of the deficiency, the hemoglobin level, and the person's overall health. Patients with
B12 deficiency usually have vague or nonspecific symptoms (eg, fatigue, irritability,
cognitive decline), which are likely to be due (at least in part) to anemia. B12 deficiency
can cause glossitis (including pain, swelling, tenderness, and loss of papillae of the
tongue). B12 deficiency can also cause subtle neurologic, cognitive, or psychiatric
changes. The most common neurologic manifestation is symmetric paresthesias or
numbness and gait problems. (See "Clinical manifestations and diagnosis of vitamin
B12 and folate deficiency".)
Laboratory features — Patients with AMAG may present with the following laboratory
abnormalities [4]:

50
 ●Elevated fasting serum gastrin – Hypergastrinemia in AMAG results from
uninhibited gastrin secretion as a consequence of parietal cell atrophy and
hypochlorhydria/achlorhydria.
●Decreased serum pepsinogen I/II ratio – Atrophy of zymogenic chief cells in
the oxyntic mucosa results in a reduction in serum pepsinogen I but not serum
pepsinogen II levels [1].
●Iron deficiency anemia – Iron deficiency anemia may be the most common
hematologic presentation of AMAG and can precede the onset of PA by several
years [11,16]. Gastric acid normally enhances iron solubility and intestinal iron
absorption by converting the ferric form of iron into the more absorbable ferrous
form. Gastric acid also facilitates peptic digestion of dietary proteins bound to iron.
Hypochlorhydria/achlorhydria and reduced peptic activity in AMAG decreases
bioavailable iron, leading to iron malabsorption and deficiency.
●Low serum vitamin B12 level – Low gastric acid and peptic activity reduce
liberation of B12 bound to dietary proteins. Also, antibodies to intrinsic factor
impair vitamin B12 absorption. Manifestations of vitamin B12 deficiency include
anemia with macrocytosis (high mean corpuscular volume), elevated
methylmalonic acid, pancytopenia, and hypersegmented neutrophils.
(See "Treatment of vitamin B12 and folate deficiencies".)
Individuals with AMAG may also have laboratory findings of other concurrent
autoimmune disorders (eg, thyroid disease, type 1 diabetes mellitus) [4-6,15,17,18].
(See 'Epidemiology' above and "Disorders that cause hypothyroidism", section on
'Chronic autoimmune (Hashimoto's) thyroiditis' and "Clinical presentation, diagnosis,
and initial evaluation of diabetes mellitus in adults", section on 'A1C'.)
Cancer risk — Patients with AMAG are at increased risk for the development of gastric
neuroendocrine tumors and adenocarcinomas.
Gastric neuroendocrine (carcinoid) tumors — Limited data on the long-term
incidence of gastric neuroendocrine tumors in patients with metaplastic (chronic)
atrophic gastritis suggest an annual incidence of 0.68 percent per person-years [19].
Neuroendocrine tumors arise from transformation of enterochromaffin-like (ECL) cells
(which are responsible for histamine secretion) within the oxyntic mucosa due to chronic
stimulation by high circulating levels of gastrin. The hypochlorhydria or achlorhydria
associated with AMAG induces hyperplasia of the antral/pyloric G cells (the cell type
responsible for producing gastrin) and hypergastrinemia. Gastrin exerts a trophic effect
on endocrine cells (eg, ECL cells) within the metaplastic and atrophic body mucosa,
which increase in number and may give rise to hyperplastic nodules both within and
around the glands [20,21]. The relationship between gastrin-induced ECL cell
hyperplasia and neuroendocrine tumor formation is supported by the observation that
antrectomy, with resultant loss of G cell mass and normalization of plasma gastrin
concentrations, can lead to reversal of endocrine hyperplasia and reduced carcinoid
tumor size [22,23]. (See "Clinical characteristics of well-differentiated neuroendocrine
(carcinoid) tumors arising in the gastrointestinal and genitourinary tracts".)
Neuroendocrine tumors in AMAG patients usually appear grossly as multiple small (<1
cm) mucosal nodules or polyps. However, most nodular and polypoid lesions
encountered in the gastric body and antrum in patients with AMAG are retained islands

51
of uninvolved mucosa (pseudopolyps) or benign epithelial hyperplastic lesions, making
biopsy of visible nodules imperative [24,25]. (See 'Endoscopy and biopsy' below.)
Gastric adenocarcinomas — Gastric adenocarcinoma develops in patients with
AMAG via intervening steps of intestinal metaplasia (IM) and dysplasia [1,26,27]. There
is evidence that pseudopyloric metaplasia is a precursor for IM [28,29]. Estimates of the
magnitude of the risk vary widely in the published literature [30-36]. A meta-analysis of
27 studies that included a total of 22417 patients showed that the calculated pooled
gastric cancer incidence rate was 0.27 percent per person-year and the overall gastric
cancer relative risk in PA was 6.8 (95% CI 2.6 to 18.1) [36]. However, patients included
in these studies may have had low vitamin B12 serum levels due to conditions other
than autoimmune gastritis. A subsequent retrospective study of 150 patients diagnosed
with AMAG on endoscopic gastric biopsy, found a high prevalence (5.3 percent) and
incidence (14.2 cases per 1000 person-years) of gastric adenocarcinoma, far exceeding
that of the general population (0.073 cases per 1000 person-years) [37].
Environmental metaplastic atrophic gastritis
Epidemiology and risk factors — Environmental metaplastic atrophic gastritis
(EMAG) is thought to be due to the adverse effects of environmental factors, such as H.
pylori infection and perhaps dietary constituents, on the gastric mucosa. However, few
studies have attempted to identify risk factors for EMAG as confirmed by biopsy.
Instead, in such epidemiologic studies, individuals with EMAG are usually identified
indirectly by screening with serum pepsinogen concentrations. A shortcoming of this
approach is that H. pylori infection, which is often common in the populations under
study, increases serum pepsinogen levels [38,39]. As a result, the prevalence of EMAG
is most likely to have been underestimated using serum pepsinogen screening [40].
However, the correlation between a low serum pepsinogen I level and histologic
demonstration of EMAG is well-documented, and pepsinogen determination is the best
available tool for population studies of this disorder [1,41,42]. (See "Gastritis: Etiology
and diagnosis", section on 'Additional tests in selected patients'.)

EMAG and intestinal-type gastric cancer are closely associated pathologically, clinically,
and epidemiologically. Thus, similar contributing factors are proposed for the two
conditions.

Helicobacter pylori — H. pylori plays a pivotal role in the development of EMAG and

gastric adenocarcinoma. Histopathologic studies indicate that chronic H. pylori infection
progresses over decades through stages of chronic superficial gastritis, gastric atrophy,
IM, dysplasia, and cancer [1,27,43]. The mechanisms whereby some patients with H.
pylori infection develop IM and carcinoma, while others do not, are multifactorial and
likely involve an interplay of bacterial virulence factors and host susceptibility factors.
(See "Association between Helicobacter pylori infection and gastrointestinal
malignancy".)
Dietary and other risk factors — Numerous studies have investigated possible dietary
causes of EMAG (ie, high salt intake), with inconclusive or inconsistent results [44].
Other possible risk factors include cigarette smoking, alcohol consumption, and chronic
bile reflux [1].
Clinical and laboratory features — Patients with EMAG may be asymptomatic, but
many of them have dyspepsia. In contrast to AMAG, fasting serum gastrin levels are not
52
markedly elevated in EMAG. Parietal cell and intrinsic factor autoantibodies and PA are
absent. (See 'Laboratory testing' below.)
Cancer risk — Patients with EMAG are at a 10- to 15-fold increased risk for gastric
cancer, particularly the intestinal type [1,27]. It is hypothesized that increased
macrophage-derived inflammatory drivers present in EMAG may be responsible for the
progression from metaplasia to carcinoma in this population [28]. (See 'Cancer
risk' above and 'Histologic features of AMAG and EMAG' below.)
DIAGNOSISThe diagnosis of metaplastic (chronic) atrophic gastritis is based on the
histologic evaluation of gastric biopsies which demonstrate atrophy of the gastric
mucosa with the loss of glandular cells and their replacement by metaplastic epithelium.
Endoscopy and biopsy — The endoscopic appearance of chronic atrophic gastritis is
normal during the early disease stages. Only in cases with extensive atrophy are the
rugal folds flattened and the submucosal vessels visible. Mucosa may appear
pseudopolypoid as polypoid areas represent islands of preserved oxyntic mucosa
adjacent to areas of atrophy. Chromoendoscopy may be superior to white-light
endoscopy [45].
Biopsy is the most reliable method to diagnose metaplastic atrophic gastritis and
diagnose H. pylori. Assessment of the severity of gastric atrophy and determining the
subtype of chronic atrophic gastritis requires gastric biopsy mapping with an adequate
number of biopsies from specific sites. We perform gastric biopsy mapping on the index
(initial) endoscopy in patients at increased risk of gastric cancer or in whom metaplastic
atrophic gastritis is suspected (eg, pernicious anemia [PA], multiple pseudopolypoid
lesions).
Biopsy mapping protocol — Biopsies of at least two topographic sites (ie, from both
the antrum and the corpus, at the lesser and greater curvature of each). The incisura
angularis is especially affected by metaplasia and atrophy and should be included in the
biopsy protocol (figure 1) [46]. Additional biopsies of suspicious-looking lesions should
also be taken. Biopsies should include adjacent flat (nonpolypoid) antral and body
mucosa. In autoimmune metaplastic atrophic gastritis (AMAG), polypoid areas may
represent islands of preserved oxyntic mucosa and atrophy will be missed unless the
flat area near the polyp is also sampled [11]. (See "Gastric intestinal metaplasia",
section on 'Upper endoscopy with biopsy'.)

Biopsies of the antrum and body should be clearly labelled in separate vials. Biopsies of
the incisura are typically included with the antral biopsies. Severe metaplastic (chronic)
atrophic gastritis can deprive pathologists of the histologic clues needed to identify the
origin of the biopsy. Thus, it is essential that biopsy specimens be taken from areas that
endoscopists believe are unequivocally within the body/fundus or antrum, and indicate
their origin clearly and place them in separate containers.

Immunostaining for endocrine cells (eg, G cells and enterochromaffin-like [ECL] cells) is
also helpful for confirming that pseudopyloric metaplasia has replaced oxyntic mucosa.
G cells are sparse or absent in pseudopyloric glands; in comparison, they are plentiful in
biopsy specimens from the true antrum in AMAG and environmental metaplastic
atrophic gastritis (EMAG), except when there is marked antral intestinal metaplasia and

53
atrophy. The presence of ECL cell hyperplasia in patients with AMAG can help confirm
that a biopsy specimen is from the body or fundus.

Histopathologic features of metaplastic (chronic) atrophic gastritis — Two main

types of metaplasia are seen in chronic atrophic gastritis: pseudopyloric metaplasia and
intestinal metaplasia (IM). Metaplasia, especially of the intestinal type, is virtually a
universal feature of chronic atrophic gastritis and is often the most dependable defining
morphologic feature. IM is a precursor lesion to dysplasia and gastric cancer.

Other types of metaplasia may be seen in chronic atrophic gastritis, including pancreatic
(acinar), ciliated cell, and squamous. These are rare and of uncertain clinical
significance.

Pseudopyloric metaplasia — Pseudopyloric metaplasia or spasmolytic polypeptide-

expressing metaplasia refers to the replacement of parietal and chief cells in the oxyntic
mucosa by epithelial mucus-secreting cells of the type normally found in the antral (ie,
pyloric) mucosa [11]. These epithelial cells appear to arise from chief cells surviving in
damaged oxyntic glands, rather than from stem cells [47]. The pseudopyloric glands
have the coiled architecture seen in normal antrum, but are smaller, less numerous,
contain less mucin, and do not have antral-type endocrine cells, especially G cells.
There is evidence that pseudopyloric metaplasia is a precursor for IM [29].
Intestinal metaplasia — IM is defined by the replacement of the surface, foveolar, and
glandular epithelium in the oxyntic or antral mucosa (or both) by intestinal epithelium,
which is easily recognized by the presence of goblet cells [1,48,49]. IM can be further
divided into three subtypes. The different subtypes of IM can be distinguished, in part,
by the mucins produced by the goblet cells:
●Type I, or complete IM, is most obvious and shows fully formed small intestinal
epithelium, including eosinophilic absorptive enterocytes with a defined brush
border, goblet cells, and Paneth cells. The goblet cells contain predominately
sialomucins.
●Type II, or incomplete IM, resembles colonic epithelium, with no defined brush
border, and consists of goblet cells containing multiple irregular mucin droplets of
varying sizes interspersed among gastric-type mucin cells (picture 1). The goblet
cells in type II IM (as in type I) contain predominately sialomucins.
●Type III is also incomplete, and the interspersed goblet cells resemble colonic
epithelium and contain predominately sulfomucins rather than sialomucins.
Type III metaplasia has been linked with gastric cancer of the intestinal type in some
studies but not others [48,49]. Thus, the clinical utility of subtyping of intestinal
metaplasia is unclear, but advocated by some experts [1,50]. The risk of gastric cancer
appears to be greatest in patients with extensive IM involving the lesser curvature from
the cardia to the pylorus or the entire stomach compared with patients with more focal
or antral-predominant IM [51]. (See "Gastric intestinal metaplasia", section on 'Risk
factors'.)
Histologic staging of severity — Histopathological staging of atrophic gastritis can be
used to identify patients with advanced atrophic gastritis and aid in prediction of cancer
risks [52-54]. High stage disease (Operative Link on Gastritis Assessment [OLGA] III/IV)
is associated with a high risk of gastric cancer (table 1) [55]. Operative Link on Gastric
54
Intestinal Metaplasia Assessment (OLGIM), another proposed staging system, shows
less interobserver variability and is prognostically useful (table 2) [4,52,53]. However,
OLGIM may be less sensitive than OLGA in identifying high-risk gastritis, and some
experts recommend using a combination of OLGA and OLGIM for staging of chronic
gastritis [56].
Determining the subtype
Histologic features of AMAG and EMAG
●Autoimmune metaplastic atrophic gastritis – In patients with AMAG, the
metaplasia, glandular atrophy, and inflammation are confined to the gastric body
and fundus. In the early stages, referred to as active autoimmune gastritis, the
oxyntic mucosa is infiltrated and destroyed by lymphocytes and plasma cells. The
uneven destruction of specialized cells (ie, parietal and chief cells) within the
oxyntic glands, with preserved islands of relatively normal oxyntic mucosa, leads
to pseudopolyposis [57,58].
Histologic features of early/evolving AMAG seen on oxyntic mucosal biopsies
include at least two of the following [59]:
•Deep/full-thickness chronic inflammation
•Oxyntic gland destruction
•Prominent eosinophils
•Intestinal/pseudopyloric/pancreatic metaplasia
•Parietal cell pseudohypertrophy
It has been suggested that autoantibodies against the hydrogen-potassium
ATPase inhibit acid secretion and cause parietal cell pseudohypertrophy because
of enlargement of canaliculi [8]. This is analogous to the morphologic effect of
proton pump inhibitors (which block the hydrogen-potassium ATPase) on parietal
cells [60].
In more advanced or end-stage AMAG, gross examination of the stomach by
endoscopy or contrast radiography may demonstrate absent or inconspicuous
rugae in the gastric body and fundus [61]. In addition, the submucosal blood
vessels may be endoscopically visible through the thin, atrophic, overlying
mucosa. Atrophy and metaplasia are usually absent in the antrum in AMAG.
In end-stage AMAG, metaplastic glands completely or almost completely replace
oxyntic glands. In extreme cases, the mucosa becomes villiform, closely
resembling normal small intestine. The epithelial cells may also appear
megaloblastic in patients who have coexisting untreated PA with vitamin B12
deficiency. Nodular or linear ECL cell hyperplasia may be evident.
The proportions of pseudopyloric metaplasia versus IM within the body and fundus
vary widely among patients with AMAG, and either form can predominate. This
variability must be taken into account when obtaining and interpreting gastric
biopsy specimens. Correct anatomic localization of the biopsy to the fundus/body
or antrum requires that the endoscopist provide accurate information about the
biopsy sites. (See 'Endoscopy and biopsy' above.)
Metaplasia is usually absent or only slight in the antrum in patients with AMAG,
but changes consistent with reactive gastropathy are often present for unclear
reasons.

55
 ●Environmental metaplastic atrophic gastritis – The principal pathologic
features in EMAG are multiple, focally distributed areas of atrophy, metaplasia,
and inflammation. As a rule, these changes are most heavily concentrated in the
antrum, but the location may vary depending upon the stage of the disease.
Changes may be more evident along the lesser curvature at the junction of the
body and antrum (ie, the transitional zone) in patients with mild or early disease. In
severe or advanced disease, metaplastic epithelium can almost completely
replace the normal antral mucosa.
The progression from early to advanced stages of EMAG is often reflected by
proximal migration of the transitional zone, which can be demonstrated by
endoscopic biopsy and in resection specimens [62]. Oxyntic mucosa (ie, mucosa
of the body and fundus) adjacent to the transitional zone is progressively replaced
by intestinal and pseudopyloric metaplastic mucosa. In addition, the intact oxyntic
mucosa may be thinned and have a reduced number of parietal cells. These
changes need to be considered when trying to make a pathologic distinction
between EMAG and AMAG on gastric biopsy specimens. A white villiform pattern
may be visible endoscopically, particularly using advanced imaging modalities
such as narrow-band imaging [63].
Gastric biopsy specimens in patients without EMAG frequently show IM in a single
or a few nests of glands. We suggest a diagnosis of EMAG should not be made
from biopsy specimens unless at least 20 percent of the available antral or
transitional mucosa is replaced by metaplastic glands, or there is unequivocal
glandular atrophy. The specimen is therefore reported as showing "focal
intestinalization," and EMAG is considered only "possible" when a small proportion
of the sampled mucosa (<20 percent) is intestinalized and glandular atrophy is
lacking.
Laboratory testing — Serologic testing for both anti-intrinsic factor and antiparietal cell
antibodies and fasting gastrin levels are recommended as an adjunct to the histologic
diagnosis of AMAG. Antibodies to intrinsic factor are highly specific for AMAG but lack
sensitivity [59]. Antibodies against parietal cells have lower specificity but are
approximately 80 percent sensitive. A combination of the two tests, in conjunction with
an elevated fasting serum gastrin level, can support the diagnosis of AMAG in patients
with of early/evolving histologic features.
Low serum pepsinogen I levels or/and a low pepsinogen I/II ratio is a non-invasive test
for detecting patients with advanced stages of metaplastic (chronic) atrophic gastritis
(EMAG and AMAG) [45]. (See 'Patients with advanced atrophic gastritis' below.)
MANAGEMENT
Endoscopic surveillance in selected patients — Endoscopic surveillance in patients
with metaplastic (chronic) atrophic gastritis is controversial and has not been uniformly
recommended [45,64].
Patients with advanced atrophic gastritis — Recommendations for endoscopic
surveillance in patients with metaplastic (chronic) atrophic gastritis are based on the
severity of atrophy, extent of intestinal metaplasia, and risk factors for gastric cancer
(eg, family history of gastric cancer).
●In patients with advanced stages of atrophic gastritis (severe atrophic
changes or intestinal metaplasia in both antrum and corpus, Operative Link on
56
 Gastritis Assessment/Operative Link on Gastric Intestinal Metaplasia Assessment
III/IV) without a family history of gastric cancer, we suggest endoscopic
surveillance every three years.
●Patients with advanced stages of atrophic gastritis and with a family history of
gastric cancer may benefit from a more intensive follow-up (eg, every one to two
years after diagnosis).
For patients with mild to moderate atrophy restricted to the antrum, surveillance is not
recommended as evidence to support it are lacking [45]. Low pepsinogen I serum levels
or/and a low pepsinogen I/II ratio, particularly when associated with H. pylori-negative
serological status, may identify patients at higher risk of gastric cancer to whom
endoscopy should be offered, however the frequency of such testing is uncertain [45].
Our recommendations are consistent with 2019 European Society of Gastrointestinal
Endoscopy (ESGE) guidelines on the management of epithelial precancerous
conditions and lesions in the stomach [45]. The American Society for Gastrointestinal
Endoscopy (ASGE) 2015 guidelines do not recommend surveillance in all patients with
gastric intestinal metaplasia, but state that patients with gastric intestinal metaplasia
who are at increased risk of gastric cancer due to ethnic background or family history
may benefit from screening [64].
Patients with pernicious anemia/autoimmune metaplastic atrophic gastritis — In
patients with pernicious anemia, an upper endoscopy should be performed to identify
prevalent lesions (carcinoid tumors and gastric cancer) at the time of diagnosis and to
stage the severity of autoimmune metaplastic atrophic gastritis (AMAG). (See 'Biopsy
mapping protocol' above and 'Histologic staging of severity' above.)
While the 2015 ASGE guidelines do not recommend routine endoscopic surveillance
after an initial endoscopic evaluation, we suggest endoscopic follow-up every three to
five years in patients with AMAG [45,65]. However, the effectiveness of endoscopy in
improving outcomes for patients with pernicious anemia is unclear. Our
recommendations are consistent with 2019 ESGE guidelines [45].
Eradication of H. pylori — There is no specific treatment for metaplastic (chronic)
atrophic gastritis. The offending agent (eg, H. pylori), if identified, should be eliminated
as early as possible. Eradication of H. pylori may lead to partial regression of atrophic
gastritis. Whether eradication of H. pylori can affect the natural history of gastric IM and
its associated cancer risk is uncertain. H. pylori eradication before the development of
extensive gastric IM may be effective in reducing gastric cancer incidence, but there
may be a point of no return during the sequence of H. pylori-induced carcinogenesis,
coinciding with the presence of extensive metaplastic (chronic) atrophic gastritis [66-69].
(See "Association between Helicobacter pylori infection and gastrointestinal
malignancy" and "Gastric intestinal metaplasia", section on 'General measures in all
patients'.)
SUMMARY AND RECOMMENDATIONS
●The term metaplastic (chronic) atrophic gastritis, also referred to as gastric
atrophy, is used to describe a form of chronic gastritis that, in addition to
inflammation, is associated with mucosal thinning, loss of specialized cells in
gastric glands, and changes in epithelial cell types (ie, metaplasia). Metaplastic
(chronic) atrophic gastritis includes two main subtypes: autoimmune and
environmental metaplastic atrophic gastritis (AMAG and EMAG). Although the
57
AMAG and EMAG may be pathogenetically and clinically distinct, they often share
histologic features and may overlap clinically. (See 'Definition' above
and 'Subtypes' above.)
●AMAG is a form of metaplastic (chronic) atrophic gastritis that results in the
replacement of the normal oxyntic mucosa in the gastric corpus by atrophic and
metaplastic mucosa, leading to a corpus predominant atrophic gastritis, reduced
or absent acid and pepsin production and loss of intrinsic factor which may
progress to a severe form of vitamin B12-deficiency anemia known as pernicious
anemia (PA). Laboratory abnormalities that are associated with AMAG include
hypergastrinemia, iron deficiency anemia, antibodies to parietal cells and intrinsic
factor, and vitamin B12 deficiency. Patients with AMAG have an increased risk for
gastric neuroendocrine tumors and gastric adenocarcinoma. (See 'Autoimmune
metaplastic atrophic gastritis' above.)
●EMAG is thought to be due to the adverse effects of environmental factors, such
as H. pylori infection and perhaps dietary constituents, on the gastric mucosa.
Patients with EMAG may be asymptomatic, but many of them have dyspepsia. In
contrast to AMAG, fasting serum gastrin levels are not markedly elevated in
EMAG, and autoantibodies to parietal cell and intrinsic factor and PA are absent.
(See 'Environmental metaplastic atrophic gastritis' above.)
●The diagnosis of metaplastic (chronic) atrophic gastritis is based on the histologic
evaluation of gastric biopsies which demonstrate atrophy of the gastric mucosa
with the loss of glandular cells and their replacement by metaplastic epithelium.
Assessment of the severity of gastric atrophy and determining the subtype of
chronic atrophic gastritis requires gastric biopsy mapping with an adequate
number of biopsies from specific sites (figure 1). (See 'Diagnosis' above.)
●AMAG is confined to the gastric body and fundus. Mucosal changes in patients
with EMAG affect both the body/fundus and the antrum in a multifocal distribution,
but with heaviest involvement of the antrum. Serologic testing for both anti-intrinsic
factor and antiparietal cell antibodies and fasting gastrin levels should be
performed as an adjunct to the histologic diagnosis of AMAG. (See 'Determining
the subtype' above.)
●There is no specific treatment for metaplastic (chronic) atrophic gastritis. The
offending agent (eg, H. pylori), if identified, should be eliminated as early as
possible. (See 'Eradication of H. pylori' above.)
●The risk of gastric cancer in patients with metaplastic (chronic) atrophic gastritis
is uncertain and probably does not warrant routine surveillance in patients with
mild to moderate atrophy restricted to the antrum. Exceptions include patients with
advanced stages of atrophic gastritis (Operative Link on Gastritis Assessment
/Operative Link on Gastric Intestinal Metaplasia Assessment III/IV) and those with
PA. (See 'Endoscopic surveillance in selected patients' above.)

58
Barrett's esophagus: Epidemiology, clinical
manifestations, and diagnosis
Author:
Stuart J Spechler, MD
Section Editor:
Nicholas J Talley, MD, PhD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Oct 19, 2021.
INTRODUCTIONBarrett's esophagus is the condition in which a metaplastic
columnar epithelium that has both gastric and intestinal features replaces the stratified
squamous epithelium that normally lines the distal esophagus. The condition develops
as a consequence of chronic gastroesophageal reflux disease (GERD) and predisposes
to the development of adenocarcinoma of the esophagus.
This topic will review the clinical manifestations and diagnosis of Barrett's esophagus.
The pathogenesis, malignant transformation, and management of Barrett's esophagus,
including surveillance for adenocarcinoma, are discussed separately. (See "Barrett's
esophagus: Pathogenesis and malignant transformation" and "Barrett's esophagus:
Surveillance and management".)
EPIDEMIOLOGY
Prevalence — Barrett's esophagus is usually discovered during endoscopic
examinations of middle-aged and older adults; the large majority of cases go
unrecognized [1]. The mean age at diagnosis of Barrett's esophagus is approximately
55 years [2]. Barrett's esophagus is uncommon in children in general and extremely rare
in children under the age of five [3]. Barrett's esophagus is two- to threefold more
common in men than in women [4].
In the United States it is estimated that as many as 5.6 percent of adults have Barrett's
esophagus. However, estimates of the prevalence of Barrett's esophagus in the general
population have varied widely ranging from 0.4 to more than 20 percent depending, in
part, upon the population studied and the criteria used to establish the diagnosis [5-9].
In studies performed in the United States, Barrett's esophagus appears to have a higher
prevalence in White individuals as compared with individuals who identify as being of
Hispanic descent or Asian descent, and prevalence appears to be lowest in Black
individuals [10,11]. The prevalence of short-segment Barrett's esophagus is
substantially higher than long-segment Barrett's esophagus [12,13]. Both conditions are
diagnosed more frequently in patients age 50 years and older. In a study that included
889 patients undergoing upper endoscopy who had protocol biopsies obtained at the
esophagogastric junction [5], the overall prevalence of specialized intestinal metaplasia
was 13.2 percent. The prevalence of long-segment and short-segment Barrett's were
6.4 and 1.6 percent, respectively. Intestinal metaplasia was limited to the
gastroesophageal junction (GEJ) in 5.6 percent. (See 'Diagnosis' below and 'Intestinal
metaplasia at GEJ' below.)

59
Although Barrett's esophagus traditionally has been considered uncommon in Asian
countries [14], a meta-analysis of 51 studies that included 453,157 individuals in Asia
revealed a surprisingly high pooled prevalence of histologically confirmed Barrett's
esophagus (1.3 percent), which was short-segment in 82 percent of cases and may
reflect misclassification error in some of the studies [15].
Risk factors
●Gastroesophageal reflux disease (GERD) – In patients with symptomatic
GERD, erosive esophagitis is an independent risk factor for Barrett's esophagus,
conferring a fivefold increased risk of Barrett's at five-year follow-up (relative risk
ratio [RRR] 5.2, 95% CI 1.2-22.9) [16]. Some studies have suggested that patients
with a peptic stricture have a higher prevalence of Barrett's esophagus than those
without strictures. This relationship is not surprising since both peptic stricture and
Barrett's esophagus are associated with more severe GERD. However, this
association has been challenged in a study of patients referred for endoscopy for
GERD in whom the prevalence of intestinal metaplasia was the same in patients
with and without strictures [17].
●Central (abdominal) obesity – Central obesity is a risk factor for GERD and for
Barrett's esophagus [18,19]. A 2009 meta-analysis that included 11 observational
studies demonstrated a small increase in the risk of Barrett's esophagus in
patients with a body mass index (BMI) >30 kg/m2 as compared with patients with a
BMI <30 kg/m2 (odds ratio [OR] 1.4, 95% CI 1.1-1.6) [18]. However, other studies
have suggested that rather than BMI, abdominal obesity as measured by a high
waist-to-hip ratio (≥0.9 in males and ≥0.85 in females) is especially associated
with an increase in risk of Barrett's esophagus [19-22].
●Family history – Familial aggregation of Barrett's esophagus and esophageal
adenocarcinoma has been described, and Barrett's esophagus has been found in
up to 28 percent of first-degree relatives of patients with esophageal
adenocarcinoma [23-25]. It is unclear if this is due to common environmental
exposures and/or an inherited predisposition. Germline mutations in
the MSR1, ASCC1, and CTHRC1 genes have been associated with the presence
of Barrett's esophagus and esophageal adenocarcinoma [26], and one large
genome-wide association study identified genetic variants associated with
Barrett's at chromosomes 6p21 and 16q24 [27]. Large cohort studies are needed
to validate these findings [28].
●Smoking – Smoking appears to have a synergistic effect with GERD in
increasing the risk of Barrett's esophagus. In a pooled analysis of five population-
based case-control studies from the International BEACON (Barrett's Esophagus
and Esophageal Adenocarcinoma Consortium), the risk of Barrett's esophagus
was 1.7 times greater in smokers than in nonsmokers without GERD, and was 1.6
times greater than in nonsmokers with GERD [29].
Cancer risk — Esophageal adenocarcinoma in patients with Barrett's esophagus is
thought to evolve through a sequence of genetic and epigenetic alterations that are
associated with dysplastic changes of progressive severity. The incidence of cancer
arising in Barrett's esophagus is described separately. (See "Barrett's esophagus:
Pathogenesis and malignant transformation", section on 'Malignant

60
transformation' and "Barrett's esophagus: Surveillance and management", section on
'Cancer risk'.)
CLINICAL FEATURESThe specialized intestinal columnar metaplasia typical of
Barrett's esophagus causes no symptoms. Most patients found to have Barrett's
esophagus are seen initially for symptoms of associated gastroesophageal reflux
disease (GERD), such as heartburn or regurgitation. Patients with complications
associated with Barrett's esophagus may have dysphagia or odynophagia from
esophageal ulceration or stricture, and rarely gastrointestinal bleeding secondary to
ulceration (picture 1) [2]. Although patients with long-segment Barrett's esophagus
frequently have prominent GERD symptoms, short-segment Barrett's esophagus is not
associated with GERD symptoms [30]. (See "Clinical manifestations and diagnosis of
gastroesophageal reflux in adults".)
SCREENING PATIENTS FOR BARRETT'S ESOPHAGUS
Whom to screen — We screen for Barrett's esophagus in patients with multiple risk
factors for adenocarcinoma. These include a hiatal hernia, age ≥50, male sex, chronic
gastroesophageal reflux disease (GERD), White individuals, central obesity, cigarette
smoking, and a confirmed history of Barrett's esophagus or esophageal
adenocarcinoma in a first-degree relative [31,32]. In patients with erosive esophagitis
found on the initial examination, we perform repeat endoscopy after a three-month
course of acid suppression to exclude the presence of Barrett's esophagus.
In patients with a negative screening upper endoscopy for Barrett's esophagus, a
routine follow-up upper endoscopy for screening for Barrett's esophagus is not indicated
[31,33,34]. In a retrospective study of 24,406 patients with GERD who had an initial
negative screening exam for Barrett's esophagus, only 2.3 percent had suspected
Barrett's esophagus at the second upper endoscopy performed within five years [35].
Our recommendations are largely consistent with the American Gastroenterological
Association (AGA) and American College of Gastroenterology (ACG) guidelines.
Multiple societies have issued guidelines regarding screening patients for Barrett's
esophagus [31,33,34,36]. It is important to appreciate that these recommendations
represent the opinions of committees, that the recommendations of different societies
vary, and that none of these approaches have been established effective in prospective
clinical studies. (See 'Society guideline links' below.)
Upper endoscopy with biopsy — The sensitivity of endoscopy for detection of
Barrett's esophagus is related to the length of involved mucosa, with detection being
more likely in patients who have long-segment Barrett's esophagus [37]. In one study of
146 patients suspected of having Barrett's esophagus on endoscopic examination, the
sensitivity and specificity of the endoscopists' diagnosis of Barrett's esophagus (using
biopsy confirmation of the endoscopic diagnosis as the gold standard) were 82 percent
(95% CI 72-92) and 81 percent (95% CI 78-84), respectively, and endoscopists
diagnosed long-segment more accurately than short-segment Barrett's esophagus (55
percent versus 25 percent) [37]. In a study that included 116 patients who were involved
in a Veterans Administration Cooperative Study that required two endoscopies six
weeks apart [38], Barrett's esophagus was confirmed by biopsy in only one of the two
endoscopies in 20 percent of patients.
Endoscopic technique

61
Identify possible Barrett's — Barrett's esophagus should be suspected when the
squamocolumnar junction is located ≥1 cm proximal to the gastroesophageal junction
(GEJ) (figure 1). On white light endoscopy, Barrett's esophagus has a characteristic
salmon-colored appearance.
●Squamocolumnar junction (Z-line) – Columnar epithelium has a reddish color
and velvet-like texture on endoscopic examination, whereas squamous epithelium
has a pale, glossy appearance (picture 2). The juxtaposition of these epithelia at
the squamocolumnar junction forms a visible line called the Z-line.
●Gastroesophageal junction (GEJ) – The GEJ is the imaginary level at which
the esophagus ends and the stomach begins anatomically. Endoscopists in the
Western hemisphere typically identify the GEJ as the level of the most proximal
extent of the gastric folds [31]. Air in the stomach should be aspirated during
evaluation of the GEJ as air can flatten gastric folds. Occasionally, the esophagus
can also form longitudinal folds and these folds should not be confused with
gastric rugal folds.
In some Asian countries (particularly Japan), the GEJ is defined as the distal extent of
the lower esophageal palisade vessels (rather than the upper end of the gastric
longitudinal folds). Palisade vessels are identified endoscopically as fine, longitudinally-
oriented blood vessels in the distal esophagus. However, one study found that using the
palisade vessels had lower interobserver reliability than using the gastric folds as a
landmark [39]. This method likely overcalls ultrashort-segment Barrett's esophagus.
Grade the extent — An endoscopic grading system (The Prague C & M Criteria) has
been recommended for reporting the extent of Barrett's esophagus [31,40].
Measurements should be obtained during withdrawal of the endoscope when the shaft
of the endoscope is straight. Distances assessed using centimeter measurements on
the shaft of the instrument can be used to measure the distance of landmarks from the
incisors more precisely.
●Circumferential extent (C value) – Depth of endoscope insertion at the GEJ –
Depth of endoscope insertion from the circumferential extent of suspected
columnar epithelium.
●Maximum extent (M value) – Depth of endoscope insertion at the GEJ – Depth
of endoscope insertion from the maximal extent of columnar epithelium.
The Prague criteria have good interobserver reliability for Barrett's esophagus involving
>1 cm of the distal esophagus [40,41].
Biopsy technique — Biopsies of the suspected Barrett's segment are needed to
establish the diagnosis [42]. At least four biopsies should be obtained for every 2-cm
segment of suspected Barrett's esophagus.
Biopsy specimens should not be obtained from a normal-appearing Z-line or irregular Z-
line that extends <1 cm up the esophagus. If such biopsies are taken and identify
intestinal metaplasia at the GEJ, there is yet no clear consensus on how to manage
those patients, but their cancer risk generally is considered to be minimal.
(See 'Histologic features' below and 'Intestinal metaplasia at GEJ' below.)
Histologic features — Traditionally, three types of columnar lining have been
described in Barrett's esophagus [43]:
Gastric cardia type mucosa — Cardiac mucosa (also called cardia-type or junctional
mucosa), which has a foveolar (pitted) surface and glands that are lined exclusively by

62
mucus-secreting cells; these cells resemble normal gastric foveolar cells. Cardiac-type
mucosa at the GEJ can be an abnormal, metaplastic lining acquired as a consequence
of chronic inflammation in the distal esophagus caused by GERD. Gastric cardiac-type
mucosa is characterized by columnar lining characterized by tortuous, tubular glands
comprised exclusively of mucus-secreting cells and devoid of acid-secreting parietal
cells [44,45]. While it was traditionally thought that cardiac-type mucosa can line up to 2
cm of the most distal esophagus, and may extend several centimeters below the GEJ to
line the most proximal stomach (the gastric cardia), it is now known to usually extend
only 1 or 2 mm beyond the Z-line [45,46]. Furthermore, some authorities now contend
that the normal Z-line is a junction between esophageal squamous and gastric fundic
(oxyntic) epithelia [47,48].
Studies also suggest that cardiac mucosa in the esophagus is not only often
metaplastic, but also the precursor of specialized intestinal metaplasia in the
esophagus. A study of 40 patients who had subtotal esophagectomy with
esophagogastrostomy, an operation frequently complicated by severe reflux esophagitis
in the esophageal remnant, supports the notion that cardiac-type mucosa is metaplastic
[49]. Endoscopic examinations performed at a median of 36 months postoperatively
showed that 19 of the 40 patients had developed columnar metaplasia in the
esophageal remnant (10 cardiac-type, 9 intestinal metaplasia). Seven patients who had
serial endoscopic examinations showed progression from cardiac-type mucosa on the
initial postoperative endoscopy to specialized intestinal metaplasia on subsequent
studies. The median time to the development of cardiac-type mucosa was 14 months,
whereas specialized intestinal metaplasia was found at a median of 27 months
postoperatively.
Atrophic gastric fundic-type epithelium — Atrophic gastric fundic-type epithelium
(also called oxyntocardiac epithelium), which has a foveolar surface lined by mucus-
secreting cells, and a deeper glandular layer that contains chief and parietal cells; these
cells resemble those in the gastric fundus.
Specialized intestinal metaplasia — Specialized intestinal metaplasia (also called
specialized columnar epithelium or simply intestinal metaplasia) comprises a number of
columnar cell types including goblet cells, gastric foveolar-type cells, small intestinal-like
cells, and colonic-like columnar cells (picture 3A) [50]. Endocrine and Paneth cells have
also been described (picture 4) [51]. It is the most common histologic type found in
Barrett's esophagus, and the only one that has a clear malignant potential [13].
Specialized intestinal metaplasia can be indistinguishable histologically from intestinal
metaplasia type II or III of the stomach. However, the presence of goblet cells is the
most useful feature for distinguishing specialized intestinal metaplasia from cardiac
mucosa. The goblet cells of specialized intestinal metaplasia contain acidic mucins
(sialomucins and sulfomucins) that can be demonstrated by staining with Alcian blue
(picture 3A-B). They may also contain colonic-like mucins that can be demonstrated
with high-iron diamine staining (picture 5). (See 'Criteria for Barrett's esophagus' below
and 'Gastric cardia type mucosa' above.)
Other methods
Novel endoscopic methods — Many endoscopic techniques have been used to
enhance identification of Barrett's metaplasia, including magnification endoscopy,
chromoendoscopy, optical coherence tomography, electronic chromoendoscopy (eg,

63
narrow band imaging), and autofluorescence endoscopy. Routine use of advanced
imaging techniques other than electronic chromoendoscopy generally is not
recommended. (See "Barrett's esophagus: Surveillance and
management" and "Barrett's esophagus: Evaluation with optical
chromoscopy" and "Optical coherence tomography in the gastrointestinal tract", section
on 'Barrett's esophagus' and "Chromoendoscopy", section on 'Barrett's
esophagus' and "Confocal laser endomicroscopy and endocytoscopy", section on
'Barrett's esophagus'.)
Nonendoscopic methods — Nonendoscopic methods for screening are also being
studied [52]. One method uses a device called a capsule sponge (Cytosponge,
SurePath; BD Diagnostics, Durham, NC) combined with an immunohistochemical
biomarker (trefoil factor 3) [53]. The patient ingests a gelatin capsule that is attached to
a string and contains a compressed mesh. The mesh is exposed when the gelatin
capsule dissolves in the stomach. The mesh is then withdrawn through the esophagus
where it collects samples of the cells lining the esophageal lumen. The biomarker is
then used to differentiate Barrett's epithelial cells from gastric columnar and esophageal
squamous cells.
In a study of 504 patients who had used acid-suppressing therapy for more than three
months during the preceding five years, 501 (99 percent) were able to swallow the
capsule [53]. The results obtained with the capsule sponge were compared with upper
endoscopy for diagnosing Barrett's esophagus. The capsule sponge had a sensitivity of
73 percent and a specificity of 94 percent for patients with at least 1 cm of
circumferential Barrett's esophagus. For patients with segments of 2 cm or more, the
sensitivity was 90 percent and the specificity was 94 percent. Cytosponge-trefoil factor 3
(TFF3) appears to improve the detection of Barrett's, however, it has a high false-
positive rate. In a randomized trial, 13,514 individuals with GERD were assigned to
usual care with upper endoscopy for detection of Barrett's if deemed necessary by their
clinician or intervention (Cytosponge-TFF3 procedure with subsequent upper
endoscopy if positive) [54]. Of the 6834 individuals in the intervention group, 1654
underwent the Cytosponge procedure and 231 (3 percent) tested positive for TFF3. Of
these, 221 individuals underwent endoscopy for positive TFF3 and 131 (59 percent) had
Barrett's esophagus or esophageal cancer. During a mean follow-up of 12 months, rates
of detection of Barrett's were significantly higher in patients in the intervention group as
compared with usual care (absolute difference 18·3 per 1000 person-years [95% CI
14·8-21·8]; rate ratio 10·6 [95% CI 6·0-18·8]). In addition, nine (<1 percent) participants
were diagnosed with dysplastic Barrett's esophagus or esophageal cancer in the
intervention group, but there were no cases in the usual care group. Further studies are
needed to determine which patient groups may benefit from screening and the cost-
effectiveness of testing.
Limitations of screening — Long-segment Barrett's esophagus can be found in 3 to 5
percent of patients who have endoscopy for chronic GERD symptoms, and 10 to 15
percent have short-segment Barrett's esophagus [13]. However, it is not clear that
screening patients with GERD symptoms reliably identifies all individuals at high risk for
esophageal adenocarcinoma [55,56]. In addition, it is estimated that more than 40
percent of patients with esophageal adenocarcinoma have no history of heartburn and
would not have qualified for screening.

64
Although patients with GERD symptoms are at increased risk for esophageal
adenocarcinoma, the absolute risk of esophageal adenocarcinoma is low and it is not
clear that screening has an impact on mortality [32,57].
It is also unclear that patients who are known to have Barrett's esophagus benefit from
surveillance and, once the diagnosis of Barrett's esophagus has been established,
patients are subject to the worry about the diagnosis, the inconvenience and risk
associated with surveillance and, potentially, a financial burden from an increase in life
insurance premiums [58]. (See "Barrett's esophagus: Surveillance and management".)
DIAGNOSISThe diagnosis of Barrett's is generally established by upper endoscopy
and biopsy [13]. (See 'Upper endoscopy with biopsy' above.)
Criteria for Barrett's esophagus — The criteria for the diagnosis of Barrett's
esophagus vary worldwide. In the United States, the diagnosis of Barrett's esophagus
requires both of the following to be present:
●Columnar epithelium lining ≥1 cm of the distal esophagus.
●Histologic examination of biopsy specimens from that columnar epithelium must
reveal intestinal metaplasia characterized with goblet cells.
In contrast, the British Society of Gastroenterology (BSG) also requires histologic proof
of metaplastic esophageal columnar mucosa, but unlike United States guidelines, the
BSG considers the finding of cardiac or oxyntocardiac mucosae (which do not contain
goblet cells) to be adequate for a diagnosis of Barrett's esophagus. Some data suggest
that gastric cardiac-type epithelium in the esophagus also might predispose to cancer
and thus might be considered "Barrett's esophagus," but most authorities still require
the presence of intestinal metaplasia for an unequivocal diagnosis [31,59]. (See 'Society
guideline links' below and 'Gastric cardia type mucosa' above.)
Barrett's esophagus that is ≥3 cm in length is termed long-segment Barrett's, whereas a
segment that is less than 3 cm is termed short-segment Barrett's esophagus [60].
Differential diagnosis
Intestinal metaplasia at GEJ — If the Z-line and the gastroesophageal junction (GEJ)
coincide and biopsy specimens at the Z-line show intestinal metaplasia, the condition is
called intestinal metaplasia at the GEJ [44].
Biopsy mapping of the body and antrum of the stomach can be useful to establish
whether intestinal metaplasia at the GEJ is merely the proximal extension of a diffuse
Helicobacter pylori (H. pylori) atrophic gastritis. Intestinal metaplasia can develop in the
stomach as a consequence of chronic H. pylori gastritis. Histologically, intestinal
metaplasia in the stomach can be indistinguishable from intestinal metaplasia in the
esophagus. Since the GEJ cannot be identified with great precision, it can be difficult to
determine whether short segments of intestinal metaplasia found in the GEJ region are
lining the esophagus (ie, short-segment Barrett's esophagus) or the proximal
stomach. H. pylori infection rarely, if ever, causes intestinal metaplasia confined to the
gastric cardia. Consequently, biopsy mapping of the body and antrum of the stomach
can be useful to establish whether intestinal metaplasia at the GEJ is merely the
proximal extension of a diffuse H. pylori atrophic gastritis. Without such mapping,
however, it can be difficult to determine whether intestinal metaplasia at the GEJ
represents an ultrashort segment of esophageal metaplasia or gastric intestinal
metaplasia due to atrophic gastritis. The morphologic and histochemical features of
gastric and esophageal intestinal metaplasia are similar, and the gross landmarks used
65
to identify the GEJ do not have the precision necessary to localize a mucosa whose
extent may be measured in only millimeters. In patients in Western countries, however,
intestinal metaplasia at the GEJ usually is associated with GERD rather than with H.
pylori infection [61].
Limited available data suggest that the esophageal cancer risk imposed by intestinal
metaplasia at the GEJ in Western patients is minimal. In a population-based cohort
study, subjects with intestinal metaplasia at the GEJ had substantially lower rates of
progression to esophageal adenocarcinoma than those with clear-cut Barrett's
esophagus (0 percent compared with 7 percent at 10 years) [62]. It seems likely that
this low risk of cancer progression is because intestinal metaplasia at the GEJ in
Western patients represent ultrashort-segment Barrett's esophagus rather than the
extension of diffuse atrophic gastritis.
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Barrett's esophagus".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Barrett's esophagus (The

Basics)" and "Patient education: Upper endoscopy (The Basics)")
●Beyond the Basics topics (see "Patient education: Barrett's esophagus (Beyond
the Basics)" and "Patient education: Upper endoscopy (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Barrett's esophagus is the condition in which a metaplastic columnar epithelium
that has both gastric and intestinal features replaces the stratified squamous
epithelium that normally lines the distal esophagus. The condition develops as a
consequence of chronic gastroesophageal reflux disease (GERD) and
predisposes to the development of adenocarcinoma of the esophagus.
(See 'Introduction' above.)
●Barrett's esophagus is usually discovered during endoscopic examinations of
middle-aged and older adults; the large majority of cases go unrecognized. The
mean age at diagnosis of Barrett's esophagus is approximately 55 years.
(See 'Epidemiology' above.)

66
●The specialized intestinal columnar metaplasia typical of Barrett's esophagus
causes no symptoms. Most patients found to have Barrett's esophagus are seen
initially for symptoms of associated GERD, such as heartburn or regurgitation.
Patients with complications associated with Barrett's esophagus may have
dysphagia or odynophagia from esophageal ulceration or stricture, and rarely
gastrointestinal bleeding secondary to ulceration. (See 'Clinical features' above.)
●Biopsies of the suspected Barrett's segment are needed to establish the
diagnosis. At least four biopsies should be obtained for every 2-cm segment of
suspected Barrett's esophagus. Biopsy specimens should not be obtained from a
normal-appearing Z-line or irregular Z-line that extends <1 cm up the esophagus.
If such biopsies are taken and identify intestinal metaplasia at the GEJ, there is
yet no clear consensus on how to manage those patients, but their cancer risk
appears to be minimal. (See 'Biopsy technique' above.)
●We screen for Barrett's esophagus in patients with multiple risk factors for
adenocarcinoma. These include a hiatal hernia, age ≥50, male sex, chronic
GERD, White individuals, central obesity, cigarette smoking, and a confirmed
history of Barrett's esophagus or esophageal adenocarcinoma in a first-degree
relative. However, the evidence to support screening for Barrett's is weak, and the
decision on when to recommend endoscopic screening should be individualized.
(See 'Screening patients for Barrett's esophagus' above.)
●The criteria for the diagnosis of Barrett's esophagus vary worldwide. In the United
States, the diagnosis of Barrett's esophagus requires both of the following to be
present:
•Columnar epithelium lining ≥1 cm of the distal esophagus.
•Histologic examination of biopsy specimens from that columnar epithelium
must reveal intestinal metaplasia characterized with goblet cells.
In contrast, the British Society of Gastroenterology (BSG) also requires histologic
proof of metaplastic esophageal columnar mucosa, but unlike United States
guidelines, the BSG considers the finding of cardiac or oxyntocardiac mucosae
(which do not contain goblet cells) to be adequate for a diagnosis of Barrett's
esophagus. (See 'Diagnosis' above.)

67
Helicobacter pylori and gastroesophageal reflux
disease
Authors:
John E Pandolfino, MD
Peter J Kahrilas, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jul 21, 2021.
INTRODUCTIONHelicobacter pylori (H. pylori) is an important risk factor for the
development of peptic ulcer disease, gastric adenocarcinoma, and primary B cell
lymphoma of the stomach. A possible role for H. pylori in the pathogenesis of
gastroesophageal reflux disease (GERD) has also been suggested in a growing number
of studies. However, the link between GERD and H. pylori is complex.
This topic review will summarize the available evidence suggesting a role for H. pylori in
GERD. The pathophysiology of GERD and the treatment of H. pylori are discussed
separately. (See "Pathophysiology of reflux esophagitis" and "Treatment regimens for
Helicobacter pylori in adults".)
EPIDEMIOLOGYSuspicion of an interaction between H. pylori and
gastroesophageal reflux disease (GERD) stems from epidemiologic data showing that
as the prevalence of H. pylori decreased in Western societies, the prevalence of GERD
and adenocarcinoma of the esophagus increased [1]. This trend led several
investigators to examine the prevalence of H. pylori in patients with GERD.
Several reports have suggested that H. pylori-positive patients were less likely to have
GERD, and, when present, the severity of esophagitis was decreased compared with
those who were H. pylori negative [2-4]. A lower prevalence of Barrett's metaplasia and
esophageal adenocarcinoma has also been described in individuals who were H.
pylori positive [5,6].
Some studies suggested that H. pylori strains positive for Cag A (strains strongly
associated with the development of corpus gastritis) may be particularly protective
against the development of esophageal adenocarcinoma [7-9]. Thus, the available data
suggest that colonization with H. pylori, particularly Cag A strains, may be protective
against the more severe forms of GERD. Unfortunately, Cag A positive strains have
also been associated with gastric adenocarcinoma. (See "Association between
Helicobacter pylori infection and gastrointestinal malignancy".)
PATHOPHYSIOLOGY OF GERD AS IT RELATES TO H.
PYLORIGastroesophageal reflux disease (GERD) is a clinical condition that results
from the reflux of caustic fluid from the stomach into the esophagus. The
pathophysiology of GERD is multifactorial with the disease ultimately related to the
balance between factors tending to damage (or sensitize) the esophageal mucosa
(excessive mucosal exposure time to caustic refluxate) and those tending to preserve it
68
(a competent esophagogastric junction [EGJ] and normal esophageal acid clearance).
GERD occurs when the balance is tipped in favor of the caustic factors.
(See "Pathophysiology of reflux esophagitis".)
For H. pylori to exert a direct pathogenetic effect on GERD, it must have an impact on
one of these primary disease determinants. However:
●H. pylori has no apparent effect on EGJ competence [3].
●There are no data suggesting that H. pylori decreases lower esophageal
sphincter (LES) pressure or the frequency of transient LES relaxations.
●Esophageal peristaltic function and acid clearance are unlikely to be affected
by H. pylori.
●There have been reports of H. pylori colonizing the esophageal mucosa. Whether
such colonization affects mucosal sensitivity has not been determined.
The primary mechanism by which H. pylori might predispose to GERD is by modifying
the gastric refluxate. However, the effects appear to be different in antral-dominant and
corpus-dominant or pangastritis (figure 1).
Antrum-dominant gastritis — H. pylori infection has a major effect on somatostatin
secreting D cells in the gastric antrum such that feedback inhibition by luminal acid on
gastrin release is interrupted. As a result, gastrin levels are higher in H. pylori-infected
individuals [10]. (See "Association between Helicobacter pylori infection and duodenal
ulcer".)
Thus, lack of feedback inhibition may be responsible for the increased acid secretion
found in patients with duodenal ulcers who are H. pylori positive and have antral-
dominant gastritis. Another theory suggests that eradication of H. pylori restores gastric
bicarbonate secretion to levels seen in patients without H. pylori infection without
affecting gastric acid secretion [11]. The net effect is the same — increased acidity of
gastric secretions.
Corpus-dominant and pangastritis — H. pylori infection associated with corpus
(fundus)-dominant gastritis is also associated with increased gastrin levels, but acid
secretion is reduced. It has been proposed that corpus-dominant gastritis decreases
acid secretion via local inflammation and increased levels of cytokines, such as tumor
necrosis factor alpha and interleukin-1-beta. These changes can eventually lead to
hypochlorhydria and gastric atrophy.
Reversal of H. pylori-induced corpus-dominant gastritis (and associated
hypochlorhydria) has the potential to increase gastric acid secretion, which could render
caustic what was previously asymptomatic reflux [11]. Return of acid secretion shortly
after successful eradication of H. pylori has been associated with increased expression
of H/K ATPase mRNA, not increased parietal cell number [12]. Whether return of
parietal cell numbers increases acid secretion in the long-term is unclear.
H. PYLORI ERADICATION AND GASTROESOPHAGEAL REFLUX
DISEASEStudies on the effect of H. pylori eradication and gastroesophageal reflux
disease (GERD) have been conflicting. The disparities can be explained by an
understanding of the variable effects of H. pylori on gastric acid secretion discussed
above. Most of the initial studies failed to differentiate patients on the basis of gastritis
severity or distribution. In addition, the period of observation in these studies may have
been too short (approximately six months) to allow for physiological equilibration to
occur. Bearing these limitations in mind, the following sections will summarize available
69
data on the effect of H. pylori eradication on GERD in specific patient groups: duodenal
ulcer, corpus-dominant gastritis, and GERD. (See "Treatment regimens for Helicobacter
pylori in adults".)
Patients with duodenal ulcers — Chronic H. pylori infection is the most important risk
factor predisposing to duodenal ulcer disease. The lifetime prevalence of peptic ulcer
disease in H. pylori-infected patients is approximately 5 to 10 percent. (See "Association
between Helicobacter pylori infection and duodenal ulcer".)
Because antrum-dominant H. pylori gastritis is associated with increased gastric acid
secretion, such patients should be at increased risk for developing GERD and duodenal
ulcer disease, and both conditions should improve with H. pylori eradication. Several
studies addressing this issue have produced variable results, which may in part be
explained in differences in study populations and endpoints measured [13-18].
Considered together, they suggest a modest benefit of H. pylori eradication on GERD-
symptoms in patients with duodenal ulcer disease and preexisting GERD. (See 'Antrum-
dominant gastritis' above.)

The following illustrates the range of findings in the largest studies:

●An analysis of eight double-blind prospective trials of H. pylori therapy in a total of

1165 patients with duodenal ulcer disease (940 active, 225 past history) found that
eradication of H. pylori did not lead to the development of erosive esophagitis or
new symptomatic GERD, but may improve symptoms (heartburn, regurgitation) in
patients with preexisting GERD [17].
●Similar conclusions were reached in a systematic review of 27 studies, which
found no evidence that H. pylori eradication in patients with duodenal ulcer
disease provoked reflux esophagitis or worsened heartburn [18].
Patients with corpus-dominant or pangastritis — Significant corpus gastritis can be
associated with decreased acid secretion. In such patients, H. pylori eradication may
result in increased acid secretion, which could be clinically relevant in patients
predisposed to GERD. Several investigators have observed that GERD was less
common in patients with severe corpus gastritis [13,19,20], while others have
demonstrated that improvement of gastritis in these patients has been associated with
an increased risk for the development of GERD [21,22].
It is important to emphasize again that H. pylori eradication does not cause GERD, it
only unmasks it. Patients with impaired defense mechanisms, such as impaired
esophageal acid clearance, hiatal hernia, and a hypotensive LES, are predisposed to
GERD but may be protected by the decreased acid secretion induced by the
inflammation and atrophy caused by H. pylori gastritis. This was demonstrated by a
study evaluating risk factors for incident esophagitis after H. pylori eradication [21]; after
eradication, the cumulative prevalence of esophagitis was significantly higher among
those with a hiatus hernia (26 versus 8 percent), a known risk factor for GERD.
Patients with GERD — Studying the effects of H. pylori eradication in patients with
GERD is difficult because these patients necessarily have preserved acid secretion and
EGJ compromise. Because they have preserved acid secretion, it is likely that these
patients have either antral-dominant gastritis or only mild corpus gastritis. Regardless,
eradication should not significantly worsen acid reflux in patients with preexisting GERD

70
and should improve reflux in patients with antral-dominant gastritis. At least three
studies support this hypothesis [23-25].
Effect of H. pylori on PPI therapy — H. pylori infection affects gastric acid secretion in
patients treated with proton pump inhibitors (PPIs). For example, inhibition of gastric
acid secretion with a PPI is greater in patients with H. pylori infection compared to H.
pylori-negative subjects [26,27]. Studies of patients with duodenal ulcers have also
demonstrated that the acid suppressing ability of PPIs is augmented in patients who are
infected with H. pylori [26,28].
Whether or not the demonstrated effect of H. pylori on acid suppression with PPIs is
clinically important has yet to be determined. In one study, PPIs appeared to be more
effective in preventing and curing ulcers in patients who are H. pylori positive with or
without NSAID use [29]. However, in a controlled trial comparing the efficacy
of esomeprazole with lansoprazole, healing rates with PPIs were not influenced by H.
pylori status [30]. In combined analysis of three large controlled trials of short- and long-
term treatment in patients with reflux disease, H. pylori was a weak independent
protective factor against relapse [31]. In another open-label study that included 971
patients with symptomatic reflux and esophagitis treated with pantoprazole, relief from
heartburn and regurgitation and the likelihood of endoscopic healing after four weeks of
therapy was significantly greater in H. pylori-positive patients compared with H. pylori-
negative patients (87 versus 76 percent) [32]. Long-term studies evaluating
maintenance therapy for GERD found no difference in the PPI dose required for H.
pylori-negative and positive patients [33,34]. Thus, it appears that although H.
pylori increases the acid suppressive effect of PPIs, there is no evidence suggesting a
need to titrate the dose according to the presence of H. pylori.
Another issue that may be of clinical relevance is acid rebound after discontinuation of
PPIs. Studies suggest that it may be more difficult to withdraw PPI therapy in H.
pylori negative patients (see "Physiology of gastric acid secretion"). The potential
significance of this effect was illustrated in a study in which basal acid output increased
by 82 percent in H. pylori-negative patients and decreased by 32 percent in H. pylori-
positive patients at day 15 after omeprazole therapy was discontinued [35]. The authors
speculated that this phenomenon was the result of persistence of corpus gastritis in H.
pylori-positive patients.
Theoretical risk of long-term PPI therapy — A major concern regarding long-term
PPI therapy has been the suggestion that patients infected by H. pylori are at increased
risk for the development of atrophic gastritis during long-term therapy with PPIs [36]. A
US Food and Drug Administration panel reviewing the data supporting this contention
concluded that the findings were flawed and inconclusive. (See "Medical management
of gastroesophageal reflux disease in adults".)
However, at least two controlled trials have demonstrated that eradication of H. pylori in
patients with reflux esophagitis receiving long-term acid suppression therapy decreases
inflammation and reverses corpus gastritis [37,38]. As a result, European consensus
guidelines suggest that H. pylori testing be performed in patients receiving long-term
acid suppression with PPIs [39].
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
71
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Acid reflux and gastroesophageal reflux

disease in adults (The Basics)" and "Patient education: H. pylori infection (The
Basics)")
●Beyond the Basics topics (see "Patient education: Gastroesophageal reflux
disease in adults (Beyond the Basics)" and "Patient education: Helicobacter pylori
infection and treatment (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Eradication of H. pylori is associated with mild worsening of gastroesophageal
reflux disease (GERD) in patients with corpus-dominant or pangastritis and
improvement in those with antral-dominant gastritis. These effects must be
balanced against the risks of continued infection (figure 1). (See 'Pathophysiology
of GERD as it relates to H. pylori' above.)
●For H. pylori diagnosed in the context of peptic ulcer disease, eradication is the
standard of care. Failure to eradicate H. pylori in the context of peptic ulcer
disease is associated with a 60 to 100 percent annual ulcer recurrence rate
compared with 10 percent after eradication. (See 'Patients with duodenal
ulcers' above.)
●For H. pylori diagnosed outside the context of peptic ulcer disease, most
authorities would advocate eradication based upon a 10 percent lifetime risk of
developing peptic ulcer disease and a two to three times higher incidence of
gastric adenocarcinoma. The mild increase in reflux associated with eliminating
corpus gastritis does not outweigh these risks. Similarly, the overall impact that H.
pylori may have on the risk of esophageal adenocarcinoma is insignificant.
(See 'Corpus-dominant and pangastritis' above and "Association between
Helicobacter pylori infection and gastrointestinal malignancy".)
●For H. pylori diagnosed in the context of GERD, the available data do not
demonstrate worsening disease with H. pylori eradication; in fact, GERD may be
improved in individuals with antral-dominant gastritis. In addition, although H.
pylori improves the ability of proton pump inhibitors to suppress acid, the dose
required to maintain remission of esophagitis is not affected. (See 'Patients with
GERD' above and 'Effect of H. pylori on PPI therapy' above.)

72
Medical management of gastroesophageal reflux
disease in adults
Author:
Peter J Kahrilas, MD
Section Editor:
Nicholas J Talley, MD, PhD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Nov 29, 2021.
What's New
Updated ACG guidelines on gastroesophageal reflux disease (November 2021)
Updated American College of Gastroenterology guidelines on gastroesophageal reflux
disease (GERD) recommend an empiric eight-week trial of proton pump inhibitors
(PPI) in patients with classic GERD symptoms without alarm features and suggest a
diagnostic endoscopy for those who do not respond adequately [1]. They also
recommend attempts at discontinuation of PPIs after eight weeks in
asymptomatic patients without Barrett's esophagus or erosive esophagitis. Our
recommendations are largely consistent with these guidelines. (See "Medical
management of gastroesophageal reflux disease in adults", section on 'PPI refractory
symptoms'.)
Read more

INTRODUCTIONThe passage of gastric contents into the esophagus

(gastroesophageal reflux) is a normal physiologic process. Most episodes are brief and
do not cause symptoms, esophageal injury, or other complications. Gastroesophageal
reflux becomes a disease when it either causes macroscopic damage to the esophagus
or causes symptoms.
This topic will review the initial management of gastroesophageal reflux disease
(GERD) and management of patients with recurrent GERD when treatment is
discontinued. Our recommendations are largely consistent with the American
Gastroenterological Association and the American College of Gastroenterology
guidelines [1,2]. The evaluation and management of refractory GERD and the role of
surgery and endoscopic therapy in patients with GERD are discussed separately.
(See "Approach to refractory gastroesophageal reflux disease in adults" and "Surgical
management of gastroesophageal reflux in adults" and "Radiofrequency treatment for
gastroesophageal reflux disease".)
PRETREATMENT EVALUATION
Assessment of clinical severity — The frequency and severity of symptoms can
guide the management of GERD. Symptoms are considered mild or moderate/severe
based on whether they impair quality of life. Symptoms may be intermittent (less than
two episodes per week) or frequent (two or more episodes per week). (See "Clinical
manifestations and diagnosis of gastroesophageal reflux in adults", section on 'Clinical
features'.)
73
Are there indications for upper endoscopy? — Upper endoscopy is not required in
the presence of typical GERD symptoms of heartburn or regurgitation [2]. We
recommend an upper endoscopy if the diagnosis of GERD is unclear and to evaluate
alarm features or abnormal imaging if not performed within the last three months. Upper
endoscopy should also be performed to screen for Barrett’s esophagus in patients with
risk factors. Patients with GERD should receive empiric acid suppressive therapy, even
if additional evaluation with an upper endoscopy is indicated. (See "Clinical
manifestations and diagnosis of gastroesophageal reflux in adults", section on
'Evaluation in selected patients'.) (Related Pathway(s): Gastroesophageal reflux
disease: Identification of adults who require upper endoscopy.)
INITIAL MANAGEMENT
Overall approach — Our management approach to patients with gastroesophageal
reflux disease (GERD) is based on the frequency and severity of symptoms and the
presence of erosive esophagitis or Barrett’s esophagus on upper endoscopy, if
previously performed (algorithm 1). (See 'Pretreatment evaluation' above and "Clinical
manifestations and diagnosis of gastroesophageal reflux in adults", section on
'Endoscopic findings'.)
In patients with mild and intermittent symptoms (fewer than two episodes per week) and
no evidence of erosive esophagitis, we suggest step-up therapy for GERD. The step-up
approach involves incrementally increasing the potency of therapy until symptom control
is achieved. In patients who are naïve to treatment, we initially recommend lifestyle and
dietary modification and, as needed, low-dose histamine 2 receptor antagonists
(H2RAs) (table 1). We suggest concomitant antacids and/or sodium alginate as needed
if symptoms occur less than once a week. For patients with continued symptoms
despite these measures, we increase the dose of H2RAs to standard dose, twice daily
for a minimum of two weeks. Further increases in the dose of H2RA, prolonging the
course of treatment, or switching to another H2RA is unlikely to control symptoms [3,4].
Therefore, if symptoms of GERD persist, we discontinue H2RAs and initiate once-daily
proton pump inhibitors (PPIs) at a low dose and then increase to standard doses if
required (table 1). We make incremental changes in therapy at four to eight-week
intervals. Once symptoms are controlled, treatment should be continued for at least
eight weeks.
In patients with erosive esophagitis, frequent symptoms (two or more episodes per
week), and/or severe symptoms that impair quality of life, we use step-down therapy in
order to optimize symptom relief. The step-down approach starts with potent
antisecretory agents and then involves incrementally, decreasing the potency of therapy
until breakthrough symptoms define the treatment necessary for symptom control. We
begin with standard-dose PPI once daily for eight weeks in addition to lifestyle and
dietary modification (table 1). We subsequently decrease acid suppression to low-dose
PPIs and then to H2RAs if patients have mild or intermittent symptoms. We discontinue
acid suppression in all asymptomatic patients with the exception of patients with severe
erosive esophagitis or Barrett's esophagus, in whom we suggest maintenance PPI
therapy. (See 'Duration of acid suppression' below.)
Patients with GERD may be managed with a step-up or step-down approach to therapy.
While the optimal strategy is controversial, both have advantages [5-7]. The step-up
approach minimizes the use of PPIs and their associated costs and side effects,
74
whereas step-down therapy provides faster symptom relief. (See "Proton pump
inhibitors: Overview of use and adverse effects in the treatment of acid related
disorders", section on 'Adverse effects'.)
Mild and intermittent symptoms
Lifestyle and dietary modification — Although several lifestyle and dietary
modifications have been used in clinical practice, a systematic review of 16 randomized
trials that evaluated the impact of these measures on GERD concluded that only weight
loss and elevation of the head end of the bed improved esophageal pH-metry and/or
GERD symptoms [8-13].
We suggest the following lifestyle and dietary measures [14,15]:
●Weight loss for patients with GERD who are overweight or have had recent
weight gain.
●Elevation of the head of the bed in individuals with nocturnal or laryngeal
symptoms (eg, cough, hoarseness, throat clearing). This can be achieved either
by putting six- to eight-inch blocks under the legs at the head of the bed or a
Styrofoam wedge under the mattress. We also suggest a corollary to this
recommendation: refraining from assuming a supine position after meals and
avoidance of meals two to three hours before bedtime [2]. (See "Clinical
manifestations and diagnosis of gastroesophageal reflux in adults", section on
'Clinical features'.)
●We suggest selective elimination of dietary triggers (caffeine, chocolate, spicy
foods, food with high fat content, carbonated beverages, and peppermint) in
patients who note correlation with GERD symptoms and an improvement in
symptoms with elimination. In a prospective study of 48,308 women, intake of
coffee, tea, and soda, but not milk, water, or juice were associated with a modest
and dose-dependent increase in risk of GERD symptoms (HR 1.34, 1.26 and
1.29) [16]. Substitution of two servings per day of coffee, tea, or soda with water
was associated with a small reduction in GERD symptoms (HR, 0.96; tea HR,
0.96; and soda HR, 0.92, respectively).
Other measures that have a physiologic basis but have not consistently been
demonstrated to improve reflux symptoms include [8,15,17,18]:
●Avoidance of tight-fitting garments to prevent increasing intragastric pressure and
the gastroesophageal pressure gradient. (See "Pathophysiology of reflux
esophagitis", section on 'Gastroesophageal junction incompetence'.)
●Promotion of salivation through oral lozenges/chewing gum to neutralize refluxed
acid and increase the rate of esophageal acid clearance.
●Avoidance of tobacco and alcohol, as both reduce lower esophageal sphincter
pressure and smoking also diminishes salivation.
●Abdominal breathing exercises to strengthen the antireflux barrier of the lower
esophageal sphincter.
In a prospective study of 42,955 females, the risk for GERD symptoms was significantly
lower for those with five antireflux lifestyle factors as compared with females without
adherence to antireflux lifestyle factors (HR 0.50; 95% CI 0.42-0.59) [19]. These
measures included eating a diet rich in whole grains, vegetables, and fruits; maintaining
a healthy body weight; not smoking; exercising for 30 minutes per day; and drinking no
more than two cups of soda, coffee, or tea per day. Up to 40 percent of weekly GERD

75
symptoms could be prevented through five dietary and lifestyle factors. The decreased
risk of GERD symptoms associated with these measures was observed even among
regular users of PPIs and H2RAs.
Antacids — As antacids do not prevent GERD, their role is limited to intermittent (on-
demand) use for relief of mild GERD symptoms that occur less than once a week [4].
Antacids usually contain a combination of magnesium trisilicate, aluminum hydroxide,
or calcium carbonate, which neutralize gastric pH, thereby decreasing the exposure of
the esophageal mucosa to gastric acid during episodes of reflux. Antacids begin to
provide relief of heartburn within five minutes, but have a short duration of effect of 30 to
60 minutes. The side effects of antacids are discussed separately. (See "Antiulcer
medications: Mechanism of action, pharmacology, and side effects", section on
'Adverse effects'.)
Surface agents and alginates — Sucralfate (aluminum sucrose sulfate), a surface
agent, adheres to the mucosal surface, promotes healing, and protects from peptic
injury by mechanisms that are incompletely understood. However, given the short
duration of action and limited efficacy as compared with PPIs, the use of sucralfate is
limited to the management of GERD in pregnancy [2,20]. (See 'Pregnancy and
lactation' below.)
Sodium alginate is a polysaccharide derived from seaweed that forms a viscous gum
that floats within the stomach and neutralizes the postprandial acid pocket in the
proximal stomach [21]. Studies evaluating the efficacy of alginates on GERD symptoms
and esophageal acid exposure suggest this may be beneficial, especially for post-
prandial symptoms in individuals with relatively mild reflux disease [22-25]. They are
also used as add on therapy in patients with refractory GERD. (See "Approach to
refractory gastroesophageal reflux disease in adults", section on 'Residual acid reflux'.)
Histamine 2 receptor antagonist — Histamine 2 receptor antagonists (H2RAs)
decrease the secretion of acid by inhibiting the histamine 2 receptor on the gastric
parietal cell. However, the development of tachyphylaxis within two to six weeks of
initiation of H2RAs limits their use in the management of GERD [26]. (See 'Mild and
intermittent symptoms' above.)
In contrast to antacids, H2RAs have a slower onset of action, reaching peak
concentrations 2.5 hours after dosing, but a significantly longer duration of action of 4 to
10 hours [27]. H2RAs are also more effective in decreasing the frequency and severity
of heartburn symptoms as compared with antacids and placebo.
However, H2RAs have limited efficacy in patients with erosive esophagitis. While
mucosal healing rates in patients with mild erosive esophagitis are 10 to 24 percent
higher with H2RA therapy as compared with placebo, H2RAs are ineffective in patients
with severe esophagitis [28-30].
Severe or frequent symptoms or erosive esophagitis — We use step-down therapy
in patients with erosive esophagitis, frequent symptoms (two or more episodes per
week), and/or severe symptoms that impair quality of life in order to optimize symptom
relief. We begin with standard-dose PPI once daily for eight weeks in addition to lifestyle
and dietary modification (table 1). (See 'Lifestyle and dietary modification' above.)
Proton pump inhibitors — PPIs should be used in patients who fail twice-daily H2RA
therapy and in patients with erosive esophagitis and/or frequent (two or more episodes
per week) or severe symptoms of GERD that impair quality of life. (See 'Mild and

76
intermittent symptoms' above and "Clinical manifestations and diagnosis of
gastroesophageal reflux in adults", section on 'Endoscopic findings'.)
PPIs are the most potent inhibitors of gastric acid secretion by irreversibly binding to
and inhibiting the hydrogen-potassium (H-K) ATPase pump. PPIs are most effective
when taken 30 to 60 minutes before the first meal of the day because the amount of H-
K-ATPase present in the parietal cell is greatest after a prolonged fast. PPIs should be
administered daily rather than on-demand because continuous therapy provided better
symptom control, quality of life, and higher endoscopic remission rates [31,32].
(See "Proton pump inhibitors: Overview of use and adverse effects in the treatment of
acid related disorders", section on 'Dose and timing of administration'.)
PPIs at standard doses for eight weeks relieve symptoms of GERD and heal
esophagitis in up to 86 percent of patients with erosive esophagitis [27,33]. There are
no major differences in efficacy among PPIs and no consistent increase in symptom
resolution or esophagitis healing rates between different doses or dosing regimens of
PPI therapy [31]. (See "Proton pump inhibitors: Overview of use and adverse effects in
the treatment of acid related disorders", section on 'Pharmacology'.)
As compared with H2RAs, PPIs provide faster symptom relief and are more effective in
relieving symptoms of GERD [34-36]. In a meta-analysis of 34 trials that included 1314
individuals, PPIs were significantly more effective than H2RAs in relieving heartburn in
patients treated empirically for GERD and in patients with nonerosive reflux disease on
upper endoscopy (relative risk [RR] 0.66 and 0.78, respectively) [36]. PPIs are also
more effective than H2RAs in healing erosive esophagitis, regardless of the severity of
esophagitis and the dose and duration of treatment [28,37,38].
Limitations of PPIs include a higher cost as compared with H2RAs, and potential side
effects. The side effects associated with PPIs are discussed separately. (See "Proton
pump inhibitors: Overview of use and adverse effects in the treatment of acid related
disorders", section on 'Adverse effects'.)
Repeat endoscopy for severe erosive esophagitis — Patients with severe erosive
esophagitis (Los Angeles classification Grade C and D) on initial endoscopy should
undergo a follow-up endoscopy after a two-month course of PPI therapy to assess
healing and rule out Barrett's esophagus. Repeat endoscopy after this follow-up
examination is not indicated in the absence of Barrett's esophagus unless patients have
bleeding, dysphagia, or a significant change in symptoms while on effective therapy for
GERD. (See "Clinical manifestations and diagnosis of gastroesophageal reflux in
adults", section on 'Endoscopic findings' and "Barrett's esophagus: Surveillance and
management", section on 'Surveillance'.)
SUBSEQUENT MANAGEMENT
PPI refractory symptoms — Patients who fail to respond to once-daily proton pump
inhibitor (PPI) therapy are considered to have refractory reflux-like symptoms. We
perform a diagnostic endoscopy along with an ambulatory esophageal pH-metry study,
ideally after PPIs are stopped for at least two weeks, in patients whose reflux-like
symptoms do not respond adequately to an eight-week empiric trial of PPIs [2,39]. This
helps phenotype them as refractory GERD, reflux hypersensitivity, or functional
heartburn [40]. The management of patients with refractory GERD is discussed
separately. (See "Approach to refractory gastroesophageal reflux disease in adults",
section on 'Initial assessment'.)
77
Duration of acid suppression
Patients with erosive esophagitis or Barrett's esophagus — Patients with erosive
esophagitis or Barrett's esophagus require maintenance acid suppression with a PPI at
standard dose as they are likely to have recurrent symptoms and complications if acid
suppression is decreased or discontinued [31,41-44]. The rationale for continuing PPI
therapy in patients with Barrett's esophagus is discussed separately. (See "Barrett's
esophagus: Surveillance and management", section on 'Management of acid reflux'.)
Patients without erosive esophagitis and Barrett's esophagus — PPIs should be
prescribed at the lowest dose and for the shortest duration appropriate to the condition
being treated [2,41]. In patients whose classic GERD symptoms respond to an eight-
week empiric trial of PPIs, an attempt should be made to discontinue treatment. In
patients on PPIs for longer than six months, we taper the PPI dose before discontinuing
it and use H2RAs for mild or intermittent symptoms. We discontinue acid suppression
completely in asymptomatic patients. (See "Proton pump inhibitors: Overview of use
and adverse effects in the treatment of acid related disorders", section on 'Discontinuing
PPIs'.)
Recurrent symptoms — Approximately two-thirds of patients with nonerosive reflux
disease relapse when acid suppression is discontinued. Patients with recurrent
symptoms should be managed with acid suppressive therapy with the medication and
dose used to previously control symptoms [42,45,46]. If necessary for symptom control,
therapy may be stepped up to medications of increasing potency as with initial therapy
(algorithm 1 and table 1).
●In patients with recurrent symptoms ≥3 months after discontinuing acid
suppression, we use repeated eight-week courses of acid suppressive therapy.
●In patients with recurrent symptoms <3 months of discontinuing acid suppression
who have not previously undergone an upper endoscopy, we perform an upper
endoscopy to rule out other etiologies and complications of GERD. Patients with
recurrent symptoms within three months of discontinuing acid suppression require
long-term maintenance therapy with a PPI for acid suppression. However, PPI
therapy should be used at the lowest effective dose necessary to control GERD
symptoms.
The role of surgery and endoscopic therapy in patients with GERD who cannot
tolerate long-term PPIs or want to discontinue therapy due to concerns about
long-term side effects are discussed separately. (See "Proton pump inhibitors:
Overview of use and adverse effects in the treatment of acid related disorders",
section on 'Adverse effects' and "Surgical management of gastroesophageal reflux
in adults", section on 'Indications for operation' and "Radiofrequency treatment for
gastroesophageal reflux disease".)
Indications for referral — Referral to a subspecialist is warranted for patients who fail
to respond to once daily PPI therapy (refractory GERD), and patients who cannot
tolerate long-term PPIs or want to discontinue therapy. (See "Approach to refractory
gastroesophageal reflux disease in adults" and "Surgical management of
gastroesophageal reflux in adults" and "Radiofrequency treatment for gastroesophageal
reflux disease".)
No role for empiric eradication of H. pylori — It is uncertain whether chronic acid
suppression with PPIs increases the risk for atrophic gastritis in patients with H. pylori.

78
Therefore, routine screening for H. pylori infection and empiric eradication of H.
pylori are not recommended in patients with GERD [14].
However, if H. pylori is diagnosed in the setting of GERD, eradication of H. pylori has
been associated with an improvement of symptoms in patients with antral-predominant
gastritis. The treatment of H. pylori in the setting of GERD is discussed separately.
(See "Helicobacter pylori and gastroesophageal reflux disease".)
PREGNANCY AND LACTATIONInitial management of gastroesophageal reflux
disease (GERD) in pregnancy consists of lifestyle and dietary modification (eg,
elevation of the head end of the bed, avoidance of dietary triggers). In patients with
persistent symptoms, pharmacologic therapy should begin with antacids followed
by sucralfate. In patients who fail to respond, similar to nonpregnant patients, histamine
2 receptor antagonists (H2RAs) and then proton pump inhibitors (PPIs) should be used
to control symptoms (table 1). (See 'Mild and intermittent symptoms' above.)
Most antacids are considered safe in pregnancy and are compatible with breastfeeding
[47]. However, antacids containing sodium bicarbonate and magnesium trisilicate
should be avoided in pregnancy [48]. (See 'Antacids' above.)
Sucralfate is likely safe during pregnancy and lactation because it is poorly absorbed
[49]. In patients who continue to have symptoms of GERD despite antacids, we suggest
sucralfate (1 g orally three times daily) [50]. In patients with continued GERD symptoms
on sucralfate, we use H2RAs [51]. (See 'Surface agents and alginates' above.)
Experience with PPIs is more limited compared with H2RAs, but suggests that PPIs are
probably safe in pregnancy [52,53]. Limited data are available on the secretion of PPIs
in breast milk. Omeprazole and pantoprazole are secreted in low concentrations in
breast milk [54,55]. However, most of this is likely destroyed by gastric acid in the
infant's stomach [56]. In pregnant patients with GERD symptoms despite H2RAs, we
suggest the use of omeprazole, lansoprazole, or pantoprazole rather than other PPIs,
as they have been more widely used in pregnancy. The safety of PPIs in pregnancy has
been evaluated in several studies. A meta-analysis of seven observational studies
found no significant difference in the risk for major congenital birth defects, spontaneous
abortions, or preterm delivery among 1530 women exposed to PPIs during pregnancy
as compared with 133,410 who were not exposed to PPIs [52]. A subsequent
observational study that evaluated PPI exposure in early pregnancy also found no
increase in the risk of major birth defects in 3651 infants exposed to PPIs during the first
trimester, as compared with 837,317 infants who had not been exposed [53].
(See 'Proton pump inhibitors' above.)
Upper endoscopy should be performed during pregnancy only if there is a strong
indication (eg, significant gastrointestinal bleeding). When possible, endoscopy should
be postponed until the second trimester [57]. Obstetrical staff should be closely involved
and the degree of maternal and fetal monitoring should be individualized.
Recommendations for procedural sedation for endoscopy in pregnant and nursing
women are discussed separately. (See "Gastrointestinal endoscopy in adults:
Procedural sedation administered by endoscopists", section on 'Special populations'.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Gastroesophageal reflux in adults".)

79
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Acid reflux and gastroesophageal reflux

disease in adults (The Basics)" and "Patient education: H. pylori infection (The
Basics)" and "Patient education: Acid reflux (gastroesophageal reflux disease)
during pregnancy (The Basics)")
●Beyond the Basics topics (see "Patient education: Gastroesophageal reflux
disease in adults (Beyond the Basics)" and "Patient education: Helicobacter pylori
infection and treatment (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●The optimal approach to the management of gastroesophageal reflux disease
(GERD) is controversial. Our management approach to patients with GERD is
based on the frequency and severity of symptoms and the presence of erosive
esophagitis or Barrett’s esophagus on upper endoscopy, if previously performed
(algorithm 1). (See 'Pretreatment evaluation' above and "Clinical manifestations
and diagnosis of gastroesophageal reflux in adults", section on 'Indications'.)
●We suggest lifestyle and dietary modification in all patients with GERD (Grade
2C). (See 'Lifestyle and dietary modification' above.)
●In patients with mild and intermittent (less than two episodes per week)
symptoms of GERD who are naïve to treatment and no evidence of erosive
esophagitis or Barrett’s esophagus, we suggest as needed low-dose histamine 2
receptor antagonists (H2RAs) (table 1) (Grade 2B). Concomitant antacids, and/or
sodium alginate can be used if symptoms occur less than once a week. In patients
with continued symptoms, we increase the dose of H2RAs to standard dose, twice
daily for a minimum of two weeks. (See 'Antacids' above and 'Overall
approach' above.)
●If symptoms of GERD persist, we discontinue H2RAs and initiate once-daily
proton pump inhibitors (PPIs) at a low dose (Grade 2B) and then increase to
standard doses if required for symptom control (table 1). Once symptoms are
controlled, therapy should be continued for at least eight weeks. (See 'Histamine 2
receptor antagonist' above and 'Overall approach' above.)

80
●In patients with erosive esophagitis, we recommend initial acid suppressive
therapy with standard-dose PPI once daily (table 1) (Grade 1A). (See 'Overall
approach' above and 'Proton pump inhibitors' above.)
●In patients with frequent (two or more episodes per week), severe symptoms that
impair quality of life or Barrett’s esophagus, we suggest initial therapy with
standard-dose PPI once daily (Grade 2B). (See 'Overall approach' above
and 'Proton pump inhibitors' above.)
●Referral to a subspecialist is warranted for patients who fail to respond to once-
daily PPI therapy and patients who cannot tolerate long-term PPIs or want to
discontinue therapy. (See 'PPI refractory symptoms' above.)
●We discontinue acid suppression in all patients with GERD whose symptoms
resolve completely with treatment, with the exception of those with erosive
esophagitis on upper endoscopy and Barrett's esophagus. (See 'Duration of acid
suppression' above.)
●In patients with recurrent symptoms within three months of discontinuing acid
suppression, we continue long-term maintenance therapy with a PPI. However, if
symptoms occur after three or more months, we use repeated eight week courses
of previously effective acid suppressive therapy (algorithm 1) (see 'Recurrent
symptoms' above).
●Management of GERD in pregnancy includes lifestyle and dietary modification
followed by pharmacologic therapy with antacids and sucralfate. Antacids
containing sodium bicarbonate and magnesium trisilicate should be avoided in
pregnancy. In patients who fail to respond, similar to nonpregnant patients, H2RAs
and then PPIs are used to control symptoms. (See 'Pregnancy and
lactation' above and 'Mild and intermittent symptoms' above.)

81
Pathophysiology of reflux esophagitis
Author:
Peter J Kahrilas, MD
Section Editor:
Nicholas J Talley, MD, PhD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jan 11, 2022.
INTRODUCTIONSome degree of reflux is physiologic [1]. Physiologic reflux
episodes typically occur postprandially, are short-lived, asymptomatic, and rarely occur
during sleep. Pathologic reflux is associated with symptoms or mucosal injury and often
occurs nocturnally. In general, the term gastroesophageal reflux disease (GERD) is
applied to patients with symptoms suggestive of reflux or complications thereof, but not
necessarily with, esophageal inflammation. Reflux esophagitis describes a subset of
patients with GERD who have endoscopic evidence of esophageal inflammation.
The pathophysiology of GERD will be reviewed here. The clinical manifestations and
diagnosis of this disorder are discussed separately. (See "Clinical manifestations and
diagnosis of gastroesophageal reflux in adults".)
MECHANISMS OF GASTROESOPHAGEAL REFLUX DISEASE The
development of gastroesophageal reflux disease (GERD) reflects an imbalance
between injurious or symptom-eliciting factors (reflux events, acidity of refluxate,
esophageal hypersensitivity) and defensive factors (esophageal acid clearance,
mucosal integrity) [2,3]. The extent of symptoms and of mucosal injury is proportional to
the frequency of reflux events, the duration of mucosal acidification, and the caustic
potency of refluxed fluid.
Esophagitis results from cytokine-triggered inflammation rather than a direct chemical
effect of prolonged exposure to acid, pepsin, and bile on the esophageal epithelium [4-
6]. This is substantiated by the observation that histopathological events in the
development of esophagitis (lymphocytic inflammation, dilated intercellular spaces)
occur deep in the epithelium, not at the luminal surface, and that regenerative changes
(basal cell hyperplasia, papillary elongation) are initiated prior to the development of
surface necrosis that was formerly hypothesized as the stimulus for those changes.
Cytokine-triggered inflammation may also cause alterations in esophageal sensitivity in
the absence of esophagitis.
Gastroesophageal junction incompetence — The antireflux barrier at the
gastroesophageal junction is anatomically and physiologically complex and vulnerable
to several potential mechanisms of reflux. The three dominant pathophysiologic
mechanisms causing esophagogastric junction (EGJ) incompetence are:
●Transient lower esophageal sphincter relaxations (TLESRs)
●A hypotensive lower esophageal sphincter (LES)
●Anatomic disruption of the gastroesophageal junction, often associated with a
hiatal hernia

82
Although examples of each have been documented in mechanistic studies, their relative
importance continues to be debated. The evolving concept is that the dominant
mechanism varies as a function of disease severity with TLESRs predominating with
mild disease, and mechanisms associated with a hiatus hernia and/or a weak sphincter
predominating with more severe disease [7].
Transient lower esophageal sphincter relaxations — TLESR is part of the
physiological mechanism of belching. TLESR is an active, vagally mediated reflex
allowing air to escape from the stomach [8,9]. A primary determinant of reflux disease is
an increased proportion of TLESRs that are associated with acid reflux rather than gas
venting alone [10-12].
TLESR involves not only LES relaxation, but also crural diaphragmatic inhibition,
esophageal shortening by contraction of its longitudinal muscle, and contraction of the
costal diaphragm [12,13]. There are several major differences between TLESR and
swallow-induced LES relaxation: TLESRs occur without an associated pharyngeal
contraction, are unaccompanied by esophageal peristalsis, and persist for longer
periods (>10 seconds) as compared with swallow-induced LES relaxations [14]. The
frequency of TLESRs is increased by distension of the stomach or by assuming an
upright posture [15].
Acid reflux has been observed with as many as 93 percent or as few as 9 percent of
TLESRs [12,16]. Acid reflux with TLESRs can result from increased compliance of the
EGJ as a consequence of weakening or dilatation of the diaphragmatic hiatus [17].
Increased compliance leads to an increased luminal cross-sectional area during
opening that in turn results in an increased volume of reflux and a reduced ability to limit
refluxate to gas. TLESRs can be inhibited by gamma aminobutyric acid receptor type B
agonists (eg, baclofen). (See "Non-acid reflux: Clinical manifestations, diagnosis, and
management", section on 'Reflux inhibitors'.)
Hypotensive lower esophageal sphincter — Only a minority of individuals with GERD
have a grossly hypotensive LES (<10 mmHg) when determined during fasting
measurements [18]. Factors that can reduce LES pressure include gastric distension,
cholecystokinin, smoking, and specific foods and medications [16]. (See 'Diet and
medications' below.)

Gastroesophageal reflux can occur with diminished LES pressure either by strain-
induced reflux or free reflux:

●Strain-induced reflux occurs when a hypotensive LES is overcome and "blown

open" by an abrupt increase of intra-abdominal pressure. Manometric data
suggest that strain-induced reflux is relatively unusual unless the LES pressure is
less than 4 mmHg [19]. However, these studies were somewhat limited by the
required instrumentation and may not be entirely reflective of normal ambulatory
circumstances.
●During free reflux, a fall in intraesophageal pH occurs without identifiable change
in either intragastric or LES pressure. Free reflux is observed only when LES
pressure is within 0 to 4 mmHg of intragastric pressure [11].
Anatomic disruption of the gastroesophageal junction — Laxity in the LES-crural
diaphragm attachment, reduced LES-crural diaphragm synergy, increased EGJ
distensibility, and increased intra-abdominal pressure can result in mechanical
83
impairment of the EGJ. Patients with hiatus hernia also have progressive disruption of
the diaphragmatic sphincter. (See 'Hiatus hernia' below.)
Physiologically, both the diaphragm and the LES contribute to gastroesophageal
sphincter competence and EGJ pressure. Recordings of LES pressure usually exhibit
inspiratory increases as a result of contraction of the diaphragmatic crus that encircles
the LES [20]. Observations of the antireflux mechanism during maneuvers such as leg
raising and abdominal compression suggest a "pinchcock" effect of crural contraction
that augments the antireflux barrier. The susceptibility to reflux under circumstances of
abrupt increases of intraabdominal pressure (eg, during bending or coughing) depends
upon both the instantaneous LES pressure and the diaphragmatic sphincter [21].
Characteristics of the refluxate — Reflux esophagitis is not usually related to
increased gastric acid secretion [22]. The intragastric pH and the amount of time the
refluxate is in contact with the mucosa are important determinants of the extent of
esophageal mucosal injury. The degree of mucosal damage is more significant if the
refluxate pH is less than two and/or if pepsin is present in the refluxate. Pepsin, bile
acids, trypsin, and food hyperosmolality increase the susceptibility of the esophageal
mucosa to acid injury [23]. At pH 2, pepsin disrupts the histologic integrity of the
mucosal barrier, increases hydrogen ion permeability, and causes hemorrhage [24]. Bile
acids have been implicated in the development of esophagitis primarily in patients with
increased duodenogastric reflux following gastric surgery [25].
Impaired esophageal acid clearance — Following reflux, esophageal acid clearance
begins with emptying the refluxed fluid from the esophagus by peristalsis and is
completed by titration of the residual acid with swallowed saliva (figure 1). Prolongation
of esophageal acid clearance time (when the esophageal pH <4) has been observed in
approximately one half of patients with esophagitis [26]. The two major causes of this
problem are impaired esophageal emptying and impaired salivary function.
Impaired esophageal emptying — The process of normal acid clearance involves
peristalsis as well as the swallowing of salivary bicarbonate. Peristalsis clears gastric
fluid from the esophagus, whereas the swallowing of saliva (pH of 7.8 to 8.0) neutralizes
any remaining acid. Both primary and secondary peristalsis are essential mechanisms
of esophageal clearance. Abnormal acid clearance improves with an erect posture,
suggesting that gravity can compensate for impaired fluid emptying.

Two mechanisms of impaired esophageal emptying have been identified:

●Ineffective esophageal motility – Ineffective esophageal motility can be

manifest as either failed or hypotensive (<30 mmHg or distal contractile integral
<450 mmHg∙cm∙s) peristaltic contractions [27]. Ineffective esophageal motility
becomes more common with increasing severity of esophagitis [18]. Whether
ineffective esophageal motility associated with peptic esophagitis is reversible is
uncertain. It is likely that acute dysfunction associated with active esophagitis is
partially reversible, while chronic dysfunction associated with stricturing or
extensive fibrosis is not.
●Re-reflux – Re-reflux or retrograde flow associated with hiatal hernias also
impairs esophageal emptying [28,29]. Re-reflux from the hernia during swallowing
occurs only with nonreducing hernias that are evident between swallows and
during peristalsis-induced esophageal shortening. In one report, for example, re-
84
 reflux was seen with almost 50 percent of test swallows in patients with
nonreducing hernias, impairing both esophageal emptying and esophageal acid
clearance [29].
The potential for re-reflux may be aggravated by the presence of an acid pocket in
the most proximal gastric cardia that escapes the buffering effects of food and
remains highly acidic during the postprandial period [30,31]. The net result is
much greater acid exposure just above the squamocolumnar junction compared to
5 cm proximal to it, likely explaining the propensity of the distal esophagus to
develop mucosal erosions [32,33].
Diminished salivary function — Reduced salivation or diminished salivary neutralizing
capacity also prolongs acid clearance. Approximately 7 ml of saliva will neutralize 1 ml
of 0.1 N HCl, with 50 percent of the neutralizing capacity being attributable to salivary
bicarbonate. The normal rate of salivation is about 0.5 ml/min; maneuvers that increase
salivation (eg, oral lozenges or gum chewing) will hasten acid clearance. Diminished
salivation during sleep, for example, explains why reflux events during sleep or
immediately prior to going to sleep are associated with markedly prolonged acid
clearance times. Cigarette smokers have prolonged esophageal acid clearance times
due to hyposalivation. In one study, smokers without symptoms of reflux disease were
found to have acid clearance times 50 percent longer than those of nonsmokers;
furthermore, the salivary titratable base content of the smokers was only 60 percent of
the age-matched nonsmokers [34]. Chronic xerostomia is associated with prolonged
esophageal acid exposure and esophagitis in a subset of patients [35,36].
Impaired defense against epithelial injury — Age and nutritional status influence the
ability of the mucosa to withstand injury. Esophageal mucosa possesses morphologic
and physiologic defenses against reflux injury at the epithelial and post-epithelial levels
[37]. These mechanisms include the following:
●Epithelial tight junctions – The main defense against reflux injury is the
epithelial barrier itself. The esophageal mucosa is a relatively "tight" epithelium,
resistant to ionic movements at the intercellular as well as the cellular level
because of the tight junctions and the lipid-rich matrix in the intercellular space
[38]. This mucosa can retard hydrogen ion penetration in the face of ion gradients
of greater than 5 pH units.
●Hydrogen ion extrusion – Two pH-activated acid extruding processes are
located on esophageal membranes: a Na+/H+ exchanger; and a sodium dependent
Cl-/HCO3- exchanger [39]. Once extruded, the hydrogen ions are buffered by
extracellular bicarbonate in equilibrium with the blood. Thus, blood flow is the main
post-epithelial defense, interacting with epithelial factors to protect against acid
injury. Blood flow increases in response to luminal acid, delivering more
bicarbonate to the intercellular space as well as providing nutrients for metabolic
activity [40].
Esophageal hypersensitivity — Increased mucosal sensitivity to acid can also
contribute to symptoms of heartburn. The cause of heartburn in patients with normal
esophageal acid exposure is uncertain, but may be related to enhanced sensory
perception of physiological reflux (visceral hyperalgesia) [41].

85
OTHER ETIOLOGIC FACTORSSeveral factors contribute to development of
pathological reflux by mechanical or functional impairment of the esophagogastric
junction (EGJ).
Hiatus hernia — The severity of reflux esophagitis correlates with the size of hiatus
hernia [42]. There are several mechanisms underlying the development of GERD in
patients with a hiatus hernia. Mechanisms include impairment of the crural
diaphragmatic component of the EGJ, low LES pressure, and a reduced threshold for
eliciting transient lower esophageal sphincter relaxation (TLESR) in response to gastric
distension [43-45]. In patients with a hiatus hernia, the supradiaphragmatic location of
the postprandial acid pocket just distal of the squamocolumnar junction also contributes
to the number of TLESRs associated with acid reflux [46]. Hiatal hernias are also
associated with prolonged recumbent acid clearance times due to ineffective
esophageal motility and re-reflux, both of which impair esophageal emptying [47].
(See 'Impaired esophageal acid clearance' above and "Hiatus hernia" and 'Transient
lower esophageal sphincter relaxations' above.)
Obesity — Obesity is a risk factor for GERD and erosive esophagitis [48]. In a cross-
sectional survey of 10,545 women, increasing body mass index was associated with a
significant increase in frequency of GERD symptoms [49]. Even moderate weight gain
in women of normal weight was associated with exacerbation of symptoms.
The mechanisms by which obesity contributes to reflux are incompletely understood. In
an observational study that included 285 patients with GERD, in whom anthropometric
variables were correlated with findings on manometry, there was a significant
correlation of body mass index and waist circumference with intragastric pressure and
the gastroesophageal pressure gradient [50]. Obesity was also associated with
disruption of the EGJ leading to a hiatal hernia and increased esophageal acid
exposure [51,52]. (See 'Hiatus hernia' above.)
Obesity has also been associated with both an increased frequency of TLESRs and an
increased proportion of TLESRs associated with acid reflux during the postprandial
period in subjects without hiatus hernia or clinical evidence of GERD. These findings
suggest that lower esophageal sphincter (LES) dysfunction might be an important
mediator of the pathogenesis of obesity-related GERD [53]. (See 'Transient lower
esophageal sphincter relaxations' above.)
Pregnancy and exogenous estrogen — Heartburn occurs with 30 to 50 percent of
pregnancies. This is likely due to hormonal (estrogen and progesterone reducing LES
tone) and possible mechanical factors (gravid uterus). Estrogen replacement therapy in
postmenopausal women also appears to modestly increase the risk of heartburn [54].
(See "Maternal adaptations to pregnancy: Gastrointestinal tract".)
Diet and medications — Specific foods (fat, chocolate, peppermint), caffeine, alcohol,
smoking, and several drugs (eg, anticholinergics, nitrates, calcium channel blockers,
tricyclic antidepressants, opioids, theophylline, diazepam, barbiturates) can cause reflux
by inducing LES hypotension.
UNCLEAR ROLE OF HELICOBACTER PYLORIThe link between
gastroesophageal reflux disease (GERD) and Helicobacter pylori is complex and, in
many cases, H. pylori may actually be protective against GERD. (See "Helicobacter
pylori and gastroesophageal reflux disease".)

86
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Hiatal hernia (The Basics)")

SUMMARY
●The development of gastroesophageal reflux disease (GERD) reflects the
balance between injurious or symptom-eliciting factors (reflux events, acidity of
refluxate, esophageal hypersensitivity) and defensive factors (esophageal acid
clearance, mucosal integrity). The extent of symptoms and of mucosal injury is
proportional to the frequency of reflux events, the duration of mucosal
acidification, and the caustic potency of refluxed fluid (algorithm 1).
(See 'Mechanisms of gastroesophageal reflux disease' above.)
●The three dominant pathophysiologic mechanisms causing gastroesophageal
junction incompetence are: transient lower esophageal sphincter relaxations
(TLESRs), a hypotensive lower esophageal sphincter (LES), and anatomic
disruption of the gastroesophageal junction, often associated with a hiatal hernia.
(See 'Gastroesophageal junction incompetence' above and 'Hiatus hernia' above.)
●The primary determinant of reflux disease is an increased proportion of TLESRs
that are associated with reflux of acid rather than gas. (See 'Transient lower
esophageal sphincter relaxations' above.)
●A minority of individuals with GERD have a grossly hypotensive LES (<10
mmHg) when determined during fasting measurements. However, periods of
gross LES hypotension may result from gastric distension, smoking and specific
foods and medications. (See 'Hypotensive lower esophageal sphincter' above
and 'Diet and medications' above.)
●Hiatus hernia is associated with progressive disruption of the diaphragmatic
sphincter, resulting in increased susceptibility to reflux with abrupt increases of
intra-abdominal pressure. Hiatus hernia is also associated with a reduced
threshold for eliciting TLESRs in response to gastric distension and malfunction of
the gastroesophageal barrier during periods of low LES pressure. (See 'Anatomic
disruption of the gastroesophageal junction' above and 'Hiatus hernia' above.)
●Esophageal acid clearance is accomplished with emptying the refluxed fluid from
the esophagus by peristalsis and the titration of the residual acid by swallowed
saliva. Reduced salivation or diminished salivary neutralizing capacity prolong

87
acid clearance. In patients with hiatus hernia, esophageal acid clearance may be
impaired due to ineffective esophageal motility and re-reflux from the hernia.
(See 'Impaired esophageal acid clearance' above.)
●Heartburn in patients with normal esophageal acid exposure may be related to
heightened esophageal sensitivity (also called visceral hyperalgesia).
(See 'Esophageal hypersensitivity' above.)

88
Proton pump inhibitors: Overview of use and
adverse effects in the treatment of acid related
disorders
Author:
M Michael Wolfe, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Nov 29, 2021.
What's New
PPI use and inflammatory bowel disease (November 2021)
Due to their ability to alter the gut microbiome, proton pump inhibitors (PPI) may impact
the risk of inflammatory bowel disease (IBD). In a study that included >600,000
individuals followed for a median of 12 years, regular PPI use was associated with an
increased risk of IBD compared with nonusers; however, the absolute risk of IBD was
low [1]. We continue to recommend prescribing PPIs at the lowest effective dose and for
the shortest duration appropriate to the condition being treated. (See "Proton pump
inhibitors: Overview of use and adverse effects in the treatment of acid related
disorders", section on 'Inflammatory bowel disease'.)
Read more

INTRODUCTIONProton pump inhibitors (PPIs) effectively block gastric acid

secretion by irreversibly binding to and inhibiting the hydrogen-potassium ATPase pump
that resides on the luminal surface of the parietal cell membrane.
This topic review will provide an overview of the mechanism of action,
pharmacokinetics, administration, and adverse effects of PPIs. The use and efficacy of
PPIs in specific acid-related disorders is presented separately. (See "Antiulcer
medications: Mechanism of action, pharmacology, and side effects".)
INDICATIONS FOR PPI THERAPYProton pump inhibitor (PPI) therapy is
indicated in the following clinical situations:
●Peptic ulcer disease – PPIs are first-line antisecretory therapy in the treatment
of peptic ulcer disease. (See "Peptic ulcer disease: Treatment and secondary
prevention", section on 'Initial antisecretory therapy'.)
●Gastroesophageal reflux disease – PPIs are indicated in patients with
gastroesophageal reflux disease, including for the treatment of erosive esophagitis
and as maintenance therapy in patients with severe erosive esophagitis or
Barrett’s esophagus. (See "Medical management of gastroesophageal reflux
disease in adults", section on 'Severe or frequent symptoms or erosive
esophagitis'.)
●Zollinger-Ellison syndrome – PPIs, often in high doses, are required to control
gastric acid hypersecretion in patients with gastrin-secreting

89
 tumors. (See "Management and prognosis of the Zollinger-Ellison syndrome
(gastrinoma)", section on 'Proton pump inhibitors'.)
●NSAID-associated ulcers – PPIs are indicated in the primary prevention of
gastroduodenal ulcers associated with NSAID use. (See "NSAIDs (including
aspirin): Primary prevention of gastroduodenal toxicity".)
●Eradication of Helicobacter pylori – PPIs are a component of several first-line
and salvage therapy regimens for H. pylori infection. (See "Treatment regimens for
Helicobacter pylori in adults".)
PHARMACOLOGY
●Mechanism of action – Proton pump inhibitors (PPIs) inhibit H-K-ATPase, the
final step of gastric acid secretion by parietal cells.
PPIs are benzimidazole prodrugs which accumulate specifically and selectively in
the secretory canaliculus of the parietal cell [1]. Within that space, they undergo
an acid catalyzed conversion to a reactive species, the thiophilic sulfonamides,
which are permanent cations. The rate of conversion varies among the
compounds and is inversely proportional to the pKa of the benzimidazole
(rabeprazole >omeprazole, esomeprazole, and lansoprazole >pantoprazole) (table
1) [2]. The reactive species interacts with the external surface of the H-K-ATPase
that faces the lumen of the secretory space of the parietal cell, resulting in
disulfide bond formation with cysteine 813 located within the alpha-subunit of the
enzyme; this is the residue that is intimately involved in hydrogen ion transport.
This covalent inhibition of the enzyme results in a specific and long-lasting
impairment of gastric acid secretion. (See "Physiology of gastric acid secretion".)
●Pharmacokinetics – PPIs are similar in structure and mechanism of action, but
PPIs differ in their pKa, bioavailability, peak plasma levels, and route of excretion
(table 1). The magnitude of these differences are small and their clinical relevance
has not been established.
PPIs are most effective when the parietal cell is stimulated to secrete acid
postprandially, a relationship that has important clinical implications for timing of
administration. Because the amount of H-K-ATPase present in the parietal cell is
greatest after a prolonged fast, PPIs should be administered before the first meal
of the day. In most individuals, once-daily dosing is sufficient to produce the
desired level of acid inhibition, and a second dose, which is occasionally
necessary, should be administered before the evening meal [1]. (See 'Dose and
timing of administration' below.)
Once-daily PPI dosing for five days inhibits maximal gastric acid output by
approximately 66 percent. Since PPIs inhibit only activated enzyme present in the
canalicular membrane, the reduction of gastric acid secretion after an initial dose
will probably be suboptimal. As inactive enzyme is recruited into the secretory
canaliculus, acid secretion will resume, albeit at a reduced level. After the second
dose is given on the next day, more H-K-ATPase will have been recruited and
subsequently inhibited, and after the third dose, additional recruitment and further
acid inhibition will probably occur. Thus, the occasional use of a PPI taken on an
"as needed" basis does not reliably provide adequate acid inhibition and does not
produce a consistent or satisfactory clinical response (in contrast to the H2
antagonists, which have a more rapid onset of action) [1].
90
 Restoration of acid secretion after discontinuing PPIs depends upon enzyme
turnover and the biological reversibility of the disulfide bond. Maximal acid
secretory capacity may not be restored for 24 to 48 hours [1]. (See 'Discontinuing
PPIs' below.)
●Metabolism – PPIs are metabolized via hepatic cytochrome P450 enzymes, with
CYP2C19 having the dominant role. However, the dominance of CYP2C19 over
other pathways varies significantly among the PPIs (table 2). The activity of
CYP2C19 is also determined to some extent by genetic polymorphisms. Two
inactivating mutations have been described as a result; the metabolism of PPIs by
this route may be delayed in these individuals [3,4]. Homozygotes for the wild type
gene rapidly metabolize these drugs, while heterozygotes are intermediate
metabolizers.
Plasma levels of PPI correlate with their metabolism, and differences may
contribute to varying dose requirements and clinical efficacy. As an example, one
study examined the effect of variable metabolism of omeprazole when using this
agent to treat H. pylori in 62 Japanese patients [3]. While eradication was
achieved in all individuals homozygous for a CYP2C19 mutation (ie, slow
metabolizers), successful treatment was achieved in only 60 and 29 percent of
heterozygotes and wild type homozygotes, respectively. In another study that
evaluated the efficacy of lansoprazole in the treatment of 65 patients with
gastroesophageal reflux disease (GERD), slow metabolizers were much more
likely to be asymptomatic as compared with heterozygotes and wild type
homozygotes (85 versus 68 and 46 percent, respectively) [5]. The response rate
in wild type homozygotes with severe GERD was only 16 percent. Wild type
homozygotes (rapid metabolizers) also had the lowest plasma lansoprazole
concentrations. (See 'Dose and timing of administration' below.)
PRETREATMENT CONSIDERATIONS AND MONITORING
Drug interactions — Clinically important drug interactions with proton pump inhibitors
(PPIs) are rare. However, PPI metabolism via hepatic cytochrome P450 enzymes may
lead to specific drug interactions in some individuals. The presence of a CYP2C19 gene
mutation results in higher plasma PPI levels in homozygous individuals. However, if this
metabolic pathway becomes saturated, the isoenzyme can become a major target for
interactions with many drugs, including warfarin, diazepam, clopidogrel,
and phenytoin (table 3). CYP3A4-mediated metabolism may also be activated under
such conditions and become the principal route of drug elimination. Furthermore,
induction of CYP1A, another P450 isoenzyme, in CYP2C19 deficient or saturated
individuals, can make them susceptible to interference with theophylline metabolism.
The specific P450 enzymes involved in PPI metabolism and the potential for
interactions among PPIs varies considerably (table 2) [6-12].

Some other important drug interactions with PPIs include the following:

●Clopidogrel - Some data suggest decreased activation of clopidogrel when used

in conjunction with omeprazole due to shared hepatic cytochrome P450-mediated
metabolism. In 2009, the United States Food and Drug Administration concluded
that patients taking clopidogrel should consult with their clinician if they are taking
or considering taking a PPI, including over-the-counter PPI preparations [13,14].
91
 However, the relevance of these data remains highly controversial. The interaction
of clopidogrel and PPIs are discussed in detail separately. (See "Clopidogrel
resistance and clopidogrel treatment failure".)
●HIV protease inhibitors - PPIs may decrease the absorption of certain HIV
protease inhibitors. PPIs are contraindicated in patients being treated
with rilpivirine. Atazanavir should not be used in patients who require a PPI dose
equivalent to >20 mg omeprazole daily. (See "Overview of antiretroviral agents
used to treat HIV", section on 'Protease inhibitors (PIs)'.)
●Methotrexate - Coadministration of PPIs with high dose methotrexate appears to
be correlated with delayed methotrexate elimination and potentially may lead to
methotrexate toxicity if not monitored appropriately.
For additional information on drug interactions, use the Lexicomp drug
interactions program provided by UpToDate. (See "Overview of the non-acute
management of unstable angina and non-ST elevation myocardial infarction", section on
'Gastrointestinal prophylaxis'.)
Laboratory testing — We limit routine laboratory testing to selected patients on PPI
therapy.
●Magnesium – We obtain serum magnesium levels prior to starting a PPI in
patients who are expected to be on long-term (≥1 year) treatment or in patients
who take PPIs in conjunction with other medications associated with
hypomagnesemia (eg, diuretics). In addition, we obtain magnesium levels
periodically in such patients while they are taking a PPI. The frequency of testing
is based on the clinical history and the presence of symptoms of
hypomagnesemia. As an example, in patients with a history of arrhythmias or QT
interval prolongation, we monitor magnesium levels every six months. The
management of hypomagnesemia is discussed in detail separately.
(See "Hypomagnesemia: Evaluation and treatment".)
●Vitamin B12 – We also obtain vitamin B12 levels yearly in patients on long-term
PPIs [15]. However, routinely monitoring vitamin B12 levels is controversial.
(See 'Magnesium malabsorption' below and 'Vitamin B12 malabsorption' below.)
There are insufficient evidence to support routine bone density monitoring or calcium
supplementation due to proton pump inhibitor use alone [16].
ADMINISTRATION
Intravenous regimen — IV PPIs are indicated prior to endoscopic evaluation in
patients with clinically significant upper gastrointestinal bleeding from a suspected
peptic ulcer. Pantoprazole and esomeprazole are the only PPIs available as an IV
formulation in the United States; IV omeprazole is available in other countries. The use
of PPIs in the treatment of bleeding peptic ulcers and the duration of treatment is
discussed in detail separately. (See "Overview of the treatment of bleeding peptic
ulcers", section on 'Oral versus intravenous dosing' and "Approach to acute upper
gastrointestinal bleeding in adults", section on 'Acid suppression'.)
Oral regimen
Selecting a PPI — The choice of a specific oral PPI and whether over-the-counter
(rather than prescription) PPIs are prescribed are often determined by patient
preference and payer coverage. A systematic review of 12 randomized trials examining

92
the relative effectiveness of different PPI doses and dosing regimens found no
consistent difference in symptom resolution and esophagitis healing rates [17].
In patients unable to swallow pills or capsules, options include an oral suspension
of lansoprazole and a powder formulation of omeprazole-sodium bicarbonate for oral
suspension.
Dose and timing of administration — PPIs should be administered 30 to 60 minutes
before breakfast for maximal inhibition of proton pumps. (See 'Pharmacology' above.)
Dose reduction, particularly for maintenance of healing of erosive esophagitis may be
possible in Asian populations. Polymorphisms in the CYP2C19 gene, which encodes
the cytochrome P450 isoenzyme that metabolizes different PPI preparations, are
common in Asian and other populations [18]. Such gene mutations would render an
individual a "slow metabolizer" and prolong the antisecretory effect of PPIs. In contrast,
the duration of acid inhibition would be decreased in a "rapid metabolizer," and
differences in PPI metabolism might account for incomplete inhibition of acid secretion
and a high prevalence of nocturnal breakthrough symptoms in gastroesophageal reflux
disease patients. (See 'Pharmacology' above and "Approach to refractory
gastroesophageal reflux disease in adults", section on 'Differences in PPI metabolism'.)
Avoidance of concurrent antisecretory agents — PPIs should not administered
concomitantly with antisecretory agents including histamine-2 receptor antagonists
(H2RAs), analogues of prostaglandin E (eg, misoprostol), and somatostatin analogues
(eg, octreotide), because of the marked reduction in acid inhibitory effects [1,19].
Antisecretory drugs can be used with a PPI provided that there is a sufficient time
interval between their administration. As an example, an H2RA can be taken before
bedtime or during the night by individuals who report nocturnal breakthrough symptoms
such as heartburn after taking a PPI in the morning or before dinner.
Switching between PPIs — Switching PPIs is a reasonable strategy in patients with
side-effects to an individual PPI and may be necessary due to cost differences.
Although there is significant interindividual and intraindividual variability in intragastric
pH control between PPIs, there are no consistent difference in relation to symptom
resolution and esophagitis healing rates [17]. Switching PPIs in patients with well-
controlled symptoms may also be associated with increased symptom severity and
decreased patient satisfaction [20]. (See "Approach to refractory gastroesophageal
reflux disease in adults", section on 'Subsequent management'.)
Discontinuing PPIs — PPIs should be prescribed at the lowest dose and for the
shortest duration appropriate to the condition being treated. (See 'Indications for PPI
therapy' above.)
We gradually taper PPI therapy in patients treated with PPIs for longer than six months.
For patients on a standard or high-dose PPI (eg, omeprazole 40 mg daily or twice daily),
we decrease the dose by 50 percent every week. For patients on twice daily dosing, the
initial reduction can be accomplished by decreasing the dosing to once in the morning
before breakfast until the patient is on the lowest dose of the medication. Once on the
lowest dose for one week, the patient is instructed to discontinue the PPI. However, no
specific method for discontinuing PPI therapy has been proven effective, and no
approach is universally accepted. [21,22].
Studies have demonstrated rebound gastric acid hypersecretion following
discontinuation of PPIs in patients with long-term use. The reasons are not entirely

93
clear, but appear to be due in part to the suppression of antral somatostatin expression,
resulting in an increase in antral gastrin release and subsequent disruption of normal
pH-related feedback inhibition of acid secretion that occurs after a meal [1].
(See "Physiology of gastric acid secretion", section on 'Tolerance and acid rebound'.)
ADVERSE EFFECTSLong-term PPI use has been associated with several safety
concerns. However, few of these concerns are supported by consistent data
demonstrating a causal relationship. (See "Physiology of gastrin", section on
'Hypergastrinemia'.)
Gastrointestinal effects
Clostridioides difficile and other enteric infections — PPI use has been associated
with an increased risk of C. difficile infection, even in the absence of antibiotic use [23-
33]. Associations with other enteric infections, including salmonellosis and
campylobacteriosis, have also been reported [34-39]. However, the pathophysiologic
mechanism involved in the increased risk of infection is unclear.
A 2017 meta-analysis of 50 observational studies found that PPI use was significantly
associated with an increased risk of C. difficile infection (relative risk [RR] 1.3; 95% CI
1.1-14). The risk of C. difficile infection appears to be greater with PPIs as compared to
H2 receptor antagonists [30,31].
PPI use has also been associated with an increased risk of recurrent C. difficile infection
[31]. In a 2017 meta-analysis of 16 observational studies that included 7703 patients
with C. difficile infection of whom 1525 (20 percent) had recurrent C. difficile infection,
gastric acid suppression was significantly associated with an increased risk of
recurrent C. difficile infection (odds ratio [OR] 1.5; 95% CI 1.2-1.9) [40]. There was
significant heterogeneity among the studies included in the meta-analysis. In adjusted
analysis using data from nine studies, PPI use was associated with an increased risk of
recurrent C. difficile infection after controlling for patient age and other co-morbid
conditions (OR 1.4; 95% CI 1.1-1.8). (See "Clostridioides difficile infection in adults:
Epidemiology, microbiology, and pathophysiology", section on 'Gastric acid
suppression'.)
Microscopic colitis — PPI use has been associated with microscopic colitis, including
lymphocytic and collagenous colitis. In a case-control study that included 95 cases of
microscopic colitis, exposure to PPIs was significantly higher in patients with
microscopic colitis as compared with controls (38 versus 13 percent, OR 4.5, 95% CI
2.0-9.5) [41]. Similar results have been reported in other case-control studies, however,
it is unclear if this association varies by PPI and if there is a dose-response relationship
in either dose or duration of use [42,43]. (See "Microscopic (lymphocytic and
collagenous) colitis: Clinical manifestations, diagnosis, and management", section on
'Medications'.)
Hypergastrinemia — Induction of hypergastrinemia has been associated with gastric
carcinoid tumors in rats. However, these observations are not generalizable to species
with gastrin physiology more analogous to humans [44]. While patients treated
with omeprazole for up to 11 years have shown some enterochromaffin-like cell
hyperplasia, no dysplasia or neoplastic changes have been observed [45]. An increased
risk of colon cancer due to hypergastrinemia has also not been established [46].
(See "Physiology of gastrin".)

94
Atrophic gastritis — Patients on long-term PPI therapy have a propensity to develop
chronic atrophic gastritis. However, the risk of atrophic gastritis is small, and in the rare
patient who develops atrophic gastritis, the clinical consequences are uncertain
[45,47,48]. (See "Risk factors for gastric cancer".)
Intestinal colonization of multi-drug resistant organisms — PPIs may increase the
risk of intestinal colonization with multi-drug resistant organisms. In a meta-analysis of
12 observational studies that included 22,305 patients, after adjusting for potential
confounders, acid suppression increased the odds of intestinal carriage of multi-drug
resistant organisms of the Enterobacterales order (producing extended-spectrum beta-
lactamases, carbapenemases, or plasmid-mediated AmpC beta-lactamases) and of
vancomycin-resistant enterococci (OR 1.74; 95% CI 1.4-2.2) [49]. Possible mechanisms
include an increase in bacteria that survive transit from the stomach to the intestine due
to reduction in gastric acid by PPIs and direct alteration of the composition of intestinal
microbiota, leading to a decrease in mean species diversity.  
Inflammatory bowel disease — In a study that pooled data from three observational
cohorts and included >600,000 individuals followed for a median of 12 years, the risk of
inflammatory bowel disease (IBD) was increased in regular PPI users as compared with
nonusers (hazard ratio [HR] 1.42; 95% CI 1.22–1.65) [50]. However, the absolute risk of
IBD was low, with a number needed to harm of 3770. In addition, absence of data on
PPI dosing precluded assessment of a dose-response relationship between PPI use
and IBD.
Malabsorption of minerals and vitamins
Magnesium malabsorption — PPIs can cause hypomagnesemia due to reduced
intestinal absorption [51]. A meta-analysis of nine observational studies that included a
total of 109,798 patients found that those who took a PPI had a significantly higher risk
(RR 1.43; 95% CI 1.08-1.88) of developing hypomagnesemia as compared with those
who did not [52]. Clinical manifestations of hypomagnesemia include neuromuscular
excitability (eg, tremor, tetany, convulsions), weakness, and apathy. Severe PPI-
induced hypomagnesemia has been associated with QT interval prolongation and
torsades de pointes [53,54]. The risk of hypomagnesemia appears to be mainly in
patients who have been on PPIs long-term (generally longer than one year) but cases
have been reported within one year of starting PPI therapy [53,55]. This potential risk
has led to recommendations to monitor serum magnesium levels in specific patients at
high risk for hypomagnesemia. Monitoring for hypomagnesemia in patients on PPIs is
discussed in detail separately. (See 'Laboratory testing' above and "Hypomagnesemia:
Clinical manifestations of magnesium depletion", section on 'Overview of clinical
manifestations'.)
Calcium and fracture risk — Although hypochlorhydria could theoretically reduce
calcium absorption, the effect appears to be relevant only for the absorption of water
insoluble calcium (eg, calcium carbonate) and can be overcome by ingestion of a
slightly acidic meal [56]. The absorption of water soluble calcium salts or calcium in
dairy products are not impacted by PPI-induced hypochlorhydria. When calcium
supplementation is necessary in patients taking PPIs, we use calcium supplements that
do not require acid for absorption, such as calcium citrate. (See 'Laboratory
testing' above and "Drugs that affect bone metabolism", section on 'Proton pump
inhibitors'.)

95
PPI-induced hypochlorhydria can augment osteoclastic activity, thereby decreasing
bone density [57,58]. Although an association between PPI use and bone fracture is
plausible, causality has not been established [59]. Nonetheless, the FDA has mandated
revised safety information on all PPIs about a possible increased risk of fractures of the
hip, wrist, and spine with the use of these medications [60]. The association between
PPIs and bone metabolism and risk of fracture are discussed in detail separately.
(See "Drugs that affect bone metabolism", section on 'Proton pump inhibitors'.)
Vitamin B12 malabsorption — Long-term therapy with PPIs has been associated with
vitamin B12 malabsorption [61,62]. However, absorption of oral B12 supplements is not
affected. (See 'Laboratory testing' above and "Treatment of vitamin B12 and folate
deficiencies", section on 'Treatment of vitamin B12 deficiency'.)
Iron malabsorption — Gastric acid plays a role in the absorption of nonheme iron, and
the use of PPIs has been associated with decreased iron absorption [63-67]. However,
in most cases the decreased absorption does not appear to be of clinical significance.
One exception may be in patients who require oral iron supplementation [66,68]. Such
patients may need a higher dose or longer duration of supplementation [66].
(See "Treatment of iron deficiency anemia in adults", section on 'Dosing and
administration (oral iron)'.)
Kidney disease — PPIs can cause acute interstitial nephritis (AIN) [69-72]. Similar to
other cases of drug-induced AIN, AIN due to PPI use is not dose-dependent, and
recurrence or exacerbation can occur with a second exposure to the same or a related
drug. (See "Clinical manifestations and diagnosis of acute interstitial nephritis", section
on 'Drugs'.)
PPI use has also been associated with an increased risk of incident chronic kidney
disease (CKD), CKD progression, and end-stage kidney disease [73-76]. However, the
mechanism underlying the association between PPI use and risk of CKD is not known,
and it is possible that the weak association observed in these studies is due to
methodological limitations (residual confounding) [73,74,77]. Further studies are needed
to help better define an etiologic relationship between PPI use and the development and
worsening of CKD.
Drug-induced lupus — In postmarketing safety surveillance, new onset of cutaneous
lupus erythematosus and systemic lupus erythematosus (SLE), and exacerbation of
existing disease have been reported in patients on PPIs [78-80]. Most cases of CLE-
associated with PPI use are subacute and occur within weeks to years after continuous
PPI therapy. PPI-associated SLE usually occurs days to years after initiating PPI
treatment and typically presents with a rash. Most patients improve within 4 to 12 weeks
of discontinuation of PPI therapy. (See "Drug-induced lupus", section on 'Causative
drugs'.)
Other associations of unclear significance
COVID-19 — It is unclear if PPI use is associated with an increased risk of COVID-19.
In a cross-sectional survey of 86,602 individuals, 53,130 reported prior abdominal pain,
acid reflux, heartburn, and regurgitation symptoms and provided data on H2RA and PPI
use. Of these, 3386 individuals (6.4 percent) self-reported a positive COVID-19 test
[81]. In analyses adjusted for socioeconomic, lifestyle, and clinical comorbidities,
patients who reported PPI use were significantly more likely to report a positive COVID-
19 test with a dose-dependent increase in odds of reporting a positive test (PPI once-

96
daily OR 2.15, 95% CI 1.9-2.4; PPI twice-daily OR 3.7, 95% CI 2.9-4.6). It is possible
that this association is due to residual confounding.
Other studies have demonstrated that patients taking PPIs are at increased risk for
severe clinical outcomes of COVID-19 but have not demonstrated an increase in
susceptibility to SARS-CoV-2 infection [82]. These data require further validation.
Dementia — Although some studies have found a significant association between use
of PPIs and incident dementia, others have not found an association between PPI use
and cognitive function [83-88]. The association between PPI use and dementia may
reflect residual confounding by factors related to both use of PPIs and the development
of dementia and is discussed in detail, separately. (See "Epidemiology, pathology, and
pathogenesis of Alzheimer disease", section on 'Medications'.)
Pneumonia — While observational studies suggest an association between PPI use
and pneumonia, the observed association may be due to confounding such that
individuals prescribed PPIs may be more likely to have other unobserved health
characteristics that predispose them to pneumonia as compared with nonusers [89-96].
(See "Epidemiology, pathogenesis, and microbiology of community-acquired pneumonia
in adults", section on 'Predisposing host conditions' and "Risk factors and prevention of
hospital-acquired and ventilator-associated pneumonia in adults", section on 'Role of
gastric pH'.)
Mortality — It is unclear if PPI use is associated with an increase in risk of death. In an
observational cohort study that included 275,977 new PPI users and 73,335 new
histamine-2 receptor antagonist (H2RA) users, over a median follow-up of 5.7 years, the
incident death rate among new PPI users was higher as compared to those receiving
H2RA (4.5 versus 3.3 per 100 person-years) [97]. New PPI users were significantly
older as compared with new users of H2RAs at the time of study entry (61.7 versus 58.5
years). However, after adjusting for potential confounders, PPI use was associated with
an increase in all-cause mortality as compared with H2RA use (HR 1.25, 95% CI 1.23-
1.28). PPI users also had an increase in risk of death as compared with individuals
without any PPI use and individuals without any acid suppression use (HR 1.15, 95% CI
1.14-1.15; HR 1.23, 95% CI 1.22-1.24, respectively). Among new PPI users, the risk of
death increased with the duration of PPI use. Limitations of the study include its
generalizability as the study cohort primarily consisted of older White males, and lack of
data on the cause of mortality. The underlying basis for this apparent increased risk of
death with PPI use are not known and further studies are needed to evaluate whether
this epidemiologic association is due to unmeasured confounding.
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

97
Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Acid reflux and gastroesophageal reflux

disease in adults (The Basics)")
●Beyond the Basics topics (see "Patient education: Gastroesophageal reflux
disease in adults (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Indications for proton pump inhibitors (PPIs) include the treatment of peptic ulcer
disease, gastroesophageal reflux disease, and Zollinger-Ellison syndrome. PPIs
are effective in the prevention of nonsteroidal anti-inflammatory drug-associated
gastroduodenal mucosal injury and are an important component of several
antimicrobial regimens used in the treatment of Helicobacter pylori infection.
(See 'Indications for PPI therapy' above.)
●PPIs effectively block gastric acid secretion by irreversibly binding to and
inhibiting the hydrogen-potassium ATPase pump that resides on the luminal
surface of the parietal cell membrane. Genetically determined variability in the PPI
metabolism can influence their efficacy. The presence of a CYP2C19 gene
mutation can results in higher plasma PPI levels in homozygous individuals. PPI
metabolism via hepatic cytochrome P450 enzymes may lead to specific drug
interactions in some individuals. However, clinically important drug interactions
with PPIs are rare. (See 'Pharmacology' above.)
●The magnitude of pharmacokinetic differences between PPIs is small and the
clinical relevance of these differences has not been established (table 1).
Intravenous PPI therapy is indicated in patients with clinically significant upper
gastrointestinal bleeding prior to endoscopy for treatment of suspected bleeding
peptic ulcers. (See 'Administration' above.)
●Oral PPIs should be administered 30 to 60 minutes before breakfast for maximal
inhibition of proton pumps. PPIs should not be administered concomitantly with
H2-receptor antagonists, and prostaglandins or somatostatin analogues. In
general, PPIs should be prescribed at the lowest dose and for the shortest
duration appropriate to the condition being treated. When discontinuing PPI
therapy, we taper the dose gradually in patients on PPIs for longer than six
months. (See 'Oral regimen' above.)
●PPI use has been associated with an increased risk
of Clostridioides difficile infection, other enteric infections, and microscopic
colitis. C. difficile infection with diarrhea may occur even in the absence of
antibiotic use. (See 'Gastrointestinal effects' above.)
●PPIs can cause hypomagnesemia due to reduced intestinal absorption. Long-
term therapy with PPIs has been associated with vitamin B12 malabsorption. We
obtain serum magnesium levels prior to starting a PPI in patients who are
expected to be on long term (≥1 year) treatment, or in patients who take PPIs in
conjunction with other medications associated with hypomagnesemia. In addition,
we also monitor magnesium and vitamin B12 levels in patients on long-term PPIs.

98
(See 'Pretreatment considerations and monitoring' above and 'Magnesium
malabsorption' above and 'Vitamin B12 malabsorption' above.)
●Although an association between PPIs and bone fracture is plausible, causality
has not been established. PPIs can decrease the absorption of water insoluble
calcium (eg, calcium carbonate). When calcium supplementation is necessary in
patients taking PPIs, we use calcium supplements that do not require acid for
absorption, such as calcium citrate. (See 'Calcium and fracture risk' above.)
●PPIs can cause acute interstitial nephritis. PPI use has also been associated with
an increased risk of incident chronic kidney disease (CKD), CKD progression, and
end-stage renal disease. However, further studies are needed to help better
define an etiologic relationship between PPI use and the development and
worsening of CKD. (See 'Kidney disease' above.)
●There are conflicting data on the association between PPI use and risk of
dementia and pneumonia. It is also unclear if PPI use is associated with an
increased risk of death. It is possible that these associations are due to residual
confounding and more studies are needed. (See 'Kidney disease' above
and 'Dementia' above and 'Mortality' above.)

99
Surgical management of gastroesophageal reflux in
adults
Author:
Steven D Schwaitzberg, MD, FACS
Section Editors:
Brian E Louie, MD, MHA, MPH, FRCSC, FACS
Nicholas J Talley, MD, PhD
Deputy Editor:
Wenliang Chen, MD, PhD
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Oct 05, 2021.
INTRODUCTIONGastroesophageal reflux disease (GERD) is defined according to
the Montreal consensus as "a condition which develops when the reflux of stomach
contents causes troublesome symptoms and/or complications" [1]. It is manifested by a
spectrum of nonspecific symptoms, including heartburn, regurgitation, dysphagia,
laryngitis, dental problems, adult-onset asthma, and/or aspiration pneumonia. The
prevalence of GERD is high and increasing, particularly in developed countries (eg, 20
percent in the United States) [2].
Lifestyle modification and medications are the first-line treatment for GERD. Surgical
management is generally reserved for patients who have persistent symptoms or
develop complications despite optimal medical therapy [3]. Additionally, patients who
are unable to tolerate, noncompliant with, or unwilling to take lifelong medications are
also surgical candidates.
Surgical management of GERD focuses on restoring a physiologic equivalent to the
normal lower esophageal sphincter (LES). Manometric studies correlate GERD with a
lower mean LES pressure, shorter mean intra-abdominal LES length, and shorter
overall sphincter length [4]. Each of these problems can be corrected by specific
surgical procedures.

The surgical management of adult patients with GERD is reviewed in this topic. The
pathophysiology, diagnosis, medical management, and complications of GERD are
discussed elsewhere:

●(See "Pathophysiology of reflux esophagitis".)

●(See "Clinical manifestations and diagnosis of gastroesophageal reflux in
adults".)
●(See "Medical management of gastroesophageal reflux disease in adults".)
●(See "Approach to refractory gastroesophageal reflux disease in adults".)
●(See "Non-acid reflux: Clinical manifestations, diagnosis, and management".)
●(See "Complications of gastroesophageal reflux in adults".)
●(See "Gastroesophageal reflux and asthma".)
●(See "Radiofrequency treatment for gastroesophageal reflux disease".)

100
Management of GERD in neonates, infants, or pediatric patients is discussed
separately:

●(See "Gastroesophageal reflux in premature infants".)

●(See "Gastroesophageal reflux in infants".)
●(See "Management of gastroesophageal reflux disease in children and
adolescents".)
INDICATIONS FOR OPERATIONAntireflux surgery is most often performed to
control gastrointestinal symptoms (eg, heartburn and regurgitation) that are refractory to
medical therapy. It may also be performed for nongastrointestinal symptoms (eg,
chronic cough, laryngeal disease, and asthma) when there is solid objective evidence to
attribute such symptoms to reflux. The gastrointestinal symptoms are also referred to as
typical symptoms and the nongastrointestinal symptoms as atypical symptoms.
Gastrointestinal indications
Failed medical management — The most frequent indication for antireflux surgery is
moderate-to-severe gastroesophageal reflux disease (GERD) incompletely controlled
by optimal medical therapy, which consists of both drug therapy and lifestyle
modifications [5]. Ten to 40 percent of patients continue to have substantial symptoms
of "reflux" despite elimination of the heartburn component of their GERD and may wish
to consider operation on this basis alone [6].
There is consensus that patients with documented pathologic acid reflux who have
complete or partial response to proton pump inhibitors (PPIs) are good candidates for
one of the antireflux procedures if continued medical therapy is not desired. The choice
of procedure depends on whether a clinically significant hiatal hernia is present
(laparoscopic fundoplication [LF] or magnetic sphincter augmentation [MSA]) or absent
(LF, MSA, or transoral incisionless fundoplication [TIF]). (See 'Choice of
procedure' below.)
However, there is less agreement on whether patients with documented pathologic acid
reflux who do not respond to maximal PPI therapy should be offered any antireflux
procedures [7]. As an example, an expert panel of 14 gastroenterologists recommended
impendence testing for all PPI nonresponders and LF or MSA only for patients with
elevated esophageal acid exposure time or those with a positive symptom-reflux
association for regurgitation and a large hiatal hernia; TIF was not recommended for
any PPI nonresponders [6]. By contrast, another expert panel of eight foregut surgeons
and seven interventional gastroenterologists recommended LF for all PPI
nonresponders with a large hiatal hernia, TIF for all PPI nonresponders without a large
hiatal hernia, and MSA for any PPI nonresponders with regurgitation-predominant
symptoms [8].
At a minimum, continued symptoms despite adequate acid suppression with double-
dose PPIs for over three months should serve as a warning that symptoms may not be
due to excess esophageal acid exposure but to another diagnosis, such as reflux
hypersensitivity, functional heartburn, a malignancy, or extraesophageal disease.
Consequently, such patients should be properly assessed before antireflux surgery.
Those patients who fail to meet objective criteria such as hypotensive lower esophageal
sphincter (LES) or objectively elevated acid exposure should most likely not be offered
a surgical procedure in this setting. (See 'Preoperative evaluation' below and "Approach

101
to refractory gastroesophageal reflux disease in adults", section on 'Diagnostic
strategies and initial management'.)
Intolerance of or noncompliance with medical therapy — Such patients may opt for
antireflux surgery due to quality of life considerations despite successful medical
management of GERD [9]. These include patients who respond to but experience
complications with PPI therapy and young patients with a hypotensive sphincter who do
not want to take medications for the rest of their lives [10].
Complications of GERD — Complications of GERD, such as severe esophagitis
(usually defined as Los Angeles class C or D) or benign peptic stricture, are diagnosed
endoscopically and may be indications for antireflux surgery if patients fail medical
therapy [9].
Surgical intervention for asymptomatic Barrett's esophagus is controversial. It remains
to be established whether surgery is beneficial for preventing esophageal
adenocarcinoma in Barrett's esophagus patients [11,12]. Patients with either Barrett-
associated dysplasia or carcinoma should not undergo antireflux surgery before the
lesion has been eradicated histologically by endoscopic therapy. (See "Barrett's
esophagus: Surveillance and management", section on 'Management of dysplasia or
intramucosal carcinoma'.)
Volume regurgitation — Patients with high-volume reflux may have typical
gastrointestinal symptoms (eg, regurgitation) or atypical, nongastrointestinal symptoms
(eg, cough or asthma due to aspiration, dental erosions). Such patients often fail PPI
therapy because of persistent reflux of weakly acidic or alkaline gastric contents through
an incompetent LES [10]. Surgical correction of the incompetent LES may alleviate their
symptoms. Preoperative evaluation of high-volume or nonacidic reflux patients is more
complicated and usually requires specialized pH testing before surgery. (See "Non-acid
reflux: Clinical manifestations, diagnosis, and management", section on 'Diagnostic
evaluation' and 'Esophageal pH testing' below.)
High-volume reflux should be distinguished from rumination, which is effortless
regurgitation of food. Heartburn may be associated with rumination, causing even more
confusion with GERD. Rumination often responds to behavioral therapy.
(See "Rumination syndrome".)
Nongastrointestinal indications — About one-half of patients with GERD report upper
respiratory symptoms, such as chronic cough, hoarseness, laryngitis, wheezing,
asthma, chronic bronchitis, aspiration, or dental erosion [13,14]. In addition, idiopathic
subglottic stenosis and even laryngeal cancer have also been implicated [15,16].
For patients whose respiratory symptoms are accompanied by typical reflux symptoms
and positive esophageal pH testing, relief of respiratory symptoms might also be
achieved by antireflux surgery. However, the outcome is less favorable for a minority of
patients who either only have respiratory symptoms [17] or have abnormal esophageal
motility [18]. In such patients, careful patient selection by objective esophageal testing
and with multidisciplinary input (otolaryngology, pulmonary, and allergy specialists) can
increase the likelihood of positive results [3,19,20]. (See 'Preoperative
evaluation' below.)
Chronic cough — After a thorough, objective medical workup to document proximal
esophageal exposure, patients with chronic cough associated with GERD can have an

102
excellent rate of resolution (77 to 81 percent) and quality-of-life outcomes with antireflux
surgery [21,22].
Laryngeal disease — A direct causal relationship between pharyngeal acid exposure
and laryngeal diseases remains controversial [23,24]. As such, antireflux surgery is not
to be used as the first-line treatment for posterior laryngitis. However, it is reasonable to
consider surgery for patients who have abnormal pharyngeal acid exposure on a double
probe pH study when medical therapy has been maximized, is not tolerated, or is
impractical [25]. (See "Laryngopharyngeal reflux".)
While some data have shown that laryngeal symptoms can be treated successfully by
laparoscopic fundoplication [26,27], others have suggested that surgery was not
consistently effective in patients who were unresponsive to aggressive PPI therapy [28];
thus, failure to respond to a PPI should serve as a warning that symptoms may not be
relieved with surgery.
Documented response rates for airway diseases are less favorable than those for
fundoplication performed for typical gastroesophageal symptoms. Approximately 70
percent of patients will experience symptom improvement, while approximately 33
percent of patients may remain on medication but often at a far reduced dosage [29].
Asthma — Gastroesophageal reflux is common among patients with asthma.
Reciprocally, respiratory symptoms, including those associated with asthma, are
increased among patients with GERD. However, consensus has not been reached on
the role of antireflux surgery in patients with asthma that is thought to be related to
GERD. (See "Gastroesophageal reflux and asthma".)
A systematic review that included 24 reports (mostly case series and uncontrolled trials)
suggested that surgery improved asthma symptoms, asthma medication use, and
pulmonary function in 79, 88, and 27 percent of patients, respectively [30]. A dramatic
reduction in steroid dependency postoperatively was documented in another study [31].
However, improvement in pulmonary function has not been documented after either
medical or surgical therapy [30,32-34].
PREOPERATIVE EVALUATIONSymptoms of gastroesophageal reflux disease
(GERD) are highly prevalent. As such, symptoms alone, with or without response to
proton pump inhibitor (PPI) therapy, are not specific enough to diagnose GERD. Before
antireflux surgery, objective data from esophageal testing are required to anatomically
and physiologically evaluate the presence and severity of GERD and determine the
indication and best operative approach for each patient.
A consensus panel recommended that upper endoscopy, standard pH testing,
esophageal manometry, and barium esophagram be performed for all patients before
antireflux surgery. Nonstandard pH testing or gastric emptying study may be required
for some patients (table 1):
Upper endoscopy — Esophageal and gastric endoscopy should be performed to
assess the esophageal and gastric mucosa for signs of malignancy prior to proceeding
with an antireflux procedure. Patients with endoscopic findings of severe (Los Angeles
grade C or D) esophagitis, biopsy-proven Barrett's esophagus ≥1 cm, or a benign peptic
stricture may forego pH testing as these criteria establish the diagnosis of pathologic
GERD [35]. (See "Clinical manifestations and diagnosis of gastroesophageal reflux in
adults", section on 'Upper gastrointestinal endoscopy'.)

103
Esophageal pH testing — Patients with typical (gastrointestinal) symptoms of GERD
should undergo standard pH testing. Patients with atypical (nongastrointestinal)
symptoms may require nonstandard pH testing, such as one with dual pH probe or
impedance in order to document proximal and/or nonacidic reflux.
Standard pH testing — Ambulatory pH testing is the gold standard for diagnosing
pathologic GERD. Prior to antireflux surgery, all patients with nonerosive GERD,
including those with Los Angeles grade A or B (mild) esophagitis and those with short-
segment (<1 cm) Barrett's esophagus, should undergo standard pH testing to document
abnormal distal esophageal acid exposure. An abnormal pH test in a PPI-dependent
patient with typical symptoms predicts successful outcomes with antireflux surgery [36].
Standard pH testing can be done via a transnasal catheter for 24 hours or a wireless pH
system for 48 hours after the patient has been off acid suppression for seven days.
Esophageal acid exposure is considered pathological if acid exposure time (AET) is >6
percent [35]. (See "Clinical manifestations and diagnosis of gastroesophageal reflux in
adults", section on 'Ambulatory esophageal pH monitoring'.)
MII (impedance) testing — Multichannel intraluminal impedance (MII) is a catheter-
based method to detect intraluminal bolus movement within the esophagus. MII is
performed in combination with manometry or pH testing. Combined MII and pH (MII-pH)
testing can detect both acid and nonacid gastroesophageal reflux. It should be
recognized that this testing is somewhat controversial and interpretation is operator
dependent.
MII is indicated to document weakly acidic or nonacidic reflux in patients who are
refractory to PPI, have high-volume regurgitation, or have nongastrointestinal
symptoms. This test can also help exclude rumination (effortless regurgitation), which
may be confused with volume reflux, or supragastric belching, which may induce reflux.
(See "Esophageal multichannel intraluminal impedance testing".)
Dual pH probe — A dual pH probe test is usually performed in patients with suspected
laryngopharyngeal reflux (LPR) to document proximal (laryngeal) reflux events [20].
(See "Laryngopharyngeal reflux".)
Symptom association — A correlation between a patient's symptoms and reflux
events can be calculated by computer software, which associates a symptom with a
reflux event if the symptom occurs within two minutes of the reflux event. Symptom
association is usually presented as either the symptom index (SI) or symptomatic
association probability (SAP). An SI >50 percent or an SAP >95 percent is considered
positive.

Symptom association scores are most useful in determining if a particular symptom (eg,
cough) is associated with acid reflux. In surgical practice, SI, and especially SAP, are
primarily used in patients who have atypical but no typical symptoms. For those with
typical symptoms and a positive acid exposure time, SI or SAP is hardly relevant.

Decisions to proceed with antireflux surgery, however, should not be made on the basis
of a positive SI or SAP alone. The calculation of SI and SAP is performed by the
computer software (of standard and MII pH testing) without manual reading of the
tracings and is highly dependent on the number of symptoms provided by the patient
during the testing period. Additionally, SI and SAP have only been validated for acid
reflux, not nonacid reflux (as detectable by MII) [37].
104
Esophageal manometry — Esophageal manometry is the most reliable way to assess
lower esophageal sphincter (LES) competence and esophageal peristalsis. Overall, the
LES is incompetent in about 60 percent of patients with GERD [38], while transient
relaxation of a competent LES is the cause of GERD in the remaining 40 percent [39].
Manometry assesses esophageal peristalsis as well and occasionally provides
alternative diagnoses, such as scleroderma or achalasia, for which antireflux surgery
may be contraindicated. Manometric findings may influence the surgical approach (eg,
partial instead of complete fundoplication for those with weak peristalsis) or a change in
management, although the need to modify the surgical procedure based upon the
manometric findings has been questioned [40].
Additionally, manometry is also used to determine the precise location of the
gastroesophageal junction for accurate pH catheter placement. (See "Overview of
gastrointestinal motility testing", section on 'Esophageal manometry'.)
Barium esophagram — Barium esophagram (ie, upper gastrointestinal series) can
demonstrate esophageal length, presence and size of any hiatal hernia, presence of
any esophageal diverticulum or stricture, and the extent of reflux with provocation.
Although primarily an intraoperative decision, the presence of a markedly shortened
esophagus or a large hiatal hernia that does not reduce in the upright position may alert
the surgeon to the necessity of a Collis gastroplasty or a transthoracic approach [41].
(See 'Shortened esophagus' below.)
Gastric emptying study — A four-hour gastric emptying study should be performed
when the history of a patient with GERD symptoms suggests gastric outlet obstruction
or gastroparesis (eg, significant nausea, repeated vomiting, severe bloating or
postprandial fullness, or retained food in the stomach after overnight fast).
A combination of antireflux procedure and significantly impaired gastric emptying will
lead to bloating and discomfort. Some surgeons will choose to add a pyloroplasty to
improve gastric emptying when serious dysfunction is revealed [42,43].
(See "Gastroparesis: Etiology, clinical manifestations, and diagnosis", section on
'Scintigraphic gastric emptying'.)
CHOICE OF PROCEDUREThere is no one best operation for all patients with
gastroesophageal reflux disease (GERD). Currently available antireflux procedures
include:
●Radiofrequency treatment (Stretta) – Endoscopic
●Transoral incisionless fundoplication (TIF) – Endoscopic
●Magnetic sphincter augmentation (MSA) – Surgical
●Laparoscopic Hill gastropexy – Surgical
●Laparoscopic partial fundoplication – Surgical
●Laparoscopic Nissen (complete) fundoplication – Surgical

These procedures vary by efficacy and durability on one hand and adverse effect
profiles on the other. At one end of the spectrum, laparoscopic Nissen fundoplication is
highly effective in relieving GERD symptoms and is the most durable amongst all the
procedures; however, it is also associated with the greatest potential for adverse
effects, such as dysphagia, difficulty in vomiting, and gas bloating. At the other end of
the spectrum, endoscopic procedures such as Stretta and TIF are the most
"physiologic" procedures and therefore the least likely to be associated with adverse

105
effects. However, their efficacy and durability are not as good as those of a complete
fundoplication. Partial fundoplications, Hill procedure, and LINX generally fall in the
middle of the spectrum both in terms of efficacy/durability and adverse effect profile.

When the requisite surgical and endoscopic expertise is available, patients should
ideally be offered a spectrum of rather than a single antireflux procedure. As such,
patients can choose the procedure that affords the right combination of
efficacy/durability and potential adverse effects/perturbation to gastrointestinal
physiology to them. The shared decision-making process also sets the correct
expectation for surgical outcomes.

In addition to patient preference, anatomic and patient/surgeon factors such as the

degree of esophageal shortening, disturbances of esophageal motility, prior operations,
and local expertise also influence the choice of operation:

Early uncomplicated disease — For patients with normal esophageal length and

motility (ie, most patients with early, uncomplicated disease), the operation of choice in
the United States is probably a laparoscopic Nissen fundoplication [44,45]. Studies
comparing partial fundoplication to the 360 degree Nissen fundoplication have
consistently reported less postoperative dysphagia with partial fundoplication but
greater long-term durability with complete fundoplication [46]. (See 'Comparison of
partial with complete fundoplication' below.)
Complicated disease — Fundoplication remains the standard of care for patients with
GERD complicated by hiatal hernia >2 cm, severe (Los Angeles class C or D) erosive
esophagitis, and/or Barrett's esophagus [47]. Such complications are contraindications
to less invasive procedures such as TIF or MSA.
Decreased esophageal motility — For patients with normal esophageal length but
decreased esophageal motility, a laparoscopic partial fundoplication (eg, Toupet) or Hill
procedure, rather than a complete (Nissen) fundoplication, should be performed. While
an incomplete wrap is often recommended in patients with poor esophageal motility
[48,49], the results of at least one trial raised questions about this recommendation [40].
Shortened esophagus — Patients with a shortened esophagus from chronic
inflammation or altered anatomy present a unique challenge. Although opinions vary
regarding what constitutes adequate esophageal mobilization, it is generally accepted
that increasing the intra-abdominal esophageal length can also be facilitated by
reduction of hiatal hernia, approximation of the diaphragmatic crura, or tethering of the
distal esophagus below the diaphragm. The bulk of a fundoplication may also keep the
gastroesophageal junction within the abdomen.
In 311 consecutive patients undergoing minimally invasive surgery for GERD and/or
hiatal hernia, the distance between the endoscopically localized gastroesophageal
junction and the apex of the diaphragmatic hiatus after maximal thoracic esophagus
mobilization was <1.5 cm in 31.8 percent of the patients who had "true" shortened
esophagus [50].
If the intra-abdominal esophagus is <2 cm despite the surgeon's best efforts at
esophageal mobilization, a Collis gastroplasty (esophageal-lengthening procedure)

106
combined with a fundoplication should be performed. A survey of about 1000 patients
reported that, at four years, those who underwent a laparoscopic Collis gastroplasty
with fundoplication had similar symptom control, satisfaction, and quality of life to those
who underwent fundoplication alone [51].
Shortened esophagus is discussed in detailed elsewhere as it is more commonly
encountered during paraesophageal hernia repair than pure antireflux surgery.
(See "Surgical management of paraesophageal hernia", section on 'Esophageal
mobilization'.)
Patients with obesity — Obesity is associated with GERD. For patients with severe
obesity, Roux-en-Y gastric bypass (RYGB) is the bariatric procedure of choice for
surgical treatment of GERD. Several small series have reported a decrease in reflux
symptoms as well as complete or partial regression of Barrett's esophagus with RYGB
[52-54]. (See "Laparoscopic Roux-en-Y gastric bypass", section on 'Gastroesophageal
reflux disease' and "Barrett's esophagus: Surveillance and management".)
Transthoracic procedure — In contemporary surgical practice, transthoracic antireflux
procedures are rarely used except in patients with concurrent pulmonary disease
requiring evaluation, extensive prior abdominal surgery, severe obesity, or a shortened
esophagus [41]. The need for adequate mobilization of the esophagus is another
reason cited for transthoracic approaches [55]. Transthoracic approaches for
esophageal reflux are exceedingly rare, except occasionally in reoperative scenarios.
The Belsey Mark IV operation involves a partial fundoplication performed by
transthoracic approach, which allows full esophageal mobilization [56]. The incomplete
fundoplication leads to fewer obstructive symptoms and is therefore recommended for
patients with poor esophageal motility who also have one of the indications for a
transthoracic approach described above.
OPERATIVE TECHNIQUESA variety of antireflux procedures have been
described for the treatment of gastroesophageal reflux disease (GERD) (figure 1A-B).
Fundoplication procedures — The original fundoplication as described by Rudolph
Nissen in 1956 involved passage of the gastric fundus behind the esophagus to encircle
the distal 6 cm of the esophagus. Since that time, many variations and modifications
have been described, and the same-named procedure may be performed differently by
different surgeons.
Important variables of a fundoplication procedure include approach (transthoracic or
abdominal), portion of stomach wall used (anterior and posterior or anterior only),
combination with other procedures (eg, vagotomy or gastroplasty), the looseness of the
wrap, the completeness of the wrap, and the length of the wrap [55].
In contemporary surgical practice, fundoplication is performed laparoscopically or
robotically in most patients. Compared with open transabdominal or transthoracic
fundoplication, laparoscopic or robotic fundoplication is endorsed by the Society of
American Gastrointestinal and Endoscopic Surgeons (SAGES) [57] because it
demonstrates comparable safety, efficacy in symptom relief [58], and patient
satisfaction as well as shorter hospital stays and recuperative times [59-64] and fewer
incisional hernias over time [65].
The reported learning curve of laparoscopic fundoplication varies among studies but
ranges from 20 to 60 [66-68]. Although robotic antireflux surgery offers no specific
advantages over other laparoscopic techniques and is more costly to perform [69], its
107
use has increased from 5.4 to 26 percent from 2012 to 2018, according to data from the
Michigan Surgical Quality Collaborative [70].
Nissen (360 degree or complete) fundoplication — Most surgeons choose to
perform a loose ("floppy") Nissen fundic wrap that is about 2 to 3 cm in length, followed
by a posterior crural repair [49,55]. Important technical points of a successful
laparoscopic Nissen fundoplication include [10]:
●Extensive mediastinal dissection should be performed, especially when a hiatal
hernia is present, to reduce >2 cm of esophagus to below the diaphragm without
tension. Reduction of hiatal hernia may also contribute to the efficacy of antireflux
surgery [71]. If the intra-abdominal esophagus is <2 cm despite the surgeon's best
efforts at esophageal mobilization, a Collis gastroplasty should be performed.
(See 'Shortened esophagus' above.)
●The short gastric vessels are typically transected to ensure a tension-free
("floppy") fundoplication. The Rosetti technique does not divide the short gastric
vessels. (See 'Rosetti-Nissen fundoplication' below.)
●The lower esophageal sphincter (LES) and fundus normally undergo vagally
mediated relaxation with swallowing. An incorrectly performed fundoplication may
prevent appropriate relaxation of the LES with swallowing. It is therefore important
that the fundus is the only part of the stomach used for reinforcing the LES; the
wrap is placed around the esophagus, not the upper stomach, and the vagal
nerves must not be injured during dissection [49].
●A posterior crural repair should be performed by approximating the right and left
pillar of the right crus with sutures. Mesh placement is optional and controversial.
A trial of 159 patients reported that, compared with suture repair alone, mesh
reinforcement did not decrease the incidence of recurrent hiatal hernia rate but
increased solid dysphagia rate at three years [72]. (See "Surgical management of
paraesophageal hernia", section on 'Closure of hiatal defect'.)
●Advocates of the importance of the gastroesophageal valve support restoration of
the angle of His (between the esophagus and gastric fundus) in order to improve
the competence of the valve [73].
●A 50 to 60 French bougie should be placed in the stomach (along the lesser
curvature) to calibrate the tightness of the fundoplication [74].
●The length of the overall sphincter can be affected by altering the length of a
fundoplication; however, too long a fundoplication may lead to obstructive
symptoms. Most surgeons would perform a wrap of 2 to 3 cm [74].
Nissen fundoplication is the most studied antireflux procedure. In many series, it has
been found superior to other antireflux procedures, with symptomatic improvement
occurring in 85 to 90 percent of patients [55].
A laparoscopic Nissen fundoplication can offer significant advantages over the open
operation with similar efficacy and safety [75] and is considered the current gold
standard of antireflux procedures against which all other procedures are benchmarked.
As an example, in the follow-up study to a randomized trial comparing laparoscopic with
open Nissen fundoplication in the Netherlands, 90 and 95 percent of patients reported
symptom relief at 17 years, with no differences in GERD symptoms or dysphagia [76].
Despite that, 43 and 49 percent of the patients used proton pump inhibitors (PPIs) after
laparoscopic and open surgery, although the indication for the PPI use was not

108
specified. Surgical reinterventions were more frequent after open surgery (18 versus 45
percent), mainly due to incisional hernia corrections (3 versus 14 percent).
In a prospective study of 150 patients followed for 15 years after laparoscopic Nissen
fundoplication, symptom control was maintained in 80 percent of patients (Visick I to II),
who also showed objective improvement according to endoscopy, esophageal
manometry, and pH monitoring [77].
Rosetti-Nissen fundoplication — A common modification is a 360 degree fundic
wrap without division of the short gastric vessels (Rosetti-Nissen). However, most
surgeons in the United States prefer to divide the short gastric vessels to allow for
greater freedom of mobilization and reduced concern about torque on the fundoplication
(which is hard to measure) and also facilitate lower esophageal relaxation [74].
A randomized trial was conducted in the 1990s to compare division versus nondivision
of short gastric vessels during laparoscopic Nissen fundoplication [78]. At 20 year
follow-up, there were no significant differences in heartburn symptoms (mean analog
scores 1.4 versus 2.1/10) and satisfaction scores (mean analog scores 8.1 versus
8.6/10) or medication use, while fewer patients in the nondivision group reported
epigastric bloating (26 versus 50 percent) [79].
The 2021 SAGES guidelines recommended choosing between the standard Nissen
fundoplication and the Rosetti modification based on the patient's values and
preferences. Those who value resolution of reflux symptoms over the risk of potential
gas bloat symptoms should choose division of the short gastric vessel as opposed to
nondivision, and vice versa [57].
Partial fundoplications — A partial 270 degree posterior wrap (Toupet) is used for
patients with severe associated motor abnormalities. A partial 180 degree anterior wrap
(Dor) has also been described [46].
Based on currently available data, the choice of anterior versus posterior or partial
versus complete (Nissen) fundoplication should be left to the individual surgeon. Many
trials and meta-analyses have shown that a well-constructed partial fundoplication can
result in similar reflux control to that of a well-performed complete fundoplication.
(See 'Comparison of partial with complete fundoplication' below and 'Comparison of
anterior with posterior fundoplication' below.)
Hill gastropexy — The Hill procedure involves imbrication of the anterior and posterior
collar sling muscular fibers at the level of the gastroesophageal junction around the
esophagus with tethering of the complex to the median arcuate ligament and closure of
the diaphragm. Intraoperative manometry is used to achieve a desired LES pressure
[56,73]. This operation has also been performed laparoscopically and is advocated by
those who support reconstruction of the angle of His and the importance of the
"gastroesophageal valve" for preventing reflux [73]. It can also be used in a patient with
a small stomach because of prior gastric resection [80].
In a randomized trial, laparoscopic Hill gastropexy and laparoscopic Nissen
fundoplication yielded similar and both excellent outcomes at one-year follow-up [81]. In
a single-center retrospective study with a median follow-up of 18.5 years, 85 percent of
patients who underwent Hill gastropexy reported good to excellent symptomatic
outcomes and quality of life [82].
Magnetic sphincter augmentation (MSA) — In 2012, the US Food and Drug
Administration (FDA) approved the LINX Reflux Management System as a treatment for

109
GERD. The device works by augmenting the LES with a ring made up of a series of rare
earth magnets. The magnets have sufficient attraction to increase the LES closure
pressure but permit food passage with swallowing. (See "Magnetic sphincter
augmentation (MSA)".)
Eligible patients must have GERD confirmed by a pH study. In addition, the
manufacturer of the LINX device also suggests that the patient have no large (>3 cm) or
paraesophageal hiatal hernias, severe (Los Angeles class C or D) esophagitis, Barrett
esophagus, obesity (body mass index >35 kg/m 2), esophageal dysmotility, or prior
upper gastrointestinal tract surgery. Patients who have allergies to titanium, stainless
steel, nickel, or iron should not receive a LINX prosthesis. Patients who have the MSA
device can undergo magnetic resonance imaging in a system up to 1.5 Tesla (1.5 T).
MSA is implanted laparoscopic or robotically. The techniques are discussed in another
topic. (See "Magnetic sphincter augmentation (MSA)", section on 'Implantation
technique'.)
MSA implantations have consistently demonstrated reduced GERD symptoms and
improved GERD-related quality of life scores, cessation of PPI use, and substantial
normalization of objective GERD measurements in the majority of patients [83].
(See "Magnetic sphincter augmentation (MSA)", section on 'Clinical improvements'.)
Dysphagia can occur early or late (>30 days) after MSA implantation by different
mechanisms and is treated differently. Device erosion occurs in 0.3 percent of patients
at four years; between 3 and 7 percent of the MSA devices are explanted by combined
endoscopic and laparoscopic means. (See "Magnetic sphincter augmentation (MSA)",
section on 'Postoperative complications'.)
Endoscopic methods — A number of methods for treating GERD endoscopically have
been developed. Their efficacy (particularly long-term) is still being defined [84].
Stretta — The Stretta procedure is the most widely studied endoscopic antireflux
procedure. The specialized catheter is placed with endoscopic assistance over a
guidewire. Using monopolar energy, a series of 56 treatments is delivered across five
levels. SAGES has performed an extensive evidence-based review of endoluminal
reflux therapies [85]. Stretta is rated as effective therapy in patients with an LES
pressure of at least 8 mmHg and hiatal hernia less than 3 cm [43,86-89].
Stretta and other endoscopic treatments of GERD are discussed in detail elsewhere.
(See "Radiofrequency treatment for gastroesophageal reflux disease".)
Transoral incisionless fundoplication — Transoral incisionless fundoplication (TIF) is
an endoscopic procedure performed under general anesthesia to create a full-thickness
serosa-to-serosa plication that is 3 to 5 cm in length and 200 to 300 degrees in
circumference (partial fundoplication).
TIF can be performed in patients with typical GERD symptoms, no or only low-grade
erosive esophagitis (grades A and B), and no or only small hiatal hernia (≤2 cm). TIF is
contraindicated in patients with high-grade erosive esophagitis, Barrett's esophagus,
atypical and extraesophageal symptoms of GERD, scleroderma, or other esophageal
pathology or surgery [90].
TIF has been compared with PPI therapy in several randomized trials [91]. In the latest
trial (TEMPO), 63 patients with GERD refractory to PPI received TIF versus maximum
standard dose PPI therapy [92]. At six months, both regurgitation and extraesophageal
symptoms were eliminated in more TIF than PPI patients (62 versus 5 percent); 90

110
percent of TIF patients were off PPIs. After six months, all patients in the PPI (control)
group elected to cross over to TIF. At three years, 90 and 88 percent of patients
reported elimination of troublesome regurgitation and all atypical symptoms,
respectively [93]. At five years, troublesome regurgitation was eliminated in 80 percent
of patients; 34 percent were on daily PPI therapy, and the average total GERD health-
related quality of life score improved from 22.2 (baseline) to 6.8 [94].
Although TIF is associated with fewer postoperative adverse effects such as gas
bloating and dysphagia compared with surgical fundoplication [95], the durability of TIF
is less certain than that of surgery. In a meta-analysis of five randomized trials and 13
prospective studies, PPI use after TIF increased over time (albeit at a reduced dose);
the satisfaction rate was 69 percent at six months [96]. This meta-analysis has been
criticized for mixing results from the first and second (current) generation of TIF devices.
The Esophyx device has been CE marked for the European market (2006) and FDA
approved for the United States market (2009) for TIF procedures. In 2017, the FDA
issued an updated clearance to include patients with hiatal hernias >2 cm when a
laparoscopic hiatal hernia repair reduces the hernia to ≤2 cm [97]. In 2016, the
American Medical Association (AMA) issued a specific CPT code for TIF. However,
many private payers still consider the TIF procedure experimental, making
reimbursement challenging.
POSTOPERATIVE SYMPTOMS AND MANAGEMENTTypical symptoms
after fundoplication include dysphagia and gas bloating, which occurred in 11 and 40
percent of patients, respectively, in one trial [98]. Management of patient symptoms
after fundoplication depends upon duration of the symptoms as well as the size and
tightness of the wrap.
In a systematic review and meta-analysis of eight randomized trials, prevalence of
postoperative dysphagia, gas bloating, inability to belch, dilatation for dysphagia, and
reoperation were higher after Nissen fundoplication than Toupet fundoplication, but
subgroup analyses showed that differences with respect to dysphagia disappeared over
time [99]. In a retrospective study of 441 patients who underwent Nissen or a partial
fundoplication, persistent postoperative dysphagia occurred in approximately one-
quarter of patients at a median follow-up of three years but did not differ between the
type of fundoplication [100].
Dysphagia — Most patients will have some degree of postoperative dysphagia after
fundoplication and require a period of modified dietary intake primarily consisting of
liquids for 2 to 12 weeks.
The type of procedure performed is a determinant of postoperative dysphagia. One
report found that dysphagia was more frequent with the laparoscopic Rosetti-Nissen
procedure (11 percent) than with either the laparoscopic Nissen (2 percent) or Toupet (2
percent) procedure [101]. Although preoperative testing cannot reliably predict
postoperative dysphagia, the most common predictor of postoperative dysphagia is the
presence of preoperative dysphagia [102]. Thus, patients with preoperative dysphagia
should be counseled appropriately before surgery. One must also consider the crural
closure as a potential source of dysphagia in the postoperative setting [103].
Dysphagia that persists for more than 12 weeks requires evaluation, which typically
begins with a barium swallow to assess the anatomic placement of the fundoplication.

111
Small migration of the fundoplication in a cephalad direction may result from inadequate
esophageal mobilization or an unrecognized shortened esophagus. However, it is not
clear whether these findings are associated with symptoms and thus can account for
the dysphagia. An asymptomatic patient can be followed expectantly. Patients with
symptoms of dysphagia or perceived reflux should undergo manometry, pH testing, and
endoscopy in order to understand the impact, if any, of cephalad migration of the
fundoplication.

At the time of the barium swallow, patients should be asked to swallow a 13 mm barium

tablet. Patients with dysphagia in whom the 13 mm barium tablet passes slowly through
the esophagus and who had normal motility preoperatively are candidates for dilation.
Approximately 2 to 12 percent of patients required dilation after fundoplication [104-
107].
There is no consensus on the optimal dilation technique (ie, bougie versus guidewire
dilation). The author performs direct bougie dilation, with or without endoscopic
guidance. In the author's experience, the procedure is well tolerated and produces good
results. Tortuous pathways through the fundic wrap are best managed by guidewire
dilation. Pneumatic dilation is very rarely needed. (See "Endoscopic interventions for
nonmalignant esophageal strictures in adults".)
Patients who have a 360 degree fundoplication may be candidates for revision to a
partial fundoplication if dysphagia persists and effective barium tablet passage cannot
be established despite dilation.

Some patients describe a "sticking" sensation in their lower or mid chest that they
mistakenly attribute to recurrent gastroesophageal reflux disease (GERD). Such
patients often resume antisecretory medications. However, it is unlikely that patients
whose fundoplication is functionally intact have persistent GERD. Thus, we suggest that
they be objectively studied prior to restarting antisecretory medications to identify those
who require dilation or a revision.

A subset of patients have persistent symptoms without objective evidence of

esophageal dysfunction radiographically or by esophageal pH studies [108,109]. The
cause of persistent symptoms in such patients is unclear but may involve anatomic,
functional, and psychological factors [110]. Care of such patients should be
individualized. Historically, this would have been a setting where a Bernstein test could
have been used, but this is rarely performed today. Esophageal pH impedance studies
may add diagnostic information in the setting of apparent surgical failure and help direct
next steps (eg, adding a proton pump inhibitor). Medical therapy aimed at treating
esophageal spasm can provide relief in some patients. (See "Clinical manifestations
and diagnosis of gastroesophageal reflux in adults".)
Gas bloat syndrome — Symptoms of gas bloat syndrome (a sensation of intestinal gas
with the inability to belch) can be elicited in a significant number of patients after
fundoplication. The pathogenesis is not well understood, although it was seen more
frequently in the past when longer and tighter fundoplications were created [111].
Symptoms may be due to aerophagia or vagal dysfunction, although this has not been
well studied. It may also be in part due to dysfunction in gastric emptying, which may
have been unrecognized preoperatively or occur secondarily to vagal damage [112].
112
The author performs a solid phase gastric emptying study to confirm the diagnosis of
long-term gastric emptying dysfunction. (See "Gastroparesis: Etiology, clinical
manifestations, and diagnosis", section on 'Scintigraphic gastric emptying'.)
In patients with only mild symptoms, we suggest empiric trials of
chewable simethicone tablets or charcoal caplets. We also instruct patients not to drink
with a straw or ingest carbonated beverages until symptoms resolve, although there is
no evidence that these strategies are effective. Although there are no randomized trials,
an empiric trial with metoclopramide (10 mg four times daily) may be helpful. Short
courses of two to three months are typical, but some patients may require long-term
treatment, placing them at risk for complications including irreversible tardive dyskinesia
(especially in older patients). Domperidone is an alternative to consider and is available
under an investigation new device (IND) program from compounding pharmacies in the
United States. Erythromycin can be considered as an alternative; however, some
patients experience significant gastric distress forcing discontinuation.
Symptoms tend to lessen over time in most patients. In patients with severe persistent
symptoms despite the above treatment approaches, pyloroplasty, pyloric Botox
injection, and pneumatic pyloric dilatation are options in select patients who have
documented gastroparesis [113,114]. Some studies have described a reduced
incidence of gas bloat with partial fundoplication [111]; conversion from a full to a partial
fundoplication has been reported but is rarely required.
MORBIDITY AND MORTALITYLong-term mortality of laparoscopic
fundoplications is very low [115-119]. The reported 30 day surgical mortality rate of
laparoscopic fundoplication is less than 0.1 to 0.2 percent [120].
The most commonly reported acute complications for laparoscopic fundoplications are
gastric or esophageal injury, splenic injury or splenectomy, pneumothorax, bleeding,
pneumonia, fever, wound infections, bloating, and dysphagia. Major acute complications
are uncommon. In a study of 2655 patients, 4.1 percent had a complication within 30
days of surgery, including 1.1 percent infection, 0.9 percent bleeding, and 0.9 percent
esophageal perforation [121].
Long-term complications could be caused by structural problem of the fundoplication
(eg, malposition of the wrap) or functional abnormalities [122]. Up to 30 percent of
laparoscopic fundoplications could be affected by a structural problem [121], which
could negatively impact the patient's quality of life [123]. Some of these long-term
problems require reoperative intervention. (See 'Revisional surgery' below.)
Failure — Antireflux surgery has a failure rate of 10 to 15 percent [124]. Operative
failure is usually defined as persistent, recurrent, or new-onset symptoms. The main
symptoms of operative failure are recurrent reflux symptoms and/or dysphagia.
(See 'Dysphagia' above.)
Consensus has not been achieved on what constitutes treatment success or failure. As
an example, while continued use of antisecretory medications may be considered
treatment failure, many patients on acid suppression medications do not have
documented pathologic reflux [109]. Furthermore, some patients who require
antisecretory medications after surgery still report high quality of life compared with
preoperative status.
In a Swedish population-based study of 2655 patients who underwent antireflux
surgery, 470 patients (17.7 percent) were assumed to have developed recurrent reflux
113
by long-term (six-month) antacid medication use (393 patients [83.6 percent]) or repeat
antireflux surgery (77 patients [16.4 percent]) at a median follow-up of 5.6 years [121].
Risk factors for developing recurrent reflux included female sex (hazard ratio [HR] 1.57,
95% CI 1.29-1.90), older age (HR 1.41, 95% CI 1.10-1.81 for age ≥61 years compared
with ≤45 years), and comorbidity (HR 1.36, 95% CI 1.13-1.65 for Charlson comorbidity
index score ≥1 compared with 0). The recurrence rate is likely underestimated due to
the fact that patients with recurrent reflux who did not take antacid medications
consistently or underwent repeat surgery would have been missed.
Revisional surgery — Approximately 5 to 10 percent of patients will need revisional
surgery after laparoscopic fundoplication [106,125]. The reoperation rate increases with
time: in a national registry study of over 4000 contemporary antireflux operations
(including hiatal hernia), the 1, 5, 10 and 15 year rates of repeat antireflux surgery were
3.1, 9.3, 11.7, and 12.8 percent, respectively [107]. In a series of 109 patients who
underwent revisional surgery after a median time of 26 months, the indications for
revisional surgery were dysphagia (48 percent), reflux (33 percent), paraesophageal
herniation (15 percent), and atypical symptoms (4 percent) [125].
Recurrence of reflux symptoms after a fundoplication requires a thorough workup before
reoperation [126,127] as fundoplication usually fails because of either the wrong
indication for surgery, flawed or insufficient preoperative workup, or technical mistakes
in executing the operation [10]. In the author's experience, the most common technical
reasons for failed antireflux surgery are disruption of the fundoplication and excessive
cephalad migration in the face of unrecognized shortened esophagus.
A reoperative fundoplication is the most commonly performed salvage procedure for
failed fundoplication if additional medical therapy fails [128]. However, the success rate
for revisional surgery is lower than primary surgery, and about 10 percent of patients will
continue to complain of reflux or dysphagia after revisional surgery. Medical therapy
remains the only option at that point. Patients with obesity may consider Roux-en-Y
gastric bypass, which is an extremely effective antireflux operation. Rare patients
undergo distal esophageal resection, but there are no outcomes data for this group.
LONG-TERM EFFICACYLong-term observational studies of laparoscopic
fundoplication performed by experienced operators generally report that 90 to 95
percent of adult patients are satisfied with the results of their surgery [129,130].
Surgical versus medical therapy — The comparison of operative management with
medical management in controlled trials has identified mixed long-term results, with
some studies finding comparable control of symptoms and others reporting better
control by a fundoplication [98,116-119,131-133]. As examples:
●Optimized, contemporary medical (esomeprazole, 20 to 40 mg/day) and surgical
(standardized laparoscopic reflux procedure [LARS]) management provides
similar five-year remission rates for controlling gastroesophageal reflux disease
(GERD) symptoms. In the LOTUS trial, patients who underwent LARS had a
slightly lower estimated five-year remission rate compared with those treated
medically (85 versus 92 percent); this difference was not statistically significant
after controlling for the effects of study dropout [98]. Dysphagia, bloating, and
flatulence were more common in patients treated with LARS (11 versus 5 percent,
40 versus 28 percent, 57 versus 40 percent, respectively), but acid regurgitation
was less common (2 versus 13 percent). Although patients who underwent LARS
114
 had lower residual 24 hour esophageal acid exposure than patients treated
medically (0.7 versus 2.1 percent at six months, and 0.7 versus 1.9 at five years),
pH parameters did not predict treatment success or failure [134].
LARS has not been shown to halt the progression from intestinal metaplasia
(Barrett's esophagus) to dysplasia; the presence of intestinal metaplasia is not an
indication for LARS. At the beginning of this trial, endoscopic findings of intestinal
metaplasia were present in 11.1 and 10.5 percent of patients in the LARS
and esomeprazole groups, respectively, and the prevalences at five years
remained stable in both groups (13.6 percent LARS; 9.3 percent esomeprazole)
[98].
●A multicenter randomized trial (total randomized = 357 patients with greater than
12 months of GERD symptoms requiring maintenance therapy with a proton pump
inhibitor [PPI]) identified better control of GERD symptoms at five years following a
fundoplication [131,132]. The REFLUX trial, which included only 246 patients with
five years of follow-up, found that patients undergoing a fundoplication (either a
total or partial fundoplication per surgeon preference, n = 127) were less likely to
require antireflux medication at five years compared with patients managed by
medical therapy alone (44 versus 82 percent) [132]. However, the study protocol
permitted crossover after randomization, and only 63 percent originally
randomized to a fundoplication had undergone the procedure, and 13 percent
randomized to medical management had undergone a fundoplication.
●In a trial conducted in Veterans Affairs (VA) gastroenterology clinics, 366 patients
with PPI-refractory heartburn who failed to respond to a two-week course of
double-dose omeprazole underwent a systematic workup including endoscopy,
esophageal biopsy, esophageal manometry, and multichannel intraluminal
impedance-pH monitoring (on PPI therapy) [135]. After excluding those who did
not have objective evidence for reflux-related heartburn, 78 patients (21 percent)
were randomly assigned to laparoscopic Nissen fundoplication, active medical
management (omeprazole plus baclofen with or without desipramine), or control
medical management (omeprazole plus placebo). At one year, more patients who
underwent surgery (67 percent) achieved a ≥50 percent decrease in
gastroesophageal reflux disease-health related quality of life score compared with
those who received medical treatments (28 percent active medical management,
12 percent control). In a prespecified subgroup analysis, patients with esophageal
visceral hypersensitivity (symptom association probability [SAP] of >95 percent
alone) responded to surgery as well as those with acid reflux (acid reflux alone or
with SAP of >95 percent).

Differences in results may be related to differences in study design. While the perfect
trial has still yet to be performed, these studies lend credence to the concept that the
optimal outcome achievable when fundoplication is performed by qualified surgeons is
at least as good as medical therapy and possibly somewhat better at five years.

A 2015 Cochrane review of four trials found that, in the short and medium term,
laparoscopic fundoplication was associated with better GERD-specific quality of life and
fewer heartburn or reflux symptoms but a higher risk of adverse events (eg, dysphagia)
compared with medical therapy [133]. However, all of the trials were at high risk of bias;
115
the overall quality of evidence was low or very low, and none of the trials reported long-
term quality-of-life data.
The 2013 American College of Gastroenterology (ACG) practice guidelines stated that
"surgical therapy is as effective as medical therapy for carefully selected patients with
chronic GERD when performed by an experienced surgeon" [136]. The 2021 Society of
American Gastrointestinal and Endoscopic Surgeons (SAGES) guidelines suggested
"managing adult patients with confirmed chronic or chronic refractory gastroesophageal
reflux with surgical fundoplication rather than continued medical treatment (conditional
recommendation based on very low certainty in the evidence of effects)" [57].
(See "Approach to refractory gastroesophageal reflux disease in adults", section on
'Subsequent management'.)
Comparison between surgical options — The comparison of surgical options is a
complex topic since there is heterogeneity of techniques and technical choices that may
impact outcome. As such, the best choice for an individual patient may be the
procedure in which the surgeon is most skilled [10]. In Europe, the two preferred
procedures are laparoscopic Nissen fundoplication and the posterior partial (Toupet)
fundoplication [124]. In North America, laparoscopic Nissen fundoplication is most
commonly performed [9].
Comparison of partial with complete fundoplication — Multiple randomized trials
have compared laparoscopic Nissen (complete) fundoplication with laparoscopic partial
(posterior or anterior) fundoplication. Although some randomized trials have shown that
partial fundoplications have fewer adverse effects (eg, dysphagia, gas bloating) than
complete fundoplication, the available trials are of limited quality and power due to
significant heterogeneity [124].
●Laparoscopic Toupet (posterior partial) fundoplication was compared with
laparoscopic Nissen (posterior complete) fundoplication in a systematic review
and meta-analysis of seven trials [137]. Toupet and Nissen fundoplication resulted
in similar reduction in esophageal acid exposure and reflux symptom control, but
Toupet fundoplication was associated with fewer cases of postoperative
dysphagia and dilation, gas-related symptoms, and reoperation. Another meta-
analysis of 13 trials reached a similar conclusion [138].
●A 180° laparoscopic anterior fundoplication was compared with Nissen
fundoplication in a 2017 systematic review and meta-analysis of six randomized
trials [139]. Anterior and Nissen fundoplications were equivalent in reflux control
and patient satisfaction. While anterior fundoplication resulted in a lower incidence
of postoperative dysphagia, Nissen fundoplication required fewer reoperations for
recurrent symptoms.
●Based on poor-quality studies, a 2014 systematic review and meta-analysis of
two trials and 12 retrospective studies failed to demonstrate superiority of either
complete or partial fundoplication [140].
The 2021 SAGES guidelines suggested choosing between a complete or partial
fundoplication based on the patient's values and preferences. Those who value
resolution of reflux symptoms over the risk of dysphagia should undergo a complete
rather than partial fundoplication and vice versa [57].
Comparison of anterior with posterior fundoplication — The laparoscopic anterior
fundoplication (LAF; 90 to 180° wrap) was proposed as an alternative to the

116
laparoscopic posterior fundoplication (LPF; 180 to 360° wrap) to reduce
postfundoplication symptoms but was reported to have higher rates of recurrence of
reflux [46,141-145].
●A meta-analysis of nine randomized trials totaling 840 patients associated LPF
with better heartburn control but LAF with lower risk of postoperative dysphagia
[146]. Similar patient satisfaction scores and reoperation rates were associated
with LAF and LPF.
Comparison of fundoplication with other surgical options — Newer minimally
invasive antireflux procedures (eg, magnetic sphincter augmentation [MSA] or transoral
incisionless fundoplication [TIF]) have not been around long enough to report long-term
efficacy and adverse effect data. As such, there is insufficient evidence to recommend
them as alternatives to fundoplication for severe GERD [3,124]. Nevertheless, such
minimally invasive procedures may be suitable for patients who wish to avoid potential
adverse effects of fundoplication (eg, dysphagia, gas bloating) but do not desire to
continue lifelong medical therapy for GERD [85].
●MSA has not been compared with fundoplication in randomized trials. In a meta-
analysis of seven observational studies, both procedures were safe and effective
in symptom control with up to one year of follow-up [147]; MSA was associated
with fewer gas bloat symptoms and increased ability to vomit and belch [148]. In a
propensity score matched retrospective study, MSA and Toupet (partial)
fundoplication had similar GERD control and side effect profiles [149].
(See "Magnetic sphincter augmentation (MSA)", section on 'MSA versus
fundoplication'.)
●TIF has not been directly compared with fundoplication in any randomized trial
either. In a network meta-analysis of TIF versus laparoscopic fundoplication
versus PPI therapy, laparoscopic fundoplication was associated with greater
sphincter augmentation than TIF. Quality-of-life improvement scores were actually
higher with TIF, but the follow-up period was shorter [150].
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Gastroesophageal reflux in adults".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

117
 ●Basics topics (see "Patient education: Hiatal hernia (The Basics)")
●Beyond the Basics topics (see "Patient education: Gastroesophageal reflux
disease in adults (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Most authorities consider antireflux procedures an equally effective alternative to
medical therapy for chronic gastroesophageal reflux disease (GERD) when
offered to appropriately selected patients by appropriately skilled
surgeons/endoscopists. (See 'Surgical versus medical therapy' above.)
•The most common indication for antireflux procedures is severe symptoms
incompletely controlled by optimal medical therapy. However, continued
symptoms despite double-dose proton pump inhibitors for over three months
should serve as a warning that symptoms may not be due to excess
esophageal acid exposure. Consequently, such patients should be properly
assessed before antireflux surgery. (See 'Failed medical management' above.)
•Antireflux procedures may also be performed for nongastrointestinal
symptoms (eg, chronic cough, hoarseness, laryngitis, wheezing, asthma,
chronic bronchitis, aspiration, or dental erosion) when there is solid objective
evidence to attribute such symptoms to reflux. However, the outcome is less
favorable for those who either only have nongastrointestinal symptoms or have
abnormal esophageal motility. (See 'Nongastrointestinal indications' above.)
•Despite successful medical management, patients who are intolerant of,
noncompliant with, or do not wish to continue lifelong medical therapy may opt
for antireflux surgery due to quality-of-life considerations. (See 'Intolerance of
or noncompliance with medical therapy' above.)
●For most patients seeking an antireflux procedure, the preoperative evaluation
should include upper endoscopy, standard pH testing, esophageal manometry,
and barium esophagram (table 1). Specialized testing may be required for unusual
symptoms. (See 'Preoperative evaluation' above.)
●Currently available antireflux procedures include:
•Radiofrequency treatment (Stretta) – Endoscopic
•Transoral incisionless fundoplication (TIF) – Endoscopic
•Magnetic sphincter augmentation (MSA) – Surgical
•Laparoscopic Hill gastropexy – Surgical
•Laparoscopic partial fundoplication – Surgical
•Laparoscopic Nissen (complete) fundoplication – Surgical
●For most patients undergoing surgical treatment for GERD, we suggest
laparoscopic Nissen fundoplication (Grade 2C). This procedure appears to be the
most effective and durable among all the procedures but is also associated with
the greatest potential for dysphagia, difficulty in vomiting, and gas bloating.
Patients with decreased esophageal motility or severe obesity may benefit from an
alternative procedure. For an individual patient, the best procedure may be the
procedure in which the surgeon is most skilled. (See 'Choice of procedure' above.)
●Postoperative symptoms of dysphagia and gas bloating occur in 11 and 40
percent of patients, respectively, after fundoplication. Dysphagia that persists for
more than 12 weeks requires evaluation with barium esophagram. Patients in
whom a 13 mm barium tablet passes slowly through the esophagus and who had
118
normal motility preoperatively are candidates for dilation. Gas bloating is managed
medically; most patients improve over time. (See 'Postoperative symptoms and
management' above.)
●Antireflux surgery has a failure rate of 10 to 15 percent. The main symptoms of
operative failure are recurrent reflux symptoms and/or dysphagia. Approximately 5
to 10 percent of patients will need revisional surgery after laparoscopic
fundoplication. (See 'Morbidity and mortality' above.)

119
Association between Helicobacter pylori infection
and duodenal ulcer
Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jun 15, 2021.
INTRODUCTIONThe majority of patients with duodenal ulcer (DU) are infected
with Helicobacter pylori (H. pylori). However, in multicenter trials, H. pylori was absent
in almost 30 percent of patients with an endoscopically documented duodenal ulcer [1].
Studies that have investigated these patients found that they have generally had a
shorter duration of symptoms and that many had regularly used nonsteroidal anti-
inflammatory drugs (NSAIDs) [2-4]. Such patients have a significantly worse outcome,
especially if treated empirically for H. pylori infection. Thus, H. pylori status should be
determined in all ulcer patients before initiating treatment [4]. High acid output may be a
cause of recurrent DU in patients in whom H. pylori has been eradicated [5]. A variety of
other causes are responsible for the remaining cases. (See "Unusual causes of peptic
ulcer disease".)
The link between H. pylori and DU will be reviewed here. The pathophysiology of H.
pylori infection as it relates to gastrointestinal disease in general is discussed
separately. (See "Pathophysiology of and immune response to Helicobacter pylori
infection".)
EVIDENCE LINKING HELICOBACTER PYLORI TO DUODENAL
ULCERSThere are several lines of evidence that implicate H. pylori as a major
etiologic factor in duodenal ulcers (DU):
●H. pylori is present in most patients who have a DU that is not related to NSAID
use
●H. pylori infection is detectable before the occurrence of DU and appears to be a
risk factor for the disorder
●Eradication of H. pylori prevents DU recurrence
Incidence of H. pylori in patients with duodenal ulcer — Early studies noted a high
incidence of H. pylori infection (then called Campylobacter pylori) in patients with DU
[6]; subsequent reviews confirmed that H. pylori is detectable in 80 to 95 percent of
these patients [2,7]. These data were supported by reports which found that the
prevalence of H. pylori is negligible in populations in which ulcer disease is rare [8].
The prevalence of H. pylori in patients with DU is changing in some parts of the world
[9]. While H. pylori infection remains very common in patients from Asia with DU, it is
becoming less common in patients from the United States and parts of Europe. In an
epidemiologic study from China, 1030 patients underwent endoscopy and 73 percent
were found to have H. pylori [10]. Peptic ulcers were present in 17 percent, more than
120
two-thirds of which were duodenal ulcers. Among those with peptic ulcers, the
prevalence of H. pylori infection was 93 percent. On the other hand, in the United
States and parts of Europe, the prevalence of H. pylori in patients with ulcer disease
appears to be falling and is now in the range of 50 to 75 percent. (See "Peptic ulcer
disease: Epidemiology, etiology, and pathogenesis", section on 'H. pylori'.)
While the association between H. pylori and DU is strong, it is not specific. As
examples, H. pylori infection is also found in patients with gastric ulcers (65 to 95
percent), dyspepsia (20 to 60 percent), gastric cancer (70 to 90 percent), and
asymptomatic patients (20 to 45 percent) [2]. (See "Approach to the adult with
dyspepsia" and "Association between Helicobacter pylori infection and gastrointestinal
malignancy".)
H. pylori occurs before manifestations of the disease — Several trials have found
that preexisting H. pylori infection is a risk factor for the development of DU [11-14].
One study, for example, reviewed the cases of more than 5000 Native Hawaiians who
had stored serum from the late 1960s [12]. Of the 65 patients who developed DU over
the next 20 years, 92 percent were H. pylori positive versus 78 percent of controls,
producing an odds ratio of 4.0. Similarly, an endoscopic study that included 526 patients
with long-term follow-up found that the odds ratio of developing DU was 5.0 for those
with preexisting H. pylori infection compared with H. pylori-negative subjects [14].
Eradication of H. pylori reduces disease recurrence — Treatment of H.
pylori infection in patients with DU decreases the incidence of ulcer recurrence. One
meta-analysis examined the recurrence rate for DU after at least six months of follow-
up. The recurrence rate was 6 percent if H. pylori was eradicated and 67 percent if it
was not [15]. A second meta-analysis found recurrence rates of 20 and 56 percent,
respectively [16]. The studies in the meta-analyses used endoscopic findings to define
ulcer recurrence. It should be noted that the rate of symptomatic recurrence is lower.
(See "Approach to refractory peptic ulcer disease".)
PATHOGENESIS OF ULCER FORMATIONThe precise mechanism by
which H. pylori contributes to duodenal ulcer (DU) formation is incompletely understood.
However, the bacterium appears to affect the following aspects of intestinal and
mucosal physiology:
●Increased gastric acid secretion
●Gastric metaplasia
●Immune response
●Mucosal defense mechanisms
There may also be a contribution from a variety of bacterial, host, and environmental
factors that have a role in the pathogenesis of ulcer formation (algorithm 1).
Increased gastric acid secretion — Acute H. pylori infection induces a short period of
hypochlorhydria. In contrast, chronic infection often leads to increases in basal and
stimulated acid output, particularly in patients who develop DUs (figure 1) [17,18]. In
addition, following H. pylori eradication, mean basal and stimulated acid output drops by
50 percent at one month and returns to normal levels by one year [17].
One mechanism by which H. pylori may enhance gastric acid secretion is via increased
release of gastrin. Gastrin is responsible for gastric acid secretion in normal subjects by
two mechanisms:

121
 ●It has a trophic action on parietal cells and histamine-secreting enterochromaffin-
like (ECL) cells
●It stimulates parietal cells largely via the release of histamine
This process is tightly controlled by a second hormone, somatostatin, which is a potent
inhibitor of gastrin synthesis, gastrin release, and gastric acid secretion.
(See "Physiology of gastric acid secretion".)
Patients with H. pylori infection have elevated basal and stimulated concentrations of
serum gastrin [17,19] and a decreased concentration of somatostatin [20]. The peak
acid output achieved after stimulation with gastrin releasing peptide, regarded as a
measure of the stomach's functional response to endogenous gastrin, is also increased
threefold in healthy H. pylori positive patients, and sixfold in those with DU [17,18].
These values completely normalize within one year of H. pylori eradication (figure 1)
[21].
The peak acid output after pentagastrin stimulation (PAOPg), a measure of the
functional parietal cell population, is also significantly increased in H. pylori-associated
DU disease. However, there are conflicting data regarding the effect of H.
pylori eradication on PAOPg [17,22,23].
The mechanism of the increased acid response in patients infected with H. pylori who
have DU disease compared with infected healthy patients is unclear. One study
suggests that it may be due to the combination of decreased sensitivity to gastrin in
healthy patients, and increased maximal acid secretory capacity in those with DU [24].
Suppression of somatostatin (rather than an increase in gastrin-secreting cells) is
probably the initial event that then leads to the hypergastrinemia that is evident in
patients with DU and H. pylori. This hypothesis is supported by a study of 28 infected
patients with DU that found that after the infection was successfully eradicated, the
median density of somatostatin-immunoreactive cells (antral D cells) increased from 9
to 19 cells/mm muscularis mucosa, while the median somatostatin mRNA/rRNA ratio
increased from 50 to 95 [25]. In contrast, the number of gastrin-immunoreactive cells
and the quantity of gastrin mRNA did not change significantly. The role of H. pylori in
this process is supported by studies that have shown eradication of infection reverses
the abnormalities in gastrin and gastric acid secretion (figure 1) [17,22].
Nevertheless, hypergastrinemia alone probably does not explain the increases in acid
output in patients infected with H. pylori; gastrin levels often return to normal within one
month after eradication, while mean basal and peak acid output remain elevated for
longer periods of time. In addition, H. pylori itself does not appear to alter the sensitivity
of gastric parietal cells to gastrin [26], although hypergastrinemia may have trophic
effects over time that result in increased parietal cell mass.
Gastric metaplasia — Gastric metaplasia refers to the presence of gastric epithelium in
the first portion of the duodenum. This abnormality appears to be a response of the
mucosa to excessive acid exposure since it only occurs when the luminal pH is less
than 2.5 [27]. In addition to acid hypersecretion, impaired duodenal bicarbonate
secretion induced by H. pylori also may contribute to the low duodenal luminal pH [28].
The metaplastic foci provide areas for H. pylori colonization, and probably have a role in
the development of duodenitis. In one study, for example, 30 of 34 (88 percent) patients
with active duodenitis had more than 5 percent gastric metaplasia in duodenal mucosal
specimens and H. pylori-associated gastritis [27]. In contrast, these two factors

122
coexisted in only 0.43 percent of patients with no duodenal inflammation. In addition,
when H. pylori was seen histologically in duodenal biopsy specimens, it was confined
only to areas of gastric metaplasia, and never occurred in the absence of a
polymorphonuclear infiltrate. A more recent study confirmed that the extent of duodenal
gastric metaplasia was fourfold greater in H. pylori-infected duodenal ulcer patients than
in infected asymptomatic controls [29]. Colonization of the duodenal bulb by H.
pylori CagA+ strains was also greater in DU patients than in infected non DU controls
(80 versus 30 percent), even though prevalence of CagA+ gastric infection was similar
in both groups. These studies support the contention that infection of gastric metaplasia
is important in the pathogenesis on DU.
H. pylori infection in areas of gastric metaplasia may weaken the mucosa, making it
more susceptible to acid injury. This hypothesis is supported by studies which have
found that gastric metaplasia increases the relative risk for ulceration fivefold; the
presence of H. pylori within these regions increases the risk 50-fold [30].
Gastric metaplasia is not the same as gastric heterotopia, in which (presumably)
congenital patches of gastric-type mucosa occur anywhere in the small or large
intestine, although the two entities may be difficult to distinguish. (See "Metaplastic
(chronic) atrophic gastritis".)
Immune response — Although H. pylori is a noninvasive organism, it stimulates a
robust inflammatory and immune response that may have a role in DU formation. The
response includes increased production of inflammatory cytokines such as interleukin
(IL)-1, IL-6, tumor necrosis factor alpha, and most notably, IL-8. (See "Pathophysiology
of and immune response to Helicobacter pylori infection".)
Mucosal defense factors — H. pylori may downregulate several important mucosal
defense factors.
●Epidermal growth factor (EGF) and transforming growth factor-alpha (TGF alpha)
are potent gastric acid inhibitors and stimuli of mucosal growth and protection.
One study found that the basal and stimulated release of EGF is significantly
increased after eradication of H. pylori, an effect that may play a role in ulcer
healing [31].
●Patients with DUs have decreased proximal duodenal mucosal bicarbonate
production. It is not clear if impaired bicarbonate secretion can be attributed to H.
pylori; however, eradication of the infection does result in normalization of
bicarbonate output [28].
●H. pylori itself releases proteases that degrade normally protective mucous
glycoproteins that overlie the mucosa [32].
Other contributing factors — Only 10 to 15 percent of patients with H. pylori infection
develop ulcer disease, suggesting that other factors are probably important in
determining the outcome of infection. One such factor is the bacterial strain; only strains
with the cytotoxin-associated gene A (cagA), coding for a 128-140 kDa protein (CagA),
coexpress vacuolating cytotoxin (VacA), a toxin that causes cell injury in vitro [33].
Approximately 85 to 100 percent of patients with DU have CagA+ strains, compared
with 30 to 60 percent of infected patients who do not develop ulcers [34]. However,
since CagA+ strains are also associated with gastric cancer, the specificity of the
association makes it less clear how the same strain can lead to two very divergent

123
conditions. (See "Pathophysiology of and immune response to Helicobacter pylori
infection".)
A specific H. pylori gene, dupA, appears to be associated with development of DU.
Those infected with dupA bacteria had more intense antral inflammation, higher levels
of IL-8 and less gastric atrophy, intestinal metaplasia, and gastric cancer, a cytokine
and histologic profile associated with DU disease [35]. One study suggested that the
dupA gene, which encompasses jhp0917 and jhp0918, was associated with an
increased risk of DU and protection against gastric cancers [36]. The 112bp region of H.
pylori dupA may be associated with an increased risk of duodenal ulcer and has the
potential to serve as a biomarker for early detection of DU [37,38].
In addition, genetic factors may help determine the susceptibility to infection with H.
pylori [39]. It has been suggested that patients with H. pylori who develop DU have an
intrinsically higher parietal cell mass or sensitivity to gastrin than H. pylori-positive
healthy adults [24,30,40]. This observation may explain the exaggerated acid response
to various stimuli in those who develop DUs. Genetic factors may also determine
duodenal cytokine response to infection. In one study, duodenal epithelial cells from
patients with a duodenal ulcer had a decreased capacity to produce inflammatory
cytokines compared with controls from patients without a DU [41].
Environmental factors such as smoking and nonsteroidal anti-inflammatory drug
(NSAID) use may also increase the risk of ulcer formation in patients with H.
pylori [32,42-44]. One report noted an increased prevalence of ulcers in H. pylori-
infected smokers compared with infected nonsmokers (73 versus 27 percent) [43].
Multiple studies of variable design have evaluated the relationship between NSAIDs
and H. pylori in the development of peptic ulcer disease. The following conclusions
were reached in a meta-analysis of 25 such studies [45]:
●Users of NSAIDs with H. pylori infection were 61 times more likely to have a
peptic ulcer than non-infected non-NSAID users.
●Either factor alone increased the risk of ulcer disease by approximately 20-fold.
●H. pylori infection and NSAID use increased the risk of ulcer bleeding by 1.8 and
4.9-fold, respectively; when present together they increased the risk of ulcer
bleeding by sixfold.
A subsequent meta-analysis of studies published between 1999 and 2008 with a total of
16,080 patients found that the mean prevalence of H. pylori infection in DUs was 81
percent [46]. If only the last five years' literature was considered, the rate was 77
percent. H. pylori-negative DUs were associated with NSAID use in 21 percent of
patients when studies with infection rates of less than 90 percent were examined. The
majority of the H. pylori-negative ulcers were attributed to false-negative H.
pylori testing, since even under optimal conditions the sensitivity of testing is below 95
percent [47]. Other causes for H. pylori-negative DUs were much less common.
NSAID users with a history of peptic ulcer disease should be tested and treated for H.
pylori if they are infected; there are also data to suggest that testing and treating (if
positive) all patients for H. pylori prior to beginning NSAID use may be warranted
[48,49], although conflicting data have also been reported [50], and this is not currently
the standard of care. (See "NSAIDs (including aspirin): Pathogenesis of gastroduodenal
toxicity".)

124
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)" and "Patient

education: H. pylori infection (The Basics)")
●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●There are several lines of evidence that implicate H. pylori as a major etiologic
factor in duodenal ulcers (DUs) (see 'Evidence linking Helicobacter pylori to
duodenal ulcers' above):
•H. pylori is present in most patients who have a DU that is not related to
nonsteroidal anti-inflammatory drug (NSAID) use
•H. pylori infection is detectable before the occurrence of DU and appears to
be a risk factor for the disorder
•Eradication of H. pylori prevents DU recurrence
●In the United States and parts of Europe, the prevalence of H. pylori in patients
with ulcer disease appears to be falling. It is now in the range of 50 to 75 percent
but remains high in Asia. (See 'Incidence of H. pylori in patients with duodenal
ulcer' above.)
●The precise mechanism by which H. pylori contributes to DU formation is
incompletely understood. However, the bacterium appears to affect the following
aspects of intestinal and mucosal physiology which may be important
(see 'Pathogenesis of ulcer formation' above):
•Increased gastric acid secretion
•Gastric metaplasia
•Immune response
•Mucosal defense mechanisms
●Only 10 to 15 percent of patients with H. pylori infection develop ulcer disease,
suggesting that other factors are probably important in determining the outcome of
infection. (See 'Other contributing factors' above.)

125
●In apparent H. pylori-negative DUs, a thorough search for infection should be
conducted as infection remains the major cause of DUs. (See 'Other contributing
factors' above.)

126
Association between Helicobacter pylori infection
and gastrointestinal malignancy
Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Feb 10, 2020.
INTRODUCTIONSince the discovery of Helicobacter pylori in the 1980s, much has
been learned about this gram-negative spiral bacteria and its associated disease states.
In 1994, the National Institutes of Health Consensus Conference recognized H.
pylori as a cause of gastric and duodenal ulcers. Later that year, the International
Agency for Research on Cancer declared H. pylori to be a group I human carcinogen for
gastric adenocarcinoma [1]. There is also evidence that H. pylori infection is a risk factor
for gastric mucosa-associated lymphomas (MALT lymphomas). (See "Clinical
presentation and diagnosis of primary gastrointestinal lymphomas".)
Despite these clear associations, there is marked individual variability in the outcomes
of H. pylori infection, with most patients having a non-neoplastic rather than neoplastic
process. H. pylori infection is associated with a complex interaction between genetic,
environmental, and bacterial factors, which potentially explains the different outcomes
possible following infection. Until these factors are better defined and their interactions
better understood, practitioners should limit testing for and treating H. pylori to those
situations where there is evidence to support a clinical benefit.
GASTRIC CANCERGastric cancer is one of the most common causes of cancer-
related death in the world [2] (see "Epidemiology of gastric cancer"). Gastric cancers
can be categorized by site of occurrence: gastroesophageal junction, proximal stomach,
and distal stomach (body and antrum). In the 1930s in the United States, distal cancers
were the most common. Over the subsequent 70 years, the incidence of gastric cancer
has fallen primarily due to a reduction in distal cancers. In comparison, an increase in
the incidence of gastroesophageal junction and proximal cancers has been noted during
the past several decades [3,4]. These observations suggest that gastroesophageal and
proximal gastric cancers share a common pathogenesis, which is distinct from that of
distal cancers [5].
Adenocarcinomas, which accounts for more than 90 percent of tumors arising in the
stomach, are of two distinct morphologic types: intestinal-type and diffuse. A sequence
of steps with phenotypic changes in the gastric mucosa has been hypothesized as a
model for carcinogenesis of intestinal type adenocarcinomas: superficial gastritis;
chronic atrophic gastritis; intestinal metaplasia (picture 1); dysplasia; and finally
carcinoma (algorithm 1) [6]. No similar sequence has been described for the diffuse
type. (See "Gastric cancer: Pathology and molecular pathogenesis".)

127
H. pylori can cause chronic active gastritis and atrophic gastritis, early steps in the
carcinogenesis sequence [7,8]. In animal models, H. pylori infection has induced gastric
adenocarcinoma [9]. Furthermore, a number of studies in humans have demonstrated a
clear association between H. pylori infection and gastric adenocarcinoma [10-12]. The
link has been demonstrated in both the intestinal and diffuse subtypes of gastric cancer
[10,13].
The relationship between H. pylori infection and gastric carcinogenesis in humans can
be illustrated by the following observations:
●H. pylori has been identified histologically in the uninvolved mucosa from
stomachs harboring cancers or precancerous changes (eg, atrophic gastritis with
or without accompanying intestinal metaplasia) [14,15].
●Epidemiologic studies demonstrate a strong correlation between H.
pylori seropositivity and gastric cancer. As an example, the EUROGAST study of
17 populations from 13 different countries (11 European countries, the United
States, and Japan) found a sixfold increased risk of gastric cancer in H. pylori-
infected populations compared with uninfected populations [16]. Similar findings
have been noted in nested-case control studies in which the stored serum of
patients with known gastric adenocarcinoma and that of matched controls were
tested for H. pylori IgG antibody. H. pylori infection was associated with odds
ratios ranging from 2.8 to 49 and attributable risks of 46 to 63 percent [12,17-20].
In a nested case control study of Japanese Americans living in Hawaii, for
example, H. pylori seropositivity was present in 94 percent of patients with gastric
cancer compared with 76 percent of matched controls; the odds ratio was 6.0 [18].
●Two meta-analyses of cohort and case control studies examining the relationship
between H. pylori seropositivity and gastric cancer found that H. pylori infection
was associated with a twofold increased risk for developing gastric
adenocarcinoma [10,11]. The relative risk for gastric cancer was greatest for
younger patients (9.29 at age less than 29) in whom the absolute risk is still quite
low [10].
One of the largest prospective studies addressing H. pylori and cancer risk included
1526 Japanese patients of whom 1246 had H. pylori infection [21]. Patients underwent
endoscopy with biopsy at enrollment and then between one and three years after
enrollment. During a mean follow-up of 7.8 years, 36 patients developed gastric cancer
(2.9 percent), all of whom were H. pylori infected. No uninfected patient developed
cancer.
The International Agency for Research on Cancer estimates that 36 and 47 percent of
all gastric cancers in developed and developing countries, respectively, are solely
attributable to H. pylori infection. This accounts for almost 350,000 gastric cancers
annually worldwide. One report indicated that of the 12.7 million new cancers occurring
in 2008, the population attributable fraction due to infections was over 16 percent
for H. pylori [22].
Despite the clear association between H. pylori and gastric adenocarcinomas, only a
minority of infected individuals will develop gastric cancer. It is thought that modulation
of the effects of infection by external, mostly environmental factors (and possibly strain
differences in H. pylori, see below) influence whether infection results in a neoplastic or
non-neoplastic process.

128
Role of H. pylori in carcinogenesis — Several hypotheses have been proposed to
explain the role of H. pylori in carcinogenesis, although the exact mechanism is
incompletely understood [23]. At present, it is believed that bacterial properties, host
response, and environmental factors all play a role.
H. pylori strain differences — The strain of H. pylori also may be a determinant of its
potential to cause cancer or ulcer disease. (See "Pathophysiology of and immune
response to Helicobacter pylori infection", section on 'Bacterial strain differences'.)
Host immune responses — Host genetics that regulate the immune response and
mucosal events that result from infection play important roles in gastric cancer
development in chronically infected individuals.
Cytokine polymorphisms — Certain polymorphisms in IL-1 beta and other cytokines
may confer an increased susceptibility to non-cardia gastric adenocarcinoma caused
by H. pylori by inducing a hypochlorhydric and atrophic response to H. pylori infection
[24-29]. An illustrative study compared IL-1 beta polymorphisms in 393 patients with
gastric cancer with 430 controls [24]. Two specific polymorphisms (IL-1B-31T and IL-
1RN*2) were associated with low acid secretion and gastric atrophy. The authors
concluded that 38 percent of H. pylori-related gastric cancer could be attributed to the
presence of these alleles. IL-1 beta, a potent inhibitor of gastric acid secretion, is
upregulated by the presence of H. pylori.
A similar report compared polymorphisms in genes for several cytokines in patients with
a variety of gastric and esophageal malignancies with a control population [25].
Proinflammatory genotypes of tumor necrosis factor alpha and IL-10 were associated
with more than a doubling of the risk of non-cardia gastric cancer. Carriage of multiple
proinflammatory polymorphisms of IL-1 beta, IL-1 receptor antagonist, tumor necrosis
factor A, and IL-10 conferred even greater risk (OR 2.8 for one, 5.4 for two, and 27.3 for
more than three). By contrast, these polymorphisms were not associated with an
increased risk of esophageal or gastric cardia cancers.
These data suggest that gene polymorphisms influence cytokine expression, gastric
inflammation, and risk for development of precancerous lesions in those infected
with H. pylori. Infection with certain virulent bacterial strain types augments inflammation
and cancer risk, supporting a complex interaction between host and bacterial in the
development of GI pathology [30]. (See "Risk factors for gastric cancer", section on
'Genetic polymorphisms'.)
Neutrophil activation — One hypothesis has been demonstrated in vitro.
CD11a/CD18- and CD11b/CD18-neutrophils, induced by H. pylori infection, interact with
intercellular adhesion molecule-1 (ICAM-1), resulting in the migration of neutrophils to
the site of infection and adhesion to the surface epithelium. The recruited neutrophils
then produce inducible nitric oxide synthase and release nitric oxide and reactive
oxygen metabolites, such as superoxide and hydroxyl ions, which in turn damage DNA.
This is followed by mutation and malignant transformation. (See "Gastric cancer:
Pathology and molecular pathogenesis", section on 'Helicobacter pylori'.) H.
pylori induces oxidative stress in epithelial cells [31].
Epithelial responses — H. pylori and the immune response induce altered rates of
gastric epithelial cell growth and death, which involve various signaling pathways
leading to apoptosis, proliferation, differentiation, and autophagy.

129
Apoptotic pathways — Two important processes in carcinogenesis are apoptosis
(programmed cell death) and hyperproliferation [32]. Following severe DNA damage,
apoptosis occurs as a protective mechanism to prevent replication of mutated DNA.
Atrophic gastritis with destruction and loss of the glands could be the result of
apoptosis. This hypothesis is supported by the finding of an increased rate of antral
apoptosis in H. pylori-infected subjects [33,34], which returns to normal following
eradication therapy [33]. The mechanism by which H. pylori induces apoptosis is
unclear. One study suggested that the organism causes apoptosis by both direct and
indirect mechanisms [35]. In the latter circumstance, H. pylori appears to sensitize
epithelial cells for apoptosis which is induced by proinflammatory stimuli (eg, tumor
necrosis factor alpha). H. pylori enhances expression of the Fas receptor on gastric
epithelial cells and may mediate apoptosis through signaling mechanisms related to the
Fas death receptor [36]. Proliferating cells may be resistant to apoptosis. This would
upset the balance between cell growth and death, leading to hyperproliferation and the
promotion of neoplasia [37]. There is evidence of an increased amount of the anti-
apoptosis protein, Bcl-2, in the setting of gastric dysplasia [38]. Other reports have
found that apoptosis may be due to plasminogen activator inhibitor (PAI)-2, the
expression of which is increased by H. pylori. PAI-2 is increased in gastric cancer [39].
An uncoupling of epithelial proliferation and apoptosis may be a strain-dependent
phenomenon. Hyperproliferation has been seen in CagA-infected patients in whom
apoptosis is not increased [40].
Cell signaling events — One report indicated that c-Src and c-Abl kinases sequentially
phosphorylate CagA [41]. The two phosphorylation events need not occur on the same
CagA molecule but are both required for the biological effects of CagA. Another study
demonstrated that vacuolating cytotoxin and variants in Atg16L1 disrupt autophagy and
promote H. pylori infection in humans. As autophagy protects against infection with H.
pylori, this could contribute to inflammation and eventual carcinogenesis [42]. A
potentially important observation is that the source of gastric cancer may not be from
gastric epithelial cells themselves but rather from bone marrow-derived cells that
differentiate into gastric epithelial cells in the presence of H. pylori [43]. If this
observation is confirmed, it would have significant implications for the treatment of H.
pylori-associated gastric cancer as well as other epithelial cancers associated with
chronic inflammation. (See "Gastric cancer: Pathology and molecular pathogenesis",
section on 'The preneoplastic cascade'.)
Environmental factors
Interaction between H. pylori and diet — The consumption of salted food appears to
increase the possibility of persistent infection with H. pylori infection [44,45]. In addition,
a synergistic interaction between H. pylori infection and salted food intake to increase
the risk of gastric cancer has also been reported in case control studies [46,47]
(see "Bacteriology and epidemiology of Helicobacter pylori infection"). Animal studies
also suggest that H. pylori infection and high-salt intake act synergistically to promote
the development of gastric cancer [48]. In one study in Mongolian gerbils, expression of
the proinflammatory mediators, inducible nitric oxide synthase (iNOS) and
cyclooxygenase-2 (COX-2), were significantly upregulated. Both iNOS and COX-2
overexpression have been demonstrated in gastric cancer [49-52]. On the other hand,
there was no significant effect on these mediators in noninfected animals [53].

130
Studies suggest that H. pylori might affect other dietary associations with gastric cancer
[54,55]. The potential protective effect of dietary antioxidants such as vitamins C and E
and beta-carotene seems to be stronger in those infected by H. pylori, even though
results are inconsistent. The risk of gastric cancer associated with red meat, processed
meat, or endogenous formation of nitrosamines appears to be observed only in patients
infected with H. pylori. A reported genetic polymorphism found to be protective
against H. pylori carcinogenesis in mice, vitamin D3 upregulated protein 1, suggests a
possible link between vitamin D deficiency and propensity for H. pylori infection to
progress into gastric cancer in humans [56].
Hypochlorhydria and ascorbic acid — Another hypothesis involves a role for
hypochlorhydria and ascorbic acid (algorithm 1) [6]. In the sequence of carcinogenesis
from atrophic gastritis to metaplasia, loss of the acid-secreting parietal cells results in an
elevated gastric pH. Nitrate-reducing bacteria proliferate in the stomach and, at the high
pH, nitrite is formed, which can interact with other nitrogen-containing compounds and
with carcinogens. Ascorbic acid may block this nitrosation reaction by scavenging
nitrates and free radicals [6].

The following observations suggest a role for the relative lack of ascorbic acid in the
pathogenesis of gastric cancer:

●The level of ascorbic acid in gastric juice is markedly reduced in the setting of
chronic gastritis, elevated gastric pH, and H. pylori infection; this may be due to
impaired secretion of ascorbic acid due to the chronic gastritis [57].
●Patients with intestinal metaplasia have lower serum levels of ascorbic acid
compared with controls [58].
●Low levels of dietary ascorbic acid can lead to progression of precancerous
lesions to dysplasia and cancer in high-risk individuals [59]. Ascorbic acid
ingestion was associated with a decreased risk of gastric cancer in case-control
studies [60,61].
Hemoglobin A1c — Other host factors may contribute to the development of gastric
cancer in H. pylori-infected individuals. A study of 2603 Japanese subjects aged ≥40
years were stratified into four groups according to baseline hemoglobin A1c (HbA1c)
levels (≤4.9 percent, 5.0 to 5.9 percent, 6.0 to 6.9 percent, and ≥7.0 percent) and
followed up prospectively for 14 years [62]. During the follow-up, 97 subjects developed
gastric cancer. The age- and sex-adjusted incidence of gastric cancer significantly
increased in the two higher HbA1c level groups. This association remained substantially
unchanged even after adjusting for the confounding factors including H.
pylori seropositivity. Among subjects who had both high HbA1c levels (≥6.0 percent)
and H. pylori infection, the risk of gastric cancer was dramatically elevated. The
mechanism whereby elevated blood sugar enhances the risk of gastric cancer is
unclear but the risk of other cancers has been shown to be increased in those with
diabetes mellitus.
Obesity — Obesity has been reported to be associated with gastric cardia
adenocarcinoma [63,64]. A mechanism explaining this association has not been
established but it may be related to H. pylori infection as there is an apparent increased
prevalence of H. pylori infection in obese patients. Another possibility is that
hyperglycemia increases the risk of developing gastric cancer [62]. Collectively, these
131
studies lead to the possibility that eradication of H. pylori in conjunction with weight loss
or better glycemic control might decrease risk of gastric cancer.
Importance of other factors — Gastric carcinogenesis cannot be explained by H.
pylori infection alone as illustrated by the following observations:
●Only a small fraction of H. pylori-infected individuals develop cancer.
●The incidence of gastric cancer varies regionally despite similar prevalence of H.
pylori worldwide [5,65].
●The gastric cancer risk is not increased in patients with H. pylori-related duodenal
ulcer disease. To the contrary, in a study from Sweden evaluating the incidence of
gastric cancer in patients previously hospitalized for gastric or duodenal ulcer, the
incidence was significantly decreased in the group with duodenal ulcer
(standardized incidence ratio 0.6 versus 1.8 in those with gastric ulcer) [66].
The explanation for this protective feature of H. pylori-induced duodenal ulcer is unclear.
One theory is that atrophic gastritis, which is an early step in gastric carcinogenesis
[67], occurs with H. pylori-related gastric ulcers but not duodenal ulcers. Host factors
may influence the susceptibility to H. pylori-induced gastric atrophy. Support for the
relationship between atrophic gastritis and H. pylori infection was derived from a study
which found that an HLA-DQA1 allele appeared to contribute to resistance against H.
pylori-associated gastric atrophy and gastric adenocarcinoma [68].
Strain differences may also provide some explanation. As noted above, infection with H.
pylori that contain a duodenal ulcer promoting gene, DupA, appears to lower the risk for
gastric cancer [69]. Another hypothesis is that duodenal ulcer may be associated with
an increased level of ascorbic acid [70], which may protect against subsequent
development of gastric cancer.

Finally, as noted above, cytokine polymorphisms associated with cancer result in more
diffuse gastritis and low acid secretion, gastric histology, and physiology unusual in
duodenal ulcer patients.

Role of family history — A family history has been associated with a 1.5- to 3-fold
increased risk of gastric cancer [71,72]. Whether this reflects clustering of H.
pylori within families with gastric cancer is uncertain. A case control study suggested
that the two risks were independent [73]. Relatives of patients with gastric cancer are
more likely to be infected by H. pylori than unrelated controls. Infected relatives are also
more likely to have low gastric acid secretion, a known marker/risk factor for gastric
cancer [74]. As noted above, this observation may be explained by hereditary
differences in inflammatory cytokine polymorphisms that determine a host's acid
secretory profile and the degree and distribution of gastric inflammation that results
from H. pylori infection. (See "Risk factors for gastric cancer", section on 'Importance of
Helicobacter pylori infection'.)
Does treatment reduce risk of gastric cancer? — Eradication of H. pylori appears to
reduce the risk of gastric cancer. The magnitude of reduction varies by the baseline
incidence of gastric cancer, but is seen even in populations with low gastric cancer
incidence. A meta-analysis of 27 studies included 48606 H. pylori infected individuals
with 715 incident gastric cancers [75]. Individuals with eradication of H. pylori had a
lower incidence of gastric cancer as compared with those who did not receive
eradication therapy (pooled incidence rate ratio 0.53; 95% CI 0.44-0.64). As compared
132
with individuals in the lowest tertile of baseline cancer incidence, those in the
intermediate and highest tertile of cancer incidence had a greater reduction in gastric
cancer incidence rate with H. pylori eradication (incidence rate ratio 44 and 38 percent,
respectively). The magnitude of benefit was not significantly different between
asymptomatic individuals and those who had undergone endoscopic resection of gastric
cancer. (See "Early gastric cancer: Treatment, natural history, and prognosis", section
on 'Helicobacter pylori infection'.)
Even if treatment does reduce the gastric cancer risk, difficulties with screening for H.
pylori and treatment arise. The cost of screening and treating would be large given the
worldwide prevalence of H. pylori infection. Nevertheless, one study that economically
modeled the cost of screening per year of life saved estimated that, in selected
populations such as Japanese American, serologic screening for H. pylori beginning at
age 50 was more beneficial than breast cancer screening [76]. Another cost-
effectiveness analysis concluded that screening and treatment could be cost-effective if
the cancer risk following eradication could be restored to that of a population that had
never been infected with H. pylori [77].
A number of major medical organizations have issued guidelines related to H.
pylori screening and eradication in high-risk populations. As examples, Asian-Pacific
guidelines and European guidelines support population-based screening in high-risk
settings [78,79]. Screening for gastric cancer is discussed in detail separately.
(See "Gastric cancer screening".)
GASTRIC LYMPHOMAPrimary gastric lymphoma accounts for 3 percent of gastric
neoplasms and 10 percent of lymphomas [80]. The stomach is the most common
extranodal site of lymphoma. Lymphoma can arise from lymph nodes or mucosal areas;
the latter is referred to as a mucosa (gut)-associated lymphoid tissue tumor (MALToma,
MALT-type lymphoma, or MALT lymphoma, now called extranodal marginal zone B-cell
lymphoma of MALT type in the REAL classification), of which the stomach is the most
common site (picture 2A-B). (See "Splenic marginal zone lymphoma".)
Presenting symptoms of gastric lymphoma include epigastric pain (which is the most
common), weight loss, anorexia, vomiting, melena, hematemesis, back pain, and
nausea. The diagnosis is based upon histologic criteria and the presence of B-cell
markers by immunocytochemistry. Histology shows lymphoepithelial changes,
polymorphic cellular content, centrocyte like cells, and reactive germinal centers. High-
grade lymphoma (eg, diffuse large B-cell lymphoma) is distinguished from low-grade
disease when the number of large blast cells exceeds 20 percent [81].
The normal stomach does not contain significant lymphoid tissue [82]. However, H.
pylori-induced gastritis leads to an aggregation of CD4+ lymphocytes and B cells in the
gastric lamina propria. Antigen presentation occurs followed by T cell activation, B cell
proliferation, and lymphoid follicle formation. The gastric follicle resembles those seen in
the ileum in Peyer's patches [83]. A follicle is characterized by a center consisting of
centroblasts and centrocytes. The center is surrounded by a B cell zone referred to as a
mantle. The mantle is enclosed by a marginal zone, which is also comprised of B cells.
A hypothesis has been proposed to describe the development of gastric B-cell
lymphoma of marginal zone type (previously called low-grade MALToma). The antigen-
presenting cell interacts with a CD4+ T-cell. The activated T-cell then binds to a B-cell
with an aberrant ability for unsuppressed proliferation. A population of centrocyte-like B-
133
cells arise to form the marginal zone, thereby representing low-grade lymphoma [83,84].
This hypothesis was supported in a report of two patients with gastric MALToma who
had a previous gastric biopsy several years before the onset of lymphoma [85]. The
lymphomas were shown to arise from a B cell clone at the site of chronic gastritis.
H. pylori infection and MALToma — Multiple studies have demonstrated an
association between H. pylori infection and MALToma, and have begun to elucidate the
mechanisms underlying this association [86-92]. As with gastric cancer, the
development of MALToma may be related to specific H. pylori strains expressing the
CagA protein. In one report, for example, serum IgG antibody to CagA was much more
common in patients with MALToma than an H. pylori-infected control group (95 versus
67 percent) [88]. It is also possible that other species of Helicobacter are involved in the
development of gastric MALT lymphomas. As an example, an association
with H. heilmannii has been described [91,93].
Efficacy of anti-H. pylori therapy — The most dramatic evidence supporting a
pathogenetic role for H. pylori in MALToma is remission of the tumor following
eradication of H. pylori with antibiotic therapy [94-101]. The role of H. pylori treatment in
the management of gastric lymphoma is discussed in detail elsewhere. (See "Treatment
of extranodal marginal zone lymphoma of mucosa associated lymphoid tissue (MALT
lymphoma)", section on 'Stage I or II H. pylori positive'.)
COLON CANCERAn association between H. pylori infection and colorectal polyps
and colorectal cancer has been described but remains controversial [102-113].
The biologic basis for such an association is uncertain. One possibility is elevated
serum gastrin levels in patients with H. pylori infection [103]. Gastrin receptors have
been identified on a variety of colon cancer cell lines, and endogenous serum gastrin
levels have been correlated with the risk of colonic neoplasms. However, studies have
not found an association of serum gastrin levels with an increased risk for colonic
neoplasia [109,112]. (See "Physiology of gastrin".)
PANCREATIC CANCERAn association between H. pylori infection and pancreatic
cancer has been reported [114-118]. In a meta-analysis that included 1083 patents with
pancreatic cancer and 1950 controls, infection with H. pylori was associated with an
increased risk of pancreatic cancer (OR 1.47, 95% CI 1.2-1.8) [119]. On subgroup
analysis, CagA positive H. pylori strains were not associated with an increased risk of
pancreatic cancer. Another report found an association between colonization with non-
CagA H. pylori strains and pancreatic cancer in patients with non-O blood types; no
association was found in patients with non-O blood types infected with CagA positive H.
pylori [116]. (See 'H. pylori strain differences' above.)
A possible mechanism proposed for the association of pancreatic cancer and H.
pylori is the link between pancreatic cancer and chronic hyperacidity [120] (see 'H.
pylori strain differences' above and "Epidemiology and nonfamilial risk factors for
exocrine pancreatic cancer"). Additional studies are needed to confirm the association
of pancreatic cancer and H. pylori infection and also to better support the putative role
of hyperacidity.
HEPATOBILIARY CANCERSeveral studies report an association between biliary
tract carcinoma and infection with H. pylori [121-125]. Although a cause and effect
relationship has not been proven, some have suggested that H. pylori may be involved

134
in the pathogenesis of biliary neoplasms through enhanced biliary cell inflammation and
proliferation [124,126].
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: H. pylori infection (The Basics)")

●Beyond the Basics topics (see "Patient education: Helicobacter pylori infection
and treatment (Beyond the Basics)")
SUMMARY
●Approximately 36 and 47 percent of all gastric cancers in developed and
developing countries, respectively, are solely attributable to H. pylori infection.
This accounts for almost 350,000 gastric cancers annually worldwide.
●Several hypotheses have been proposed to explain the role of H. pylori in
carcinogenesis, although the exact mechanism is incompletely understood.
(See 'Role of H. pylori in carcinogenesis' above.)
●Multiple studies have demonstrated an association between H. pylori infection
and mucosa-associated lymphoid tissue lymphoma (MALToma). The most
dramatic evidence supporting a pathogenetic role for H. pylori in MALToma is
remission of the tumor following eradication of H. pylori with antibiotic therapy.
(See 'Gastric lymphoma' above.)

135
Bacteriology and epidemiology of Helicobacter
pylori infection
Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jul 31, 2020.
INTRODUCTIONGastric organisms were first observed more than 100 years ago
and their association with gastritis has been recognized since the 1970s [1]. The true
implication of these microbes was not fully realized, however, until 1982 when Marshall
and Warren identified and subsequently cultured the gastric bacterium, Campylobacter
pyloridis, later reclassified as Helicobacter pylori (H. pylori) [2]. This organism is now
known to cause chronic gastritis, most peptic ulcers, and gastric adenocarcinoma and
lymphoma. (See appropriate topic reviews.)
Understanding its bacteriology and the unique features that enable it to survive in an
environment as hostile as the stomach are important in appreciating how H.
pylori causes tissue injury and clinical disease [3,4]. The epidemiology of this infection
sheds light on the geographic, ethnic, and racial differences in prevalence of H. pylori-
associated diseases and the changing incidence of these conditions throughout the
world.
This topic reviews the bacteriology and epidemiology of H. pylori. The pathophysiology
and immune response to H. pylori, clinical manifestations of H. pylori infection,
indications for testing for H. pylori, and its treatment are discussed separately.
(See "Pathophysiology of and immune response to Helicobacter pylori
infection" and "Association between Helicobacter pylori infection and duodenal
ulcer" and "Helicobacter pylori and gastroesophageal reflux disease" and "Indications
and diagnostic tests for Helicobacter pylori infection in adults" and "Association between
Helicobacter pylori infection and gastrointestinal malignancy" and "Acute and chronic
gastritis due to Helicobacter pylori" and "Peptic ulcer disease: Epidemiology, etiology,
and pathogenesis".)
BACTERIOLOGY
Microbiology — H. pylori is a spiral shaped, microaerophilic, gram negative bacterium
measuring approximately 3.5 microns in length and 0.5 microns in width. In vitro, it is a
slow growing organism that can be cultured on blood agar or selective media such as
Skirrow's media incubated at 37ºC in a 5 percent oxygen atmosphere for three to seven
days [5]. Small, uniformly sized, translucent bacterial colonies form and the organisms
can be morphologically characterized by Gram stain and their typical spiral or rod
shaped appearance. High power microscopy reveals that the organism has two to
seven unipolar sheathed flagella which enhance its mobility through viscous solutions.

136
If the growth environment is less than ideal, coccoid forms of H. pylori can occasionally
be seen in culture [5]. These coccoid forms are thought to represent an adaptation to
hostile surroundings; they appear to be more resistant and may enable the organism to
survive for periods of time outside the human host in feces or in drinking water.
In addition to morphologic characterization, the organism can be biochemically
characterized as catalase, oxidase, and urease positive. Urease appears to be vital for
its survival and colonization; it is produced in abundance, making up more than 5
percent of the organism's total protein weight. Bacterial urease activity is clinically
important because it forms the basis for several invasive and noninvasive tests to
diagnose infection. (See "Indications and diagnostic tests for Helicobacter pylori
infection in adults".)
Gastric adaptation of H. pylori — The organism's urease, motility, and ability to
adhere to gastric epithelium are factors that allow it to survive and proliferate in the
gastric milieu [6]. Disruption of urease activity, bacterial mobility, or attachment
prevents H. pylori colonization [4]. The interactions of H. pylori proteins underlying these
activities are rapidly being uncovered [7,8].
●Bacterial urease hydrolyzes gastric luminal urea to form ammonia that helps
neutralize gastric acid and form a protective cloud around the organism, enabling
it to penetrate the gastric mucus layer [9]. A specific gene (ureI) exists within H.
pylori's urease gene cluster that encodes for a pH dependent urea channel [10].
As the pH outside the organism drops, the urea channel permits internal
movement of urea to maintain a favorable intracellular pH, allowing the bacterium
to survive in an acid milieu.
●Its spiral shape, flagella, and the mucolytic enzymes which it produces facilitate
its passage through the mucus layer to the gastric surface epithelium [5]. On the
other hand, gastric mucin appears to serve as a natural antibiotic, protecting the
host against H. pylori infection [11].
●H. pylori then attaches to gastric epithelial cells by means of specific receptor-
mediated adhesion (picture 1) [4,12]. Although attachment is dependent upon
binding of bacterial surface adhesins to specific epithelial cell receptors, host
factors can modulate this process. As an example, certain individuals may
express specific surface receptors or greater numbers of receptors, making them
more susceptible to H. pylori attachment and colonization [13].
EPIDEMIOLOGYH. pylori is the most common chronic bacterial infection in humans
[14,15]. Studies involving genetic sequence analysis suggest that humans have been
infected with H. pylori since they first migrated from Africa around 58,000 years ago
[16]. H. pylori has been demonstrated worldwide and in individuals of all ages.
Conservative estimates suggest that 50 percent of the world's population is affected.
Infection is more frequent and acquired at an earlier age in resource-limited countries
compared with industrialized nations [15]. Once acquired, infection persists and may or
may not produce gastroduodenal disease.
In resource-limited nations, where the majority of children are infected before the age of
10, the prevalence in adults peaks at more than 80 percent before age 50 [15,17]. In
developed countries, such as the United States, evidence of infection in children is
unusual but becomes more common during adulthood. Serologic evidence of H. pylori is
rarely found before age 10 but increases to 10 percent in those between 18 and 30
137
years of age and to 50 percent in those older than age 60 [15]. Within any age group,
infection appears to be more common in Black and Hispanic populations compared to
White populations; these differences are probably in part related to socioeconomic
factors [18,19].
The increased prevalence of infection with age was initially thought to represent a
continuing rate of bacterial acquisition throughout one's lifetime. However,
epidemiologic evidence now indicates most infections are acquired during childhood
even in developed countries [15,20]. Most infections were acquired before five years of
age with a declining incidence thereafter in one report from Ireland [21]. Thus, the
frequency of H. pylori infection for any age group in any locality reflects that particular
cohort's rate of bacterial acquisition during childhood years [20]. The organisms can be
cultured from vomitus or diarrheal stools suggesting the potential for transmission
among family members during periods of illness [22,23].
The risk of acquiring H. pylori infection is related to socioeconomic status and living
conditions early in life. Factors such as density of housing, overcrowding, number of
siblings, sharing a bed, and lack of running water have all been linked to a higher
acquisition rate of H. pylori infection [24-26]. More recent studies continue to show that
in resource-limited countries such as Iran, childhood hygiene practices and family
education determine the prevalence of H. pylori infection [27]. The association of H.
pylori infection with level of education, income, and race/ethnicity is not unique to H.
pylori since similar associations have been described with other chronic infections
including cytomegalovirus, herpes simplex virus-1 and hepatitis B [28]. Within a
particular country, a decline in prevalence of H. pylori appears to parallel economic
improvement. In Japan, for example, 70 to 80 percent of adults born before 1950, 45
percent of those born between 1950 and 1960, and 25 percent of those born between
1960 and 1970 are infected [29]. This rapid decline of infection has been attributed to
Japan's post-war economic progress and improvement in sanitation.
The consumption of salted food appears to increase the possibility of persistent
infection with H. pylori infection [30,31]. In addition, a synergistic interaction between H.
pylori infection and salted food intake to increase the risk of gastric cancer has also
been reported in case-control studies [32,33]. (See "Risk factors for gastric cancer".)
Possible hereditary susceptibility — Hereditary susceptibility to H. pylori infection has
not been proven. However, studies do suggest that members of certain ethnic and racial
groups, including Hispanics and African Americans, have a higher rate of infection than
White Americans. These differences are not entirely explained by differences in
socioeconomic status [34].
Twin studies also support genetic susceptibility to infection [35,36]. Monozygotic twins
raised in different households have a greater concordance of H. pylori infection than do
dizygotic twins raised apart. However, twins raised together have a higher concordance
of H. pylori status than twins raised separately, emphasizing the role of childhood
environment in H. pylori acquisition.
H. pylori transmission — The route by which infection occurs remains unknown
[14,37]. Person-to-person transmission of H. pylori through either fecal/oral or oral/oral
exposure seems most likely [23,37]. Humans appear to be the major reservoir of
infection; however, H. pylori has been isolated from primates in captivity and from
domestic cats [38]. It is unclear how these animals originally acquired H. pylori infection.

138
However, at least in the case of cats, isolation of viable organisms from saliva and
gastric juice samples suggests that transmission to humans might occur [39]. One
report described the identification of H. pylori in milk and gastric tissue of sheep
suggesting that sheep may be a natural host for the organism [40]. This may explain the
higher infection rate that has been observed among shepherds compared to their
siblings [41].
Fecal/oral transmission of bacteria is also possible. Contaminated water supplies in
resource-limited countries may serve as an environmental source of bacteria. The
organism remains viable in water for several days and, using polymerase chain reaction
techniques, evidence of H. pylori can be found in most samples of municipal water from
endemic areas of infection [42-44]. Children who regularly swim in rivers, streams,
pools, drink stream water, or eat uncooked vegetables are more likely to be infected
[45]. H. pylori has been cultured from diarrheal stools of children in The Gambia, West
Africa where almost all inhabitants are infected by five years of age [46].
Intrafamilial clustering of infection further supports person-to-person transmission.
Infected individuals are more likely to have infected spouses and children than
uninfected individuals [26,47]. A study of children in Colombia found that the risk of
infection correlated directly with the number of children aged two to nine in the
household while younger children were more likely to be infected if older siblings were
also infected [48]. Isolation of genetically identical strains of H. pylori from multiple
family members [49] as well as custodial patients in the same institution [50] further
supports transmission among persons sharing the same living environment. In addition
to the familial type of transmission that occurs in developed and other nations,
horizontal transmission between persons who do not belong to a core family also
appears to take place in countries where the prevalence of infection is high [51].
Oral/oral transmission of bacteria has yet to be confirmed. Organisms have been
identified in dental plaque [52], but the prevalence can be quite low [53] so it is not
known if this location can serve as a source or reservoir. Dentists and oral hygienists
who have continual occupational exposure to dental plaque do not have a higher
prevalence of H. pylori [54]. On the other hand, infected gastric secretions can serve as
a source of bacterial transmission [55,56].
Iatrogenic infection has been documented following the use of a variety of inadequately
disinfected gastric devices, endoscopes, and endoscopic accessories [56]. In addition,
gastroenterologists and nurses appear to be at increased risk for acquiring H. pylori; this
is presumably due to occupational exposure to infected gastric secretions [24].
Mandated universal precautions, standardized equipment disinfection, and use of video-
endoscopes which remove the instrument channel away from the mouth should reduce
occupational and iatrogenic H. pylori transmission [56].
Reinfection — Reinfection with H. pylori following successful bacterial cure is unusual.
Recurrence of infection most commonly represents recrudescence of the original
bacterial strain. In adults, reacquisition of bacteria occurs in less than 2 percent of
persons per year [57,58], a rate that is similar to primary adult acquisition of infection
[20]. The low reinfection rate in adults supports the lower risk for infection during
adulthood, although acquired immunity conferred by primary infection may also be
important.

139
It has been hypothesized that the reinfection rates may be higher in children, in
resource-limited countries, and in those from low socioeconomic status. However, a
study from China found that the reinfection rates were comparable to reports from the
west (annual recurrence rate of 1 percent) [59]. Similarly, another report found a low
annual reinfection rate (2 percent) in children older than five, regardless of
socioeconomic status [60].
In a systematic review and meta-analysis of132 studies that included 53, 934 patient-
years, the global annual recurrence, reinfection and recrudescence rate of H.
pylori were 4.3, 3.1 and 2.2 percent, respectively [61]. The H. pylori recurrence rate was
inversely to the human development index and directly related to H. pylori prevalence.
Areas with an H. pylori recurrence of 5 percent or greater were Alaska, Mexico, many
countries in Central America, Colombia, Peru, and Chile, Ireland, Greece, Turkey, Iran,
Yemen, Pakistan, India, Bhutan, China, Taiwan, Thailand, Malaysia, and South Korea.
Many countries do not have data of recurrence, but overall it appears H.
pylori recurrence remains a problem, one that is closely associated with socioeconomic
status and sanitary conditions.
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: H. pylori infection (The Basics)")

●Beyond the Basics topics (see "Patient education: Helicobacter pylori infection
and treatment (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Helicobacter pylori (H. pylori) is a spiral shaped, microaerophilic, gram negative
bacterium measuring approximately 3.5 microns in length and 0.5 microns in
width. The organism has two to seven unipolar sheathed flagella which enhance
its mobility through viscous solutions. (See 'Bacteriology' above.)
●The organism can be biochemically characterized as catalase, oxidase, and
urease positive. Urease appears to be vital for its survival and colonization; it is
produced in abundance, making up more than 5 percent of the organism's total
protein weight. Bacterial urease activity is clinically important because it forms the
basis for several invasive and noninvasive tests to diagnose infection.
(See 'Microbiology' above.)

140
●H. pylori has been demonstrated worldwide and in individuals of all ages.
Conservative estimates suggest that 50 percent of the world's population is
affected. Infection is more frequent and acquired at an earlier age in resource-
limited countries compared with industrialized nations. Once acquired, infection
persists and may or may not produce gastroduodenal disease.
(See 'Epidemiology' above.)
●The risk of acquiring H. pylori infection is related to socioeconomic status and
living conditions early in life. Factors such as density of housing, overcrowding,
number of siblings, sharing a bed, and lack of running water have all been linked
to a higher acquisition rate of H. pylori infection. (See 'Epidemiology' above.)
●The route by which infection occurs remains unknown. Person-to-person
transmission of H. pylori through either fecal/oral or oral/oral exposure seems
most likely. (See 'H. pylori transmission' above.)
●Reinfection with H. pylori following successful bacterial cure is unusual.
Recurrence of infection most commonly represents recrudescence of the original
bacterial strain. (See 'Reinfection' above.)

141
Indications and diagnostic tests for Helicobacter
pylori infection in adults
Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Apr 02, 2020.
INTRODUCTIONHelicobacter pylori (H. pylori) is the most prevalent chronic
bacterial infection and is associated with peptic ulcer disease, chronic gastritis, gastric
adenocarcinoma, and gastric mucosa associated lymphoid tissue (MALT) lymphoma [1-
4].
This topic will review the clinical indications for testing for H. pylori, diagnostic tests
for H. pylori, and their interpretation. Our recommendations are largely consistent with
2017 guidelines from the American College of Gastroenterology and the Maastricht V
consensus report [5,6]. The epidemiology, pathophysiology, and immune response
to H. pylori and treatment regimens for H. pylori infection are discussed separately.
(See "Pathophysiology of and immune response to Helicobacter pylori
infection" and "Acute and chronic gastritis due to Helicobacter pylori" and "Association
between Helicobacter pylori infection and gastrointestinal malignancy" and "Association
between Helicobacter pylori infection and duodenal ulcer" and "Bacteriology and
epidemiology of Helicobacter pylori infection" and "Treatment regimens for Helicobacter
pylori in adults".)
INDICATIONS FOR TESTINGTesting for H. pylori should be performed only if the
clinician plans to offer treatment for positive results. Indications for testing include:
●Low grade gastric mucosa associated lymphoid tissue (MALT) lymphoma.
●Active peptic ulcer disease or past history of peptic ulcer if cure of H.
pylori infection has not been documented.
●Early gastric cancer.
Other indications for testing for H. pylori are more controversial. These
indications include [2,7-10]:
●Uninvestigated dyspepsia in patients <60 years without alarm features –
Patients with dyspepsia who are <60 years of age and do not have any alarm
features (table 1) should be tested for H. pylori [11,12]. (See "Approach to the
adult with dyspepsia", section on 'Patient age <60 years'.)
●Prior to chronic treatment with NSAIDs or long-term, low-
dose aspirin use – H. pylori infection is a risk factor for the development of ulcers
and for ulcer bleeding in patients on low-dose aspirin treatment. H. pylori also
increases the risk of NSAID-related peptic complications. (See "NSAIDs (including
aspirin): Pathogenesis of gastroduodenal toxicity" and "NSAIDs (including aspirin):
Primary prevention of gastroduodenal toxicity".)

142
 ●Unexplained iron deficiency – H. pylori can cause iron deficiency and iron
deficiency anemia by interfering with absorption of oral iron. (See "Causes and
diagnosis of iron deficiency and iron deficiency anemia in adults", section on
'Celiac disease/atrophic gastritis/H. pylori' and "Treatment of iron deficiency
anemia in adults", section on 'Approaches to lack of response'.)
●Adults with immune thrombocytopenia – Limited evidence suggests that
eradication of H. pylori infection improves platelet counts in some adult patients
with idiopathic thrombocytopenic purpura. (See "Initial treatment of immune
thrombocytopenia (ITP) in adults", section on 'H. pylori testing'.)
There are insufficient data to support routine testing for H. pylori in all asymptomatic
individuals with a family history of gastric cancer or in patients with lymphocytic gastritis,
hyperplastic gastric polyps, and hyperemesis gravidarum. (See "Gastric polyps", section
on 'Hyperplastic polyps' and "Gastric cancer screening", section on 'Helicobacter pylori
eradication'.)
APPROACH TO DIAGNOSTIC TESTINGThe choice of test used to
diagnose H. pylori depends on whether a patient requires an upper endoscopy for
evaluation of symptoms or surveillance. Endoscopy is not indicated solely for the
purpose of establishing H. pylori status. Other important determinants include the recent
use of medications that can suppress the bacterial load of H. pylori (eg, proton pump
inhibitor therapy [PPI], antibiotics, and bismuth), the prevalence of H. pylori, test
availability, and cost. A suggested approach to initial diagnostic testing in a patient with
suspected H. pylori is outlined in the algorithm (algorithm 1).
Medications that should be discontinued prior to testing — PPI use within one to
two weeks and bismuth/antibiotic use within four weeks of testing can decrease the
sensitivity of all endoscopy-based tests and non-invasive tests of active H.
pylori infection (stool antigen and urea breath test). Patients should be advised to stop
PPI therapy one to two weeks prior to testing. If feasible, testing should performed at
least four weeks after completion of bismuth/antibiotic treatment.
Patient undergoing upper endoscopy — Among patients undergoing upper
endoscopy, the choice of test to diagnose H. pylori and extent of testing varies based on
the clinical presentation and endoscopic findings.
●No active peptic ulcer bleeding
•No recent PPI/bismuth/antibiotic use and no indication for gastric
biopsy – In patients without recent PPI/bismuth/antibiotic use who do not
require biopsies of the stomach for histology, the diagnosis of H. pylori can be
established with a biopsy urease test. (See 'Biopsy urease testing' below.)
•Recent PPI/bismuth/antibiotic use or indication for gastric biopsy – In
patients with visible endoscopic abnormalities (eg, gastric ulcer, gastropathy),
and patients with recent antisecretory/bismuth/antibiotic use, we perform
histology to diagnose H. pylori. Since recent PPI/bismuth/antibiotic use may
diminish numbers of bacteria detected by histology, we perform a urea breath
or stool antigen assay to confirm a negative test result once these medications
have been held for an appropriate length of time. (See 'Medications that should
be discontinued prior to testing' above and 'Histology' below and 'Urea breath
testing' below and 'Stool antigen assay' below.)

143
 •Prior antibiotic treatment failures – In patients with H. pylori that is
refractory to two courses of antibiotic therapy we perform culture and antibiotic
sensitivity testing on gastric biopsies in order to guide treatment.
(See 'Bacterial culture and sensitivity testing' below and 'Confirmation of
eradication' below and "Treatment regimens for Helicobacter pylori in adults".)
●Active peptic ulcer bleeding – In patients with bleeding duodenal or gastric
ulcer on upper endoscopy, we perform a gastric mucosal biopsy at the time of the
initial endoscopy unless it is impractical or difficult, such as with a blood-filled
stomach. A negative biopsy result does not exclude H. pylori in the setting of an
active upper gastrointestinal bleed, and another test (ideally a urea breath test)
should be performed to confirm a negative result [13,14].
If a gastric mucosal biopsy is not obtained at the time of endoscopy, we perform a
urea breath test or stool antigen assay to diagnose H. pylori.
Testing for H. pylori does not need to be performed immediately, and can be
performed once the patient has stopped bleeding and can safely be off PPI
therapy for one to two weeks. (See "Peptic ulcer disease: Clinical manifestations
and diagnosis", section on 'Exclude Helicobacter pylori (H. pylori) infection'.)
Patients not undergoing upper endoscopy — In patients who do not require
endoscopic evaluation, the diagnosis of H. pylori should be made with a test for active
infection (stool antigen assay or urea breath test). In clinical situations where patients
are unable or unwilling to stop PPI therapy one to two weeks prior to testing, positive
test results are true positives; negative results may represent false negatives and
should be confirmed with repeat testing after stopping PPI therapy for one to two weeks.
Serologic testing for H. pylori should be avoided altogether. If performed in areas of low
prevalence of H. pylori, positive results should be confirmed with a test for active
infection prior to initiating eradication therapy. (See 'Serology' below.)
DIAGNOSTIC TESTSDiagnostic testing for H. pylori can be divided into invasive
(endoscopic) and noninvasive techniques, based upon the use of upper endoscopy.
Endoscopy-based tests, the urea breath test, and the stool antigen assay test for
active H. pylori infection.
Endoscopic testing — The diagnosis of H. pylori can usually be established during
endoscopy by one of three methods: biopsy urease test, histology, and much less
commonly by bacterial culture. (See 'Patient undergoing upper endoscopy' above
and 'Confirmation of eradication' below.)
Biopsy urease testing — Gastric biopsy specimens are placed in a medium that
contains urea and a pH reagent. Urease cleaves urea to liberate ammonia, producing
an alkaline pH and a resultant color change. Commercially available urease testing kits
differ in the use of medium (agar versus membrane pad) and testing reagents and the
time to obtain a final result (1 to 24 hours).
The sensitivity and specificity of biopsy urease testing is approximately 90 and 95
percent, respectively [3]. With the rapid urease testing kits, results are obtained in one
hour. One hour sensitivity and specificity of rapid urease tests are comparable (89 to 98
percent and 89 to 93 percent, respectively) to those seen with agar gel tests at 24 hours
and superior to agar gel tests at one hour [15-18]. False-negative results can occur in
patients with acute upper gastrointestinal bleeding or with the use of PPIs, antibiotics, or
bismuth-containing compounds [19]. In such patients it is recommended that samples
144
be taken from the both the gastric antrum and the fundus to increase the sensitivity of
the test [20]. Increasing the number of gastric biopsy specimens from one to four also
increases the sensitivity of the test [21].

Although biopsy urease kits are inexpensive, the cost is greater than that of a simple
diagnostic endoscopy due to the addition of the biopsy and the resultant "upcoding" of
the procedure to esophagogastroduodenoscopy with biopsy. Nevertheless, biopsy
urease testing is less expensive as compared with histology.

Histology — Gastric biopsies can diagnose of H. pylori infection and associated lesions

(eg, atrophic gastritis, intestinal metaplasia, dysplasia, and MALT lymphoma). Biopsies
for histology should be taken from both the antrum and body of the stomach, especially
when looking for evidence of multifocal atrophic gastritis (picture 1A-B) and/or intestinal
metaplasia. The accuracy of histologic diagnosis of H. pylori infection can be improved
by using special stains such as Giemsa or specific immune stains [22]. (See "Acute and
chronic gastritis due to Helicobacter pylori" and "Metaplastic (chronic) atrophic
gastritis".)
The sensitivity and specificity of histology for the diagnosis of H. pylori infection are 95
and 98 percent, respectively. The sensitivity of histology may be decreased in patients
with acute peptic ulcer bleeding and in patients on PPI therapy due to the proximal
migration of H. pylori to the corpus from the antrum. Even in the absence of PPI use,
the density of H. pylori can vary at different sites and interpretation of histologic slides is
associated with interobserver variability [23,24]. (See 'Patient undergoing upper
endoscopy' above.)
Bacterial culture and sensitivity testing — Biopsies for culture should be obtained
before the forceps are contaminated with formalin. The tissue should be placed into a
container with a few drops of saline. This preparation will allow for culture and antibiotic
sensitivity testing. While bacterial culture has high specificity, it has low sensitivity as H.
pylori is difficult to culture. Microcapillary culturing methods and improved transport
media may improve sensitivity [25-27]. (See 'Confirmation of eradication' below
and "Treatment regimens for Helicobacter pylori in adults", section on 'Treatment
failure'.)
Noninvasive testing — Noninvasive tests for the diagnosis of H. pylori include urea
breath testing (UBT), stool antigen testing, and serology. Of these, UBT and stool
antigen assay are tests of active infection. H. pylori serology can be positive in patients
with an active or prior infection.
Urea breath testing — UBT is based upon the hydrolysis of urea by H. pylori to
produce CO2 and ammonia. Urea with a labeled carbon isotope (non-radioactive 13C or
radioactive 14C) is given by mouth; H. pylori liberate labeled CO2 that can be detected
in breath samples. The tests can be performed in 15 to 20 minutes and have similar
cost and accuracy. The dose of radiation in the 14C test is approximately 1 microCi
which is equivalent to one day of background radiation exposure [28]. Even though this
dose of radiation is small, the non-radioactive 13C test is preferred in young children
and pregnant women. (See "Radiation-related risks of imaging", section on 'Special
populations'.)
The sensitivity and specificity of UBT are approximately 88 to 95 percent and 95 to 100
percent, respectively [29]. Thus, false-positive results are uncommon. False-negative
145
results may be observed in patients who are taking PPIs, bismuth, or antibiotics and in
the setting of active peptic ulcer bleeding [19,30]. The sensitivity but not specificity of
the UBT is reduced in the setting of an active peptic ulcer bleed (67 and 93 percent,
respectively) [13]. It is unclear if the reduction in sensitivity is induced by the bleeding
process or by changes in the microenvironment of H. pylori to decrease the activity of
urease.
The effect of PPIs, which is presumably due to suppression of H. pylori, was illustrated
in a series of 93 patients who had H. pylori infection documented by UBT [30].
Treatment with lansoprazole was associated with a negative result on subsequent UBT
in 33 percent of patients. Repeat breath test results 3, 7, and 14 days after stopping
lansoprazole were positive in 91, 97, and 100 percent, respectively. In another study of
60 patients with biopsy-proven H. pylori infection who underwent UBT testing 7 and 14
days after treatment with lansoprazole false-negative rates with lansoprazole and
bismuth were 40 and 55 percent, respectively [31]. It is controversial whether H2RAs
affect the sensitivity of the UBT [19,32-34].
Stool antigen assay — The detection of bacterial antigen indicates an ongoing H.
pylori infection. Stool antigen testing can therefore be used to establish the initial
diagnosis of H. pylori and to confirm eradication [3]. Of the available tests, stool antigen
testing is the most cost-effective in areas of low to intermediate prevalence of H.
pylori [35]. (See 'Patients not undergoing upper endoscopy' above and 'Confirmation of
eradication' below.)
The sensitivity and specificity of the laboratory-based monoclonal enzyme
immunoassay (94 and 97 percent, respectively) are comparable to the UBT [36-47].
Stool antigen testing is affected by the recent use of bismuth compounds, antibiotics,
and PPIs. Although some data suggest that stool antigen test is predictive of eradication
as early as seven days after completion of therapy, to reduce false-negative results,
patients should be off antibiotics for four weeks, and PPIs for one to two weeks, prior to
testing [19,31,39,48,49]. In the setting of active bleeding from peptic ulcers, the
specificity of the stool antigen testing may decrease [50,51]. However, the sensitivity of
the monoclonal enzyme immunoassay remains high in the setting of a recent peptic
ulcer bleed. This was illustrated in a study in which 34 patients underwent inpatient stool
antigen testing a mean of 2.8 days (range 0 to 8 days) after hospitalization for a
bleeding peptic ulcer. The sensitivity of the monoclonal enzyme immunoassay in this
study was 94 percent, significantly higher than a polyclonal enzyme immunoassay and
a rapid monoclonal immunochromatographic stool test (74 and 60 percent,
respectively). (See 'Patient undergoing upper endoscopy' above and 'Medications that
should be discontinued prior to testing' above.)
Stool antigen testing using the polyclonal enzyme immunoassay is no longer used given
its low sensitivity. The rapid in-office monoclonal immunochromatographic stool antigen
tests has high specificity but its adoption has been limited by its low sensitivity (96 and
50 percent, respectively) [52]. (See 'Confirmation of eradication' below.)
Serology — Laboratory-based ELISA test to detect immunoglobulin G (IgG) antibodies
is inexpensive and noninvasive. However, serologic tests require validation at the local
level, which is impractical in routine practice. In addition, concerns over its accuracy
have limited its use. Guidelines recommend that serologic testing should not be used in

146
low prevalence populations as the low accuracy of serology would result in
inappropriate treatment in significant numbers of patients [2,3,11].
One meta-analysis that evaluated the performance of several commercially available
serologic assays found an overall sensitivity and specificity of 85 and 79 percent,
respectively [53]. Inaccurate serological tests are more common in older adults and in
patients with cirrhosis in whom specificity can be decreased [54,55]. Local prevalence
of H. pylori affects the positive predictive value of antibody testing. In areas where the
prevalence of H. pylori is less than 20 percent, as in much of the United States, a
positive serologic test is more likely to be a false positive. As a result, secondary testing
with a stool or breath test to confirm the initial result is appropriate before initiating
treatment. However, a negative test in a patient with a low pretest probability for
infection, is helpful to exclude infection. An example would be a young individual with
dyspepsia and no evidence of peptic ulcer disease, especially in an area where the
prevalence of H. pylori is low.

Serology does not reliably distinguish between active and past infection. A quantitative
ELISA test is generally used in research settings. It allows for determination of IgG titers
of paired sera from the acute and convalescent (three to six months or longer) phase
and can confirm eradication of the infection. However, this is not performed in clinical
practice.

Other infrequently used tests

●13C-urea assay – A serodiagnostic test using a 13C-urea assay is a noninvasive
tool for diagnosis of H. pylori infection, but is rarely, if ever, used in clinical
practice. This test is performed by measuring two serum specimens; one is taken
before and the second 60 minutes after ingestion of a 13C-urea-rich meal. The
test has a reported sensitivity of 92 to 100 percent and a reported specificity of 96
to 97 percent [56,57].
●Polymerase chain reaction – Quantitative polymerase chain reaction (PCR)
testing on gastric biopsies can be used to detect low bacterial loads. It can also be
used to identify specific mutations associated with antimicrobial resistance.
However, the use of PCR based testing is limited by its high cost.
CONFIRMATION OF ERADICATIONConfirmation of eradication should be
performed in all patients treated for H. pylori because of increasing antibiotic resistance
[2,5]. Eradication can be confirmed with a urea breath test, stool antigen testing, or
endoscopy-based testing. The choice of test depends on the need for an upper
endoscopy (eg, follow-up of bleeding peptic ulcer) and local availability. Endoscopy with
biopsy for culture and sensitivity should be performed in patients with persistent H.
pylori infection after two courses of antibiotic treatment. Tests to confirm eradication
should be performed at least four weeks after completion of antibiotic treatment [5].
PPIs should be held for one to two weeks prior to testing to reduce false-negative
results. Serologic testing should not be performed to confirm eradication as patients
may continue to have antibodies after eradication [58]. (See 'Biopsy urease
testing' above and 'Stool antigen assay' above and 'Endoscopic testing' above
and 'Histology' above and 'Bacterial culture and sensitivity testing' above
and 'Serology' above.)

147
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Helicobacter pylori".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: H. pylori infection (The

Basics)" and "Patient education: Gastritis (The Basics)")
●Beyond the Basics topics (see "Patient education: Helicobacter pylori infection
and treatment (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Testing for H. pylori should be performed only if the clinician plans to offer
treatment for positive results. Indications for testing include:
•Low grade gastric MALT lymphoma
•Active peptic ulcer disease or past history of peptic ulcer if cure of H.
pylori infection has not been documented
•Early gastric cancer
●Other indications that are supported more limited evidence of benefit include:
•Uninvestigated dyspepsia in patients <60 years without alarm features (table
1)
•Prior to chronic treatment with NSAIDs and/or long-term, low-dose aspirin
•Unexplained iron deficiency anemia
•Adults with immune thrombocytopenia
●The choice of test used to diagnose H. pylori depends on whether a patient
requires an upper endoscopy for evaluation of symptoms or surveillance.
Endoscopy is not indicated solely for the purpose of establishing H. pylori status.
Other important determinants include the recent use of medications that can
suppress the bacterial load of H. pylori (eg, proton pump inhibitor therapy [PPI],
antibiotics, and bismuth), concurrent peptic ulcer bleeding, the prevalence of H.
pylori, test availability, and cost (see 'Approach to diagnostic testing' above)
(algorithm 1).
●Patients should be advised to stop PPI therapy one to two weeks prior to testing.
If feasible, testing should performed at least four weeks after completion of

148
bismuth/antibiotic treatment. (See 'Approach to diagnostic testing' above
and 'Medications that should be discontinued prior to testing' above.)
●In patients who do not require endoscopic evaluation, initial diagnosis of H.
pylori should be made with a test for active infection (stool antigen or urea breath
test). The urea breath test and stool antigen assay both have high sensitivity and
specificity for identifying active H. pylori infection. (See 'Patients not undergoing
upper endoscopy' above and 'Urea breath testing' above and 'Stool antigen
assay' above.)
●In patients undergoing upper endoscopy who have no recent
PPI/bismuth/antibiotic use and do not require biopsies of the stomach for
histology, the diagnosis of H. pylori can be established with a biopsy urease test.
(See 'Patient undergoing upper endoscopy' above and 'Biopsy urease
testing' above.)
●In patients with visible endoscopic abnormalities (eg, gastric ulcer, gastropathy)
and patients with recent PPI/bismuth/antibiotic use, we obtain gastric biopsies for
histology to diagnose H. pylori. In patients with recent PPI/bismuth/antibiotic use,
we perform a urea breath test or stool antigen assay to confirm a negative test
result. (See 'Histology' above and 'Urea breath testing' above and 'Stool antigen
assay' above.)
●Urea breath test and stool antigen assay have high sensitivity and specificity for
active H. pylori infection. Serology has low sensitivity and specificity for H.
pylori and cannot differentiate between active and past infection. Serology should
not be performed in areas of low H. pylori prevalence. If performed, positive
results should be confirmed with a test for active infection prior to initiating
eradication therapy. (See 'Noninvasive testing' above.)
●Confirmation of eradication should be performed in all patients treated for H.
pylori. Eradication can be confirmed with a urea breath test, stool antigen assay,
or endoscopy-based testing. Tests to confirm eradication should be performed at
least four weeks after completion of antibiotic treatment. Endoscopy with biopsy
for culture and sensitivity should be performed in patients with persistent H.
pylori infection after two courses of antibiotic treatment. (See 'Confirmation of
eradication' above.)

149
Pathophysiology of and immune response to
Helicobacter pylori infection
Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Apr 27, 2021.
INTRODUCTIONHelicobacter pylori (H. pylori) is highly adapted to the gastric
environment where it lives within or beneath the gastric mucous layer (see "Bacteriology
and epidemiology of Helicobacter pylori infection"). The bacterium generally does not
invade gastroduodenal tissue. Instead, it renders the underlying mucosa more
vulnerable to acid peptic damage by disrupting the mucous layer, liberating enzymes
and toxins, and adhering to the gastric epithelium. In addition, the host immune
response to H. pylori incites an inflammatory reaction which further perpetuates tissue
injury.
The chronic inflammation induced by H. pylori upsets gastric acid secretory physiology
to varying degrees and leads to chronic gastritis which, in most individuals is
asymptomatic and does not progress (picture 1). In some cases, however, altered
gastric secretion coupled with tissue injury leads to peptic ulcer disease, while in other
cases, gastritis progresses to atrophy, intestinal metaplasia, and eventually gastric
carcinoma or rarely, due to persistent immune stimulation of gastric lymphoid tissue,
gastric lymphoma. (See "Association between Helicobacter pylori infection and
duodenal ulcer" and "Acute and chronic gastritis due to Helicobacter pylori" and "Gastric
intestinal metaplasia" and "Metaplastic (chronic) atrophic gastritis" and "Association
between Helicobacter pylori infection and gastrointestinal malignancy".)
As a result, the pathophysiology of H. pylori infection and its eventual clinical outcome
should be viewed as a complex interaction between the host and the bacterium. This
interaction is influenced by the environment and modulated by a number of largely as
yet unidentified factors [1,2].
The pathophysiology of H. pylori infection as it relates to gastrointestinal disease in
general will be reviewed here. The role of H. pylori in specific disease processes is
discussed separately. (See "Association between Helicobacter pylori infection and
duodenal ulcer" and "Acute and chronic gastritis due to Helicobacter
pylori" and "Association between Helicobacter pylori infection and gastrointestinal
malignancy".)
BACTERIAL FACTORSTissue injury induced by H. pylori depends upon bacterial
attachment and the subsequent release of enzymes and other microbial products that
can cause cellular damage.
Bacterial attachment — H. pylori exclusively colonizes gastric type epithelium, which
suggests specific recognition of cell type by the bacterium [3]. Electron microscopy has

150
confirmed the presence of tight adherence of H. pylori to the gastric cell surface through
formation of membrane attachment pedestals similar to those described with
enteropathogenic Escherichia coli [4,5]. This process requires that bacterial adhesins
recognize and specifically bind to host receptors expressed on the cell surface [3]. The
attachment process may morphologically or functionally alter the epithelial cell or
activate certain bacterial functions making them more toxic. At the site of adherence,
bacterial membrane proteins coded by genes contained in the cag pathogenicity island
(PAI) open channels in the epithelial cell membrane that enable direct contact of
bacterial factors with the cytoplasm [6].
Bacterial attachment is partially mediated by a number of adhesins and outer
membrane proteins. Three Hop proteins have been implicated in the pathogenesis of H.
pylori infection, BabA (HopS), OipA (HopH), and SabA (HopP) [7]. BabA, the best
characterized of the three adhesin proteins, mediates binding to fucosylated Lewis b
(Le(b)) blood group antigens on host cells [8]. OipA may serve as an adhesin but it also
promotes inflammation by increasing IL-8 expression [9]. SabA mediates binding to
glycoconjugates containing sialic acid [10]. Replacement of nonsialylated Lewis
antigens by sialylated Le(x) or Le(a) has been associated with H. pylori induced gastric
inflammation and cancer [10,11]. Thus, the role of Lewis antigen expression in bacterial
attachment is unclear. Nevertheless, the homologous structures of H.
pylori lipopolysaccharide and host LewisX antigen may lead to an autoimmune
response with subsequent cell injury [12,13]. H. pylori can also bind to class II MHC
molecules on the surface of gastric epithelial cells and induce apoptosis [14]. In fact,
binding of the organism's urease to surface class II MHC is itself sufficient to induce
apoptosis [15]. H. pylori modifies its lipid A in the outer membrane by removal of
phosphate groups from the 1- and 4'-positions of the lipid A backbone. The enzyme
responsible for dephosphorylation of the lipid A 4'-phosphate group in H. pylori,
Jhp1487 (LpxF), has been identified and mutants created to demonstrate that
dephosphorylation of the lipid A domain of H. pylori LPS by LpxE and LpxF is key to its
ability to colonize the mammalian host [16].
Release of enzymes — H. pylori elaborates several enzymes that can cause cellular
damage by direct or indirect mechanisms.
●Urease accounts for over 5 percent of the organism's total protein weight [17].
Urea, when hydrolyzed by bacterial urease, can form compounds such
as ammonium chloride and monochloramine that can directly damage epithelial
cells [18]. In addition, the urease enzyme itself is antigenic, activates the host
immune system, and indirectly produces injury by stimulating inflammatory cells
[18].
●Bacterial phospholipases can alter the phospholipid content of the gastric
mucosal barrier, changing its surface tension, hydrophobicity, and permeability
[19]. The conversion of lecithin to lysolecithin (a toxic compound) by
phospholipase A2 can lead to cell injury [19], while lipolysis can disrupt the
structure and integrity of gastric mucus [20].
●H. pylori produces more catalase enzyme than most other bacteria. This enzyme,
an antioxidant, may protect the organism from toxic oxygen metabolites liberated
by activated neutrophils and allow it to survive and proliferate in an inflamed and
damaged gastric mucosa [19,21].

151
 ●Bacterial proteolytic enzyme activity can further degrade mucus. However, the
importance of proteolysis remains controversial [19].
Bacterial strain differences — Functional differences exist between strains of H.
pylori that may relate to virulence and tissue damage [22].
CagA and VacA — One such difference is expression of an 87 kilodalton (kd)
vacuolating cytotoxin (VacA) which causes cell injury in vitro and gastric tissue damage
in vivo [22-24]. All H. pylori contain the gene coding for VacA; however, only those
strains that encode the cytotoxin-associated gene (cag) pathogenicity island (PAI),
including cytotoxin-associated gene A (CagA), coding for a 128 to 140 kd protein
(CagA), coexpress VacA [23]. VacA behaves as a passive urea transporter that is
potentially capable of increasing the permeability of the gastric epithelium to urea,
thereby creating a favorable environment for H. pylori infectivity [25]. Virulence of VacA
appears to depend upon the function of a tyrosine phosphatase receptor in gastric
epithelial cells [26]. H. pylori strains with different VacA alleles have differing toxicity
[27].
CagA is not cytotoxic but is antigenic and can be detected serologically [28]. Its function
is unknown but, since it is necessary for VacA expression, it may play a role in
transcription, excretion, or function of the VacA cytotoxin [28]. H. pylori can translocate
its CagA protein into gastric epithelial cells via a type IV secretory apparatus. There it is
tyrosine phosphorylated and possibly plays a role in host cell responses [29-31].
Virulent strains of H. pylori encode cag PAI, which expresses a type IV secretion system
(T4SS). This T4SS forms a syringe-like pilus structure for the injection of virulence
factors such as the CagA effector protein into host target cells. This is achieved by a
number of T4SS proteins, including CagI, CagL, CagY and CagA, which by itself binds
the host cell integrin member β(1) followed by delivery of CagA across the host cell
membrane. CagA is not cytotoxic but is antigenic and can be detected serologically [28].
A role of CagA interaction with phosphatidylserine has also been shown to be important
for the injection process. After delivery, CagA becomes phosphorylated by oncogenic
tyrosine kinases and mimics a host cell factor for the activation or inactivation of some
specific intracellular signalling pathways [28-32].
Strains producing VacA and CagA cause more intense tissue inflammation and induce
cytokine production [23,24,33]. Two other genes (picA and picB, now termed CagE)
which are cotranscribed and genetically linked to cagA share homology with genes
coding for toxins in other known pathogenic bacteria [23,34]. The gene product of picB
(CagE) induces the release of epithelial cytokines, including interleukin-8 (IL-8) [23,34].
This effect appears to be mediated by nuclear factor kappa B, which activates
transcription of IL-8 MRNA [35]. In addition, bacteria that express CagA are potent
inducers of IL-8 [36,37].

The clinical significance of CagA positivity is demonstrated in two different disorders:

●Approximately 85 to 100 percent of patients with duodenal ulcers have CagA+

strains, compared to 30 to 60 percent of infected patients who do not develop
ulcers [38]. Cag E positivity has also been associated with gastroduodenal
disease in adults and children [39,40].
●CagA strains are associated with a higher frequency of precancerous lesions and
gastric cancer [41]. The risk of malignancy may be related to specific amino acid
152
 sequences (EPIYA) in the CagA protein [42]. (See "Association between
Helicobacter pylori infection and gastrointestinal malignancy".)
Other virulence factors — In addition to CagA, several other H. pylori virulence factors
have been described [43-49]. The strength of these associations has not been well
defined in large populations.
●"Induced by contact with epithelium" (iceA) has been associated with peptic
ulcers [43,46,47].
●"Blood group antigen-binding adhesin" (babA2) has been associated with
duodenal ulcers and gastric cancer [48].
●"Outer inflammatory protein" (oipA) has been associated with duodenal ulcers
[49].
Many of the virulence factors described above can coexist in the same H. pylori strains,
making it unclear as to which factors might be most important. In addition, as noted
above, expression of CagA is associated with both gastric cancer and duodenal ulcer,
yet these two disorders rarely, if ever, coexist thereby minimizing the significance of this
"virulence factor" in understanding disease pathogenesis.
One study suggested that the oipA status may be a better predictor of H.
pylori virulence than any of the other previously described virulence factors. The study
included 247 H. pylori infected patients (86 with gastritis, 86 with a duodenal ulcer, and
75 with gastric carcinoma) in whom H. pylori isolates were tested for the other H.
pylori virulence factors discussed above. On multivariate analysis, only oipA status
remained an independent predictor of H. pylori density, mucosal inflammation, and high
mucosal IL-8 levels. However, the actual biological significance of these observations is
unknown. Adaptation to gastric conditions is enabled by phase variation of genes
encoding outer membrane proteins [50,51].
INFLAMMATORY RESPONSEAlthough H. pylori is a noninvasive organism, it
stimulates a robust inflammatory and immune response [52,53]. A variety of factors may
contribute to these changes, which are described below. Bacterial colonization,
persistence and virulence, and resulting innate and adaptive host immune responses
are all important in the pathogenesis of H. pylori related disease [7,54,55].
●The organism produces a number of antigenic substances, including heat shock
protein, urease, and lipopolysaccharide, all of which can be taken up and
processed by lamina propria macrophages and activate T-cells [53,56]. Cellular
disruption, especially adjacent to epithelial tight junctions, undoubtedly enhances
antigen presentation to the lamina propria and facilitates immune stimulation. The
net result is increased production of inflammatory cytokines such as IL-1, IL-6,
tumor necrosis factor-alpha (TNF-alpha), and most notably, IL-8 [36,52,57,58].
(See 'Interleukin-8 and other cytokines' below.)
●A B cell response to H. pylori (with production of IgG and IgA antibodies) occurs
locally in the gastroduodenal mucosa and systemically. The role of local
antibodies in producing tissue injury or modulating inflammation in H.
pylori infection remains controversial [52,53]. Prolonged stimulation of gastric B
cells by activated T cells can lead to MALT lymphoma in rare cases.
(See "Association between Helicobacter pylori infection and gastrointestinal
malignancy".)

153
T cells are also activated during infection and their cytokines boost bacterial binding (by
inducing class II MHC). While T cells are recruited to the infected gastric mucosa, they
appear to be hyporesponsive. B7-H1 (programmed death-1 ligand 1), a member of B7
family of proteins associated with T cell inhibition, appears to be involved in the
suppression of T cell proliferation and IL-2 synthesis during H. pylori infection, and thus
may contribute to its chronicity [59].
Different T helper cell subsets can be distinguished by characteristic profiles of cytokine
secretion. Th1 cells promote cell-mediated immune responses through elaboration of
TNF-alpha and IFN gamma. Th2 cells produce Il-4, IL-10 and TGF beta. It appears that
during H. pylori infection the T-cell immunity is inappropriately skewed toward a Th1
response that promotes epithelial cell inflammatory cytokine production (IL-8 stimulated
by IFN gamma and TNF-alpha) and directly impacts epithelial apoptosis [60,61].
●H. pylori infection induces a marked increase in the flux of leukocytes and in the
appearance of platelet and leukocyte-platelet aggregates in gastric venules in a
murine model [62]. Circulating platelet aggregates and activated platelets were
also detected in patients infected with H. pylori, suggesting that platelet activation
and aggregation contribute to the associated microvascular dysfunction and
inflammatory cell recruitment. Platelet aggregation mediated by an H.
pylori interaction with von Willebrand factor is speculated to contribute to infection
related ulcer disease but also possibly non-GI manifestations of infection such as
cardiovascular disease and idiopathic thrombocytopenia [63,64].
●Not all H. pylori infected individuals develop clinical disease. Host genetics are
important in determining the physiologic and clinical response to infection. Host
polymorphism of IL-1 beta (and possibly IL-10) appears to determine the degree
of inflammatory response to infection, resulting alteration in acid secretion (hyper
or hypo secretion), and risk for subsequent gastric cancer [65,66]. One series of
meta-analyses investigated genes coding for the interleukin (IL) proteins (IL1B,
IL1RN, IL8, and IL10) and for TNF-alpha. Gastric cancers were stratified by
histologic subtype and anatomic subsite, by H. pylori infection status, by
geographic location (Asian or non-Asian study population), and by a quantitative
index of study quality. Results consistently supported increased cancer risk for
IL1RN2 carriers; the increased risk was specific to non-Asian populations and was
seen for intestinal and diffuse cancers, distal cancers, and, to a lesser extent,
cardia cancers. In Asian populations, reduced risk was observed in association
with IL1B-31C carrier status. These results indicate the importance of stratification
by anatomic site, histologic type, H. pylori infection, and country of origin. Study
quality considerations, both laboratory and epidemiologic, can also affect results
and may explain, in part, the variability in results published to date [67].
Interleukin-8 and other cytokines — Research has centered on epithelial IL-8
production induced by different strains of H. pylori [36,68]. IL-8 is a potent chemotactic
factor, activates neutrophils, and recruits acute inflammatory cells into the mucosa. H.
pylori appears to activate transcription factor NF-kB (nuclear factor kappa B), which in
turn increases IL-8 production [35,69]. NF-kB also regulates the expression of additional
inflammatory response genes, and may play a role in the mucosal epithelial response to
other bacterial infections in addition to H. pylori.

154
Bacteria that express CagA and VacA are more potent inducers of IL-8; however, the
gene primarily responsible for IL-8 induction is picB (now renamed CagE), which is
located upstream of the CagA gene [34]. CagA/VacA-positive strains are also more
often found in patients with clinical manifestations of H. pylori infection, indirectly
suggesting that IL-8 may play an important pathophysiologic role in gastroduodenal
disease.
TNF-alpha can also augment IL-8 production by the inflamed mucosa. Following
successful eradication of H. pylori, mucosal levels of mRNA for both TNF-alpha and IL-8
are reduced in parallel with the decline in local inflammation [70].
H. pylori infection increases IL-17 in the gastric mucosa of humans and experimental
animals. IL-17 induces the secretion of IL-8 by activating the ERK 1/2 MAP kinase
pathway. IL-23 is also increased in patients with H. pylori related gastritis and regulates
IL-17 secretion via the STAT3 pathway. The early events in the immune response of
immunized and challenged mice include the recruitment of T cells and the production of
IL-17. Neutrophil attracting chemokines are released, and the bacterial load is
considerably reduced [71]. IL-17 plays a dual role in infection and vaccination. In
infection, T regulatory cells (Tregs) suppress the inflammatory reaction driven by IL-17,
thereby favoring bacterial persistence. Immunization produces Helicobacter-specific
memory T-helper cells that can possibly alter the ratio between T-helper 17 and Treg
responses so that the IL-17-driven inflammatory reaction can overcome the Treg
response leading to bacterial clearance [71].
Virulent H. pylori strains that specifically activate signaling in epithelial cells via the
innate immune molecule, nucleotide oligomerization domain 1 (NOD1), are more
frequently associated with IFN-gamma-dependent inflammation and with severe clinical
outcomes (ie, gastric cancer and peptic ulceration). H. pylori activation of the NOD1
pathway causes enhanced proinflammatory signaling in epithelial cells in response to
IFN-gamma stimulation through the direct effects of H. pylori on two components of the
IFN-gamma signaling pathway, STAT1 and IFN regulatory factor 1 (IRF1) [72].
Survival of H. pylori — H. pylori itself is in part able to survive this inflammatory
onslaught by producing the enzyme, catalase. This enzyme neutralizes the damaging
reactive oxygen metabolites liberated by neutrophils [19]. With time, the host appears to
downregulate the acute inflammatory response, making it easier for the organism to
persist and proliferate [73].
ANTIBODY RESPONSEMost infected individuals systemically produce specific
antibodies to a variety of H. pylori antigens. The antibody response changes as infection
progresses from an acute to a chronic stage [74].
●Detection of IgM antibodies is an insensitive indicator of acute infection and
generally is not clinically useful, even in children [75].
●IgA and IgG antibodies are produced in response to infection, remain present as
long as infection is active, and quantitatively decrease after infection is cured [76].
●Antibodies to CagA protein are detectable in gastric tissue and serum and permit
the identification of infection with presumably more virulent organisms [77].
The role of local antibodies in the immunopathogenesis of gastroduodenal mucosal
injury is unclear [52]. Virtually all infected persons have a specific gastric mucosal IgA
and IgG response. IgA antibodies may modulate mucosal injury by inhibiting antigen
uptake, disrupting bacterial adherence and motility, and neutralizing various toxins. IgG
155
presumably augments inflammatory injury by activating complement and facilitating
neutrophil activation.
An antibody response may also be seen against autoantigens, including IL-8 [78], antral
epithelium [79,80], and homologous host and bacterial epitopes (eg, LewisX,
lipopolysaccharide, and heat shock protein) [52]. The immunoglobulin specificity of
MALT lymphoma may be for such autoantigens [81].
VACCINATIONWhile the mucosal immune response to H. pylori leads to tissue
injury, it is also key to vaccination strategies [71,82,83]. One could question if effective
vaccination is possible since the organism can successfully evade the immune
response to natural infection [84]. Nevertheless, preliminary studies suggest that
preventive vaccination may be feasible. Immunization with crude sonicates of bacteria
[85,86] and recombinant subunits of urease [87,88] and catalase [89] protect animals
from H. pylori exposure. Human vaccines have undergone [88,90] and continue to
undergo clinical testing [91]. In a randomized phase 3 vaccine trial, 4464 H.
pylori uninfected children (ages 6 to 15 years) were assigned to a three-dose oral
recombinant H. pylori vaccine or placebo [92]. The vaccine contained fusion proteins
composed of the B subunits of H. pylori urease and heat-labile toxin of Escherichia
coli. At one year, the incidence of H. pylori infection was significantly lower in the
vaccine group as compared with placebo (0.7 versus 2.4 events per 100 person-years),
with a vaccine efficacy of 72 percent (95% CI 48.2-85.6). Among patients who
completed extended follow-up, H. pylori acquisition continued to be lower in vaccinated
as compared with unvaccinated children, but protection levels were lower in the second
and third year (vaccine efficacy 55 [95% CI 0.9-81.0] and 56 [95% CI -24.7-86.2]
percent, respectively). There were no serious adverse events related to the vaccine.
Additional studies with long-term follow-up are needed to validate these results.
Studies on vaccination have expanded to the area of therapeutic immunization [84,93].
A therapeutic vaccine could boost the natural immune response and facilitate
spontaneous bacterial eradication or enhance the effectiveness of antibiotic regimens. A
successful therapeutic vaccine would obviate the widespread use of antibiotics and
minimize the development of antibiotic-resistant H. pylori organisms. Unfortunately, the
promise of vaccines has not been met after more than two decades of research. A
renewed or expanded commitment to exploit advances in our understanding of the host
immune response to H. pylori is necessary for the advancement of an H. pylori vaccine
[94].
SUMMARY AND RECOMMENDATIONS
●The pathophysiology of H. pylori infection and its eventual clinical outcome
should be viewed as a complex interaction between the host and the bacterium.
This interaction is influenced by the environment and modulated by a number of
largely as yet unidentified factors.
●Tissue injury induced by H. pylori depends upon bacterial attachment and the
subsequent release of enzymes and other microbial products that can cause
cellular damage. (See 'Bacterial factors' above.)
●Functional differences exist between strains of H. pylori that may relate to
virulence and tissue damage. However, many of the virulence factors can coexist
in the same H. pylori strains, making it unclear as to which factors might be most
important. (See 'Bacterial strain differences' above.)
156
 ●Although H. pylori is a noninvasive organism, it stimulates a robust inflammatory
and immune response. Bacterial colonization, persistence and virulence, and
resulting innate and adaptive host immune responses are all important in the
pathogenesis of H. pylori related disease. (See 'Inflammatory response' above.)

Treatment regimens for Helicobacter pylori in adults

Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: May 20, 2021.
INTRODUCTIONMultiple antibiotic regimens have been evaluated for Helicobacter
pylori (H. pylori) therapy [1-6]. However, few regimens have consistently achieved high
eradication rates. There are also limited data on H. pylori antibiotic resistance rates to
guide therapy. The treatment regimen that is selected must consider local antibiotic
resistance patterns (if known), previous exposure and allergies to specific antibiotics,
cost, side effects, and ease of administration.
This topic will review treatment regimens for H. pylori. The bacteriology, epidemiology,
and diagnostic tests for H. pylori infection are discussed elsewhere. (See "Indications
and diagnostic tests for Helicobacter pylori infection in adults" and "Bacteriology and
epidemiology of Helicobacter pylori infection".) (Related Pathway(s): Helicobacter pylori:
Initial treatment for adults.)
INDICATIONS FOR TREATMENTAll patients with evidence of active infection
with H. pylori should be offered treatment. Indications for testing for H. pylori infection
are discussed in detail separately. (See "Indications and diagnostic tests for
Helicobacter pylori infection in adults", section on 'Indications for testing'.)
INITIAL ANTIBIOTIC THERAPY
Approach to selecting an antibiotic regimen — The choice of initial antibiotic
regimen to treat H. pylori should be guided by the presence of risk factors for macrolide
resistance and the presence of a penicillin allergy [7]. In patients with one or more risk
factors for macrolide resistance, clarithromycin-based therapy should be avoided. A
suggested approach to the selection of antibiotics for initial treatment of H.
pylori infection is outlined in the algorithm (algorithm 1 and table 1).
(See 'Clarithromycin-based therapy' below.) (Related Pathway(s): Helicobacter pylori:
Initial treatment for adults.)

Risk factors for macrolide resistance include:

●Prior exposure to macrolide therapy for any reason

●High local clarithromycin resistance rates ≥15 percent or eradication rates with
clarithromycin triple therapy ≤85 percent

157
A resistance threshold of ≥15 percent is commonly used for choosing alternative empiric
antibiotic regimen for H. pylori [8,9]. In the United States, given the limited information
on antimicrobial resistance rates, we generally assume clarithromycin resistance rates
are greater than 15 percent unless local data indicate otherwise [10]. Data suggest
that H. pylori antibiotic resistance rates are high worldwide. In a systematic review and
meta-analysis that included 178 studies, comprising 66,142 isolates from 65 countries,
primary and secondary resistance to clarithromycin, metronidazole,
and levofloxacin were high (≥15 percent) in the majority of WHO regions [11]. The
pooled prevalence of primary clarithromycin resistance was >15 percent in European,
Eastern Mediterranean and Western regions but were lower in the Americas (10
percent, 95% CI 4-16) and the South East Asia region (10 percent, 95% CI 5-16).
Resistance to clarithromycin was significantly associated with failure of H.
pylori eradication with a clarithromycin-containing regimen (odds ratio, 6.97; 95% CI,
5.2-9.3). However, the study was limited by significant heterogeneity and 10 of the 13
studies contributing to the pooled data for the Americas region were derived from South
America. Local surveillance data are needed guide the choice of eradication regimens.
Patients with risk factors for macrolide resistance — In patients with risk factors for
macrolide resistance, we use bismuth quadruple therapy (algorithm 1 and table 1) [7-
9,12-16]. (See 'Bismuth quadruple therapy' below and 'Levofloxacin based
therapy' below.)
Patients without risk factors for macrolide resistance — In patients without risk
factors for macrolide resistance, we use clarithromycin-based triple therapy with a
proton pump inhibitor (PPI), amoxicillin, and clarithromycin (algorithm 1 and table 1).
Other first-line antibiotic regimens for these patients include bismuth quadruple therapy
and clarithromycin-based concomitant therapy. In penicillin-allergic
individuals, metronidazole can be substituted for amoxicillin. In patients with
metronidazole exposure within the past few years, we use bismuth quadruple therapy.
(See 'Clarithromycin-based therapy' below and 'Bismuth quadruple therapy' below
and 'Concomitant therapy' below.)
Other potential first-line treatment regimens include clarithromycin-based hybrid or
sequential therapy [7]. However, the clarithromycin-based hybrid therapy has not been
universally endorsed as an option for first-line therapy given its complexity [8]. In
addition, some North American guidelines recommend against the use of sequential
therapy as a first-line regimen given the lack of data from North American trials [8].
Duration of therapy — We recommend clarithromycin-based triple therapy and
bismuth quadruple treatment regimens for H. pylori be administered for 14 days. Our
recommendations are largely consistent with current guidelines which recommend
extended (10 to 14 days) treatment with all antibiotic regimens for H. pylori (table 1)
[8,9,12].
Tolerability and compliance — Side effects are reported in up to 50 percent of
patients taking one of the triple therapy regimens [3,17]. The adverse effects are usually
mild; fewer than 10 percent of patients stop treatment due to side effects [17].
Clarithromycin-based triple therapy and bismuth quadruple therapy appear to have
similar efficacy, compliance, and tolerability [18]. While the tolerability and compliance
of sequential, hybrid, and concomitant therapies appears to be similar to triple therapy
in clinical trials, these regimens are more complex. Side effects of individual drugs are

158
discussed in detail separately. (See "Metronidazole: An overview" and "Azithromycin
and clarithromycin", section on 'Adverse reactions' and "Tetracyclines", section on
'Adverse reactions' and "Fluoroquinolones", section on 'Adverse effects' and "Proton
pump inhibitors: Overview of use and adverse effects in the treatment of acid related
disorders", section on 'Adverse effects' and "Antiulcer medications: Mechanism of
action, pharmacology, and side effects", section on 'Adverse effects'.)
Antibiotic regimens — Initial antibiotic regimens for H. pylori can be broadly divided
into bismuth, clarithromycin, and levofloxacin containing regimens (table
1 and algorithm 1).
Bismuth quadruple therapy — Bismuth quadruple therapy consists of bismuth
subsalicylate, metronidazole, tetracycline, and a PPI given for 14 days [19]. A
combination capsule containing bismuth subcitrate, metronidazole, and tetracycline
(Pylera) has been approved by the United States Food and Drug Administration. A
regimen using the combination capsule (three capsules four times daily plus PPI twice
daily) is somewhat simpler than standard quadruple therapy (four to eight pills four
times daily and a PPI twice daily). For details, refer to bismuth subcitrate-metronidazole-
tetracycline in Lexicomp. If tetracycline is not available, doxycycline (100 mg twice daily)
may be substituted [20,21].
In North American trials, the mean eradication rate with bismuth quadruple therapy
administered for 10 days was 91 percent [12,22,23]. A 2013 meta-analysis of 12
randomized trials reported comparable eradication rates with bismuth quadruple therapy
and clarithromycin triple therapy (78 and 69 percent, respectively) [24]. However, there
was significant heterogeneity in treatment duration, drug dosing, and the meta-analysis
included trials performed in North America, Europe, and Asia. No significant differences
in efficacy were noted when both regimens were administered for 10 to 14
days. Metronidazole resistance has a limited impact on eradication success rate in
patients treated with bismuth quadruple therapy and can be overcome by increasing the
dose, duration, or frequency of therapy [25]. (See 'Factors associated with antibiotic
treatment failure' below.)
Clarithromycin-based therapy
Triple therapy — Clarithromycin triple therapy consists of clarithromycin, amoxicillin,
and a PPI, all given twice daily (table 1). We suggest treatment for 14 days, as longer
duration of treatment may be more effective in curing infection
[7,8,26]. Metronidazole can be used instead of amoxicillin in penicillin-allergic
individuals. PPI-clarithromycin-metronidazole and PPI-clarithromycin-amoxicillin
regimens are equivalent [9,27].
Eradication rates for clarithromycin triple therapy in the United States are below 80
percent [7]. The eradication rates of clarithromycin triple therapy is significantly
impacted by the presence of clarithromycin resistance [18,24]. In a meta-analysis of two
trials in which patients were treated with clarithromycin triple therapy, eradication rates
for clarithromycin-sensitive H. pylori strains and clarithromycin-resistant strains were 90
and 22 percent respectively [18]. Addition of bismuth to 14-day clarithromycin-based
triple therapy may improve eradication rates in areas with high antimicrobial resistance
[28]. However, further studies are needed.
Concomitant therapy — Concomitant therapy consists of a clarithromycin, amoxicillin,
a nitroimidazole (tinidazole or metronidazole), and a PPI administered together (table

159
1). If concomitant therapy is used to treat H. pylori, the regimen should be continued for
10 to 14 days.
While efficacy data from the North America are lacking, in a meta-analysis of 19
randomized trials that included 2070 individuals in Europe, Asia and Latin America,
eradication rates were significantly higher with concomitant quadruple therapy as
compared with clarithromycin triple therapy (90 and 78 percent, respectively) [29]. The
efficacy of concomitant therapy was decreased in patients with clarithromycin-
resistant H. pylori infection but to a smaller degree as compared with clarithromycin
triple therapy. In this meta-analysis, longer durations of therapy (7 to 10 versus 3 to 5)
were associated with a trend toward higher cure rates. However, additional studies are
needed to assess whether extending the duration of concomitant therapy to 14 days
results in improved eradication rates.
Hybrid therapy — Hybrid therapy consists of amoxicillin and a PPI for seven days
followed by amoxicillin, clarithromycin, a nitroimidazole, and a PPI for seven days (table
1). Hybrid therapy has been suggested as an alternative to clarithromycin triple therapy.
However, the complexity of the treatment regimen has limited its use as a first-line
regimen in the treatment of H. pylori.
In a meta-analysis that included six randomized trials, which compared hybrid therapy
with sequential and/or concomitant therapy, the eradication rate with hybrid therapy was
89 percent [30]. The efficacy and tolerability of hybrid therapy is comparable to
concomitant and sequential regimens [31]. Hybrid therapy has not been directly
compared with clarithromycin-based triple therapy. However, in one randomized trial in
which 440 patients were assigned to 12 days of triple therapy or reverse hybrid therapy
(amoxicillin and pantoprazole for 12 days and clarithromycin plus metronidazole for the
initial seven days), eradication rates were significantly higher with reverse hybrid
therapy (96 versus 89 percent) [32]. In contrast to patients who received clarithromycin
triple therapy, clarithromycin resistance did not significantly impact eradication rates in
patients treated with reverse hybrid therapy.
Sequential therapy — The 10-day clarithromycin-containing sequential therapy
regimen consists of amoxicillin and a PPI for five days, followed by clarithromycin and
nitroimidazole (eg, metronidazole) plus a PPI for five days (table 1) [33]. Given the
complexity of the sequential therapy regimen and the lack of superiority to 14 day
clarithromycin triple therapy in North America, clarithromycin-containing sequential
therapy has not been uniformly endorsed by guidelines as a first-line therapy [8].
(See 'Patients without risk factors for macrolide resistance' above.)
In a 2013 meta-analysis of 46 randomized trials that included 13,532 patients who were
assigned to sequential therapy or other regimens, the overall eradication rate for
sequential therapy was 84 percent [34]. Eradication rates with sequential therapy were
significantly higher as compared with clarithromycin triple therapy administered for 7 or
10 days. However, there was no significant difference in eradication rates between
sequential therapy and 14 days of clarithromycin-based triple therapy or 10 to 14 days
of bismuth quadruple therapy. The efficacy of sequential therapy varies widely by region
[7]. Randomized trials in Latin America and Asia have also demonstrated lower
eradication rates with sequential therapy as compared with clarithromycin triple therapy;
however, the efficacy of sequential therapy may be higher in Europe [35-39].

160
Levofloxacin based therapy — Due to rising rates of levofloxacin resistance,
levofloxacin should not be used for treatment unless the H. pylori strain is known to be
sensitive to it or if the population levofloxacin resistance rates are known to be less than
15 percent [7,11,40,41]. Studies evaluating the efficacy of levofloxacin containing
regimens in North America are lacking. Limited data suggest that fluoroquinolone
resistance rates in North America are high [11]. Levofloxacin resistance decreases the
eradication success rate of levofloxacin containing regimens by 20 to 40 percent [7].
(See 'Approach to selecting an antibiotic regimen' above and 'Salvage
regimens' below.)
●Levofloxacin triple therapy – Levofloxacin triple therapy consists of
levofloxacin, amoxicillin, and a PPI for 10 to 14 days. In a network meta-analysis
eradication rates with levofloxacin triple therapy for 10 to 14 days were
significantly higher than clarithromycin triple therapy for seven days (90 versus 73
percent) [31]. Although not directly compared, the pooled eradication rate of
levofloxacin triple therapy was also higher than clarithromycin triple therapy for 10
to 14 days (81 percent, 95% CI, 78 to 84 percent). Metronidazole can be
substituted for amoxicillin in penicillin-allergic individuals.
●Levofloxacin quadruple therapy – Limited data support the use of quadruple
therapy with levofloxacin, omeprazole, nitazoxanide, and doxycycline (LOAD). In
an open label prospective trial, H. pylori treatment-naïve patients randomized to
LOAD for 7 or 10 days had significantly higher eradication rates as compared
with clarithromycin triple therapy for 10 days (89, 90, and 73 percent, respectively)
[42]. However, additional studies are needed to confirm these results and
determine whether this more expensive regimen is cost-effective. Other
levofloxacin-based quadruple therapy regimens that have been used as salvage
therapy include PBLA (PPI, bismuth, levofloxacin, amoxicillin), PBLT (PPI,
bismuth, levofloxacin, tetracycline), and PBLM (PPI, bismuth,
levofloxacin, metronidazole) [41].
●Levofloxacin sequential therapy – Levofloxacin sequential therapy consists
of amoxicillin and a PPI for five to seven days followed by levofloxacin, amoxicillin,
a nitroimidazole and a PPI for five to seven days. A meta-analysis of six
international trials compared the efficacy of fluoroquinolone sequential therapy for
10 to 14 days and either clarithromycin triple therapy for 7 to 14 days or standard
sequential therapy for 10 days [43]. The pooled eradication rate with
fluoroquinolone sequential therapy was significantly higher as compared with
clarithromycin triple or standard sequential therapies combined (88 versus 71
percent).
CONFIRMATION OF ERADICATIONTests to confirm eradication should be
performed in all patients treated for H. pylori. Eradication may be confirmed by a urea
breath test, fecal antigen test, or upper endoscopy performed four weeks or more after
completion of antibiotic therapy. PPI therapy should be withheld for one to two weeks
prior to testing [12,40,44]. Endoscopy with biopsy for culture and sensitivity should be
performed in patients with persistent H. pylori infection after two courses of antibiotic
treatment [41]. (See "Indications and diagnostic tests for Helicobacter pylori infection in
adults", section on 'Confirmation of eradication'.)

161
TREATMENT FAILUREApproximately 20 percent of patients fail an initial attempt
at H. pylori eradication [45]. Such patients require salvage therapy (algorithm
2 and table 2).
Factors associated with antibiotic treatment failure — Factors associated with
treatment failure include poor patient compliance and resistance of the patient’s H.
pylori strain to prescribed antibiotics. H. pylori is naturally resistant to several commonly
used antibiotics, including vancomycin, trimethoprim, and sulfonamides [46]. A specific
mutation leading to clarithromycin resistance appears to be associated with a reduced
likelihood of eradication [47]. Prior use of macrolide antibiotics,
and levofloxacin increases the risk of H. pylori resistance to these antibiotics [48].
Clarithromycin resistance has a greater effect on treatment efficacy as compared
with metronidazole resistance [25]. Resistance rates to amoxicillin, tetracycline,
and rifabutin are low (<5 percent), and these can be considered for subsequent
therapies in refractory H. pylori infection [41]. Inadequate acid suppression is also
associated with H. pylori eradication failure.
Salvage therapy for persistent H. pylori infection
Suggested approach — In patients with persistent H. pylori infection, the choice of
antibiotic therapy should be guided by the patient’s initial treatment regimen, the use of
other antibiotics, and the presence of relevant antibiotic allergies [41]. Antibiotics
included in the initial regimen should generally be avoided [49].
However, amoxicillin can be reused as resistance rarely develops. Patients with a
reported history of penicillin allergy should be referred to an allergist to determine if they
have a true penicillin allergy. A suggested approach to the selection of antibiotics for
persistent H. pylori infection is outlined in the algorithm (algorithm 2 and table 2).
(See "An approach to the patient with drug allergy".)
Culture with antibiotic sensitivity testing should be performed to guide antibiotic
treatment in patients who have failed two prior treatment regimens. Compliance with
medications should also be reinforced. We reserve the use of rifabutin-containing
regimens for patients with ≥3 previous antibiotic failures. However, other experts have
suggested its use as a second-line agent [41]. The impact of metronidazole resistance
can be overcome by increasing the dose (1.5 to 2 g daily in divided doses), duration, or
frequency of administration of metronidazole [41]. Resistance rates
to amoxicillin, tetracycline, and rifabutin are low (<5 percent), and these can be
considered for subsequent therapies in refractory H. pylori infection even if previously
used [41].
The use of high-dose PPIs (double the standard dose), use of more potent PPIs and
those less dependent on metabolism by CYP2C19 (eg, esomeprazole or rabeprazole),
or potassium-competitive acid blockers where available, can lower intragastric acidity in
patients with refractory H. pylori infection [41,50].
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults", section
on 'Bacterial culture and sensitivity testing'.)
Salvage regimens — Salvage regimens in patients who have failed initial antibiotic
therapy include (table 2):
●Bismuth quadruple therapy – Bismuth quadruple therapy should be used for 14
days when used as salvage regimen. In randomized trials performed in Europe,
United States, and Asia eradication rates with 14-day salvage bismuth quadruple
162
therapy were approximately 80 percent [12]. Eradication rates were significantly
higher in studies performed in Asia as compared with Europe and the United
States (82 versus 74 percent) [51-53]. The overall eradication rate for 14-day
bismuth quadruple therapy in these trials was higher in patients who had
previously failed clarithromycin-based regimens without bismuth as compared with
bismuth quadruple treatment (100 versus 53 percent).
●Levofloxacin-based therapy – Levofloxacin-based triple therapy has
demonstrated efficacy as a salvage regimen in patients who have failed
initial clarithromycin triple therapy or bismuth quadruple
therapy. Levofloxacin triple therapy has also demonstrated efficacy in patients
who have failed two prior attempts at treatment. In a pooled analysis from six
European cohort studies, when used as a salvage regimen in patients who had
failed two previous eradication attempts, levofloxacin triple therapy administered
for 10 days has a pooled eradication rate of 73 percent [54]. Most patients in these
studies were treated with clarithromycin triple therapy followed by bismuth
quadruple therapy.
Other levofloxacin-based quadruple therapy regimens that have been used as
salvage therapy include PBLA (PPI, bismuth, levofloxacin, amoxicillin), PBLT (PPI,
bismuth, levofloxacin, tetracycline), and PBLM (PPI, bismuth,
levofloxacin, metronidazole) [41].
●High-dose dual therapy – High-dose dual therapy with amoxicillin (at least 2 g
divided three or four times per day to avoid low trough levels) and proton pump
inhibitor (PPI) for 14 days is a salvage treatment option, particularly in patients in
whom dual metronidazole/clarithromycin resistance or levofloxacin resistance is
suspected [41]. The pooled eradication rate of high-dose dual therapy with
amoxicillin and PPI as a salvage regimen in three randomized trials performed in
Europe and Asia was 78 percent [12].
The role of high-dose dual therapy as first-line treatment is unclear, as studies
evaluating the efficacy of this regimen have been conflicting. In studies conducted
in the United States and Korea, eradication rates in treatment-naive patients were
low (72 and 79 percent, respectively) [55,56]. However, in a randomized trial in
China in which 232 treatment-naive patients were assigned to high-dose dual
therapy or bismuth quadruple therapy, there was no significant difference in
eradication rates between the two groups [57]. While high-dose dual therapy had
lower treatment-related adverse effects as compared with bismuth quadruple
therapy, it is important to note that there may have been reporting bias due to the
open-label study design.
●Rifabutin triple therapy – The rifabutin-based triple regimen consists of
rifabutin, amoxicillin, and a PPI twice daily for 14 days. In a meta-analysis of
cohort studies that evaluated the efficacy of rifabutin triple therapy as salvage
treatment, pooled eradication rates as a second, third, or fourth/fifth line agent
were 79, 66, and 70 percent respectively [58]. The role of rifabutin-based triple
therapy as a first-line treatment option is unclear. In a randomized trial in which
455 H. pylori treatment-naïve patients were assigned to treatment with rifabutin-
based triple therapy or high-dose dual therapy with amoxicillin and omeprazole,
eradication rates with the rifabutin-containing regimen were significantly higher (84

163
 versus 58 percent) [59]. Eradication rates were unaffected by resistance
to clarithromycin or metronidazole. However, the study was conducted in the
United States and excluded persons of Asian descent due to a higher prevalence
of poor CYP 2C19 metabolizers. Rifabutin-based triple therapy is expensive and
can cause reversible myelotoxicity. It also has the potential to increase the
prevalence of rifabutin-resistant mycobacteria.
●Clarithromycin-based therapy – Clarithromycin-based therapy (eg, PPI,
bismuth, clarithromycin, tetracycline), can be used as a salvage regimen in
patients with no risk factors for macrolide resistance (no prior macrolide exposure
and local clarithromycin resistance known to be <15 percent) [41,60,61].
(See 'Clarithromycin-based therapy' above.)
ADJUVANT THERAPIES WITH UNCLEAR ROLEA number of potential
adjuvant therapies for H. pylori have been evaluated, but additional studies are needed
to support their use.
●Statins – Addition of statin therapy as an adjuvant to triple therapy has been
associated with a reduction in H. pylori mediated inflammation and an increase
in H. pylori eradication rates [62-64]. However, large trials are needed to confirm
these findings.
●Probiotics – Probiotics may have an inhibitory effect on H. pylori. In addition,
they may improve compliance with treatment by reducing antibiotic side effects. A
meta-analysis that included 10 clinical trials of adjuvant probiotics in patients
with H. pylori infection demonstrated higher cure rates and a reduction in the
incidence of side effects in patients who received probiotic supplementation
(pooled OR 2.1 and 0.3, respectively). However, studies included in this meta-
analysis were at high risk of bias due to lack of blinding and inadequate allocation
concealment. In addition, there was significant variability in the probiotics used
and antibiotic treatment regimens to eradicate H. pylori. (See "Probiotics for
gastrointestinal diseases".)
●Vonoprazan – Limited data from retrospective studies suggest that the use of
vonoprazan, an oral potassium-competitive acid blocker (PCAB), rather than a
proton pump inhibitor, with amoxicillin and clarithromycin may increase H.
pylori eradication rates [65,66]. However, PCABs are not available outside of Asia,
and data on safety and comparative efficacy with other antisecretory agents are
limited [65]. (See "Antiulcer medications: Mechanism of action, pharmacology, and
side effects", section on 'Potassium-competitive acid inhibitors'.)
TREATMENT DURING PREGNANCY AND LACTATIONWhen peptic ulcer
disease is diagnosed in a woman who is pregnant, the mainstay of treatment is typically
acid suppression [67]. If H. pylori is present, treatment is typically deferred until after
delivery. However, with the exception of bismuth, fluoroquinolones, and tetracycline, the
other medications used for H. pylori eradication are low risk in pregnancy, especially
after 14 weeks. This includes clarithromycin, amoxicillin, and probably metronidazole.
Moreover, there is some evidence that H. pylori can cause severe nausea and vomiting
in pregnancy, including hyperemesis gravidarum [68,69]. Thus, if indicated, H.
pylori treatment should be considered in pregnancy.
Some of the medications typically used for the treatment of H. pylori are possibly unsafe
for nursing infants (eg, bismuth, metronidazole, levofloxacin). (See "Medical
164
management of gastroesophageal reflux disease in adults", section on 'Pregnancy and
lactation' and "Prenatal care: Patient education, health promotion, and safety of
commonly used drugs", section on 'Antibiotics'.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Helicobacter pylori".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: H. pylori infection (The

Basics)" and "Patient education: Gastritis (The Basics)")
●Beyond the Basics topics (see "Patient education: Helicobacter pylori infection
and treatment (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●All patients with evidence of active infection with H. pylori should be offered
treatment. The choice of initial antibiotic regimen to treat H. pylori should be
guided by the presence of risk factors for macrolide resistance and the presence
of a penicillin allergy. In patients with risk factors for macrolide resistance,
clarithromycin-based therapy should be avoided (algorithm 1 and table 1).
(See 'Approach to selecting an antibiotic regimen' above.)
Risk factors for macrolide resistance include:
•Prior exposure to macrolide therapy for any reason.
•High local clarithromycin resistance rates ≥15 percent or eradication rates with
clarithromycin-based triple therapy ≤85 percent.
In the United States, given the limited information on antimicrobial resistance
rates, we generally assume clarithromycin resistance rates are greater than 15
percent unless local resistance data indicate otherwise.
●For initial therapy in patients without risk factors for macrolide resistance, we
suggest triple therapy with a proton pump inhibitor (PPI), amoxicillin (1 g twice
daily), and clarithromycin (500 mg twice daily) for 14 days (Grade 2B). We
suggest substitution of amoxicillin with metronidazole only in penicillin-allergic
individuals since metronidazole resistance is common and can reduce the efficacy
of treatment (Grade 2B). (See 'Clarithromycin-based therapy' above.)

165
●We suggest bismuth quadruple therapy as initial treatment in patients with risk
factors for macrolide resistance (algorithm 1 and table 1). Quadruple therapy
consists of a PPI, bismuth subsalicylate, and two antibiotics
(metronidazole and tetracycline) given four times daily for 14 days. Alternatively, a
commercially available combination capsule containing bismuth subsalicylate,
metronidazole, and tetracycline may be used in conjunction with a PPI.
(See 'Salvage therapy for persistent H. pylori infection' above and 'Bismuth
quadruple therapy' above.)
●Tests to confirm eradication should be performed in all patients treated for H.
pylori. Eradication may be confirmed by a urea breath test, fecal antigen test, or
upper endoscopy performed four weeks or more after completion of antibiotic
therapy. PPI therapy should be withheld for one to two weeks prior to testing.
(See 'Confirmation of eradication' above and "Indications and diagnostic tests for
Helicobacter pylori infection in adults", section on 'Diagnostic tests'.)
●In patients with persistent H. pylori infection, the choice of antibiotic therapy
should be guided by the patient's initial treatment regimen and the presence of
relevant antibiotic allergies (algorithm 2 and table 2). For patients failing a course
of H. pylori treatment, we suggest an alternate regimen using a different
combination of medications (Grade 2B). In general, clarithromycin and antibiotics
used previously should be avoided if possible. (See 'Salvage therapy for persistent
H. pylori infection' above.)
●Culture with antibiotic sensitivity testing should be performed to guide antibiotic
treatment in patients who have failed two prior treatment regimens. Compliance
with medications should also be reinforced. We reserve the use of rifabutin-
containing regimens for patients with ≥3 prior antibiotic failures. (See 'Salvage
therapy for persistent H. pylori infection' above.)

166
Antiulcer medications: Mechanism of action,
pharmacology, and side effects
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Apr 13, 2020.
INTRODUCTIONEradication of Helicobacter pylori (H. pylori), withdrawal of
nonsteroidal anti-inflammatory drugs, and antisecretory drugs are the mainstays of
treatment for peptic ulcer disease.
The pharmacology of antiulcer drugs, excluding the antibiotics used to treat H. pylori,
will be reviewed here. For detailed prescribing information, readers should refer to the
individual drug information topics within UpToDate. Comprehensive information on
drug-drug interactions can be determined using the drug interactions tool (Lexicomp
drug interactions). This tool can be accessed from the UpToDate online search page or
through the individual drug information topics in the section on drug interactions.
Complete information on United States Food and Drug Administration (FDA) labeling for
each drug can be accessed using the FDA searchable database.
Treatment regimens for H. pylori and management of peptic ulcer disease are
discussed elsewhere. (See "Treatment regimens for Helicobacter pylori in
adults" and "Peptic ulcer disease: Treatment and secondary prevention".)
ANTISECRETORY AGENTS
Histamine-2 receptor antagonists
Mechanism of action — Histamine-2 receptor antagonists (H2RAs)
(eg, cimetidine, famotidine, and nizatidine) inhibit acid secretion by blocking H2
receptors on the parietal cell (figure 1).

H2RAs are well absorbed after oral dosing; peak serum concentrations occur within one
to three hours. Absorption is reduced 10 to 20 percent by concomitant antacid
administration, but not by food.

H2RAs are eliminated by a combination of hepatic and renal metabolism [1]. Their
bioavailability is reduced 30 to 70 percent by first-pass hepatic metabolism [1]. Renal
clearance is generally greater than accounted for by glomerular filtration, reflecting the
importance of renal tubular secretion [1]. The dose of all the H2RAs is generally
reduced by 50 percent in patients with severe renal failure [1-3]. The half-life
of cimetidine is prolonged with liver failure, but dose reduction is necessary only if renal
failure accompanies severe hepatic disease [1,4].
Adverse effects — Side effects of H2RAs are rare [5].
●Gynecomastia and impotence – Gynecomastia and impotence occur
with cimetidine in a dose- and time-dependent fashion and resolve when
167
 discontinued [6]. This effect is relatively specific for cimetidine and rarely occurs if
treatment is limited to standard doses for eight weeks or less [7].
●Hematopoietic and immune effects – Long-term H2RA use is also associated
with B12 deficiency [8]. H2RAs have also been implicated in idiosyncratic cases of
myelosuppression, thrombocytopenia, neutropenia, anemia, and pancytopenia [9-
13]. Other rare diseases that have been associated with H2RAs include
polymyositis and interstitial nephritis, an immune complex rash, and fever
[9,14,15]. Some immunomodulatory effects may reflect a unique action of the
imidazole ring of cimetidine rather than actions at H2 receptors [15].
●Central nervous system (CNS) effects – H2RAs have been associated with
CNS side-effects including confusion, restlessness, somnolence, agitation,
headaches, dizziness, and, with prolonged therapy, hallucinations, focal twitching,
and seizures [13,14,16]. Although these symptoms are generally reversible upon
discontinuation of the drug, cases with more persistent CNS symptoms have been
reported [17].
Mental status changes appear to be most common in older adults in the intensive
care unit with renal or hepatic dysfunction and are rare in outpatients
[16,18]. Cimetidine has been implicated as the most frequent cause of these CNS
symptoms, but similar side effects have also been described
with famotidine [14,16].
●Hepatic effects – H2RAs are metabolized in the liver by the cytochrome P450
system. Of the H2RAs, cimetidine is the most potent inhibitor of the P450 system
(CYP1A2, 2C9, and 2D6) and can cause significant drug interactions. H2RAs can
rarely cause acute hepatic injury [4,19,20]. The clinical presentation of the hepatic
injury may be cholestatic, hepatocellular, or mixed with features of both cholestatic
and hepatocellular injury [20]. Patients may rarely have features of hypersensitivity
with rash, fever, and/or eosinophilia. The acute hepatitis is rapidly reversible after
withdrawal of the drug. Hepatotoxicity has recurred with rechallenge in a sufficient
number of cases to firmly establish this rare, but important, association [20].
(See "Drug-induced liver injury", section on 'Clinical manifestations'.)
●Cardiac effects – Sinus bradycardia, hypotension, atrioventricular block,
prolongation of the QT interval, and sinus and cardiac arrest have occurred with
the rapid infusion of an H2RA [1,21,22]. Oral therapy also has been reported to
cause cardiac toxicity, although clinically significant effects upon sinus rhythm or
conduction are rare. Possible risk factors for cardiac events include rapid
intravenous infusion, high dose, conditions that would delay drug clearance (eg,
renal or hepatic dysfunction), and underlying cardiac disease [1].
●Renal effects – Mild increases in serum creatinine have been observed
with cimetidine. However, immune-mediated interstitial nephritis is rare [20].
●Other – Contamination with nitrosodimethylamine (NDMA), a probable human
carcinogen, has been found in ranitidine. This has led to its withdrawal by the US
Food and Drug Administration (FDA) [23]. NDMA impurities were found to have
been introduced during the manufacturing processes and as the result of product
degradation during storage. FDA's testing has not found NDMA
in famotidine and cimetidine.
Proton pump inhibitors

168
Mechanism of action — The proton pump inhibitors (PPIs)
(eg, omeprazole, lansoprazole, dexlansoprazole, rabeprazole, pantoprazole,
and esomeprazole) effectively block acid secretion by irreversibly binding to and
inhibiting the hydrogen-potassium ATPase pump that resides on the luminal surface of
the parietal cell membrane (figure 1). (See "Proton pump inhibitors: Overview of use
and adverse effects in the treatment of acid related disorders", section on
'Pharmacology'.)
A complete discussion of the pharmacology, clinical efficacy, and safety of the PPIs is
presented separately. (See "Proton pump inhibitors: Overview of use and adverse
effects in the treatment of acid related disorders".)
Adverse effects — Adverse effects of PPIs are discussed in detail, separately.
(See "Proton pump inhibitors: Overview of use and adverse effects in the treatment of
acid related disorders", section on 'Adverse effects'.)
Indications and comparative efficacy — As a result of stronger acid suppression as
compared with H2RA, PPI use results in faster control of peptic ulcer disease symptoms
and higher ulcer healing rates [24-26]. PPIs are also more effective in preventing
nonsteroidal anti-inflammatory drug (NSAID)-induced gastroduodenal toxicity and in
healing gastroduodenal ulcers associated with NSAIDs when they cannot be
discontinued [27]. PPIs are a component of H. pylori antibiotic regimens and are used in
the treatment of hypersecretory states (eg, gastrinoma). (See "Proton pump inhibitors:
Overview of use and adverse effects in the treatment of acid related disorders", section
on 'Indications for PPI therapy' and "Management and prognosis of the Zollinger-Ellison
syndrome (gastrinoma)" and "Treatment regimens for Helicobacter pylori in adults",
section on 'Initial antibiotic therapy' and "Treatment regimens for Helicobacter pylori in
adults".)
Unlike H2RAs, tolerance does not occur for PPIs, but variable PPI metabolism may
account for some difference in efficacy. (See "Physiology of gastric acid
secretion" and "Proton pump inhibitors: Overview of use and adverse effects in the
treatment of acid related disorders".)
ANTACIDSAntacids usually contain a combination of magnesium
trisilicate, aluminum hydroxide, or calcium carbonate.
Mechanism of action — Antacids can neutralize gastric acid and reduce acid delivery
to the duodenum. The following are hypothesized mechanisms for the acid-independent
actions of antacids; however it is unclear which, if any, of these actions facilitate peptic
ulcer healing [28,29]:
●Aluminum hydroxide binds growth factors and enhances their binding to
experimental ulcers, possibly serving to deliver growth factors to injured mucosa.
●Antacids promote angiogenesis in injured mucosa [30].
●Antacids bind bile acids and also inhibit peptic activity [31].
●Heavy metals are well known to suppress, but generally not eradicate, H. pylori.
Indication — Given the effectiveness of PPIs, antacids are not used in the treatment of
peptic ulcer disease. The role of antacids is limited to the treatment of heartburn
associated with mild intermittent gastroesophageal reflux disease. (See "Medical
management of gastroesophageal reflux disease in adults", section on 'Antacids'.)
Adverse effects — Antacid side effects depend upon the quantity consumed and the
duration of therapy. Magnesium-containing antacids cause diarrhea and
169
hypermagnesemia; the latter only becomes important in patients with renal insufficiency.
Some antacids may also contain sodium, and volume overload can occur in susceptible
patients.
Ingestion of large amounts of calcium and absorbable alkali, particularly calcium
carbonate, can lead to hypercalcemia, alkalosis, and acute or chronic renal injury, a
constellation known as the milk-alkali syndrome [32]. (See "The milk-alkali syndrome".)
Significant aluminum retention only occurs in patients with renal failure and may result
in neurotoxicity and anemia following prolonged treatment with aluminum hydroxide.
Aluminum hydroxide blocks intestinal absorption of phosphate; two weeks of therapy
with moderate doses can result in significant hypophosphatemia, especially if the
patient is on a low phosphate diet or is phosphate depleted for other reasons [33].
(See "Aluminum toxicity in chronic kidney disease", section on 'Other medications'.)
SUCRALFATE
Mechanism of action — Sucralfate is a sulfated polysaccharide, sucrose octasulfate,
complexed with aluminum hydroxide. It prevents acute, chemically-induced mucosal
damage and heals chronic ulcers without altering gastric acid or pepsin secretion or
significantly buffering acid [28,34]. Similar to aluminum-containing antacids, sucralfate
stimulates angiogenesis and the formation of granulation tissue, possibly due to growth
factor binding [28]. Sucralfate also binds to the injured tissue, thereby delivering growth
factors and reducing access to pepsin and acid. Aluminum hydroxide mediates some of
the actions of sucralfate, but the sucrose octasulfate moiety may also have a role by
contributing sulfhydryl groups to reduce oxidant damage to epithelial cells.
Indication — Sucralfate has been reported to suppress H. pylori and inhibit acid
secretion in infected patients with duodenal ulcers [35]. However, sucralfate is not used
to treat peptic ulcers as proton pump inhibitors heal ulcers more rapidly and to a greater
extent [36]. The use of sucralfate is limited to the initial management of
gastroesophageal reflux disease in pregnancy. (See "Medical management of
gastroesophageal reflux disease in adults", section on 'Pregnancy and lactation'.)
Adverse effects — Sucralfate has few adverse effects [34]. It can bind to other drugs if
taken simultaneously. Similar to antacids, significant aluminum retention only occurs in
patients with renal failure [3,37,38]. Sucralfate can also bind to phosphate and lead to
hypophosphatemia [39]. Combining sucralfate and antacids can potentially amplify
these effects. (See "Aluminum toxicity in chronic kidney
disease" and "Hypophosphatemia: Causes of hypophosphatemia".)
BISMUTH
Mechanism of action — Bismuth does not inhibit or neutralize gastric acid. The most
dramatic action of bismuth salts is the suppression of H. pylori [40]. Other actions that
may promote ulcer healing include the following:
●Inhibition of peptic activity but not pepsin secretion [41].
●Bismuth from colloidal bismuth subcitrate (CBS) may bind to ulcer craters [42,43].
●Macrophages, recruited to the edge of the ulcer crater in CBS-treated rats, may
promote healing [42].
●CBS may increase mucosal prostaglandin production and mucus and
bicarbonate secretion.
Indication — The role of bismuth preparations (eg, bismuth subcitrate and bismuth
subsalicylate [BSS]) is limited to part of a quadruple antibiotic therapy regimen in H.
170
pylori-positive ulcers [44]. (See "Treatment regimens for Helicobacter pylori in adults",
section on 'Bismuth quadruple therapy'.)
Adverse effects — In the colon, bismuth salts react with hydrogen sulfide to form
bismuth sulfide, which blackens the stools [40].
Bismuth toxicity is rare with CBS or BSS [40,45]. Bismuth absorption varies with the
specific form of bismuth; absorption is much greater with CBS than with BSS or bismuth
subnitrate [46,47].
Coadministration of histamine-2 receptor antagonists increases bismuth absorption from
CBS, but not from BSS or bismuth subnitrate [48]. Bismuth should be avoided or serum
bismuth concentrations monitored in patients with renal failure [3].
The subsalicylate moiety in BSS is converted to salicylic acid and absorbed; however,
salicylate in the absence of the acetyl group does not inhibit platelet function or appear
to share the same high risk of aspirin for gastrointestinal bleeding [49-52]. However, the
salicylate from BSS will raise serum salicylate levels and can cause salicylate toxicity,
and combination with other salicylate products should therefore be avoided.
MISOPROSTOL
Mechanism of action — Misoprostol is a 15-deoxy-15-hydroxy-16-methyl analogue of
prostaglandin E1. Prostaglandins, particularly of the E and I group, inhibit acid secretion
by selectively reducing the ability of the parietal cell to generate cyclic adenosine
monophosphate in response to histamine [53]. Prostaglandins also enhance mucosal
defense mechanisms [54]. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity", section on 'Misoprostol'.)
Indication — The role of misoprostol in peptic ulcer disease is limited to the prevention
of nonsteroidal anti-inflammatory drug-induced gastroduodenal ulcers. It is
contraindicated in women of childbearing potential who are not on contraception.
(See 'Pregnancy and lactation' below.)
Adverse effects — The most frequent side effects of misoprostol are dose-dependent
cramping, abdominal pain, and diarrhea [55,56]. These side effects interfere with
compliance in many patients. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity", section on 'Misoprostol'.)
POTASSIUM-COMPETITIVE ACID INHIBITORSPotassium-competitive acid
inhibitors (PCABs) work by competing for potassium on the luminal side of the parietal
cell, and cause rapid and reversible inhibition of pumps and therefore acid secretion
[57]. After oral doses, PCABs rapidly achieve high plasma concentrations and have
linear, dose-dependent pharmacokinetics; therefore, they have a rapid onset of action
and achieve a full effect with the first dose. The effect is also seen with repeated doses.
PCABs are not available outside of Asia [58-61]. In randomized controlled trials, the
efficacy with regard to ulcer healing and prevention of NSAID-induced ulcers has been
similar to proton pump inhibitor therapy with a comparable safety profile [60-62]. The
PCAB vonoprazan has been approved for the prevention of nonsteroidal anti-
inflammatory drug-induced ulcers and treatment of peptic ulcer disease in Japan.
Another PCAB, revaprazan, has been approved in Korea. Studies also suggest that
vonoprazan may be effective in combination with antibiotics for the eradication of H.
pylori [63,64].
SPECIAL POPULATIONS

171
Pregnancy and lactation — All histamine-2 receptor antagonists (H2RAs) appear to
be safe in pregnancy, with cimetidine having the most safety data available [65].
Cimetidine are concentrated in breast milk but are compatible with breastfeeding.
Experience with proton pump inhibitors (PPIs) is more limited compared with H2RAs,
but suggests that PPIs are probably safe in pregnancy [66-68]. Limited data are
available on the secretion of PPIs in breast milk. Omeprazole and pantoprazole are
secreted in low concentrations in breast milk [69,70]. However, a large portion of them
is likely to be destroyed by gastric acid in the infant's stomach [71]. (See "Medical
management of gastroesophageal reflux disease in adults", section on 'Pregnancy and
lactation'.)
Most antacids are considered safe in pregnancy and are compatible with breastfeeding
[72]. However, antacids containing sodium bicarbonate and magnesium trisilicate
should be avoided in pregnancy [73]. (See "Medical management of gastroesophageal
reflux disease in adults", section on 'Antacids'.)
Sucralfate is likely safe during pregnancy and lactation because it is poorly absorbed
[74].
Prostaglandins of the E group are uterotropic. Misoprostol has been given with or
without mifepristone to induce abortion [68,71]. As a result, it is contraindicated in
women of childbearing potential who are not on contraception. All patients should be
informed of this risk to minimize the drug being inadvertently given by the patient to a
pregnant woman.
SUMMARY
●Eradication of H. pylori, withdrawal of nonsteroidal anti-inflammatory drugs
(NSAIDs), and antisecretory drugs are the mainstays of treatment for peptic ulcer
disease. (See 'Introduction' above.)
●Histamine-2 receptor antagonists (H2RAs) inhibit acid secretion by blocking H2
receptors on the parietal cell (figure 1). Proton pump inhibitors (PPIs) effectively
block acid secretion by irreversibly binding to and inhibiting the hydrogen-
potassium ATPase pump that resides on the luminal surface of the parietal cell
membrane. (See 'Histamine-2 receptor antagonists' above and 'Proton pump
inhibitors' above.)
●PPI use results in faster control of peptic ulcer disease symptoms and higher
ulcer healing rates as compared with H2RAs. PPIs are also more effective in
preventing NSAID-induced gastroduodenal toxicity and in healing gastroduodenal
ulcers associated with NSAIDs when they cannot be discontinued. PPIs are a
component of H. pylori antibiotic regimens and are used in the management of
hypersecretory states (eg, Zollinger-Ellison syndrome). The use of H2RAs is
largely limited to the management of mild gastroesophageal reflux disease.
(See 'Indications and comparative efficacy' above.)
●Antacids and sucralfate have no role in the treatment of peptic ulcer disease. The
role of bismuth preparations (eg, bismuth subcitrate and bismuth subsalicylate) is
as part of a quadruple antibiotic therapy regimen in H. pylori-positive ulcers.
(See 'Antacids' above and 'Sucralfate' above and 'Bismuth' above.)
●Prostaglandin analogues (eg, misoprostol) are effective for preventing NSAID-
induced ulcers, but they have no established role for healing ulcers. Misoprostol is

172
contraindicated in women of childbearing potential who are not on contraception.
(See 'Misoprostol' above.)
●Potassium-competitive acid inhibitors (PCABs) work by competing for potassium
on the luminal side of the parietal cell, and cause rapid and reversible inhibition of
pumps and therefore acid secretion. PCABs are not widely available outside of
Asia, and data on safety and comparative efficacy with antisecretory agents are
limited.

173
Approach to refractory peptic ulcer disease
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jan 11, 2022.
INTRODUCTIONA peptic ulcer is an excavated defect in the gastric or duodenal
mucosa that extends through the muscularis mucosa into the deeper layers of the wall.
Most peptic ulcers respond to treatment with antimicrobial therapy for Helicobacter
pylori, withdrawal of nonsteroidal antiinflammatory drugs, or treatment with antisecretory
drugs. However, in some individuals, the ulcer is either refractory to conventional
therapy or recurs following successful initial treatment.
This topic will review the factors associated with refractory ulcer disease, and their
diagnosis and management. The clinical manifestations, diagnosis, and initial
management of peptic ulcer disease are discussed in detail, separately. (See "Peptic
ulcer disease: Clinical manifestations and diagnosis" and "Peptic ulcer disease:
Treatment and secondary prevention".)
TERMINOLOGY
Refractory peptic ulcer — A refractory peptic ulcer is defined as an endoscopically
proven ulcer greater than 5 mm in diameter that does not heal after 8 to 12 weeks of
treatment with a proton pump inhibitor.
EPIDEMIOLOGYIn the absence of continued nonsteroidal antiinflammatory drug
use, acid suppression heals >90 percent of peptic ulcers. However, approximately 5 to
10 percent of ulcers are refractory to 12 weeks of antisecretory therapy with a proton
pump inhibitor (PPI). Even with continued PPI use, approximately 5 to 30 percent of
peptic ulcers recur within the first year based on whether Helicobacter pylori has been
successfully eradicated [1,2].
ETIOLOGY AND RISK FACTORSThere is significant overlap in the risk factors
for refractory and recurrent peptic ulceration (table 1 and table 2) [3-6].
Persistent H. pylori infection — Persistent Helicobacter pylori can underlie refractory
peptic ulceration because this infection was not initially considered, testing was falsely
negative, or eradication therapy failed [7]. Failure of eradication therapy may be due to
the selection of an inappropriate regimen, antibiotic resistance (particularly
to clarithromycin), or poor patient compliance. (See "Treatment regimens for
Helicobacter pylori in adults", section on 'Treatment failure'.)
Use of culprit medications
Continued nonsteroidal anti-inflammatory drug (NSAID) use — Continuing NSAID
use is a leading cause of refractory peptic ulceration [7-9]. In an illustrative study that
included 60 patients with refractory peptic ulcer and 54 non-refractory matched controls,
NSAID and analgesic use was a significant predictor of ulcer refractoriness, present in

174
40 percent of those with nonhealing ulcers [10]. Forty-four percent of NSAID use was
surreptitious and was detected by measuring platelet cyclooxygenase activity.
Other medications/substance use — Several medications can cause or exacerbate
peptic ulcer disease when used alone or in combination with NSAIDs (eg,
glucocorticoids, cytotoxic agents, alendronate, olmesartan) [11]. Substances (eg,
cocaine) can cause peptic ulceration from ischemia due mucosal vasoconstriction
[12,13]. (See "Unusual causes of peptic ulcer disease", section on 'Non-NSAID
medications' and "Unusual causes of peptic ulcer disease", section on 'Non-occlusive
ischemia'.)
Impaired healing
Ulcer characteristics — In a subset of patients, refractory ulcers may be caused by an
intense inflammatory response, dense scarring, or low mucosal blood flow, which impair
angiogenesis and tissue repair [14]. Ulcer size also impacts the healing time. Some
studies suggest that large and small gastric ulcers heal at the same rate of
approximately 3 mm per week, and thus larger ulcers will require more time to heal [12].
However, another explanation for slow healing in patients with large ulcers is that they
are often associated with fibrosis, which in turn adversely affects ulcer healing [13].
Comorbid diseases — Risk factors associated with impaired wound healing include
uremia, respiratory failure, organ transplantation, cirrhosis, and a critical illness [15].
Smoking — Cigarette smoke and its active ingredients can suppress mucosal cell
proliferation and induce apoptosis during ulceration and adversely affect the healing
processes [16].
Ineffective antisecretory therapy
Limited adherence or rapid metabolism — Limited adherence with antisecretory
therapy and tolerance to histamine 2 receptor antagonists can contribute to refractory
peptic ulceration. Although tolerance does not develop with proton pump inhibitors
(PPIs), rapid P450 mediated metabolism might account for incomplete inhibition of acid
secretion and failure of PPI therapy in a small subset of patients. Other mechanisms of
PPI resistance are rare and still poorly defined. (See "Proton pump inhibitors: Overview
of use and adverse effects in the treatment of acid related disorders", section on
'Pharmacology'.)
Acid hypersecretory states — Acid hypersecretion is associated with refractory peptic
ulceration in patients with Zollinger Ellison syndrome (gastrinoma) [17]. Increased acid
production and ulcer disease have been described in patients with primary
hyperparathyroidism [18]. (See "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis" and "Unusual causes of peptic ulcer disease", section on
'Acid hypersecretory states'.)
A small number of patients with acid hypersecretion have neither of these risk factors.
Smoking and genetic factors may play an important role in the pathogenesis of these
ulcers. A proportion of these patients have idiopathic acid hypersecretion. Idiopathic
acid hypersecretion has been defined as a condition with a high basal acid output of 10
mEq/hour or more with a normal basal serum gastrin or a negative secretin test [19,20].
However, all refractory ulcers in these individuals subsequently heal with high-dose
antisecretory therapy. (See "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis".)

175
Other underlying disease — Other rare causes of gastric and duodenal ulceration
include Crohn disease, sarcoid, lymphoma, ischemia, eosinophilic gastroenteritis,
tuberculosis, syphilis, cytomegalovirus, IgG4-related sclerosing disease, and mesenteric
ischemia. (See "Unusual causes of peptic ulcer disease" and "Clinical manifestations,
diagnosis, and prognosis of Crohn disease in adults", section on 'Other gastrointestinal
features' and "Eosinophilic gastrointestinal diseases", section on 'Clinical
manifestations' and "Extrapulmonary manifestations of sarcoidosis", section on
'Gastrointestinal' and "Clinical presentation and diagnosis of primary gastrointestinal
lymphomas", section on 'Gastric lymphoma' and "Abdominal tuberculosis", section on
'Intestinal tuberculosis' and "Epidemiology, clinical manifestations, and treatment of
cytomegalovirus infection in immunocompetent adults", section on 'Gastrointestinal
manifestations' and "Syphilis: Epidemiology, pathophysiology, and clinical
manifestations in patients without HIV", section on 'Gastrointestinal findings'.)
DIAGNOSISRefractory ulcers are suspected in patients with peptic ulcer disease with
persistent or recurrent dyspepsia. They are diagnosed at upper endoscopy performed
for evaluation of symptoms or on routine endoscopic surveillance after initial therapy for
peptic ulcer disease. (See "Peptic ulcer disease: Treatment and secondary prevention",
section on 'Repeat upper endoscopy in selected patients'.)
Upper endoscopy in patients with refractory ulcers can also help determine the
underlying etiology. Biopsies of gastric ulcers should be performed to exclude an
underlying malignancy and other causes of ulceration (eg, Crohn disease, sarcoid,
eosinophilic gastroenteritis). If possible, we obtain biopsies from four quadrants of the
ulcer. If endoscopic features suspicious for malignancy (eg, nodularity at the edges of
the ulcer or infiltration of the surrounding tissue creating a heaped up appearance at the
edge of the ulcer), we obtain biopsies using jumbo forceps with more extensive
sampling along the edges of the ulcer. In addition, we biopsy the gastric antrum and
body for Helicobacter pylori. (See "Indications and diagnostic tests for Helicobacter
pylori infection in adults", section on 'Endoscopic testing'.)
MANAGEMENT
Address the etiology and risk factors
●Re-evaluate risk factors – In patients with refractory ulcers, it is important to
reevaluate the underlying etiology and risk factors for peptic ulcer disease. Key
elements of the history include the following:
•Compliance with antisecretory therapy
•Continued nonsteroidal anti-inflammatory drug (NSAID) use
•Use of medications/substances associated with peptic ulcers or that may
impact healing
•Risk factors associated with poor ulcer healing (smoking, co-morbid diseases)
(see 'Comorbid diseases' above)
In patients for whom the etiology remains uncertain based on the history and
biopsy results from the upper endoscopy that established their diagnosis, selected
laboratory testing (fasting serum gastrin and serum calcium levels) may help
elucidate the underlying etiology. (See 'Diagnosis' above and "Unusual causes of
peptic ulcer disease", section on 'Evaluation of H. pylori and NSAID negative
ulcers'.)

176
 ●Eradicate Helicobacter pylori (H. pylori) – Most patients with refractory
ulceration have H. pylori. Eradication of H. pylori improves ulcer healing rates in
patients with peptic ulcers [7,21-24]. In patients with refractory ulcers who do not
have evidence of H. pylori on gastric biopsies, we perform additional testing to
confirm the negative result. In patients treated for H. pylori, eradication of infection
should be confirmed four or more weeks after the completion of therapy.
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults",
section on 'Confirmation of eradication'.)
●Avoid culprit medications and tobacco – Patients should be advised to avoid
NSAIDs. We also advise cessation of tobacco use given that it affects gastric
microcirculation adversely and impairs ulcer healing [25-29]. (See "Peptic ulcer
disease: Epidemiology, etiology, and pathogenesis", section on 'Smoking'.)
●Treat the underlying cause – Patients with other causes of ulceration (eg,
Crohn disease, sarcoidosis) require concurrent treatment to facilitate ulcer healing
and prevent recurrence. (See "Unusual causes of peptic ulcer disease", section on
'Etiology'.)
Antisecretory therapy
Twice-daily dosing (eg, omeprazole 20 mg twice daily) is usually effective in inducing
healing in patients with ulcers that were refractory to once daily standard dose proton
pump inhibitor (PPI) [3]. After 12 additional weeks of PPI therapy, we perform a repeat
upper endoscopy to assess ulcer healing and obtain additional biopsies. Long-term
maintenance acid inhibitory therapy (eg, omeprazole 20 mg daily) should be routinely
offered to these patients once the ulcer has healed. (See 'Repeat upper
endoscopy' below.)
PPIs are the preferred therapy for refractory ulcer disease. Healing rates for duodenal
ulcers, and to a lesser extent gastric ulcers, are associated with the degree of inhibition
of acid secretion [30,31]. In a report of patients with refractory ulcers after three months
of H2RA therapy, a 40 mg dose of omeprazole produced better healing than continued
standard H2RA (96 versus 57 percent) [32]. With antisecretory therapy with a PPI >90
percent of refractory ulcers heal with an additional eight weeks of treatment. Potassium
competitive acid inhibitors have are comparable to PPIs in preventing NSAID-related
ulcer disease but are not available in many countries [33].
Repeat upper endoscopy — In patients with refractory ulcers, endoscopy should be
performed after 12 weeks of additional treatment with a PPI. This endoscopy serves to
document ulcer healing.
We biopsy the ulcer scar if the ulcer has healed as some neoplastic ulcers can heal with
antisecretory therapy. If the ulcer has not healed, we obtain biopsies from four
quadrants of the ulcer ideally using jumbo forceps with more extensive sampling along
the edges of the ulcer. In addition, we biopsy the gastric antrum and body for H. pylori.
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults", section
on 'Endoscopic testing'.)
Biopsies for culture and sensitivity should be performed to guide antibiotic therapy in
patients with persistent H. pylori infection after two courses of antibiotic treatment.
(See "Treatment regimens for Helicobacter pylori in adults", section on 'Confirmation of
eradication'.)

177
Surgery in selected patients — Surgical management is rarely required and is
reserved for the peptic ulcers that fail to heal after twice-daily antisecretory therapy with
a PPI for 24 weeks in whom other correctable factors (eg, medication noncompliance,
NSAID use, and H. pylori infection) have been addressed. Indications for surgical
management of peptic ulcer disease and surgical treatment options are discussed in
detail separately. (See "Surgical management of peptic ulcer disease", section on
'Natural history of peptic ulcer disease'.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Peptic ulcer disease".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)" and "Patient

education: H. pylori infection (The Basics)")
●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●A refractory peptic ulcer is defined as an endoscopically proven ulcer greater
than 5 mm in diameter that does not heal after 12 weeks of treatment with a
proton pump inhibitor (PPI). (See 'Terminology' above.)
●In the absence of continued nonsteroidal anti-inflammatory drug (NSAID) use,
acid suppression heals >90 percent of peptic ulcers. However, approximately 5 to
10 percent of ulcers are refractory to 12 weeks of antisecretory therapy with a PPI.
Even with continued PPI use, approximately 5 to 30 percent of peptic ulcers recur
within the first year. While there are several factors that are associated with
refractory gastric/duodenal ulceration, persistent Helicobacter pylori (H. pylori)
infection and NSAID use are the two main causes. (See 'Epidemiology' above.)
●Refractory ulcers are suspected in patients with peptic ulcer disease with
persistent or recurrent dyspepsia. They are diagnosed at upper endoscopy
performed for evaluation of symptoms or on routine endoscopic surveillance after
initial therapy for peptic ulcer disease. Biopsies of the ulcer to exclude an
underlying malignancy and other causes of ulceration (eg, Crohn disease,
178
sarcoid). In addition, we obtain biopsies of the antrum and body to diagnose or
exclude H. pylori. In patients with refractory ulcers who do not have evidence of H.
pylori on gastric biopsies, we perform additional testing to confirm the negative
result. (See 'Diagnosis' above.)
●Patients with refractory ulcers should be evaluated for continued NSAID use, the
use of other medications associated with peptic ulcers, and comorbidities and
other factors that may contribute to poor ulcer healing and recurrence. Fasting
serum gastrin and total calcium levels should be measured to exclude Zollinger-
Ellison syndrome and hyperparathyroidism, respectively.
(See 'Epidemiology' above.)
●Patients should be advised to avoid NSAIDs and tobacco. Patients with H.
pylori should be treated with a goal of eradication of H. pylori infection. Eradication
of infection should be confirmed after the completion of therapy.
(See 'Management' above.)
●In patients with refractory peptic ulcer disease, the mainstay of therapy is
additional antisecretory therapy. For patients who fail to respond to standard
doses of PPI (eg, omeprazole 40 mg daily), twice-daily dosing is usually effective
in inducing healing. We perform a repeat endoscopy to document ulcer healing
following 12 additional weeks of PPI therapy. Surgical management is reserved for
the gastric ulcers that fail to heal after twice-daily antisecretory therapy with a PPI
for 24 weeks in whom other correctable factors (eg, medication noncompliance,
NSAID use, and H. pylori infection) have been addressed.
(See 'Management' above.)

179
Association between Helicobacter pylori infection
and duodenal ulcer
Author:
J Thomas Lamont, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jun 15, 2021.
INTRODUCTIONThe majority of patients with duodenal ulcer (DU) are infected
with Helicobacter pylori (H. pylori). However, in multicenter trials, H. pylori was absent
in almost 30 percent of patients with an endoscopically documented duodenal ulcer [1].
Studies that have investigated these patients found that they have generally had a
shorter duration of symptoms and that many had regularly used nonsteroidal anti-
inflammatory drugs (NSAIDs) [2-4]. Such patients have a significantly worse outcome,
especially if treated empirically for H. pylori infection. Thus, H. pylori status should be
determined in all ulcer patients before initiating treatment [4]. High acid output may be a
cause of recurrent DU in patients in whom H. pylori has been eradicated [5]. A variety of
other causes are responsible for the remaining cases. (See "Unusual causes of peptic
ulcer disease".)
The link between H. pylori and DU will be reviewed here. The pathophysiology of H.
pylori infection as it relates to gastrointestinal disease in general is discussed
separately. (See "Pathophysiology of and immune response to Helicobacter pylori
infection".)
EVIDENCE LINKING HELICOBACTER PYLORI TO DUODENAL
ULCERSThere are several lines of evidence that implicate H. pylori as a major
etiologic factor in duodenal ulcers (DU):
●H. pylori is present in most patients who have a DU that is not related to NSAID
use
●H. pylori infection is detectable before the occurrence of DU and appears to be a
risk factor for the disorder
●Eradication of H. pylori prevents DU recurrence
Incidence of H. pylori in patients with duodenal ulcer — Early studies noted a high
incidence of H. pylori infection (then called Campylobacter pylori) in patients with DU
[6]; subsequent reviews confirmed that H. pylori is detectable in 80 to 95 percent of
these patients [2,7]. These data were supported by reports which found that the
prevalence of H. pylori is negligible in populations in which ulcer disease is rare [8].
The prevalence of H. pylori in patients with DU is changing in some parts of the world
[9]. While H. pylori infection remains very common in patients from Asia with DU, it is
becoming less common in patients from the United States and parts of Europe. In an
epidemiologic study from China, 1030 patients underwent endoscopy and 73 percent
were found to have H. pylori [10]. Peptic ulcers were present in 17 percent, more than
180
two-thirds of which were duodenal ulcers. Among those with peptic ulcers, the
prevalence of H. pylori infection was 93 percent. On the other hand, in the United
States and parts of Europe, the prevalence of H. pylori in patients with ulcer disease
appears to be falling and is now in the range of 50 to 75 percent. (See "Peptic ulcer
disease: Epidemiology, etiology, and pathogenesis", section on 'H. pylori'.)
While the association between H. pylori and DU is strong, it is not specific. As
examples, H. pylori infection is also found in patients with gastric ulcers (65 to 95
percent), dyspepsia (20 to 60 percent), gastric cancer (70 to 90 percent), and
asymptomatic patients (20 to 45 percent) [2]. (See "Approach to the adult with
dyspepsia" and "Association between Helicobacter pylori infection and gastrointestinal
malignancy".)
H. pylori occurs before manifestations of the disease — Several trials have found
that preexisting H. pylori infection is a risk factor for the development of DU [11-14].
One study, for example, reviewed the cases of more than 5000 Native Hawaiians who
had stored serum from the late 1960s [12]. Of the 65 patients who developed DU over
the next 20 years, 92 percent were H. pylori positive versus 78 percent of controls,
producing an odds ratio of 4.0. Similarly, an endoscopic study that included 526 patients
with long-term follow-up found that the odds ratio of developing DU was 5.0 for those
with preexisting H. pylori infection compared with H. pylori-negative subjects [14].
Eradication of H. pylori reduces disease recurrence — Treatment of H.
pylori infection in patients with DU decreases the incidence of ulcer recurrence. One
meta-analysis examined the recurrence rate for DU after at least six months of follow-
up. The recurrence rate was 6 percent if H. pylori was eradicated and 67 percent if it
was not [15]. A second meta-analysis found recurrence rates of 20 and 56 percent,
respectively [16]. The studies in the meta-analyses used endoscopic findings to define
ulcer recurrence. It should be noted that the rate of symptomatic recurrence is lower.
(See "Approach to refractory peptic ulcer disease".)
PATHOGENESIS OF ULCER FORMATIONThe precise mechanism by
which H. pylori contributes to duodenal ulcer (DU) formation is incompletely understood.
However, the bacterium appears to affect the following aspects of intestinal and
mucosal physiology:
●Increased gastric acid secretion
●Gastric metaplasia
●Immune response
●Mucosal defense mechanisms
There may also be a contribution from a variety of bacterial, host, and environmental
factors that have a role in the pathogenesis of ulcer formation (algorithm 1).
Increased gastric acid secretion — Acute H. pylori infection induces a short period of
hypochlorhydria. In contrast, chronic infection often leads to increases in basal and
stimulated acid output, particularly in patients who develop DUs (figure 1) [17,18]. In
addition, following H. pylori eradication, mean basal and stimulated acid output drops by
50 percent at one month and returns to normal levels by one year [17].
One mechanism by which H. pylori may enhance gastric acid secretion is via increased
release of gastrin. Gastrin is responsible for gastric acid secretion in normal subjects by
two mechanisms:

181
 ●It has a trophic action on parietal cells and histamine-secreting enterochromaffin-
like (ECL) cells
●It stimulates parietal cells largely via the release of histamine
This process is tightly controlled by a second hormone, somatostatin, which is a potent
inhibitor of gastrin synthesis, gastrin release, and gastric acid secretion.
(See "Physiology of gastric acid secretion".)
Patients with H. pylori infection have elevated basal and stimulated concentrations of
serum gastrin [17,19] and a decreased concentration of somatostatin [20]. The peak
acid output achieved after stimulation with gastrin releasing peptide, regarded as a
measure of the stomach's functional response to endogenous gastrin, is also increased
threefold in healthy H. pylori positive patients, and sixfold in those with DU [17,18].
These values completely normalize within one year of H. pylori eradication (figure 1)
[21].
The peak acid output after pentagastrin stimulation (PAOPg), a measure of the
functional parietal cell population, is also significantly increased in H. pylori-associated
DU disease. However, there are conflicting data regarding the effect of H.
pylori eradication on PAOPg [17,22,23].
The mechanism of the increased acid response in patients infected with H. pylori who
have DU disease compared with infected healthy patients is unclear. One study
suggests that it may be due to the combination of decreased sensitivity to gastrin in
healthy patients, and increased maximal acid secretory capacity in those with DU [24].
Suppression of somatostatin (rather than an increase in gastrin-secreting cells) is
probably the initial event that then leads to the hypergastrinemia that is evident in
patients with DU and H. pylori. This hypothesis is supported by a study of 28 infected
patients with DU that found that after the infection was successfully eradicated, the
median density of somatostatin-immunoreactive cells (antral D cells) increased from 9
to 19 cells/mm muscularis mucosa, while the median somatostatin mRNA/rRNA ratio
increased from 50 to 95 [25]. In contrast, the number of gastrin-immunoreactive cells
and the quantity of gastrin mRNA did not change significantly. The role of H. pylori in
this process is supported by studies that have shown eradication of infection reverses
the abnormalities in gastrin and gastric acid secretion (figure 1) [17,22].
Nevertheless, hypergastrinemia alone probably does not explain the increases in acid
output in patients infected with H. pylori; gastrin levels often return to normal within one
month after eradication, while mean basal and peak acid output remain elevated for
longer periods of time. In addition, H. pylori itself does not appear to alter the sensitivity
of gastric parietal cells to gastrin [26], although hypergastrinemia may have trophic
effects over time that result in increased parietal cell mass.
Gastric metaplasia — Gastric metaplasia refers to the presence of gastric epithelium in
the first portion of the duodenum. This abnormality appears to be a response of the
mucosa to excessive acid exposure since it only occurs when the luminal pH is less
than 2.5 [27]. In addition to acid hypersecretion, impaired duodenal bicarbonate
secretion induced by H. pylori also may contribute to the low duodenal luminal pH [28].
The metaplastic foci provide areas for H. pylori colonization, and probably have a role in
the development of duodenitis. In one study, for example, 30 of 34 (88 percent) patients
with active duodenitis had more than 5 percent gastric metaplasia in duodenal mucosal
specimens and H. pylori-associated gastritis [27]. In contrast, these two factors

182
coexisted in only 0.43 percent of patients with no duodenal inflammation. In addition,
when H. pylori was seen histologically in duodenal biopsy specimens, it was confined
only to areas of gastric metaplasia, and never occurred in the absence of a
polymorphonuclear infiltrate. A more recent study confirmed that the extent of duodenal
gastric metaplasia was fourfold greater in H. pylori-infected duodenal ulcer patients than
in infected asymptomatic controls [29]. Colonization of the duodenal bulb by H.
pylori CagA+ strains was also greater in DU patients than in infected non DU controls
(80 versus 30 percent), even though prevalence of CagA+ gastric infection was similar
in both groups. These studies support the contention that infection of gastric metaplasia
is important in the pathogenesis on DU.
H. pylori infection in areas of gastric metaplasia may weaken the mucosa, making it
more susceptible to acid injury. This hypothesis is supported by studies which have
found that gastric metaplasia increases the relative risk for ulceration fivefold; the
presence of H. pylori within these regions increases the risk 50-fold [30].
Gastric metaplasia is not the same as gastric heterotopia, in which (presumably)
congenital patches of gastric-type mucosa occur anywhere in the small or large
intestine, although the two entities may be difficult to distinguish. (See "Metaplastic
(chronic) atrophic gastritis".)
Immune response — Although H. pylori is a noninvasive organism, it stimulates a
robust inflammatory and immune response that may have a role in DU formation. The
response includes increased production of inflammatory cytokines such as interleukin
(IL)-1, IL-6, tumor necrosis factor alpha, and most notably, IL-8. (See "Pathophysiology
of and immune response to Helicobacter pylori infection".)
Mucosal defense factors — H. pylori may downregulate several important mucosal
defense factors.
●Epidermal growth factor (EGF) and transforming growth factor-alpha (TGF alpha)
are potent gastric acid inhibitors and stimuli of mucosal growth and protection.
One study found that the basal and stimulated release of EGF is significantly
increased after eradication of H. pylori, an effect that may play a role in ulcer
healing [31].
●Patients with DUs have decreased proximal duodenal mucosal bicarbonate
production. It is not clear if impaired bicarbonate secretion can be attributed to H.
pylori; however, eradication of the infection does result in normalization of
bicarbonate output [28].
●H. pylori itself releases proteases that degrade normally protective mucous
glycoproteins that overlie the mucosa [32].
Other contributing factors — Only 10 to 15 percent of patients with H. pylori infection
develop ulcer disease, suggesting that other factors are probably important in
determining the outcome of infection. One such factor is the bacterial strain; only strains
with the cytotoxin-associated gene A (cagA), coding for a 128-140 kDa protein (CagA),
coexpress vacuolating cytotoxin (VacA), a toxin that causes cell injury in vitro [33].
Approximately 85 to 100 percent of patients with DU have CagA+ strains, compared
with 30 to 60 percent of infected patients who do not develop ulcers [34]. However,
since CagA+ strains are also associated with gastric cancer, the specificity of the
association makes it less clear how the same strain can lead to two very divergent

183
conditions. (See "Pathophysiology of and immune response to Helicobacter pylori
infection".)
A specific H. pylori gene, dupA, appears to be associated with development of DU.
Those infected with dupA bacteria had more intense antral inflammation, higher levels
of IL-8 and less gastric atrophy, intestinal metaplasia, and gastric cancer, a cytokine
and histologic profile associated with DU disease [35]. One study suggested that the
dupA gene, which encompasses jhp0917 and jhp0918, was associated with an
increased risk of DU and protection against gastric cancers [36]. The 112bp region of H.
pylori dupA may be associated with an increased risk of duodenal ulcer and has the
potential to serve as a biomarker for early detection of DU [37,38].
In addition, genetic factors may help determine the susceptibility to infection with H.
pylori [39]. It has been suggested that patients with H. pylori who develop DU have an
intrinsically higher parietal cell mass or sensitivity to gastrin than H. pylori-positive
healthy adults [24,30,40]. This observation may explain the exaggerated acid response
to various stimuli in those who develop DUs. Genetic factors may also determine
duodenal cytokine response to infection. In one study, duodenal epithelial cells from
patients with a duodenal ulcer had a decreased capacity to produce inflammatory
cytokines compared with controls from patients without a DU [41].
Environmental factors such as smoking and nonsteroidal anti-inflammatory drug
(NSAID) use may also increase the risk of ulcer formation in patients with H.
pylori [32,42-44]. One report noted an increased prevalence of ulcers in H. pylori-
infected smokers compared with infected nonsmokers (73 versus 27 percent) [43].
Multiple studies of variable design have evaluated the relationship between NSAIDs
and H. pylori in the development of peptic ulcer disease. The following conclusions
were reached in a meta-analysis of 25 such studies [45]:
●Users of NSAIDs with H. pylori infection were 61 times more likely to have a
peptic ulcer than non-infected non-NSAID users.
●Either factor alone increased the risk of ulcer disease by approximately 20-fold.
●H. pylori infection and NSAID use increased the risk of ulcer bleeding by 1.8 and
4.9-fold, respectively; when present together they increased the risk of ulcer
bleeding by sixfold.
A subsequent meta-analysis of studies published between 1999 and 2008 with a total of
16,080 patients found that the mean prevalence of H. pylori infection in DUs was 81
percent [46]. If only the last five years' literature was considered, the rate was 77
percent. H. pylori-negative DUs were associated with NSAID use in 21 percent of
patients when studies with infection rates of less than 90 percent were examined. The
majority of the H. pylori-negative ulcers were attributed to false-negative H.
pylori testing, since even under optimal conditions the sensitivity of testing is below 95
percent [47]. Other causes for H. pylori-negative DUs were much less common.
NSAID users with a history of peptic ulcer disease should be tested and treated for H.
pylori if they are infected; there are also data to suggest that testing and treating (if
positive) all patients for H. pylori prior to beginning NSAID use may be warranted
[48,49], although conflicting data have also been reported [50], and this is not currently
the standard of care. (See "NSAIDs (including aspirin): Pathogenesis of gastroduodenal
toxicity".)

184
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)" and "Patient

education: H. pylori infection (The Basics)")
●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●There are several lines of evidence that implicate H. pylori as a major etiologic
factor in duodenal ulcers (DUs) (see 'Evidence linking Helicobacter pylori to
duodenal ulcers' above):
•H. pylori is present in most patients who have a DU that is not related to
nonsteroidal anti-inflammatory drug (NSAID) use
•H. pylori infection is detectable before the occurrence of DU and appears to
be a risk factor for the disorder
•Eradication of H. pylori prevents DU recurrence
●In the United States and parts of Europe, the prevalence of H. pylori in patients
with ulcer disease appears to be falling. It is now in the range of 50 to 75 percent
but remains high in Asia. (See 'Incidence of H. pylori in patients with duodenal
ulcer' above.)
●The precise mechanism by which H. pylori contributes to DU formation is
incompletely understood. However, the bacterium appears to affect the following
aspects of intestinal and mucosal physiology which may be important
(see 'Pathogenesis of ulcer formation' above):
•Increased gastric acid secretion
•Gastric metaplasia
•Immune response
•Mucosal defense mechanisms
●Only 10 to 15 percent of patients with H. pylori infection develop ulcer disease,
suggesting that other factors are probably important in determining the outcome of
infection. (See 'Other contributing factors' above.)

185
●In apparent H. pylori-negative DUs, a thorough search for infection should be
conducted as infection remains the major cause of DUs. (See 'Other contributing
factors' above.)

186
COX-2 inhibitors and gastroduodenal toxicity: Major
clinical trials
Author:
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jan 25, 2021.
INTRODUCTIONThe primary effect of nonsteroidal anti-inflammatory drugs
(NSAIDs) is to inhibit cyclo-oxygenase (prostaglandin H synthase, or PGHS), thereby
impairing the ultimate transformation of arachidonic acid to prostaglandins, prostacyclin,
and thromboxanes. Two related isoforms of the cyclo-oxygenase (COX) enzyme have
been described: COX-1 (PGHS-1) and COX-2 (PGHS-2). COX-1 is involved in gastric
cytoprotection. It had been proposed that the ideal NSAID would inhibit the inducible
COX-2 isoform (thereby decreasing inflammation) without having any effect on the
constitutive COX-l isoform (thereby minimizing gastric toxicity) [1]. (See "NSAIDs:
Pharmacology and mechanism of action".)
One NSAID that selectively inhibits COX-2, celecoxib, is currently approved by the US
Food and Drug Administration (FDA). Rofecoxib is a COX-2 inhibitor that was removed
due to an increased risk of stroke and myocardial infarction with long-term use.
Valdecoxib was removed because of concerns of cardiovascular risk and reports of
Stevens-Johnson syndrome. An increased risk of cardiovascular events has also been
observed with celecoxib, especially in higher doses. These observations led the FDA to
issue warnings regarding the use of these drugs. (See "NSAIDs: Adverse
cardiovascular effects".)
Other COX-2 inhibitors (parecoxib, etoricoxib, and lumiracoxib) are available in some
countries. Parecoxib is a prodrug that is converted to valdecoxib. These drugs have at
least a 200- to 300-fold selectivity for inhibition of COX-2 over COX-1. (See "Overview
of COX-2 selective NSAIDs".)
This topic review will summarize the major clinical trials that have focused on the
gastroduodenal protective effects of the COX-2 inhibitors as compared to nonselective
NSAIDs. An approach to patients at risk for gastroduodenal toxicity is presented
separately. (See "NSAIDs (including aspirin): Primary prevention of gastroduodenal
toxicity".) An overview of COX-2 inhibitors is also available. (See "Overview of COX-2
selective NSAIDs".)
CELECOXIBCelecoxib was approved by the FDA based upon the results of clinical
trials involving more than 5200 patients with osteoarthritis (OA) or rheumatoid arthritis
(RA) in which its efficacy and toxicity were compared with that of nonselective
nonsteroidal anti-inflammatory drugs (NSAIDs) and placebo. These data demonstrated
that celecoxib produced comparable analgesic and anti-inflammatory effects to
nonselective NSAIDs.

187
Two trials in patients with RA that used a primary endpoint of endoscopically detected
ulcers found a lower ulcer incidence in patients on celecoxib compared with
nonselective NSAIDs [2,3]:
●One of the published studies included 688 patients with RA who completed a
controlled trial comparing various doses of celecoxib to naproxen or placebo for
12 weeks [2]. All doses of celecoxib and naproxen improved symptoms and signs
of RA compared with placebo. The incidence of endoscopically determined
gastroduodenal ulcers among patients taking celecoxib was similar to placebo
(approximately 4 percent) and was much lower than observed with naproxen (26
percent).
●Similar results were found in a second study of 655 patients with RA, which
compared the efficacy and gastrointestinal toxicity
of celecoxib versus diclofenac [3].

Subsequent trials that focused on more clinically relevant endpoints of symptomatic

ulcers and ulcer complications have been conflicting:

●The CLASS study, involving 8059 patients with OA or RA, found

that celecoxib was associated with significantly fewer symptomatic ulcers and
ulcer complications than ibuprofen or diclofenac during the first six months of
therapy [4]. However, during the second six months of treatment (a time period
that was not included in the published report) the incidence of ulcer complications
was higher in the celecoxib than in the ibuprofen or diclofenac treated groups. As
a result, at the end of a year of observation, there were no significant differences
in incidence of ulcer complications among the three groups [5]. Possible reasons
for the apparent failure of celecoxib to reduce the risk of ulcer complications in this
study include the use of low-dose aspirin in approximately 20 percent of patients,
the use of a high (less COX-2 selective) dose of celecoxib (800 mg per day), and
the use of comparator NSAIDs with relatively low potential for causing such
adverse GI effects. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity".)
●A second trial (the SUCCESS-I study) included over 13,000 patients with OA
from 39 countries, 8,800 of whom were randomly assigned to celecoxib (100 or
200 mg twice daily for 12 weeks) and 4394 of whom were assigned to a
nonselective NSAID (diclofenac 50 mg or naproxen 500 mg twice daily for 12
weeks) [6]. The drugs were similarly effective in reducing OA symptoms. The
number of clinically significant ulcers was higher in the NSAID group, but there
were only nine ulcer complications in this study (seven and two, respectively, OR
7.0; 95% CI 1.5-33.8).
●The CONDOR study compared celecoxib with diclofenac plus omeprazole. The
primary endpoint was a composite of clinically significant upper or lower
gastrointestinal events. In this study, the diclofenac and omeprazole group was
associated with an increased risk of gastrointestinal events compared with the
celecoxib group (hazard ratio 4.3, 95% CI 2.6-7.0) [7].
PARECOXIBThe safety of short-term parenteral administration of parecoxib, which is
converted to valdecoxib, was assessed in an endoscopic study of 92 healthy elderly
people ages 65 to 75 who were randomly assigned to receive one of three interventions
188
for seven days: intravenous parecoxib sodium (40 mg twice daily),
intravenous ketorolac (15 mg four times daily), or placebo [8]. Endoscopic evaluation
revealed no gastric or duodenal ulceration among those who received placebo or
parecoxib compared with ulcers in 23 percent of subjects in the ketorolac group.

Parecoxib is available in much of Europe.

ETORICOXIBControlled trials have compared etoricoxib to nonselective nonsteroidal

anti-inflammatory drugs (NSAIDs) [9]. A combined analysis evaluated ten phase II/III
randomized trials of etoricoxib in the treatment of osteoarthritis, rheumatoid arthritis, and
chronic low back pain [10]. In this analysis, the number of confirmed gastroduodenal
events (defined as a clinical, not just an endoscopic, diagnosis of peptic ulcer disease)
was significantly lower in patients who received etoricoxib compared with those treated
with nonselective NSAIDs (0.95 versus 2.24 percent, RR for etoricoxib 0.42; 95% CI
0.26-0.67).

Etoricoxib is approved in many countries but not in the United States.

LUMIRACOXIBLumiracoxib, a highly selective COX-2 inhibitor, is a phenylacetic

acid analog of diclofenac and not a tricyclic compound like the other coxibs. A combined
report of two studies with identical design (the "TARGET" trial) included 18,325 patients
with osteoarthritis who were randomly assigned to lumiracoxib (400 mg once
daily), naproxen (500 mg twice daily), or ibuprofen (800 mg three times daily) for 52
weeks [11]. In patients who were not taking aspirin, the cumulative incidence of ulcer
complications (ie, bleeding, perforation, or obstruction) was significantly lower with
lumiracoxib than with naproxen or ibuprofen (hazard ratio 0.21, 95% CI 0.12-0.37).
However, there was no reduction in ulcer complications in patients concurrently taking
aspirin.
In the "TARGET" trial, there were more myocardial infarctions in the lumiracoxib group
compared with naproxen (0.38 versus 0.21 percent), but the differences were not
statistically significant. However, the study excluded most patients with known,
symptomatic coronary artery disease, a group at increased risk for adverse
cardiovascular events. Furthermore, the study may have been underpowered to detect
a difference in the rates of myocardial infarction [12].
In another study that did not address cardiovascular risks [13], the risk of
gastroduodenal ulceration with lumiracoxib was similar to celecoxib and significantly
less than ibuprofen [13].

Lumiracoxib has been associated with hepatotoxicity and is therefore not widely
available.

ROFECOXIBRofecoxib was found to have a 50 percent relative risk reduction in

adverse gastrointestinal events compared with naproxen in rheumatoid arthritis patients
in the VIGOR trial [14]. However, in a three-year placebo-controlled trial for colorectal
adenoma chemoprevention, rofecoxib was associated with an increased incidence in
clinically relevant upper gastrointestinal bleeding compared with placebo (RR 4.9, 95%
CI 1.98-14.54) [15]. These findings underscore that the COX-2 inhibitors may be safer
189
than conventional nonsteroidal anti-inflammatory drugs (NSAIDs) for reduction in the
risk of gastrointestinal bleeding, but are still associated with an increased risk.

Rofecoxib was removed from the market because of an increased risk of myocardial
infarctions and stroke associated with long-term use.

CONCURRENT ASPIRINThe potential gastroduodenal sparing effect of selective

COX-2 inhibitors may be reduced by concurrent low-dose aspirin therapy for primary or
secondary prevention of cardiovascular or cerebrovascular disease [4,6,16].
(See "Aspirin in the primary prevention of cardiovascular disease and
cancer" and "Aspirin for the secondary prevention of atherosclerotic cardiovascular
disease".)
Loss of the gastroduodenal sparing effect of selective COX-2 inhibitors with
concurrent aspirin use is most likely related directly to blockade of COX-1 in the upper
gastrointestinal tract [17].
●In the CLASS study, similar rates of ulcers and ulcer complications were noted
among those receiving celecoxib plus low-dose aspirin as among patients
receiving a nonselective nonsteroidal anti-inflammatory drug (NSAID) plus low-
dose aspirin (2.02 and 2.12 percent, respectively) [4]. Thus, patients receiving
both aspirin and a selective COX-2 inhibitor may require prophylactic antiulcer
therapy if they are at increased risk for gastroduodenal toxicity. (See "NSAIDs
(including aspirin): Primary prevention of gastroduodenal toxicity".)
●Another study included 1615 patients with osteoarthritis who were randomly
assigned to placebo, enteric-coated aspirin (81 mg daily), enteric-coated aspirin
(81 mg daily) plus rofecoxib (25 mg daily), or ibuprofen (800 mg three times daily)
[18]. The 12-week cumulative incidence of endoscopically detected ulcers was
significantly higher in the rofecoxib plus aspirin group than in the aspirin alone
group (16.1 percent versus 7.3 percent). The endoscopic ulcer rate of 16.1
percent with rofecoxib plus aspirin was similar to the rate in the ibuprofen group
(17.1 percent). The rate of endoscopic ulcers with low-dose aspirin alone (7.3
percent) was numerically higher than the incidence with placebo (5.8 percent), but
the difference was not statistically significant. The authors concluded that the
addition of rofecoxib to low-dose aspirin increased the incidence of endoscopic
ulcers to a rate comparable to nonselective NSAIDs.
Conversely, other studies suggest that gastrointestinal (GI) toxicity of COX-2 inhibitors
may be less than that of nonselective NSAIDs even with concurrent use of aspirin, as
illustrated by the following examples:
●A randomized controlled trial addressing (as the primary endpoint) cardiovascular
toxicity of the COX-2 inhibitor etoricoxib versus the nonselective
NSAID diclofenac showed less GI toxicity in the etoricoxib arm regardless of
proton pump inhibitor (PPI) or ASA co-administration [19]. However, this benefit
only held true for uncomplicated GI events; there was no difference between the
two groups for complicated GI events.
●Two randomized controlled endoscopic studies assessed the occurrence of
gastroduodenal ulcers in healthy volunteers given aspirin together with
either celecoxib, naproxen, or placebo [20,21]. Although the incidence of ulcers

190
 was increased with celecoxib compared with placebo, significantly fewer ulcers
were found with celecoxib than with naproxen.
●A Cochrane review concluded that COX-2 inhibitors appear to offer greater upper
GI safety than nonselective NSAIDs, but that the coadministration of aspirin might
reduce this safety advantage [22].
COMBINED USE OF COX-2 INHIBITORS AND PPIsIn patients considered
to be at exceptionally high risk for peptic ulcer disease, combining a selective COX-2
inhibitor with a proton pump inhibitor (PPI) may confer added protection against
gastrointestinal (GI) toxicity. A randomized controlled trial of Helicobacter pylori-
negative patients who had been hospitalized for upper GI bleeding associated with
nonselective nonsteroidal anti-inflammatory drugs (NSAID) use found that after 13
months none of the patients placed on celecoxib in combination with a PPI had
recurrent ulcer bleeding compared with 9 percent of patients placed on celecoxib and
placebo [23].
The PRECISION trial addressed (as the primary endpoint) cardiovascular toxicity
of celecoxib versus the nonselective NSAIDs naproxen and ibuprofen in patients who
were at increased risk for cardiovascular disease (nearly half taking low dose aspirin)
and whose arthritis pain was resistant to acetaminophen [24]. Patients with GI
ulceration within 60 days of randomization or perforation, obstruction, or ulcer bleeding
within six months of randomization were excluded from the trial. This randomized trial
compared the safety of celecoxib (100 to 200 mg twice daily) with naproxen (375 to 500
mg twice daily) and ibuprofen (600 to 800 mg three times daily) in 24,081 patients with
arthritis. All patients were provided esomeprazole (20 or 40 mg per day). During a
median follow up period of 34 months, clinically significant GI events were uncommon
and occurred in 0.3 percent, 0.7 percent, and 0.7 percent of patients, respectively. It is
unclear what proportion of patients were at high-risk of GI bleeding as the proportion of
patients with a remote peptic ulcer or ulcer bleeding were not reported. (See "NSAIDs:
Adverse cardiovascular effects".)
In another trial, 514 patients with a history of upper GI bleeding who were on
nonselective NSAIDs and required low dose aspirin, were randomly assigned to
continuous low dose aspirin (80 mg daily), esomeprazole (20 mg daily), and
either celecoxib (100 mg twice daily) or naproxen (500 mg twice daily) [25]. The
cumulative incidence of recurrent upper gastrointestinal bleeding from gastroduodenal
ulcers or bleeding erosions at the 18-month follow-up was significantly lower in the
celecoxib group as compared with the naproxen group (5·6 percent, 95% CI 3·3-9·2
versus 12·3 percent, 95% CI 8·8-17·1). There was no significant difference in the
number of serious cardiovascular events between the two groups but the study was not
powered to detect such a difference.
SUMMARY
●The primary effect of nonsteroidal anti-inflammatory drugs (NSAIDs) is to inhibit
cyclo-oxygenase, thereby impairing the ultimate transformation of arachidonic acid
to prostaglandins, prostacyclin, and thromboxanes. There are two isoforms of the
cyclo-oxygenase enzyme. The COX-1 isoform is involved in gastric cytoprotection,
and the COX-2 isoform is increased during states of inflammation.
(See 'Introduction' above.)

191
●COX-2 inhibitors appear to be safer than conventional NSAIDs for reduction in
the risk of gastrointestinal (GI) bleeding, but are still associated with an increased
risk of bleeding compared with placebo. The potential gastroduodenal sparing
effect of selective COX-2 inhibitors is probably reduced by concurrent low-
dose aspirin therapy. (See 'Concurrent aspirin' above.)
●In patients considered to be at exceptionally high risk for peptic ulcer disease,
combining a selective COX-2 inhibitor (eg, celecoxib) with a proton pump inhibitor
(PPI) (eg, esomeprazole) confers added protection against GI toxicity.
(See 'Combined use of COX-2 inhibitors and PPIs' above.)
●Celecoxib is currently approved by the United States Food and Drug
Administration. Parecoxib, etoricoxib, and lumiracoxib are available in some
countries.
●Celecoxib produces comparable analgesia and anti-inflammatory effects to
nonselective NSAIDs. Compared with nonselective NSAIDs, celecoxib has been
associated with fewer endoscopically detectable, but not necessarily fewer
clinically significant gastric and duodenal ulcers. (See 'Celecoxib' above.)
●Short-term parenteral administration of parecoxib has been associated with a
lower risk of gastric or duodenal ulceration compared with ketorolac.
●Etoricoxib has been associated with an approximately 50 percent lower incidence
of investigator-reported and confirmed adverse upper GI events compared with
treatment with nonselective NSAIDs in phase II/III clinical studies.
(See 'Etoricoxib' above.)
●Lumiracoxib is a highly selective COX-2 inhibitor. In one study, the cumulative
incidence of ulcer complications (ie, bleeding, perforation, or obstruction) was
significantly lower with lumiracoxib than with naproxen or ibuprofen, in patients
who were not taking aspirin. Reductions in ulcer complications were not apparent
in patients concurrently taking aspirin.
●Rofecoxib was found to have a 50 percent relative risk reduction in GI events
compared with naproxen in rheumatoid arthritis patients in the VIGOR trial.
However, in a three-year, placebo-controlled trial for colorectal adenoma
chemoprevention, rofecoxib was associated with an increased incidence in
clinically relevant upper GI bleeding compared with placebo. Rofecoxib was
removed from the market because of an increased risk of MIs and stroke
associated with long-term use. (See 'Rofecoxib' above.)

192
Causes of upper gastrointestinal bleeding in adults
Author:
Don C Rockey, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Anne C Travis, MD, MSc, FACG, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jul 27, 2020.
INTRODUCTIONUpper gastrointestinal bleeding (UGIB) is a common medical
condition that results in substantial morbidity, mortality, and medical care cost. It
commonly presents with hematemesis (vomiting of blood or coffee ground-like material)
and/or melena (black, tarry stools). In 5 to 10 percent of patients with severe UGIB, it
may present as hematochezia.

This topic will review the different causes of UGIB. The approach to the diagnosis and
management of patients with UGIB, as well as more detailed discussions of some
specific causes of UGIB, are discussed separately:

●(See "Approach to acute upper gastrointestinal bleeding in adults".)

●(See "Peptic ulcer disease: Clinical manifestations and diagnosis".)
●(See "Overview of the treatment of bleeding peptic ulcers".)
●(See "Overview of the management of patients with variceal bleeding".)
●(See "Portal hypertensive gastropathy".)
●(See "Angiodysplasia of the gastrointestinal tract".)
●(See "Mallory-Weiss syndrome".)
EPIDEMIOLOGYThe annual incidence of hospitalization for acute UGIB in the
United States is approximately 65 per 100,000 individuals and is more common than
lower gastrointestinal (GI) bleeding [1]. The hospitalization rate for UGIB is estimated to
be sixfold higher than for lower GI bleeding [2,3]. The incidence of UGIB is higher in
men than in women (128 versus 65 per 100,000 in one study) and increases with age.
The reported frequencies of specific causes of UGIB vary and have changed over time
[4-9]. (See 'Differential diagnosis' below.)
Older studies suggested that peptic ulcer disease was responsible for approximately
one-half of UGIB, but more recent studies suggest that while still prominent, ulcer
disease is now a less common cause (approximately 20 to 25 percent of cases) [4,5,7]
and other disorders such as esophagitis are becoming comparatively more common [1].
Among patients with bleeding peptic ulcers, gastric ulcers are more common than
duodenal ulcers [4,7].
DIFFERENTIAL DIAGNOSISUGIB can be classified into several broad
categories based on anatomic and pathophysiologic factors (table 1). From a
pathophysiologic perspective, ulcerative and erosive lesions (gastric or duodenal ulcers,
esophagitis, and gastritis) are far more common than vascular lesions (varices,
angiodysplasia), mass lesions (adenocarcinoma, polyps), or traumatic lesions (Mallory-
193
Weiss tear). Of note, the source of bleeding cannot be identified in 10 to 15 percent of
patients with UGIB, probably because the culprit lesion is either difficult to identify (such
as a Dieulafoy's lesion), obscured by retained blood clot at endoscopy, or because the
culprit lesion healed by the time endoscopy was performed.
The most common causes of UGIB include the following (in approximate descending
order of frequency) [1,4-6,9,10]:
●Gastric and/or duodenal ulcers
●Severe or erosive gastritis/duodenitis
●Severe or erosive esophagitis
●Esophagogastric varices
●Portal hypertensive gastropathy
●Angiodysplasia (also known as vascular ectasia)
●Mallory-Weiss syndrome
●Mass lesions (polyps/cancers)
●No lesion identified (10 to 15 percent of patients)

Other less common causes of UGIB include:

●Dieulafoy's lesion
●Gastric antral vascular ectasia
●Hemobilia
●Hemosuccus pancreaticus
●Aortoenteric fistula
●Cameron lesions
●Ectopic varices
●Iatrogenic bleeding after endoscopic interventions
SPECIFIC CAUSES
The most common causes of UGIB, gastroduodenal ulcer disease, severe or erosive
esophagitis, severe gastritis/duodenitis, and esophagogastric varices, remain pervasive
in medicine. Accompanying symptoms (ulcer disease may be associated with pain or
anorexia, while esophagitis may be associated with reflux symptoms), severity of
bleeding (ulcers with bleeding arteries or varices typically bleed most aggressively), and
type of bleeding (hematochezia plus hematemesis typically is caused by vascular
lesions), may provide clues as to the diagnosis (table 1). However, in most patients, it is
not possible to predict the cause of bleeding based on clinical grounds alone, and
endoscopy is required to make a definitive diagnosis [11]. (See "Approach to acute
upper gastrointestinal bleeding in adults", section on 'Diagnostic studies'.)
Acute on chronic bleeding is a unique clinical syndrome characterized by a typical
presentation with acute bleeding (hematemesis, melena, or hematochezia), and also by
the presence of chronic GI bleeding manifested by iron deficiency anemia. Given the
diverse types of lesions that can be found in the upper GI tract, virtually any lesion may
cause acute on chronic bleeding. It has been reported that portal hypertensive lesions
of the upper GI tract (with portal hypertensive enteropathy/gastropathy causing chronic
bleeding, and esophageal varices causing acute bleeding) are a common scenario for
this type of bleeding [12]. Peptic ulcer disease and vascular lesions also commonly
present with acute on chronic bleeding.

194
Peptic ulcer disease — Gastroduodenal ulcers are a common cause of UGIB (picture
1A-B and picture 2 and movie 1) [13,14]. The four major risk factors for bleeding peptic
ulcers are as follows [15,16] (see "Peptic ulcer disease: Epidemiology, etiology, and
pathogenesis" and "Unusual causes of peptic ulcer disease"):
●Helicobacter pylori infection
●Nonsteroidal anti-inflammatory drugs (NSAIDs)
●Physiologic stress
●Excess gastric acid
Reduction or elimination of these risk factors reduces ulcer recurrence and rebleeding
rates [17-20].
●H. pylori – H. pylori is a spiral bacterium that infects the superficial gastric
mucosa and disrupts the mucous layer, making the mucosa more susceptible to
acid damage. (See "Pathophysiology of and immune response to Helicobacter
pylori infection".)
The chronic inflammation induced by H. pylori upsets gastric secretory physiology
to varying degrees and leads to chronic gastritis that, in most individuals, is
asymptomatic and does not progress. In some cases, however, altered gastric
secretion coupled with tissue injury leads to peptic ulcer disease [21-24].
(See "Treatment regimens for Helicobacter pylori in adults".)
●NSAIDs – NSAIDs, including low-dose aspirin, commonly predispose to
ulceration in the GI tract [3,25,26]. NSAID-induced injury results from both local
effects and systemic prostaglandin inhibition. The majority of these ulcers are
asymptomatic and uncomplicated. However, older adults with a prior history of
bleeding ulcer disease are at increased risk for recurrent ulcers and complications
[27-29]. NSAIDs have also been implicated as an important factor for nonhealing
ulcers [30]. (See "NSAIDs (including aspirin): Pathogenesis of gastroduodenal
toxicity".)
●Stress – Stress-related ulcers are a common cause of acute UGIB in patients
who are hospitalized for life-threatening nonbleeding illnesses. Patients with these
secondary episodes of bleeding have a higher mortality than those admitted to the
hospital with primary UGIB [31]. The risk of stress ulcer bleeding is increased in
patients with respiratory failure and those with a coagulopathy. (See "Stress ulcers
in the intensive care unit: Diagnosis, management, and prevention".)
The psychologic stress that is commonly associated with life events and/or
psychological factors has not clearly been associated with gastroduodenal ulcer
disease, and acid reduction therapy is not routinely indicated in this situation.
●Gastric acid – Gastric acid and pepsin are essential cofactors in the
pathogenesis of peptic ulcers. Impairment of mucosal integrity by factors such
as H. pylori, NSAIDs, or physiologic stress leads to increased cell membrane
permeability to back diffusion of hydrogen ions, resulting in intramural acidosis,
cell death, and ulceration. Rarely, hyperacidity is the sole cause of peptic
ulceration, as in patients with Zollinger-Ellison syndrome. Control of gastric acidity
is considered an essential therapeutic maneuver in patients with active UGIB.
(See "Zollinger-Ellison syndrome (gastrinoma): Clinical manifestations and
diagnosis" and "Approach to acute upper gastrointestinal bleeding in adults".)

195
A sizeable body of literature suggests that gastroduodenal ulcers associated with
different risk factors behave different clinically [13,14,32,33]. In a recent large study of
patients from North America with gastroduodenal ulcers, approximately one-half had
evidence of H. pylori infection and one-half did not [34]. The lowest rate of rebleeding
and mortality was observed in patients with H. pylori-positive ulcers. Patients with H.
pylori-negative ulcers had poorer outcomes, regardless of use of NSAIDs. Patients with
ulcers negative for H. pylori and no history of NSAID use had the worst outcomes and
more severe systemic disease.
Esophagitis — Esophagitis is a common cause of upper GI pathology and bleeding. In
a large cohort of patients with UGIB, 13 percent of patients had esophagitis [6], similar
to the number of patients with duodenal ulcers (12 percent). Patients with erosive
esophagitis often have a history of gastroesophageal reflux disease (GERD) [35]. Other
risk factors include medication use (eg, NSAIDS, oral bisphosphonates, tetracycline)
and infections (eg, Candida, herpes simplex virus). (See "Clinical manifestations and
diagnosis of gastroesophageal reflux in adults" and "Medication-induced
esophagitis" and "Herpes simplex virus infection of the esophagus" and "Esophageal
candidiasis in adults" and "AIDS-related cytomegalovirus gastrointestinal disease",
section on 'Esophagitis'.)
Hematemesis appears to be more common in patients with bleeding from esophagitis
than in patients with bleeding from other sources (86 versus 55 percent in one study)
[35]. Conversely, melena appears to be less common (38 versus 68 percent,
respectively).
Compared with patients with UGIB due to other causes, patients with UGIB due to
esophagitis often have a more benign course, with shorter hospital stays, lower
rebleeding rates, and lower mortality rates [35]. Treatment of bleeding esophagitis
involves acid suppression, endoscopic therapy (rarely required), and treatment/removal
of risk factors for GERD. (See "Medical management of gastroesophageal reflux
disease in adults".)
Gastritis/gastropathy and duodenitis/duodenopathy — Gastritis and duodenitis are
predominantly inflammatory processes. The terms "gastritis" and "duodenitis" are used
to denote inflammation-associated mucosal injury. However, epithelial cell injury and
regeneration are not always accompanied by mucosal inflammation. This distinction has
caused considerable confusion since the terms "gastritis" and "duodenitis" are often
used to describe endoscopic or radiologic characteristics of the mucosa rather than
specific histologic findings. Epithelial cell damage and regeneration with minimal or no
associated inflammation are properly referred to as "gastropathy" or "duodenopathy."
(See "Gastritis: Etiology and diagnosis".)

Gastritis and duodenitis are commonly identified at the time of endoscopy, but rarely
lead to significant UGIB (note that they are, in fact, often identified at the time of
endoscopy and may be commented on). Such commonly identified trivial endoscopic
abnormalities are typically not the cause of (significant) UGIB in the absence of other
factors such as anticoagulation or a coagulopathy. A careful examination for commonly
missed lesions is essential in patients with clinically significant blood loss in whom the
endoscopic examination reveals only gastritis or duodenitis. The examination should
include careful inspection for ulcers in unusual places, varices, and Dieulafoy's lesions,
all of which may be difficult to identify.
196
A multitude of disorders and exposures are associated with gastritis and duodenitis.
Many of the associated disorders or exposures are also risk factors for peptic ulcer
disease [36,37] (see 'Peptic ulcer disease' above). Other risk factors include excessive
alcohol consumption, radiation injury, obesity surgery, and chronic bile reflux. Gastritis
may also be found in certain autoimmune diseases. Bleeding in patients with
gastritis/gastropathy or duodenitis/duodenopathy is more common in the setting of
anticoagulant use. (See "Gastritis: Etiology and diagnosis" and "Acute and chronic
gastritis due to Helicobacter pylori".)
The differential diagnosis of acute gastritis includes gastric antral vascular ectasia
(GAVE) and, in particular, portal hypertensive gastropathy (PHG). PHG should be
suspected in patients with cirrhosis with or without overt evidence of portal
hypertension, in patients with proximal greater than distal gastric involvement, and in
patients with concomitant varices. GAVE typically has a unique endoscopic appearance
with linear aggregates of petechiae in the antrum, giving the antral area the appearance
of a watermelon rind. (See "Portal hypertensive gastropathy" and 'Gastric antral
vascular ectasia' below.)
The diagnosis of gastritis and duodenitis is made almost exclusively by endoscopic
evaluation of the mucosa. Although typically not necessary in the setting of acute
bleeding, the diagnosis can be confirmed by biopsy and histologic evaluation. Histologic
findings can vary over a wide spectrum, ranging from epithelial hyperplasia to extensive
epithelial cell damage with infiltration by inflammatory cells. (See "Gastritis: Etiology and
diagnosis", section on 'Diagnosis'.)
Bleeding from gastritis or duodenitis is typically self-limited. Treatment includes
removing the causative agent, a limited course of acid suppression with a proton pump
inhibitor, withholding anticoagulants when they may be contributing (if possible), and, if
bleeding is severe, endoscopic therapy with modalities such as argon plasma
coagulation. (See "Treatment regimens for Helicobacter pylori in adults" and "NSAIDs
(including aspirin): Treatment of gastroduodenal toxicity".)
Portal hypertension — Several causes of UGIB are the result of portal hypertension,
including esophageal varices, PHG, gastric varices, and ectopic varices. However, it
should be kept in mind that patients with portal hypertension can develop UGIB from
sources unrelated to portal hypertension (eg, peptic ulcer disease). In one study,
approximately 40 percent of patients with cirrhosis and UGIB had a cause unrelated to
portal hypertension [38].
While most patients with portal hypertension have cirrhosis, portal hypertension can
also occur in the absence of cirrhosis, a condition referred to as "noncirrhotic portal
hypertension." Causes of noncirrhotic portal hypertension include portal vein
thrombosis, schistosomiasis, and idiopathic noncirrhotic portal hypertension.
Additionally, portal hypertension and associated variceal bleeding can occur as a result
of thrombosis of mesenteric vessels (see "Noncirrhotic portal hypertension").
Mesenteric thrombosis should be considered in all patients with variceal bleeding from
atypical locations (such as gastric or small bowel varices).
Varices — Varices (picture 3A-B and picture 4) develop as a consequence of portal
hypertension in approximately 50 percent of patients with cirrhosis, and variceal
hemorrhage occurs at an annual rate of 5 to 15 percent [39]. The onset of UGIB from
varices usually signifies significant portal hypertension, which is typically associated

197
with advanced liver disease (Child-Pugh class B or C). Patients who develop bleeding
while being treated with a beta blocker have a poor prognosis [40]. The clinical
presentation of patients with varices, which most commonly manifests as hematemesis
and/or melena, may be similar to that seen in patients with bleeding from nonvariceal
lesions [38]. Indeed, differentiating variceal from nonvariceal bleeding based on clinical
grounds, even in patients with known cirrhosis, is often difficult if not impossible.
Varices may be identified in the esophagus and/or the stomach. They may also be seen
at sites other than the esophagus or stomach, such as the small bowel (ectopic
varices). Isolated gastric varices can result from segmental portal hypertension due to
splenic vein thrombosis, which results from injury to the splenic vein due to pancreatitis,
pancreatic carcinoma, or trauma in the left upper quadrant (picture 4). Risk factors for
variceal hemorrhage include increasing severity of liver disease, increasing Child-Pugh
class, variceal size, and the presence of red wale markings on varices [41,42].
(See "Prediction of variceal hemorrhage in patients with cirrhosis", section on 'Predictive
factors'.)
As with most other causes of UGIB, endoscopy is the diagnostic modality of choice for
esophagogastric varices. Ectopic varices may be detected with methods such as
angiography or video capsule endoscopy. (See "Approach to acute upper
gastrointestinal bleeding in adults", section on 'Diagnostic studies'.)
Various treatments are available for acute hemostasis of esophageal variceal
hemorrhage [39]. Endoscopic band ligation is the standard treatment, with sclerotherapy
used in situations where band ligation is not technically feasible. Early transjugular
intrahepatic portosystemic shunt (TIPS) placement may be appropriate for some
patients with gastroesophageal variceal bleeding [43]. (See "Methods to achieve
hemostasis in patients with acute variceal hemorrhage", section on 'Management of
esophageal varices'.)
The only mechanical technique considered to be highly effective for treating gastric
varices is injection with some form of cyanoacrylate (glue). Techniques used on other
lesions, such as injection with epinephrine and cautery are generally not effective and
should be avoided. TIPS and balloon-occluded retrograde transvenous obliteration
(BRTO) techniques may also be effective in some patients [44]. Banding may be
possible if the varices are near the gastroesophageal junction, but is generally not
effective in gastric varices. (See "Methods to achieve hemostasis in patients with acute
variceal hemorrhage", section on 'Management of gastric varices'.)
TIPS, BRTO, or surgical shunting are typically needed to treat bleeding ectopic varices
and should only be performed in consultation with experts. (See "Methods to achieve
hemostasis in patients with acute variceal hemorrhage", section on 'Management of
ectopic varices'.)
Portal hypertensive gastropathy — PHG (congestive gastropathy), while extremely
common in patients with portal hypertension, is an uncommon cause of significant
bleeding in these patients. When PHG is the sole cause of bleeding, there is diffuse
mucosal oozing with no other lesions, such as varices, to account for the GI bleeding
and anemia. The mucosa is friable, and bleeding presumably occurs when the ectatic
vessels rupture. The severity of gastropathy is related to the level of portal pressure, the
level of hepatic vascular resistance, and the degree of reduction in hepatic blood flow.

198
PHG, which most often causes occult bleeding [45], is discussed in detail elsewhere.
(See "Portal hypertensive gastropathy".)
PHG may be confused with GAVE, which may also be seen in patients with cirrhosis.
(See 'Gastric antral vascular ectasia' below.)
Vascular lesions — Vascular lesions in the GI tract that may cause bleeding include
angiodysplasias, Dieulafoy's lesions, and GAVE.
Angiodysplasia — Angiodysplasias are the most common vascular anomalies
encountered in the GI tract. Angiodysplasia of the stomach or duodenum has been
incriminated as the cause of blood loss in 4 to 7 percent of patients with GI bleeding
[46-48]. Such patients may present with either occult bleeding or overt bleeding [45].
Angiodysplasia is usually diagnosed by endoscopy, but in some cases, radiographic
imaging or surgery may be required for detection.
Angiodysplasias of the GI tract are discussed in detail elsewhere. (See "Angiodysplasia
of the gastrointestinal tract".)
Dieulafoy's lesion — A Dieulafoy's lesion is a dilated aberrant submucosal vessel that
erodes the overlying epithelium in the absence of a primary ulcer (picture 5) [10]. The
submucosal artery does not undergo normal branching within the wall of the stomach.
As a result, the caliber of the artery is in the range of 1 to 3 mm, approximately 10 times
the normal caliber of mucosal capillaries. Dieulafoy's lesions are usually located in the
proximal stomach along the lesser curvature, near the esophagogastric junction
(typically within 5 cm), although they have been found in all areas of the GI tract,
including the esophagus, duodenum, and colon [10,49-52].
The etiology of Dieulafoy's lesion is unknown. Additionally, events triggering bleeding
are not well understood. Patients who bleed from Dieulafoy's lesions are typically men
with comorbidities including cardiovascular disease, hypertension, chronic kidney
disease, diabetes, or alcohol abuse [52]. The use of NSAIDS is also common among
patients with Dieulafoy's lesions; one theory is that NSAIDS incite bleeding by causing
mucosal atrophy and ischemic injury [10]. Bleeding episodes are often self-limited,
although bleeding can be recurrent and profuse.
Endoscopy is the diagnostic modality of choice to detect a Dieulafoy's lesion and is
particularly helpful when performed during acute bleeding [53,54]. This is because
active arterial pumping may be visualized in an area without an associated ulcer or
mass lesion. In the absence of active bleeding, a Dieulafoy's lesion may appear as a
raised nipple or visible vessel without an associated ulcer; however, the aberrant vessel
may not be seen unless there is active bleeding from the site. Because a Dieulafoy's
lesion can be difficult to identify, it should be considered in the differential diagnosis of
bleeding without a clear source. Endoscopic ultrasonography may be useful in
confirming the diagnosis [54]. (See "Evaluation of suspected small bowel bleeding
(formerly obscure gastrointestinal bleeding)", section on 'Etiology'.)
A number of approaches have been shown to be effective for the treatment of
Dieulafoy's lesions. Endoscopic hemostasis may be achieved with a combination of
epinephrine injection followed by bipolar probe coagulation (picture 6), heater probe
thermal coagulation, or endoscopic clip placement [55-57].
Other approaches that have been used successfully to treat Dieulafoy's lesions include
endoscopic band ligation, argon plasma coagulation, and cyanoacrylate injection
[56,58-62]. However, endoscopic band ligation should be used with caution in this

199
setting because it has been associated with perforation (especially when performed
where the gastric wall is thin, such as the fundus) and bleeding from the resulting ulcer
once the band falls off; a fatality from such an ulcer has been reported [55,58]. Similarly,
band ligation is probably unsafe in the small bowel and right colon, since it may entrap
the serosa and thereby lead to perforation [63]. (See "Overview of the treatment of
bleeding peptic ulcers", section on 'Endoscopic therapy'.)
Doppler ultrasound has been used to confirm ablation of a Dieulafoy's lesion by
documenting the absence of blood flow following treatment [64]. Endoscopic tattooing is
helpful for locating the lesion for further endoscopic retreatment or intraoperative wedge
resection (picture 6). (See "Tattooing and other methods for localizing gastrointestinal
lesions".)
If rebleeding recurs after one endoscopic treatment, therapeutic options include repeat
endoscopic hemostasis, angiographic embolization, or surgical wedge resection of the
lesion. A combined endoscopic and laparoscopic approach has been described; this
approach allows precise location of the lesion with intraoperative endoscopy, followed
by a limited laparoscopic surgical wedge resection. There is no further risk of rebleeding
from a Dieulafoy's lesion after surgical wedge resection. (See "Angiographic control of
nonvariceal gastrointestinal bleeding in adults".)

Few studies have compared the various treatment approaches, so treatment should be
based upon local experience and expertise. In most instances, the initial approach to
rebleeding should be aggressive endoscopy and/or angiography. Surgical resection
should be reserved for difficult-to-control bleeding.

Gastric antral vascular ectasia — GAVE, or watermelon stomach, is an uncommon

cause of UGIB that is often confused with PHG, both of which can occur in patients with
cirrhosis [65]. (See "Portal hypertensive gastropathy".)
The term "watermelon stomach" is derived from the characteristic endoscopic
appearance of longitudinal rows of flat, reddish stripes radiating from the pylorus into
the antrum that resemble the stripes on a watermelon (picture 7) [9]. The red stripes
represent ectatic and sacculated mucosal vessels. A punctate form (in which the red
stripes are not apparent) has also been described and appears to be more common in
patients with underlying cirrhosis [66]. While acute bleeding may occur, low-grade GI
bleeding is more common, often with iron deficiency anemia. It is uncommon for
patients to present with acute and massive bleeding.
GAVE is usually an isolated problem but has been associated with cirrhosis and
systemic sclerosis [65]. In one series of 744 consecutive patients with nonvariceal
UGIB, bleeding was due to GAVE in 4 percent [67]. Portal hypertension was present in
31 percent of the patients with GAVE in this cohort. The most common clinical profile of
a patient with GAVE is an older (>70 years old) woman. In the series described above,
for example, the median age was 74 years, and 80 percent of the patients were women
[67].
Patients may also present with acute bleeding. The clinical presentation is similar
whether portal hypertension is present or not, except that those with portal hypertension
may have diffuse antral angiomas rather than the classic linear pattern [67].
The diagnosis is based on the classic endoscopic appearance. It may be confirmed with
endoscopic biopsy, endoscopic ultrasound, tagged red blood cell scan, or computed
200
tomography (CT) scan [68]. Histopathologically, GAVE is characterized by vascular
ectasia, spindle cell proliferation, and fibrohyalinosis (picture 8) [65].
Episodic transfusions are required in some patients. Endoscopic coagulation with a
heater probe, bipolar probe, argon plasma coagulator, laser therapy, or radiofrequency
ablation obliterates the vascular ectasia and decreases the degree of bleeding (picture
9) [67,69]. (See "Argon plasma coagulation in the management of gastrointestinal
hemorrhage".)
Portal decompression with TIPS does not reliably reduce bleeding, underscoring the
uncertain relationship of GAVE to portal hypertension [70,71]. Antrectomy prevents
recurrent bleeding but is usually reserved for patients who fail endoscopic therapies.
Combination estrogen/progesterone therapy may decrease bleeding, although the
ectatic vessels appear to persist [72].
Trauma or iatrogenic — Traumatic or iatrogenic causes of UGIB include Mallory-
Weiss syndrome, Cameron lesions in patients with a hiatal hernia, aortoenteric fistulas,
foreign body ingestion, postsurgical anastomotic bleeding, and postpolypectomy
bleeding.
Mallory-Weiss syndrome — Mallory-Weiss syndrome is characterized by longitudinal
mucosal lacerations (intramural dissections) in the distal esophagus and proximal
stomach that are usually associated with forceful retching. The lacerations often lead to
bleeding from submucosal arteries. The prevalence of such tears among patients
presenting with UGIB is approximately 5 percent [73-75]. The amount of blood loss is
usually small and self-limited. However, massive hemorrhage requiring transfusions and
even leading to death can occur [76].
Mallory-Weiss tears are usually secondary to a sudden increase in intra-abdominal
pressure. Precipitating factors include vomiting, straining at stool or lifting, coughing,
seizures, hiccups under anesthesia, closed-chest massage, blunt abdominal injury,
colonoscopic preparation with polyethylene glycol electrolyte lavage solution, and
gastroscopy [73-75,77-80]. (See "Mallory-Weiss syndrome", section on 'Etiology and
pathogenesis'.)
Endoscopy is the diagnostic modality of choice to document the presence of a
gastroesophageal tear. Most tears heal spontaneously [81,82]. Endoscopic therapy is
the first-line treatment for actively bleeding lacerations. Several hemostatic methods
have been used to control bleeding, including injection of epinephrine, thermal
coagulation, endoscopic clip placement, and endoscopic band ligation. (See "Mallory-
Weiss syndrome", section on 'Diagnosis' and "Mallory-Weiss syndrome", section on
'Treatment of persistent and recurrent bleeding'.)
Cameron lesions — Cameron lesions are erosions or ulcers occurring in the sac of a
hiatal hernia [83-85]. They have been described in up to 5 percent of patients with a
hiatal hernia who undergo upper endoscopy [85]. They are usually an incidental finding,
but they rarely cause acute or massive UGIB. They may also cause chronic bleeding
leading to iron deficiency anemia [86-90]. Although their pathogenesis is incompletely
understood, potential contributing factors include reflux esophagitis and mechanical
trauma.

The diagnosis is made by visualizing the lesion at the time of endoscopy; here, careful
inspection of the hiatal hernia is required, as is familiarity with the appearance of the

201
lesion (linear ulcers or erosions on the mucosal folds of a hiatal hernia at the
diaphragmatic impression).

Management depends on the clinical setting. Acute bleeding can be treated

endoscopically with standard hemostatic techniques [91]. Patients with iron deficiency
from chronic bleeding can be treated with a proton pump inhibitor after iron repletion,
which may help prevent recurrence of anemia [88]. Surgery to repair the hiatal hernia
can be considered in patients with recurrent bleeding despite the above measures.
(See "Overview of the treatment of bleeding peptic ulcers", section on 'Endoscopic
therapy'.)
Aortoenteric fistulas — Aortoenteric fistula is a rare cause of acute UGIB and is most
often iatrogenic. Because it is associated with a high mortality rate, it represents a true
medical emergency. The third or fourth portion of the duodenum is the most common
site for aortoenteric fistulas, followed by the jejunum and ileum [92,93].
(See "Aortoenteric fistula: Recognition and management".)
Most patients present with an initial "herald bleed" that is manifested by hematemesis
and/or hematochezia; this may be followed up to several weeks later by massive
bleeding and exsanguination. Intermittent bleeding can be seen if a blood clot
temporarily seals the fistula. Other signs and symptoms may include abdominal or back
pain, fever, and signs of sepsis. Infrequently, an abdominal mass is palpable or an
abdominal bruit is heard. (See "Aortoenteric fistula: Recognition and management",
section on 'Clinical manifestations'.)

A high index of suspicion is needed to establish the diagnosis of an aortoenteric fistula.

This disorder should be considered in all patients with massive or repetitive UGIB and a
history of a thoracic or abdominal aortic aneurysm or a prosthetic vascular graft.

The diagnosis is typically made with imaging, such as CT or CT angiography. The

reason that imaging is important diagnostically is that it has the potential to reveal the
phlegmon that is most often present and leads to the formation of the aortoenteric
fistula. It should be noted that angiography alone is not recommended because it often
does not reveal the defect in the aorta/duodenum. Endoscopy is important primarily to
exclude other, more common causes of acute UGIB, such as peptic ulcer disease.
Endoscopy with an enteroscope or side-viewing endoscope may reveal a graft, an ulcer
or erosion at the site of an adherent clot, or an extrinsic pulsatile mass in the distal
duodenum (or rarely, the esophagus). Exploratory laparotomy may be considered in
patients with a suspected aortoenteric fistula and severe ongoing bleeding.
(See "Aortoenteric fistula: Recognition and management", section on 'Diagnosis'.)
The mortality rate of an untreated aortoenteric fistula that presents with upper GI
hemorrhage is nearly 100 percent. Surgical repair of the aortic aneurysm and fistula is
the standard treatment (though often difficult), regardless of the cause [93,94].
(See "Aortoenteric fistula: Recognition and management", section on 'Management'.)
Therapy of an aortoenteric fistula due to an infected graft consists of intravenous
antibiotics and urgent surgical intervention [92,95]. (See "Aortoenteric fistula:
Recognition and management", section on 'Management'.)
Other traumatic or iatrogenic causes — Other traumatic or iatrogenic causes of UGIB
include foreign body ingestion, postsurgical anastomotic bleeding, postpolypectomy
202
bleeding, and bleeding after sphincterotomy. (See "Post-endoscopic retrograde
cholangiopancreatography (ERCP) bleeding".)
Foreign body ingestion may be seen in the setting of psychiatric disorders, dementia, or
in patients with loose dentures. Plain radiographs of the neck, chest, and abdomen may
reveal a radiopaque foreign body or signs of perforation. (See "Ingested foreign bodies
and food impactions in adults".)
Patients with gastroenteric or enteroenteric anastomoses may develop bleeding from
marginal ulcers. Causes of marginal ulcers include [96-99]:
●Poor tissue perfusion due to tension or ischemia at the anastomosis
●Presence of foreign material such as staples or nonabsorbable suture
●Excess exposure to acid
●NSAID use
●H. pylori infection
●Cigarette smoking
The diagnosis of a gastroenteric marginal ulcer is established by upper endoscopy, and
initial medical treatment is with gastric acid suppression. Depending on the location,
diagnosis of an enteroenteric marginal ulcer may require deep small bowel enteroscopy.
(See "Late complications of bariatric surgical operations", section on 'Marginal
ulcers' and "Overview of deep small bowel enteroscopy".)
Upper gastrointestinal tumors — Neoplasms of the upper GI tract account for less
than 3 percent of all cases of severe UGIB [100], but bleeding can be the initial
manifestation of the tumor [101]. Bleeding may occur with both benign and malignant
lesions. Bleeding can result from diffuse mucosal ulceration or from erosion into an
underlying vessel. Virtually any tumor type may bleed, including adenocarcinomas, GI
stromal tumors, lymphomas, and Kaposi sarcomas. While a less common cause of
upper GI bleeding than many other tumors (eg, adenocarcinoma, GI stromal tumors),
Kaposi sarcoma is particularly vascular in nature and should be considered in patients
with HIV infection or in those who are immunosuppressed.
Endoscopic findings suggestive of gastric malignancy include irregular ulcer margins
and an exophytic or fungating ulcerated mass (picture 10). Endoscopic biopsy,
brushing, or needle aspiration for histologic or cytologic examination is performed for
definitive diagnosis. Endoscopic ultrasound is useful for staging local disease, while CT
scanning can be helpful for staging and detection of distant metastases. (See "Clinical
features, diagnosis, and staging of gastric cancer" and "Clinical presentation, diagnosis,
and prognosis of gastrointestinal stromal tumors".)
Patients with severe bleeding secondary to malignant upper GI tumors have a poor
prognosis, and the majority of patients die within 12 months [100,101]. Surgical
resection for cure or palliation is the treatment of choice. Medical therapy is most often
palliative and consists of chemotherapy and/or radiation therapy. (See "Surgical
management of invasive gastric cancer" and "Initial systemic therapy for locally
advanced unresectable and metastatic esophageal and gastric cancer" and "Adjuvant
and neoadjuvant treatment of gastric cancer".)
Endoscopic treatment for bleeding upper GI tumors includes standard techniques such
as injection therapy, thermal contact probes (bipolar or heater probes), and argon
plasma coagulation (APC). A more novel technique is the use of hemostatic
nanopowder [1,101]. Preliminary data suggest that hemostatic nanopowder may be

203
effective in some patients with bleeding due to GI tract cancer [102]. For all hemostatic
techniques in patients with malignancy, control of active bleeding can be successful, but
rebleeding is common, especially from nonhealing, malignant ulcers [103].
(See "Overview of the treatment of bleeding peptic ulcers", section on 'Hemostatic
sprays'.)
Hemobilia — Hemobilia, or bleeding from the hepatobiliary tract, is a rare cause of
acute UGIB. It should be considered in any patient with acute UGIB and a recent history
of hepatic parenchymal or biliary tract instrumentation and/or injury, including
percutaneous or transjugular liver biopsy [104-107], percutaneous transhepatic
cholangiogram [108,109], cholecystectomy [110], endoscopic biliary biopsies or
stenting, TIPS placement [111], angioembolization [112], or blunt or penetrating
abdominal trauma. Other causes of hemobilia include gallstones, cholecystitis [113],
hepatic or bile duct tumors, intrahepatic stents [114], hepatic artery aneurysms [115],
and hepatic abscesses [115].
The classic triad of hemobilia (biliary colic, obstructive jaundice, and occult or acute GI
bleeding [116]) is uncommonly present. Obstructive jaundice, with or without biliary
sepsis, may occur if blood clots within the biliary system.
The diagnosis of hemobilia is often overlooked if the papilla is not carefully examined
endoscopically. A side-viewing duodenoscope is helpful for visualizing the ampulla and
may be useful for performing diagnostic endoscopic retrograde
cholangiopancreatography (ERCP) in this setting [117]. Cross-sectional imaging (with
CT or magnetic resonance imaging, including magnetic resonance
cholangiopancreatography [MRCP]) is often helpful in patients with hemobilia and
should be considered early in the course of the evaluation. Technetium-tagged red
blood cell scan or selective hepatic artery angiography may reveal the source of
hemobilia [118]. (See "Evaluation of suspected small bowel bleeding (formerly obscure
gastrointestinal bleeding)", section on 'Radiographic imaging'.)
Treatment is directed at the primary cause of bleeding, such as embolization or surgical
resection of a hepatic tumor [119]. In the case of bleeding following liver biopsy
[106,120], laparoscopic cholecystectomy [121], or percutaneous transhepatic
cholangiogram [122], arterial embolization is often employed and is typically effective.
(See "Angiographic control of nonvariceal gastrointestinal bleeding in adults", section on
'Embolization'.)
Hemosuccus pancreaticus — Hemosuccus pancreaticus, or bleeding from the
pancreatic duct, is another rare cause of UGIB. It is most often found in patients with
chronic pancreatitis, pancreatic pseudocysts, or pancreatic tumors. Bleeding occurs
when a pseudocyst or tumor erodes into a vessel, forming a direct communication
between the pancreatic duct and vessel. The vessel involved is often a splenic artery
pseudoaneurysm [123]. Hemosuccus pancreaticus may also be found after therapeutic
endoscopy of the pancreas or pancreatic duct, including pancreatic stone removal,
pancreatic duct sphincterotomy, pseudocyst drainage, or pancreatic duct stenting.
Bleeding is often aggressive and may be life-threatening.
Hemosuccus pancreaticus should be suspected clinically when UGIB occurs in one of
the above settings in which pancreatic injury is present. The diagnosis is usually best
confirmed with cross-sectional imaging (ie, abdominal CT scanning and/or MRCP,

204
although ERCP or intraoperative exploration may also be used) [124-126]. Cross-
sectional imaging is usually performed first because it is the least invasive modality.
Mesenteric arteriography with coil embolization can control acute bleeding and is
usually the preferred treatment [124,126]. If bleeding persists or is massive, treatment is
with pancreaticoduodenectomy and ligation of the bleeding vessel, which definitively
prevents rebleeding [127,128]. (See "Angiographic control of nonvariceal
gastrointestinal bleeding in adults", section on 'Embolization'.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Gastrointestinal bleeding in adults".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Upper endoscopy (The

Basics)" and "Patient education: Peptic ulcers (The Basics)" and "Patient
education: Acid reflux and gastroesophageal reflux disease in adults (The
Basics)")
●Beyond the Basics topics (see "Patient education: Upper endoscopy (Beyond the
Basics)" and "Patient education: Peptic ulcer disease (Beyond the
Basics)" and "Patient education: Gastroesophageal reflux disease in adults
(Beyond the Basics)")
SUMMARY
●Upper gastrointestinal bleeding (UGIB) is a common medical condition that
results in high patient morbidity and medical care costs, with an annual incidence
in the Unites States of approximately 65 per 100,000 adults.
(See 'Epidemiology' above.)
●The causes of UGIB can be classified into several broad categories based on
anatomic and pathophysiologic factors (table 1). The most common causes of
UGIB include peptic ulcer disease, esophagogastric varices, and erosive
esophagitis. (See 'Differential diagnosis' above and 'Specific causes' above.)
●Most causes of UGIB are diagnosed endoscopically, and in many patients,
endoscopic therapy is the treatment of choice. (See "Approach to acute upper
gastrointestinal bleeding in adults".)

205
Management and prognosis of the Zollinger-Ellison
syndrome (gastrinoma)
Author:
Emily Bergsland, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: May 26, 2020.
INTRODUCTIONPatients with Zollinger-Ellison syndrome (ZES) have gastrin-
secreting tumors and the associated clinical consequences. This disorder can occur
sporadically, or as a manifestation of multiple endocrine neoplasia type 1 (MEN 1).
Medical therapy is the current standard of care for most patients with ZES as part of the
MEN 1 syndrome. By contrast, many patients with sporadic ZES are candidates for
surgical therapy. (See "Zollinger-Ellison syndrome (gastrinoma): Clinical manifestations
and diagnosis" and "Multiple endocrine neoplasia type 1:
Treatment" and 'Surgery' below.)
Prior to the development of effective acid suppression therapy, the major morbidity and
mortality of ZES were related to complications of fulminant peptic ulcer disease; total
gastrectomy was the only effective measure to protect patients from these problems [1].
The development of H2 antagonists and the more powerful proton pump inhibitors has
resulted in a significant decrease in morbidity and mortality from ulcer disease and has
obviated the need for gastrectomy [2]. Of 212 patients with ZES studied prospectively
for a mean of 13.8 years, a ZES-related cause of death could be identified in only one-
half of the 31 percent who died. All of the ZES-related deaths were due to tumor spread;
none were due to hypersecretory complications [3].
This topic review will discuss the two current goals of therapy in ZES [4]:
●Control of the complications resulting from autonomous release of gastrin
●Control of the tumor itself
MEDICAL MANAGEMENTThe goal of medical management in Zollinger-Ellison
syndrome (ZES) is to limit the clinical manifestations and complications of peptic ulcer
disease. Because peptic symptoms in patients with ZES are often a poor marker of acid
secretion, formal acid secretory studies have been advocated to guide the dosage of
acid suppressants, with the goal of reducing gastric acid secretion to below 10 mEq/h
prior to the next dose [5]. However, such studies are often not available even in major
medical centers, and it has become customary among many gastroenterologists to
initiate and maintain proton pump inhibitors at high dosage, using symptoms alone as a
signal to increase the dose. To the author's knowledge, there have been few if any
episodes of major peptic diathesis associated with this approach.
Proton pump inhibitors — Proton pump inhibitors
(eg, omeprazole, lansoprazole, dexlansoprazole, pantoprazole, rabeprazole,
and esomeprazole) effectively block acid secretion by irreversibly binding to and

206
inhibiting the hydrogen/potassium ATPase that resides on the luminal surface of the
parietal cell. (See "Proton pump inhibitors: Overview of use and adverse effects in the
treatment of acid related disorders".)
Their effects last for more than 24 hours; as a result, many patients can be treated with
a once a day regimen [5].
Patients with ZES should be started on a high dose of a PPI (eg, omeprazole 40 mg
twice daily, pantoprazole 80 mg twice daily) [4,6,7]. PPIs have been generally safe,
even when used in high doses. Some patients require an early upward titration of these
doses; however, once control of acid output has been achieved, a gradual dose
reduction is usually possible [8]. In a study of 37 patients who had received high-dose
omeprazole for almost two years, nearly 50 percent were able to titrate the maintenance
dose down to 20 mg daily [9]. Overall, 95 percent of patients without MEN 1, severe
gastroesophageal reflux, or previous partial gastrectomy had safe reductions in their
medication dose. PPIs are generally well tolerated and can control hypergastrinemia in
ZES for >10 years (although some patients experience low vitamin B12 levels) [10].
(See "Proton pump inhibitors: Overview of use and adverse effects in the treatment of
acid related disorders", section on 'Adverse effects'.)
When PPIs are unable to control gastric acid secretion, somatostatin analogs such
as octreotide and lanreotide can inhibit secretion of gastrin [11,12]. However, due to the
unpredictability of the response and requirement for parenteral administration, they are
not first-line agents for symptomatic patients with hypergastrinemia.
SURGERYPatients with a sporadic gastrinoma who do not have evidence of
metastatic spread of disease should be offered exploratory laparotomy and resection
with curative intent, even in the event of negative imaging studies in approximately 17
percent of patients [1,4,13,14]. This recommendation stems from the fact that 60 to 90
percent of gastrinomas are malignant and, in addition to eliminating (or at least
decreasing) the need for antisecretory medical therapy, successful resection of sporadic
gastrinomas protects against the possibility of eventual morbidity and death from
metastatic spread of the tumor (see below). In the hands of an experienced surgeon, up
to 50 percent of these patients will be cured [13,15]. Lymphadenectomy exceeding
more than 10 lymph nodes at the time of surgery has been shown to achieve a higher
biochemical cure as compared with selective or no lymphadenectomy [16,17]. The
number of positive lymph nodes (or lymph node ratio) appears to have prognostic
significance in gastrinoma and other panNETs. Recognizing the potential for cure in
patients undergoing complete tumor removal (particularly with modern imaging tools
and our ability to control acid hypersecretion throughout the perioperative period), the
merits (and risks) of surgery need to be considered in the context of the life expectancy
and co-morbidities of each patient [10].
The likelihood for surgical cure is especially high for extrapancreatic gastrinomas (eg,
those in the duodenum or peripancreatic lymph nodes). In contrast, laparotomy is not
routinely recommended for patients with Zollinger-Ellison syndrome (ZES) as part of
MEN 1 since the multifocal nature of the tumors in this disorder almost uniformly
precludes cure of gastrin hypersecretion [4,15]. (See "Multiple endocrine neoplasia type
1: Treatment".) However, because a minority of MEN-1 tumors can have aggressive
growth patterns, some recommend imaging techniques or even surgical exploration to
identify those exceeding 2 cm with the intent of resecting them. Of note, while not
207
routine, some groups have taken a more aggressive approach in MEN1 patients with
gastrinomas, pursuing surgical resection based on localization with the selective arterial
secretagogue injection test to achieve a biochemical cure [18].
Eighty percent of curable gastrinomas lie within the gastrinoma triangle comprised of
the head of the pancreas and the duodenal sweep. Sporadic gastrinomas are often
solitary and >2 cm and located in the pancreas; tumors arising in the setting of MEN 1
most commonly arise in the duodenum and are typically small (<2 cm) and multiple [19].
The combination of preoperative localization techniques makes it possible for the
experienced surgeon to identify over 90 percent of sporadic gastrinomas [4,20].
Intraoperative transduodenal illumination and duodenotomy are of particular value in
detecting very small gastrinomas arising in the wall of the duodenum [14,21].
Enucleation is preferred when feasible, but local resection is often required for
pancreatic head lesions and a distal pancreatectomy may be necessary for large tail
lesions [10]. Whipple resections are not routine in ZES, but are reserved for pancreatic
head or duodenal lesions that cannot be removed by enucleation. In such cases, the
potential benefits, including improved lymph node retrieval, need to be weighed against
the risks of complications from the Whipple procedure. In the unlikely event that a
sporadic gastrinoma cannot be identified at surgery, we suggest deferring a Whipple's
procedure in favor of closure, with the intent of serial imaging every six months to try to
localize the neoplasm [22]. Laparoscopic surgery is controversial in ZES compared with
other panNETs, owing to the need for more extensive exploration and
lymphadenectomy in sporadic tumors. A minimally invasive approach may be
reasonable, however, in the setting of a pancreatic tail gastrinoma occurring in the
setting of MEN1 [10]. (See "Classification, epidemiology, clinical presentation,
localization, and staging of pancreatic neuroendocrine neoplasms".)
Gastric secretion may not return to the normal range following gastrinoma resection
because of a residual excess of gastric parietal cells, a consequence of the trophic
effect of chronically elevated gastrin levels. Up to 40 percent of patients will require
prolonged antisecretory therapy to control hyperacidity following curative resection, and
such patients need continued monitoring for acid hypersecretion [23,24]. Of 50 patients
who underwent curative resection for ZES, gastric hypersecretion was observed for a
mean of eight years in 62 percent of patients and was judged to be extreme in 28
percent despite normal blood gastrin levels [25].
A parietal cell (proximal gastric) vagotomy performed at the time of tumor resection has
been advocated to reduce (and in some cases obviate) the need for postoperative
medical therapy, particularly when complete resection of the gastrinoma tissue cannot
be accomplished [23,26]. However, it is currently uncommonly performed because of
the efficacy of proton pump inhibitors. Furthermore, relatively few surgeons currently
perform this procedure.
The reduction in mortality associated with surgical therapy for patients without
metastatic disease was illustrated in a prospective study of 124 patients with gastrinoma
presumed to be free of metastasis by imaging studies [27]. Only 3 percent of the 98
patients who underwent resection developed liver metastases during a mean follow-up
period of 6.3 years [27]. By contrast, 23 percent of 26 patients treated medically
developed metastatic disease over a slightly longer follow-up period (8.7 years). Two

208
deaths due to metastatic gastrinoma occurred in the medically treated group compared
with no disease-specific deaths in the operative group.
Reoperation for recurrence — Although surgery decreases the incidence of hepatic
metastases and improves survival, long-term biochemical cure is achieved in less than
30 percent. Reoperation may be of benefit in those with recurrent ZES in whom the
tumor can be identified and localized. In one study, for example, 17 patients with
recurrent disease that was unequivocally imaged underwent 18 reoperations [28]. Five
patients were disease free after operation, with a median follow-up of 28 months. There
were no deaths in the cured group; two patients in the group with persistent disease
died during a median follow-up of 34 months [29].
RADIATION THERAPY FOR NONSURGICAL CANDIDATES Experience
with external beam radiotherapy (RT) in the management of gastrinomas is limited.
Although pancreatic neuroendocrine tumors were previously considered to represent a
radioresistant neoplasm, data from published case reports and small case series
suggest that RT can produce high rates of symptomatic palliation and freedom from
local progression in patients who are not candidates for surgical resection [30-36].
THERAPY OF METASTATIC DISEASEThe liver is the major metastatic site for
gastrinomas, as it is with other islet cell tumors. The second most common site is bone
(7 percent of patients in one series), almost all of which occur in patients who also have
liver metastases [37]. The axial skeleton (spine or sacrum) is the primary site of bone
metastasis, but other sites can be involved [37]. Historically, somatostatin receptor
scintigraphy and MRI have been thought to be the best imaging modalities to detect
these lesions; the former being preferred because extra-axial lesions can occur.
Because of its greater sensitivity, 68-Ga DOTATATE PET/CT may be preferable to
conventional somatostatin receptor scintigraphy, if available [38,39].
Metastatic gastrinoma is now the most common cause of morbidity and mortality in
patients with Zollinger-Ellison syndrome (ZES). Unfortunately, current treatment
modalities are of limited benefit. A general algorithmic approach to therapy for patients
with metastatic disease is outlined in the figure (algorithm 1).
Somatostatin analogs — Somatostatin analogs like octreotide and lanreotide are
highly effective in controlling the symptoms associated with hormone hypersecretion in
other pancreatic islet cell tumors that express somatostatin receptors such as
glucagonomas and VIPomas, as well as carcinoid tumors; its efficacy is less predictable
for gastrinomas [40-42]. Octreotide can reduce gastrin levels, and may slow tumor
growth, but objective evidence of antitumor activity is rare [42-45]. As an example, in a
report of 15 patients treated with octreotide for malignant gastrinoma and progressive
hepatic metastases, seven had stabilization of tumor growth, and one an objective
decrease in tumor size [45]. The median duration of benefit was 25 months. In the
United States, octreotide is approved for control of hormone-mediated symptoms in
patients with neuroendocrine tumors.
Lanreotide appears to have similar clinical efficacy as octreotide, and is also available in
a long-acting depot form (Lanreotide-SR) [40,46]. It is approved for use in the United
States based on the results of the randomized phase III CLARINET study showing a
statistically significant improvement in progression-free survival in patients with
nonfunctional gastroenteropancreatic neuroendocrine tumors treated with lanreotide
compared to those treated with placebo [47]. (See "Metastatic well-differentiated
209
pancreatic neuroendocrine tumors: Systemic therapy options to control tumor growth
and symptoms of hormone hypersecretion", section on 'Benefits'.)
Liver-directed therapy
Resection — Hepatic resection is indicated for the treatment of metastatic liver disease
in the absence of diffuse bilobar involvement, compromised liver function, or extensive
extrahepatic metastases (eg, pulmonary, peritoneal). Although the majority of cases will
not be cured by surgery, prolonged survival is often possible, given the slow-growing
nature of these tumors [48,49].
In general, resection should be considered only for patients with a limited number of
hepatic metastases and is most successful when undertaken with curative intent.
(See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to
control tumor growth and symptoms of hormone hypersecretion", section on 'Surgical
resection'.)
Hepatic artery embolization — Liver metastases derive most of their blood supply
from the hepatic artery, whereas healthy hepatocytes derive approximately 70 percent
of their blood supply from the portal vein. This provides the rationale for therapeutic
embolization of the hepatic artery, with the goal of inducing necrosis of the metastases
with minimal damage to normal liver parenchyma.
Hepatic arterial embolization with or without selective hepatic artery infusion of
chemotherapy is frequently applied as a palliative technique in patients with
symptomatic hepatic metastases who are not candidates for surgical resection [50-53].
Response rates, as measured by a decrease in hormonal secretion or by radiographic
regression, are generally over 50 percent. Randomized trials have not yet been
performed, thus it is not known with certainty if one type of embolization is preferable to
another. Radioembolization with selective internal radiation therapy using Yttrium
microspheres is also used, although prospective studies comparing one type of
embolization with another have not been completed [53]. (See "Metastatic
gastroenteropancreatic neuroendocrine tumors: Local options to control tumor growth
and symptoms of hormone hypersecretion", section on 'Hepatic arterial embolization'.)
RFA and cryoablation — Other approaches to the treatment of hepatic-predominant
disease include radiofrequency ablation (RFA) and cryoablation, either alone or in
conjunction with surgical debulking [54-56]. These procedures, which can be performed
using percutaneous or laparoscopic approaches, appear to be less morbid than either
hepatic resection or hepatic artery embolization. However, both techniques are
applicable only to smaller lesions, and their long-term efficacy is uncertain.
(See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to
control tumor growth and symptoms of hormone hypersecretion", section on 'Ablation'.)
Liver transplantation — The number of patients with liver-isolated metastatic disease
in whom orthotopic liver transplantation (OLT) has been attempted is small, and follow-
up data are insufficient to judge whether complete cure has truly been achieved. The
limited availability of donor organs in many regions has restricted investigation of this
procedure.
Until more data become available, most clinicians consider that liver transplantation is
an investigational approach for metastatic islet cell tumors, including gastrinoma.
(See "Metastatic gastroenteropancreatic neuroendocrine tumors: Local options to

210
control tumor growth and symptoms of hormone hypersecretion", section on 'Liver
transplantation'.)
Chemotherapy and novel treatment approaches — Experience with systemic
chemotherapy for metastatic gastrinoma is limited. The traditional regimen of choice has
been streptozocin and doxorubicin. Although objective response rates as high as 69
percent were initially reported for metastatic neuroendocrine tumors, [57] with
decreases in endocrine hyperfunction, the true radiologic response rate is probably
between 10 and 40 percent [58,59]. Uncertainty as to efficacy, as well as the toxicity of
this regimen, which can include nausea, prolonged myelosuppression, and renal failure,
has prevented its widespread acceptance as a standard first-line therapy for patients
with metastatic neuroendocrine tumors, including gastrinoma.
Antitumor activity has also been shown for regimens containing the orally active
alkylating agent temozolomide. A retrospective review of 143 patients treated
with capecitabine plus temozolomide reported that 54 percent of patients experienced a
radiographic response to therapy [60]. Response to chemotherapy was not influenced
by O(6)-methylguanine DNA methyltransferase expression, proliferative activity, or ALT
pathway activation. More recently, the results from a prospective randomized study
(ECOG2211) of capecitabine plus temozolomide compared with temozolomide alone in
pancreatic neuroendocrine tumors revealed similar response rates in both arms
(approximately 30 percent), but the median progression-free survival was longer in the
combination arm (22.7 months versus 14.4 months, HR 0.58, p=0.023) [61]. As a result
of this study, use of capecitabine plus temozolomide has become routine for advanced
panNET. (See "Metastatic well-differentiated pancreatic neuroendocrine tumors:
Systemic therapy options to control tumor growth and symptoms of hormone
hypersecretion", section on 'Cytotoxic chemotherapy'.)
The modest efficacy of conventional cytotoxic chemotherapy has prompted the
development of novel therapeutic approaches for patients with advanced pancreatic
neuroendocrine tumors. These include molecularly targeted therapy with inhibitors of
the mechanistic target of rapamycin (mTOR), small molecule vascular endothelial
growth factor (VEGF) receptor tyrosine kinase inhibitors, and lutetium Lu177 dotatate
peptide receptor radioligand therapy [62-66]. (See "Metastatic well-differentiated
pancreatic neuroendocrine tumors: Systemic therapy options to control tumor growth
and symptoms of hormone hypersecretion", section on 'Molecularly targeted
therapy' and "Metastatic well-differentiated pancreatic neuroendocrine tumors: Systemic
therapy options to control tumor growth and symptoms of hormone hypersecretion",
section on 'Peptide receptor radioligand therapy'.)
PROGNOSISMortality from gastrinomas largely depends upon whether the tumor is
benign or malignant, and the extent of disease involvement. In an illustrative series, 185
patients with Zollinger-Ellison syndrome (ZES) were followed prospectively for a mean
of 12.5 years [67]. The following results were noted:
●Liver metastases were found in 24 percent of patients at the time of diagnosis;
the majority of these patients had a primary pancreatic neoplasm, and 67 percent
had primary tumors that were greater than 3 cm in size.
●Patients with liver metastases had a 10-year survival of only 30 percent
compared with a 15-year survival of 83 percent in those without liver metastases.

211
 ●Patients with lymph node metastases had the same mortality as those who were
free of visceral metastases.
●Patients with MEN 1 had a significantly lower rate of metastasis at the time of
initial diagnosis (6 percent); their high overall survival rate (100 percent at 20
years) reflected this fact.
The level of fasting serum gastrin (FSG) at the time of initial diagnosis may provide an
indication of disease extent and estimated prognosis in patients with sporadic ZES. In a
follow-up report of 239 patients with ZES, the level of preoperative FSG correlated with
tumor size and presence of lymph nodes and liver metastases (as found at exploration),
as well as primary site (pancreas tumors associated with highest levels of FSG) [68].
The five-year survival rates for patients with mild (0 to 499 pg/mL), moderate (500 to
1000 pg/mL), or severe elevations (>1000 pg/mL) of FSG were 94, 92, and 86 percent,
respectively. The corresponding 10-year survival rates were 86, 87, and 73 percent.
POSTTREATMENT SURVEILLANCEThere is limited evidence from which to
make recommendations for follow-up after resection of a gastrinoma. Guidelines from
the National Comprehensive Cancer Network based upon expert consensus include the
following recommendations for follow-up after resection of pancreatic neuroendocrine
tumor [69]:
●3 to 12 months postresection – History and physical examination, serum gastrin,
and abdominal multiphasic computed tomography or magnetic resonance imaging
and chest CT (+/- contrast) as clinically indicated.
●Long-term – History and physical examination with tumor markers every 6 to 12
months for a maximum of ten years. Imaging studies with abdominal multiphasic
computed tomography or magnetic resonance imaging and chest CT (+/- contrast)
as clinically indicated.
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Well-differentiated gastroenteropancreatic
neuroendocrine tumors".)
SUMMARY AND RECOMMENDATIONS
●Medical therapy is the current standard of care for most patients with Zollinger-
Ellison syndrome (ZES) as part of the MEN 1 syndrome. (See "Multiple endocrine
neoplasia type 1: Treatment".) By contrast, we recommend that (in addition to
medical therapy) patients with a sporadic gastrinoma and without evidence of
metastatic spread of disease be treated with exploratory laparotomy and resection
with curative intent (Grade 1B). In addition to eliminating or at least decreasing
the need for antisecretory medical therapy, successful resection of sporadic
gastrinomas reduces the risk of eventual morbidity and death from metastatic
spread of the tumor. (See 'Medical management' above.)
●The goal of medical management in ZES is to limit the clinical manifestations and
complications of peptic ulcer disease. We recommend that patients with ZES
should be started on a high dose proton pump inhibitor (eg, omeprazole 40 mg
twice daily) (Grade 1B). Subsequent lowering of dosage without recurrence of
symptoms is usually achievable. (See 'Medical management' above.)
●Mortality from gastrinomas depends largely upon whether the tumor is benign or
malignant, and the extent of disease involvement. Metastatic gastrinoma is the
212
 most common cause of morbidity and mortality in patients with ZES.
Unfortunately, the current modalities for treatment of metastatic disease are not
curative. (See 'Prognosis' above.)
●For patients with limited, resectable liver-isolated metastatic gastrinoma, we
recommend surgical resection of the hepatic metastases along with the primary
tumor (Grade 1B). Although the majority of cases will not be cured by surgery,
given the slow-growing nature of the tumor, extended survival is sometimes
possible. (See 'Resection' above.)
●Other treatment options for patients with unresectable hepatic-predominant
metastatic disease include bland embolization, chemoembolization, selective
internal radiation therapy (radioembolization), RFA, and cryoablation. (See 'Liver-
directed therapy' above.)
●The efficacy of somatostatin analogs for patients with symptomatic metastatic
gastrinoma is unpredictable, but some patients with somatostatin receptor-positive
tumors may benefit. (See 'Somatostatin analogs' above.)
●Other systemic therapy approaches (chemotherapy, molecularly targeted agents,
and peptide receptor radioligand therapy) to control symptoms and tumor growth
are discussed in detail elsewhere. (See "Metastatic well-differentiated pancreatic
neuroendocrine tumors: Systemic therapy options to control tumor growth and
symptoms of hormone hypersecretion".)
ACKNOWLEDGMENTThe UpToDate editorial staff thank Dr. Stephen E.
Goldfinger, MD, for his contributions as author to prior versions of this topic review.

213
Multiple endocrine neoplasia type 1: Clinical
manifestations and diagnosis
Author:
Andrew Arnold, MD
Section Editor:
Marc K Drezner, MD
Deputy Editor:
Jean E Mulder, MD
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Aug 30, 2021.
INTRODUCTIONMultiple endocrine neoplasia type 1 (MEN1) is an autosomal
dominant predisposition to tumors of the parathyroid glands (which occur in the large
majority of patients by age 50 years), anterior pituitary, and enteropancreatic endocrine
cells; hence, the mnemonic device of the "3 Ps" [1]. However, the clinical spectrum of
this disorder has been expanded. The duodenum is a common site of tumors
(gastrinomas) in these patients, and carcinoid tumors, adrenal adenomas, and lipomas
are more common than in the general population (table 1).
The clinical manifestations and diagnosis of MEN1 will be reviewed here. The genetics
of this disorder, its distinction from other multiple endocrine neoplasia (MEN)
syndromes, and its treatment are discussed separately. (See "Multiple endocrine
neoplasia type 1: Definition and genetics" and "Multiple endocrine neoplasia type 1:
Treatment" and "Clinical manifestations and diagnosis of multiple endocrine neoplasia
type 2".)
DEFINITION OF MEN1Multiple endocrine neoplasia type 1 (MEN1) is a rare
heritable disorder classically characterized by a predisposition to tumors of the
parathyroid glands, anterior pituitary, and pancreatic islet cells [1,2]. The presence of
MEN1 is defined clinically as the occurrence of two or more primary MEN1 tumor types,
or in family members of a patient with a clinical diagnosis of MEN1, the occurrence of
one of the MEN1-associated tumors (table 1). It should be noted that these are clinical
definitions and do not necessarily indicate that mutation of the MEN1 gene will be
identifiable or responsible (see "Multiple endocrine neoplasia type 1: Definition and
genetics"). In addition, for diagnosis of MEN1, there are situations where genetic criteria
can be used. (See 'Diagnosis' below.)
CLINICAL MANIFESTATIONS
Primary hyperparathyroidism — Multiple parathyroid tumors causing
hyperparathyroidism are the most common manifestation of multiple endocrine
neoplasia type 1 (MEN1), displaying almost 100 percent penetrance overall and at least
75 percent penetrance by age 50 years [1-3]. In most cases, it is the initial manifestation
of MEN1. Reliable incidence figures do not exist, but it has been estimated that the
incidence of MEN1 ranges from 1 to 18 percent in patients with primary
hyperparathyroidism [1] and probably much closer to the lower end of this range [4].
Primary hyperparathyroidism in the setting of familial MEN1 has a number of different
features from the common sporadic (non-familial) form of the disease [1,5]:
214
 ●The male-to-female ratio is even in MEN1 in contrast to the female predominance
in sporadic hyperparathyroidism.
●Hyperparathyroidism in MEN1 typically presents in the second to fourth decade
of life, approximately two decades earlier than in sporadic hyperparathyroidism.
●Multiple gland disease is typical in MEN1 and, given sufficient time, perhaps
universal. In comparison, approximately 80 to 85 percent of patients with sporadic
disease have single parathyroid adenomas. There can be marked asymmetry in
size among the distinct glands and, upon initial neck exploration, some
parathyroid glands in MEN1 may appear to be grossly normal. However, even the
smaller glands will generally exhibit hypercellularity on histologic examination.
●A strong and seemingly inexorable proliferative drive in parathyroid cells appears
to exist in classical MEN1, as indicated by the high rate of recurrent
hyperparathyroidism after apparently successful subtotal parathyroidectomy. One
report from the National Institutes of Health (NIH), as an example, found a
recurrence rate above 50 percent at 12 years [6]; most other studies have had
shorter follow-up periods. The high recurrence rate clearly distinguishes the
hyperparathyroidism of MEN1 from that seen in sporadic disease. It has also
resulted in differences of opinion with respect to optimal surgical management of
this disorder. (See "Multiple endocrine neoplasia type 1: Treatment".)
Similar to sporadic primary hyperparathyroidism, the majority of patients are
asymptomatic or minimally symptomatic, and hypercalcemia is detected by routine or
surveillance-driven biochemical screening. If clinical manifestations of primary
hyperparathyroidism are present, they may include decreased bone mineral density,
kidney stones, and symptoms of hypercalcemia (eg, polyuria, polydipsia, constipation).
The biochemical diagnosis of primary hyperparathyroidism is based, as it is in other
patients, on the demonstration of hypercalcemia with inappropriately high serum
parathyroid hormone (PTH) concentrations. (See "Primary hyperparathyroidism: Clinical
manifestations" and "Primary hyperparathyroidism: Diagnosis, differential diagnosis,
and evaluation".)
Pituitary adenomas — Clinically apparent pituitary adenomas have been found in
approximately 15 to 20 percent of patients with MEN1 when sought by computed
tomography (CT) or magnetic resonance imaging (MRI) [7] and 42 percent in a
multicenter study of 324 MEN1 patients [8]. The pathological prevalence in one series
was over 60 percent [9]. Pituitary tumors have been reported to be the first
manifestation of MEN1 in 13 percent of patients [10]. The range of pituitary cell types is
similar to that found in sporadic pituitary adenomas. Thus, the most common type of
pituitary adenoma in MEN1 is lactotroph, but somatotroph, corticotroph, gonadotroph,
and clinically nonfunctioning tumors can also occur (table 1). Multiple pituitary tumors
are rarely present in MEN1.
The phenotypic presentation of pituitary disease is variable. In one large kindred, as an
example, lactotroph adenomas predominated, and none of the 165 patients had
acromegaly [7,11]. Furthermore, the distribution of lactotroph adenomas was not even;
the prevalence was more than 50 percent in some branches of the kindred and very low
in others [12].
A multicenter study of 324 MEN1 patients (42 percent of whom had pituitary tumors),
compared with 110 non-MEN patients with pituitary adenomas, revealed that [8]:

215
 ●Of the 136 MEN1 patients with pituitary adenomas, 85 percent had
macroadenomas (versus 42 percent in non-MEN1 patients).
●In the same patients, hormonal hypersecretion was normalized in 42 percent
after treatment versus 90 percent in non-MEN1 patients, reflecting the finding that
the MEN1 patients had adenomas that were larger and more aggressive than
those in non-MEN patients.
In contrast, another study, which examined the results of systematic presymptomatic
screening for pituitary adenomas over a significant follow-up period (median 6 years),
showed that such screening primarily detected nonfunctioning microadenomas that
grew only occasionally and without clinical consequence; detected prolactinomas
responded well to medical treatment [13].

Whether a program of routine and lifelong surveillance by imaging would decrease

morbidity from pituitary disease in MEN1 remains unknown.

The clinical manifestations, approach to diagnosis, and therapy of pituitary adenomas in

patients with MEN1 is similar to that in patients with sporadic adenomas. (See "Causes,
presentation, and evaluation of sellar masses", section on 'Pituitary adenomas'.)
Pancreatic islet cell/gastrointestinal endocrine tumors — Effective treatment is
usually available for the hyperparathyroidism and pituitary disease in MEN1; as a result,
the malignant potential of enteropancreatic neuroendocrine tumors is now the primary
life-threatening manifestation of MEN1.
Functioning pancreatic islet cell or gastrointestinal endocrine cell tumors become
clinically apparent in one-third to two-thirds of patients with MEN1 (table 1). The most
common cause of symptomatic disease is the Zollinger-Ellison (gastrinoma) syndrome
(ZES), leading to multiple peptic ulcers or diarrhea. It has been estimated that 60
percent of patients with MEN1 have either ZES or asymptomatic elevation in serum
gastrin concentrations; on the other hand, MEN1 is present in approximately 25 percent
of patients with ZES [14,15]. Symptomatic insulinomas also occur with moderate
frequency, while VIPomas and glucagonomas are rare (table 1). (See "VIPoma: Clinical
manifestations, diagnosis, and management" and "Glucagonoma and the glucagonoma
syndrome".)
The prevalence of radiographically confirmed, nonfunctioning tumors is similar to that of
gastrinomas, ranging from 30 to 80 percent [16-19]. Like hormonally active
enteropancreatic tumors in MEN1, clinically "nonfunctioning" pancreatic neuroendocrine
tumors may be malignant and capable of causing liver metastases.
(See 'Nonfunctioning pancreatic tumors' below.)
Zollinger-Ellison syndrome — Historically, attempts at surgical cure of the
hypergastrinemia in ZES in patients with MEN1 were uniformly unsuccessful. It is now
apparent that the basis for the failure of these approaches, namely resection of palpable
tumors and/or partial pancreatectomy, is due to the biological nature and characteristics
of the tumors in MEN1. In contrast to sporadically occurring gastrinomas, the
gastrinomas in MEN1 patients are multifocal, often exceedingly small, and easily
overlooked. In addition, the duodenum is a common site of gastrinomas both in MEN1
and in sporadic gastrinoma [14,20,21]; in comparison, in MEN1, the tumors that are
found in the pancreas do not usually secrete gastrin [14].

216
The risk of death from malignant spread of MEN1-associated gastrinoma appears to be
less than that for sporadic gastrinoma. Local lymph node metastases are common but
are not necessarily associated with a poor prognosis or a high likelihood that clinically
important metastases will occur [22].
In one large series, investigators at the NIH prospectively followed 107 patients with
MEN1 and ZES and reviewed 1009 cases from the literature [23]. Their findings were as
follows:
●Approximately 25 percent of MEN1/ZES patients had no family history of MEN1.
●ZES was the initial clinical manifestation of MEN1 in only 8 percent of patients
with MEN1/ZES if careful testing was done.
●The onset of ZES symptoms preceded the diagnosis of hyperparathyroidism in
45 percent of patients.
●The diagnosis of ZES was delayed for three to six years after its onset.
●Pituitary disease occurred in 60 percent of patients.
●In patients without a family history of MEN1, ZES and other MEN1 manifestations
occurred later and were less severe.
Hypersecretion of gastrin in ZES in MEN1 may be suspected clinically by the presence
of multiple peptic ulcers (image 1) or symptoms like diarrhea. The diagnosis is
confirmed by the same biochemical and gastric acid output criteria as are used in the
sporadic cases [22,24,25] (see "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis"). Hypercalcemia from coexisting hyperparathyroidism can
significantly exacerbate the symptoms of ZES, and parathyroidectomy to correct
hypercalcemia can reduce fasting and secretin stimulated gastrin levels and basal acid
secretion [26].
The incidence of Cushing's syndrome has been reported to be increased in patients
with ZES. When Cushing's syndrome occurs in patients with nonfamilial gastrinoma, the
usual cause is ectopic corticotropin (ACTH) release from the islet-cell tumor. These
cases are associated with severe symptoms. In contrast, patients with familial MEN1
and ZES who develop Cushing's syndrome usually have a corticotroph adenoma of the
pituitary and relatively mild symptoms of cortisol excess [27].
Insulinoma — Insulin-producing pancreatic islet cell adenomas in MEN1 are often
small, may be multiple, and may be associated with the simultaneous presence of other
islet cell tumors. Insulinoma in MEN1 typically presents in the second to fourth decade
of life, earlier than in sporadic insulinoma, which usually occurs in individuals older than
40 years [1]. The diagnosis of insulinoma depends, as in nonfamilial causes, upon the
documentation of hypoglycemia with characteristic symptoms that are rapidly reversed
by the administration of glucose, and inappropriately high serum insulin concentrations.
(See "Insulinoma".)
Nonfunctioning pancreatic tumors — It is important to recognize that pancreatic
neuroendocrine tumors in MEN1 often synthesize multiple hormones. But hormone
synthesis does not always have clinical consequences, suggesting that many such
tumors may be defective in their peptide hormone processing apparatus or have an
inefficient secretory mechanism [28]. Like hormonally active enteropancreatic tumors in
MEN1, clinically "nonfunctioning" pancreatic neuroendocrine tumors may be malignant
and can metastasize to the liver. Nonfunctioning pancreatic neuroendocrine tumors

217
have been detected as early as ages 12 to 14 in asymptomatic children with MEN1
[29,30].
Nonfunctioning pancreatic neuroendocrine tumors are among the most common tumor
of the pancreaticoduodenal region in patients with MEN1 [16-19]. In a report of 579
MEN1 patients, 108 patients with isolated nonfunctioning pancreatic neuroendocrine
tumors were identified with the following clinical characteristics and course [17]:
●The penetrance of nonfunctioning pancreatic neuroendocrine tumors was 34
percent at age 50 years.
●The risks of metastasis and death were low for patients with tumors ≤20 mm.
●Average life expectancy for patients with nonfunctioning pancreatic
neuroendocrine tumors was similar to that for gastrinoma patients (69 to 70 years)
and shorter than that for patients without pancreatic tumors (77 years).
The best way to detect these nonfunctioning tumors is unclear. Assays for tumor
markers like chromogranin A have low value [31,32]. A limited amount of data suggests
that endoscopic ultrasound (EUS) outperforms CT scanning in this setting, and a
combination of MRI plus EUS has been recommended [2,30]. 68Gallium-DOTATATE
PET/CT scanning has been reported to have especially high sensitivity for detecting
neuroendocrine tumors in MEN1, at times leading to a change in management [33].
Such sensitive imaging methods increase detection of indolent tumors as well as
potentially aggressive lesions. In a retrospective study, 18F-FDG PET/CT imaging was
useful for predicting the malignant potential of pancreatic neuroendocrine tumors in
MEN1 [34]. (See "Classification, epidemiology, clinical presentation, localization, and
staging of pancreatic neuroendocrine neoplasms", section on 'Endoscopic
ultrasonography'.)
Other tumors — A number of other tumors also occur with increased frequency in
MEN1. These include carcinoid, cutaneous tumors, adrenal tumors (especially
nonfunctional adrenocortical adenomas), gastric enterochromaffin-like cell carcinoids,
pheochromocytoma (very rarely), angiomyolipomas, meningiomas, and spinal cord
ependymomas (table 1).
Carcinoid tumors — Thymic carcinoid tumors occur with increased frequency in MEN1
(2.6 to 8 percent in retrospective series of patients with MEN1), mostly in men [35-37].
Heavy smoking may be a risk factor [35]. Carcinoids in women with MEN1 are most
often bronchial [38].
Thymic carcinoids, the most common cause of anterior mediastinal masses in MEN1,
are typically nonfunctional (in contrast to the substantial incidence of ectopic Cushing's
syndrome in patients with sporadic thymic carcinoid) and tend to be aggressive.
(See "Pathology of mediastinal tumors".)
A prospective study of thymic carcinoids in 85 patients with MEN1 evaluated for
pancreatic endocrine tumors and followed for a mean of eight years (with serial chest
CT, MRI, and somatostatin receptor scintigraphy [SRS]) reported the following results
[39]:
●Seven patients (8 percent) developed thymic carcinoids, all of which were
hormonally inactive.
●All seven patients were male, and ZES was present in six.
●Five of the seven were asymptomatic, one had cough, and one had chest pain.

218
 ●CT and MRI were more sensitive than SRS for detecting the tumors initially or
with recurrence.
●All patients underwent surgical resection. All four patients followed for more than
one year postoperatively had tumor recurrence.
Some have recommended regular screening, by chest imaging studies, for this tumor in
men with MEN1 [1,35,37]. Given the rarity of these tumors and the unproven survival
benefits of this approach, we consider such routine surveillance reasonable but not
mandatory. Certainly, it seems prudent to strongly advise men with definite or possible
MEN1 against smoking, to take into consideration a strong family history of carcinoid
tumors, and to perform prophylactic thymectomy in patients undergoing
parathyroidectomy, although even this measure does not fully prevent subsequent
development of thymic neoplasia [36,40]. (See "Multiple endocrine neoplasia type 1:
Treatment".)
Gastric carcinoids and enterochromaffin-like cell proliferation (a precursor lesion of
gastric carcinoid) occur with substantial frequency in patients with MEN1 and ZES. In a
prospective study of 57 patients with MEN1 and ZES, advanced enterochromaffin-like
cell proliferation and gastric carcinoid were detected in 53 and 23 percent, respectively
[41]. Long duration of ZES, long duration of medical treatment, high fasting serum
gastrin levels, and the presence of gastric nodules on gastroscopy were associated with
a higher risk of gastric carcinoid. Such patients may benefit from regular monitoring for
gastric carcinoid. (See "Clinical characteristics of well-differentiated neuroendocrine
(carcinoid) tumors arising in the gastrointestinal and genitourinary tracts", section on
'Stomach' and "Multiple endocrine neoplasia type 1: Treatment".)
Cutaneous tumors — Cutaneous tumors are common in MEN1 (table 1) [1,2,42]; their
presence in patients with pancreatic endocrine tumors suggest the diagnosis of MEN1.
This was illustrated in a prospective study of 110 consecutive patients with gastrinoma
(48 with MEN1 and 62 without MEN1) with the following findings [43]:
●Angiofibromas and collagenomas were more common in MEN1 patients than in
those without MEN1 (64 versus 8 percent, and 62 versus 5 percent, respectively).
●These cutaneous tumors were multiple in 77 to 81 percent of MEN1 patients;
lipomas were present in 17 percent.
●The combination criterion of more than three angiofibromas and any
collagenomas had a sensitivity of 75 percent and a specificity of 95 percent for the
diagnosis on MEN1. The sensitivity and specificity of this criterion compares
favorably to the finding of hyperparathyroidism in patients who present initially with
gastrinomas [23].
Similarly, the presence of angiofibromas or collagenomas can be helpful clinically in
suggesting the diagnosis of MEN1 in selected patients with primary
hyperparathyroidism. Melanoma and hibernoma have also been reported in MEN1
patients [2], but this association and potential menin-related pathogenesis require
further investigation.
Breast cancer — The risk of breast cancer in female patients with MEN1 has been
reported to be almost double, and with earlier mean onset, compared with the general
population [44,45]. Early screening (eg, beginning age 40) has been reasonably
suggested but evidence for effectiveness remains to be demonstrated.

219
DIAGNOSISThe clinical diagnosis of multiple endocrine neoplasia type 1 (MEN1) is
based upon the occurrence of two or more primary MEN1 tumor types (parathyroid
gland, anterior pituitary, and enteropancreatic). In family members of a patient with a
clinical diagnosis of MEN1, the occurrence of one of the MEN1-associated tumors is
consistent with familial MEN1 [1].
The diagnosis of MEN1 (or at least a determination that an individual is genetically
predisposed to developing MEN1 clinically) can also be made by identifying a
germline MEN1 mutation in an individual in whom the clinical diagnosis of MEN1 is not
clearly established or in an asymptomatic family member who has not yet developed the
serum biochemical or radiological abnormalities associated with tumor development.
(See 'Confirming the diagnosis of MEN1' below and 'Screening of family members in
MEN1 kindreds' below.)

Given the complexity of decision-making and specialized skills needed in the diagnosis,
management, and treatment of MEN1, it is strongly recommended that this be done in
centers with established multidisciplinary teams experienced in the care of MEN1
patients.

MEN1 MUTATIONAL ANALYSIS

Potential benefits — The optimal role of DNA testing in the context of multiple
endocrine neoplasia type 1 (MEN1) is not clear cut [1,46]. In large part, this is due to a
dearth of solid data showing that preclinical detection of MEN1-related tumors leads to
interventions that improve morbidity or mortality. This situation contrasts, for example,
with the established value of RET DNA testing in multiple endocrine neoplasia type 2
(MEN2) kindreds (table 2) (see "Clinical manifestations and diagnosis of multiple
endocrine neoplasia type 2"). Nonetheless, there are circumstances in which DNA
testing for MEN1 can be helpful, and this option should be seriously considered. We
make determinations regarding MEN1 DNA testing on a case-by-case basis.
Direct DNA testing for MEN1 gene mutations is available in academic and commercial
laboratories (ie, Genetic Testing Registry). Generally speaking, DNA testing can have
utility in several linked ways including [46,47]:
●Confirming the clinical diagnosis of the syndrome in a proband
●Examining a clearly affected proband to determine if mutation-specific carrier
testing can be offered to relatives in that family
●Definitively determining whether or not asymptomatic or other relatives of a
proband carry the mutant gene
●Prenatal/preimplantation diagnosis
Based upon these potential benefits, guidelines from an international group of
endocrinologists recommend offering MEN1 mutational analysis to [1]:
●Any index patient with clinical MEN1 (two or more primary MEN1 tumor types)
●All first-degree relatives of known MEN1 mutation carriers
●Individuals with suspicious or atypical MEN1 (eg, multiple parathyroid tumors,
gastrinoma, or multiple pancreatic neuroendocrine tumors)
Confirming the diagnosis of MEN1 — We make determinations regarding multiple
endocrine neoplasia type 1 (MEN1) DNA testing on a case-by-case basis, but
discussions with patient and genetic counselor more often than not lead to pursuit of

220
such testing. Situations in which DNA testing can be helpful may arise when the
diagnosis of MEN1 is unable to be clearly established on clinical grounds and would
alter management. Examples may include some patients with a suggestive family
history who present with isolated primary hyperparathyroidism [46] or those with
apparently sporadic Zollinger-Ellison syndrome (ZES), some of whom will
have MEN1 mutation and would therefore be managed differently. (See "Multiple
endocrine neoplasia type 1: Treatment".)
Relevant factors to consider in these situations include the expected yield of testing, or
likelihood of a positive result, which can vary markedly depending on the specific clinical
presentation. For example, while MEN1 gene mutation is detectable in approximately 70
percent of kindreds with classic familial MEN1, the yield of testing drops to 7 percent in
individuals with a sporadic presentation of combined hyperparathyroidism and pituitary
adenoma [48]. Approximately 10 percent of kindreds with familial isolated
hyperparathyroidism have a detectable MEN1 mutation, and yields can be even lower
when less stringent criteria are selected, such as sporadic isolated hyperparathyroidism
with age under 40 [49].
Cost-benefit considerations can importantly influence decision-making. Sequencing
costs have generally dropped but can remain substantial in some settings (and are
variably covered by insurance in the United States). It is also important to recognize that
a negative result (mutation not detected) does not rule out the diagnosis of MEN1 nor
the possibility that unidentified pathologic disruption of the MEN1 gene is responsible.
Beyond assessing whether detection of an MEN1 mutation would impact a patient's
immediate clinical management, factors influencing the decision to test include an
examination of the potential utility of a positive finding for the purposes of family
screening and one's approach to the prospective surveillance for MEN1-related tumors
in the proband and family. (See 'Monitoring for MEN1-associated tumors' below.)
Screening of family members in MEN1 kindreds
Candidates for screening — We have a discussion about DNA testing with the index
patient and appropriate family members, making decisions on a case-by-case basis.
Involvement of a genetic counselor can be very helpful. Proper informed consent must
be obtained for each individual to be tested.

When a patient is diagnosed as having MEN1, the issue of screening family members
who are at risk often arises. In general, the primary and most compelling purpose of
such screening in human tumor predisposition syndromes is to prevent disease-related
morbidity and mortality that would otherwise occur. However, there is at present little
evidence that early, preclinical detection actually reduces overall morbidity or mortality
in MEN1. Nonetheless, because benefit seems likely in some instances and because
other helpful information can potentially result, screening may be pursued and DNA-
based testing merits serious consideration.

Screening approach — If DNA-based family screening is to be pursued, the initial step

is to test the MEN1 gene, usually from a sample of peripheral blood or buccal cells, from
the affected index case, if not already performed. If MEN1 sequencing of the affected
patient does reveal a pathologic mutation, the presence or absence of this family-
specific mutation can then be determined in at-risk relatives. Thus, a significant potential
benefit of such testing is the identification of family members who do not have the
221
mutation and therefore do not need regular surveillance. The value of this benefit is
enhanced to the extent that the clinician tends to opt for one of the more costly or
intensive approaches to surveillance to detect MEN1-associated conditions in at-risk
individuals. (See 'Monitoring for MEN1-associated tumors' below.)

The presence of the mutation in an asymptomatic family member does not indicate the
need for a major intervention but does focus the need for regular surveillance (eg,
assessment of symptoms, signs, biochemical/imaging tests) on these individuals. It is
possible that asymptomatic individuals’ knowledge that they definitely carries the
disease gene may increase compliance with surveillance visits and testing. Other issues
include genetic discrimination, which, in the United States, remains a potential concern,
despite certain protections in the Genetic Information Nondiscrimination Act of 2008.
Approaches to DNA testing and screening can vary in different nations.

Finally, knowledge of a family's specific MEN1 mutation can resolve the small potential

for diagnostic confusion attributable to rare MEN1 phenocopies within MEN1 kindreds,
namely individuals who can initially be classified as having the syndrome when they
develop a typical tumor (eg, prolactinoma) but may then be proven by DNA testing to
have not inherited the pathologic mutation [50].
Alternative to DNA screening — If DNA testing is not employed for screening
asymptomatic family members in known or suspected MEN1 kindreds, one low-cost
option is measurement of serum calcium [7]. This approach exploits the high
penetrance of hyperparathyroidism in MEN1. In addition, adding measurements of
serum parathyroid hormone (PTH) and/or ionized calcium and assuring the absence of
vitamin D deficiency may improve sensitivity and specificity of screening. Also, as noted
above, the presence of angiofibromas or collagenomas can be useful in this context.
MONITORING FOR MEN1-ASSOCIATED TUMORSFor patients with
multiple endocrine neoplasia type 1 (MEN1), known MEN1 carriers, and family
members whose risk has not been eliminated by DNA testing, we monitor for MEN1-
associated tumors as follows:
●We maintain clinical vigilance for symptoms or signs that could be due to MEN1-
associated tumors. These include symptoms of nephrolithiasis, amenorrhea,
galactorrhea, growth abnormalities, cushingoid changes, headache, vision issues,
cough, erectile dysfunction, peptic ulcer disease, diarrhea, and neuroglycopenic or
sympathoadrenal symptoms from hypoglycemia.
●We typically measure serum calcium, PTH, and prolactin annually to detect
asymptomatic hyperparathyroidism and prolactinoma, respectively.
●We tend toward conservatism in performing imaging studies, given the absence
of prospective evidence for improved survival outcomes, and taking patient
preferences into account regarding the frequency and nature of such imaging is
reasonable. Often we will initially perform an imaging study for enteropancreatic
neoplasia, favoring modalities and subsequent intervals that minimize radiation
exposure (eg, endoscopic ultrasound and magnetic resonance imaging [MRI]),
with a follow-up study one or two years later, and address factors like patient
anxiety.
The extent to which additional surveillance for endocrine tumors, employing biochemical
and/or radiographic methods, should be used can be debated since evidence for their
222
efficacy in improving outcomes is not strong [1,51]. Nevertheless, some published
guidelines have opted for pointing the clinician to a more aggressive screening protocol
for MEN1-associated risks beginning at very early ages [1,48]. A 2012 paper, for
example, while acknowledging weaknesses in available supporting data, suggested
routine annual measurement of serum calcium, parathyroid hormone (PTH), gastrin,
fasting glucose, insulin, insulin-like growth factor-1 (IGF-1), prolactin, and
chromogranin-A, starting in childhood and continuing for life. Imaging tests (magnetic
resonance imaging [MRI] of the pituitary and MRI/computed tomography [CT]
scan/endoscopic ultrasound (EUS) to evaluate for enteropancreatic tumors) were
suggested every one to three years [1]. Others have recommended more limited
biochemical testing and somewhat different imaging approaches [13,51].
We believe that cost-effectiveness and risk-benefit considerations (including those
related to diagnostic radiation exposure) can be taken into account in determining the
prospective preclinical surveillance program of an individual with MEN1 or a family
member at risk, beyond the maintenance of disease-focused clinical vigilance. For
example, annual measurement of serum calcium offers the opportunity to inexpensively
detect asymptomatic hyperparathyroidism, which might be treated surgically. Other
combinations of biochemical and imaging surveillance, including those in published
protocols, can reasonably be used but are not mandatory given the absence of support
by high-quality evidence [1]. The surveillance approach will also be reasonably informed
by one's potential use of tumor size criteria in the decision to operate on
enteropancreatic endocrine tumors (see "Multiple endocrine neoplasia type 1:
Treatment"). New advances in treatment could dramatically alter these
recommendations; for example, a future demonstration that an aggressive surgical
approach to gastrinoma clearly improves disease-related mortality would provide a
rationale for intensive biochemical and anatomic screening, which is capable of
detecting gastrointestinal or pancreatic disease in asymptomatic family members [16].
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Well-differentiated gastroenteropancreatic
neuroendocrine tumors".)
SUMMARY AND RECOMMENDATIONS
●General principles – Given the complexity of decision-making and specialized
skills needed in the diagnosis, monitoring, and treatment of multiple endocrine
neoplasia type 1 (MEN1), it is strongly advised that patients are evaluated and
managed in centers with established multidisciplinary teams experienced in the
care of MEN1 patients. (See 'Diagnosis' above.)
●MEN1 diagnosis – The clinical diagnosis of MEN1 is based upon the occurrence
of two or more primary MEN1 tumor types (parathyroid gland, anterior pituitary,
and enteropancreatic). In family members of a patient with a clinical diagnosis of
MEN1, the occurrence of one of the MEN1-associated tumors is consistent with
familial MEN1 (table 1). (See 'Definition of MEN1' above and 'Diagnosis' above.)
The diagnosis of MEN1 (or at least a determination that an individual is genetically
predisposed to developing MEN1 clinically) can also be made by identifying a
germline MEN1 mutation in an individual in whom the clinical diagnosis of MEN1
is not clearly established or in an asymptomatic family member who has not yet
223
developed the serum biochemical or radiological abnormalities associated with
tumor development.
●Primary hyperparathyroidism – In most cases, multiple parathyroid tumors
causing primary hyperparathyroidism are the initial manifestation of MEN1, and
they are found in the large majority of patients by age 50 years (table 1). Similar to
sporadic adenomas causing primary hyperparathyroidism, most patients are
asymptomatic or minimally symptomatic, and hypercalcemia is detected by routine
(or surveillance-based) biochemical screening. The biochemical diagnosis of
primary hyperparathyroidism is based, as it is in all patients with primary
hyperparathyroidism, upon the demonstration of hypercalcemia with
inappropriately high serum parathyroid hormone (PTH) concentrations.
(See 'Primary hyperparathyroidism' above.)
●Pituitary adenomas – The most common type of pituitary adenoma in MEN1 is
a lactotroph adenoma, but somatotroph, corticotroph, gonadotroph, and clinically
nonfunctioning adenomas can also occur. The approach to diagnosis and therapy
of pituitary adenomas in patients with MEN1 is similar to that in patients with
sporadic adenomas. (See 'Pituitary adenomas' above and "Causes, presentation,
and evaluation of sellar masses", section on 'Evaluation of a sellar mass'.)
●Pancreatic islet cell/gastrointestinal endocrine tumors – Functioning
pancreatic islet cell or gastrointestinal endocrine tumors become clinically
apparent in approximately one-third of patients with MEN1. The most common
cause of symptomatic disease is the Zollinger-Ellison (gastrinoma) syndrome
(ZES) (table 1). (See 'Pancreatic islet cell/gastrointestinal endocrine
tumors' above.)
●Screening of family members – DNA testing for MEN1 gene mutations is
available commercially and can provide valuable information in specific situations,
although its results generally do not dictate use of a major intervention established
to improve morbidity or mortality. We make determinations regarding MEN1 DNA
testing on a case-by-case basis, but discussions with patient and genetic
counselor more often than not lead to pursuit of such testing. Biochemical
screening (ie, serum calcium) of family members can be considered as a less
costly alternative to genetic screening, given the high penetrance of primary
hyperparathyroidism in MEN1, although its negative predictive value at younger
ages is limited. (See 'MEN1 mutational analysis' above and 'Screening of family
members in MEN1 kindreds' above.)
●Monitoring for MEN-1 associated tumors – We carefully monitor all patients
with MEN1, known MEN1 mutation carriers, and at-risk family members with
unknown carrier status for symptoms or signs of MEN1-associated tumors, such
as nephrolithiasis, amenorrhea (women), galactorrhea, erectile dysfunction (men),
peptic ulcer disease, diarrhea, and neuroglycopenic or sympathoadrenal
symptoms from hypoglycemia. We typically measure serum calcium, PTH, and
prolactin annually to detect asymptomatic hyperparathyroidism and prolactinoma,
respectively. Often, we perform additional surveillance using biochemical and
imaging modalities. Others routinely use more aggressive screening protocols for
MEN1-associated risks, beginning at very early ages. Differences in approaches

224
to surveillance in large part relate to the poor quality of supportive evidence in this
area. (See 'Monitoring for MEN1-associated tumors' above.)

225
Multiple endocrine neoplasia type 1: Definition and
genetics
Author:
Andrew Arnold, MD
Section Editor:
Benjamin A Raby, MD, MPH
Deputy Editor:
Jean E Mulder, MD
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Aug 30, 2021.
INTRODUCTIONThe multiple endocrine neoplasia (MEN) syndromes are rare, but
recognition is important both for treatment and for evaluation of family members.
This topic will review the classification and genetics of the MEN type 1 (MEN1)
syndrome (OMIM ID #131100). The clinical manifestations, diagnosis, and therapy of
MEN type 1 and the MEN type 2 (MEN2) syndromes are discussed separately.
(See "Multiple endocrine neoplasia type 1: Clinical manifestations and
diagnosis" and "Multiple endocrine neoplasia type 1: Treatment" and "Classification and
genetics of multiple endocrine neoplasia type 2" and "Clinical manifestations and
diagnosis of multiple endocrine neoplasia type 2" and "Approach to therapy in multiple
endocrine neoplasia type 2".)
DEFINITIONMultiple endocrine neoplasia type 1 (MEN1) is a rare heritable disorder
classically characterized by a predisposition to tumors of the parathyroid glands,
anterior pituitary, and pancreatic islet cells (table 1) [1,2]. The presence of MEN1 is
defined clinically as the occurrence of two or more primary MEN1 tumor types, or in
family members of a patient with a clinical diagnosis of MEN1, the occurrence of one of
the MEN1-associated tumors. Multiple parathyroid tumors causing primary
hyperparathyroidism are the most common component of MEN1, occurring in the large
majority of patients by age 50 years, and is the initial manifestation of the disorder in
most patients [2-4]. In one series of 220 patients with MEN1, parathyroid, pituitary
glands, and pancreatic islet cell tumors occurred in 95, 30, and 41 percent of affected
patients, respectively [3].
The prevalence of MEN1 is approximately 2 per 100,000 [2]. The incidence ranges from
1 to 18, 16 to 38, and less than 3 percent in patients with parathyroid adenomas,
gastrinomas, and pituitary adenomas, respectively [1].
Patients with MEN1 may have tumors other than those in the parathyroid and pituitary,
and pancreatic islet cells. The duodenum is a common site of tumors (gastrinomas) in
these patients, and thymic or bronchial carcinoid tumors, enterochromaffin cell-like
gastric tumors, adrenocortical adenomas, and lipomas are more frequent than in the
general population. Other associated tumors include angiofibromas, angiomyolipomas,
spinal cord ependymomas (table 1), and an increased risk of breast cancer has been
reported [5].
GENETICS

226
MEN1 gene — The inheritance of classical multiple endocrine neoplasia type 1 (MEN1)
follows an autosomal dominant pattern, indicating that Mendelian inheritance of a single
mutant gene is responsible for transmitting the tumor predisposition within a given
family. In 1988, genetic linkage analysis implicated a region on the long arm of
chromosome 11 (11q13) as the site of the "MEN1 gene" [6]. A decade later, the critical
gene in this region was identified, given the gene symbol designation MEN1, and its
protein product termed "menin" [7]. Across multiple studies, mutations in
the MEN1 gene have been detected in 70 to 90 percent of unrelated MEN1 kindreds [8-
11], although this figure is variable across studies and is sensitive to the mode of case
selection. It has been reasonably hypothesized that most typical MEN1 kindreds without
detectable MEN1 gene mutations nonetheless have inactivating germline mutations in
(or near and in "cis" with) the same gene but outside the coding region that is typically
sequenced in diagnostic and research labs. Somatic mosaicism of pathogenic MEN1
mutations in tumor-prone tissues could be occurring in some cases as well. However, it
is clear that mutations in the MEN1 gene are not responsible for all individuals, or even
kindreds, with an MEN1 phenotype (see 'Other genes' below). Furthermore, rare
phenocopies have been reported, ie, individuals within MEN1 kindreds who were initially
classified as having the syndrome when they developed a typical tumor (eg,
prolactinoma) but were then proven by DNA testing to have not inherited the pathologic
mutation [12].
Much has been learned about the biochemical and cellular functions of menin, but the
precise way(s) in which these functions relate to tumorigenesis is still not well
established. However, it is clear that most of the pathogenic MEN1 gene mutations
found in MEN1 patients would be expected to inactivate or disrupt menin function.
Typical of a classical tumor suppressor gene, the spectrum of reported
germline MEN1 mutations occur throughout the gene and yield no strong
genotype/phenotype relationships [11]. The genotype-phenotype correlations are often
unclear, even within a family [13]. In addition, biallelic somatic mutations within this
gene have been found in 12 to 17 percent of typical nonfamilial parathyroid adenomas
[14-16] and some sporadic gastrinomas and insulinomas [17], sporadic neuroendocrine
tumors of the foregut [18], sporadic carcinoid tumors of the lung [19], and sporadic
pituitary tumors [20], further supporting the relationship between the mutations and
tumorigenesis (see "Pathogenesis and etiology of primary hyperparathyroidism").
However, the large majority of non-MEN1-associated pituitary tumors, whether sporadic
or familial, do not have an MEN1 mutation [21,22].
Other genes — Syndromes clinically related to but genetically distinct from MEN1 do
exist. At least one family with an unusual expression of MEN1 (eg, a lower than
expected incidence of hyperparathyroidism and higher than expected incidence of
pituitary tumors) was reported to have a predisposing gene at a location distinct from
the chromosome 11q13 site [23,24]. Germline AIP mutation, in the absence
of MEN1 gene mutation, has been reported in the setting of pituitary plus parathyroid
neoplasia [25]. Furthermore, mutations in the MEN1 gene are infrequent in kindreds
with familial isolated hyperparathyroidism [26,27] and were not found in three kindreds
with familial pituitary adenoma [28] and one with isolated familial acromegaly [29].
Cyclin-dependent kinase inhibitor genes — An inherited mutation of the p27 cyclin-
dependent kinase (CDK) inhibitor gene, CDKN1B, was reported in one kindred whose

227
proband had hyperparathyroidism and acromegaly due to a growth hormone-producing
pituitary tumor; the proband's father had acromegaly, and the sister had a renal
angiomyolipoma [30]. Germline CDKN1B mutation was also reported in a few other
cases of MEN1 that collectively exhibited features including hyperparathyroidism,
Cushing's disease, cervical carcinoid tumor, bilateral nonfunctioning adrenal masses,
and Zollinger-Ellison syndrome with duodenal and pancreatic masses [31,32]. MEN1-
like disease caused by germline CDKN1B mutation has been termed MEN4 (OMIM ID
#610755) and may account for 1 to 3 percent of unrelated MEN1-like cases without
identifiable MEN1 mutations [32-34].
Rare germline mutations in three other CDK inhibitor genes, CDKN2B, CDKN2C,
and CDKN1A, encoding the p15, p18, and p21 proteins, respectively, have also been
identified in this setting and may collectively account for another 1 to 2 percent of
MEN1-like cases without detectable MEN1 mutations [32].
Overall, because patients presenting with the combination of sporadic parathyroid plus
pituitary tumor have a much lower yield of detectable MEN1 mutation than in typical
MEN1 kindreds or in sporadic cases of parathyroid plus pancreatic tumor [35], it seems
likely that this phenotype may often be due to mutation in a gene other than MEN1, or
possibly the coincidental presence of sporadic tumors absent any major genetic
predisposition. However, as noted above, the extent to which mutation in CDKN1B or
other CDK inhibitor genes is responsible for MEN1-like phenotypes, sporadic or familial,
in the absence of MEN1 mutation appears to be small [32,34,36].
DNA testing — Direct DNA testing for MEN1 mutation is available for clinical use, has a
useful role in certain settings, and should be considered on an individual basis [1].
However, in contrast to testing for RET gene mutations in MEN2, presymptomatic DNA
diagnosis has not been shown to yield equally clear benefit in preventing morbidity and
mortality in individuals at risk for MEN1. (See "Multiple endocrine neoplasia type 1:
Clinical manifestations and diagnosis", section on 'MEN1 mutational
analysis' and "Clinical manifestations and diagnosis of multiple endocrine neoplasia
type 2", section on 'Evaluation'.)
PATHOGENESISPatients with classical multiple endocrine neoplasia type 1
(MEN1) have often inherited one inactivated copy of the MEN1 gene from an affected
parent [7,9]. The actual outgrowth of a tumor is thought to require the subsequent
somatic inactivation, often by gross deletion, of the remaining normal copy of the gene
in one cell (so-called "two-hit" effect described by Knudson). Such a parathyroid cell, as
an example, would then be devoid of the MEN1 gene's normal tumor suppressor
function, and could gain a selective advantage over its neighbors, resulting in a clonal
proliferation (figure 1). The high incidence of endocrine tumors in MEN1 (which has over
90 percent penetrance) and the common multiplicity of these tumors implies that
somatic inactivation of the remaining normal copy of the gene occurs at an appreciable
frequency and contributes importantly to tumorigenesis in the clinically affected tissues.
This model also appears to apply to some of the nonendocrine tumors that occur in
patients with MEN1.
Further functional studies of the MEN1 gene and its product are certain to shed light on
these processes. In addition, the identification and analysis of other genes whose
somatic alteration is also important in the emergence of clonal tumors in this syndrome
should further clarify the relationship between genotype and phenotype in MEN1.
228
SUMMARY
●Definition – Multiple endocrine neoplasia type 1 (MEN1) is a rare autosomal
dominant disorder with a prevalence of approximately 2 per 100,000. MEN1 is
defined clinically as the presence of two of the three main MEN1 tumor types
(parathyroid, enteropancreatic endocrine, and pituitary tumors), or in family
members of a patient with a clinical diagnosis of MEN1, the occurrence of one of
the MEN1-associated tumors. In addition, patients with MEN1 may have tumors
other than those in the parathyroid, pituitary glands, and in the pancreatic islet
cells, including duodenal gastrinomas, thymic or bronchial carcinoid tumors,
enterochromaffin cell-like gastric tumors, adrenocortical adenomas, lipomas,
angiofibromas, angiomyolipomas, and spinal cord ependymomas (table 1).
(See 'Definition' above.)
●MEN1 gene – The MEN1 tumor suppressor gene is located on the long arm of
chromosome 11 (11q13). Its protein product is termed "menin." Over
1000 MEN1 gene mutations have been detected that inactivate or disrupt menin
function. Inactivation of menin results in loss of tumor suppression. Families with
the same types of mutations do not necessarily have the same clinical phenotype.
(See 'MEN1 gene' above.)
●Other genes – Syndromes clinically related to but genetically distinct from MEN1
do exist, and mutations in the MEN1 gene are not responsible for all individuals, or
even kindreds, with an MEN1 phenotype. (See 'Other genes' above.)

229
NSAIDs (including aspirin): Pathogenesis of
gastroduodenal toxicity
Author:
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Feb 01, 2021.
INTRODUCTIONNonsteroidal anti-inflammatory drugs (NSAIDs) are in use
throughout the world. NSAIDs are popular because of their versatile effectiveness as
analgesics, antipyretics, and anti-inflammatory agents. Aspirin is also used in low doses
as an anti-platelet agent [1]. Unfortunately, aspirin (even in very low doses) and other
NSAIDs can injure the gastric and duodenal mucosa, with considerable morbidity and
mortality [2].
The pathogenesis and some clinical aspects of gastroduodenal toxicity attributed to the
use of NSAIDs and aspirin will be reviewed here. Other topics, such as other side
effects, including injury to the small and large intestine, recommendations for the
prevention and treatment of NSAID-induced gastroduodenal injury, and an overview of
selective COX-2 inhibitors are discussed elsewhere. (See "NSAIDs: Adverse effects on
the distal small bowel and colon" and "Nonselective NSAIDs: Overview of adverse
effects" and "NSAIDs (including aspirin): Primary prevention of gastroduodenal
toxicity" and "Overview of COX-2 selective NSAIDs".)
SYSTEMIC VERSUS TOPICAL EFFECTSAspirin and many other nonsteroidal
anti-inflammatory drugs (NSAIDs), (eg, ibuprofen, indomethacin, naproxen,
and ketorolac) are carboxylic acids [3]. Their pKa values range from 3.50 (aspirin) to
4.85 (ibuprofen). As such, they are not ionized at the acidic pH ordinarily found in the
gastric lumen and thus can be absorbed across the gastric mucosa. Once these drugs
move from the acidic environment of the gastric lumen into the pH–neutral mucosa, the
drugs ionize and are trapped temporarily in gastric epithelial cells where it may damage
these cells.
However, this "topical" epithelial injury by many NSAIDs does not appear to be of prime
importance in the pathogenesis of clinically important endpoints (symptomatic ulcers)
[4]. The pathogenesis of symptomatic peptic ulcer disease caused by repeated
exposure to NSAIDs is mainly a consequence of systemic (post-absorptive) inhibition of
gastrointestinal mucosal cyclo-oxygenase (COX) activity. Even intravenous or
intramuscular administration of aspirin or NSAIDs such as ketorolac can cause gastric
or duodenal ulcers in animals and humans [4-7].
THE CENTRAL PROTECTIVE ROLE OF PROSTAGLANDINS Cyclo-
oxygenase (COX), the rate-limiting enzyme in prostaglandin (PG) synthesis, converts
the unsaturated fatty acid arachidonic acid (C20:4) (derived from phospholipids in cell
membranes) into PGG2 and then to PGH2 (figure 1). The gastric and duodenal mucosa
230
proceed to convert PGH2 to various prostanoids (prostaglandins and thromboxane A2).
PGs such as PGE2 protect the mucosal lining from injury by luminal acid-pepsin.
There are two functional forms of COX in the human body, COX-1 and COX-2 [3].
These two proteins are 50 to 60 percent homologous and are coded on chromosomes 9
and 1, respectively. COX-1 is a constitutive enzyme with a fairly steady rate of
expression in most cells of the body. In contrast, COX-2 is produced in many cells only
when bacterial polysaccharides (endotoxin), pro-inflammatory cytokines such as TNFa
or IL-1b, or growth factors (mitogens) induce COX-2 expression.
The healthy gastric and duodenal mucosa constitutively use COX-1 to produce its
mucosal-protective PGs [8]. Many NSAIDs block COX-1 and COX-2 more or less
equally (ie, are non-selective) and thus may impair gastric PG production at low
concentrations (table 1). Selective inhibitors of COX-2, may better preserve PG-
mediated gastrointestinal mucosal protection (table 1). Examples
include celecoxib and etodolac. However, COX-2 selective inhibitors (eg, celecoxib)
may still block COX-1 in the stomach and duodenum at clinically recommended doses
and thus have the potential to cause damage [9]. (See "Overview of COX-2 selective
NSAIDs".)
MECHANISMS OF GASTRODUODENAL PROTECTION BY
ENDOGENOUS PROSTAGLANDINSMany mucosal functions are altered by
endogenous prostaglandins (PGs) and by exogenously administered PGs; these
changes may contribute to the mucosal protective effects of PGs. While it is true that
certain PGs such as PGE2 reduce gastric acid secretion, hypochlorhydria does not
entirely explain the mucosal protection observed with PGE2. In animals, for example,
doses of PGE2 far too low to inhibit gastric acid secretion profoundly protect against
gastric injury induced by aspirin, alcohol, and other gastric irritants. This non-
antisecretory effect of PGs was referred to as "cytoprotection". Some of the
cytoprotective mechanisms of PGs include:
●Stimulation of mucin secretion by epithelial cells
●Stimulation of bicarbonate secretion by epithelial cells
●Stimulation of phospholipid secretion by epithelial cells
●Enhancement of mucosal blood flow and oxygen delivery to epithelial cells via
local vasodilation
●Increased epithelial cell migration towards the luminal surface (restitution)
●Enhanced epithelial cell proliferation
The first two mechanisms, stimulation of epithelial mucin and bicarbonate secretion,
combine to form an alkaline, unstirred water layer on the surface of the gastric mucosa,
which retards diffusion of acid-pepsin from the lumen into the mucosa. Gastric and
duodenal injury by acid and pepsin may occur when these protective functions are
compromised as a consequence of PG deficiency induced by COX-1 inhibiting
nonsteroidal anti-inflammatory drugs (NSAIDs) or, experimentally, by antibodies to PGs
[10]. This damage may eventually lead to gastric and/or duodenal ulcer formation, with
or without serious ulcer complications (bleeding, perforation, and obstruction).
ROLE OF NITRIC OXIDEGeneration of nitric oxide (NO) by NO synthase (NOS)
may be a key intermediate in cytoprotection. Similar to the role of COX-1, constitutive
NO synthase (NOS) is important in the maintenance of an intact mucosal lining. Two
enzymes contribute to the basal, constitutive NOS activity: neuronal NOS (nNOS; type
231
I) and endothelial NOS (eNOS; type III). The cytoprotective mechanisms of NO parallel
prostaglandin (PG) effects and include:
●Stimulation of gastric mucin secretion
●Stimulation of bicarbonate secretion
●Enhancement of mucosal blood flow
●Maintenance of epithelial barrier function
In some but not all models, inducible NO synthase (iNOS; type II) produces high levels
of NO leading to physiological responses that are often quite different than seen in the
basal state. Inducible NOS is generally associated with inflammatory states, similar to
COX-2 [11]. However, the relationship between the various NOS forms and COX
enzymes has not been fully elucidated. Most studies implicate both constitutive and
inducible enzymes in the maintenance of gastric mucosal integrity as well as in
epithelial restitution [12,13]. Manipulation of NO levels in models of nonsteroidal anti-
inflammatory drug (NSAID)-induced mucosal damage suggests a protective role for
both NO and PGs [14,15].
Because of the importance of NO in mucosal defense, aspirin or NSAIDs coupled to NO
themselves or to NO donors (some of which also release hydrogen sulfide) have been
developed [16].
SPECTRUM OF GASTRODUODENAL MUCOSAL INJURY Injury to the
stomach or duodenum by nonsteroidal anti-inflammatory drugs (NSAIDs) can range
from subtle alterations in gastric mucosal barrier function through microscopic damage
to surface cells to gross injury visible through an endoscope or at the time of surgery for
an ulcer complication.

The most subtle change is disruption of mucosal barrier function, manifest as increased
mucosal permeability to hydrogen ions (which then diffuse more rapidly from the lumen
into the mucosa) and to sodium ions (which then diffuse more rapidly from the mucosa
into the lumen). Disruption of the gastric mucosal barrier and the resultant damage to
surface cells by gastric acid may result in macroscopic injury over time if repair
mechanisms are ineffective.

Repair mechanisms include rapid migration of deeper epithelial cells lining gastric pits to
cover the damaged surface (restitution) and less rapid regeneration of new epithelial
cells from progenitor cells (proliferation). Epithelial restitution and proliferation both
require adequate amounts of well-oxygenated blood at a pH close to 7.4.
Prostaglandins (PGs) and nitric oxide (NO) have important roles in these repair
mechanisms, while COX inhibitors can disrupt these PG-dependent reparative
processes. Gastric restitution has been associated with induction of COX-2 [17].
Aspirin-induced gastric injury is also associated with inhibition of vascular endothelial
growth factor (VEGF), which may reduce angiogenesis, inhibit autophagy (an adaptive
response to cell stress that can promote cell survival), and enhance gastric cell
apoptosis [18,19].
Macroscopic injury by NSAIDs ranges from edema, erythema, subepithelial
hemorrhage, erosions (mucosal breaks, without visible depth to the lesion), and ulcers
(mucosal breaks, with visible depth to the lesion). Only erosive/ulcerative lesions are
considered clinically important, although lesser lesions may be precursors of erosive
lesions if reparative mechanisms fail. Development of the full spectrum of lesions in the
232
PG-depleted stomach or duodenum requires acid and pepsin; potent acid inhibition (eg,
with PPIs) markedly protects against their development. (See "NSAIDs (including
aspirin): Primary prevention of gastroduodenal toxicity".)
Gastric damage — The gastrointestinal mucosa uses COX-1 to generate mucosal-
protective PGs. Aspirin doses as low as 10 mg/day inhibit gastric PG generation
considerably and can damage the stomach [20]. Epidemiologic and placebo-controlled
studies indicate that the risk of serious, clinically-relevant gastrointestinal damage
increases as the aspirin dose is raised [21-23].
After low-dose aspirin therapy is stopped, the human stomach requires five to eight
days to recover its COX-1 activity and its ability to synthesize protective PGs,
suggesting a very slow turnover of gastric COX-1 [24]. Thus, gastric mucosa somewhat
resembles the megakaryocyte/platelet, which requires 10 to 14 days to recover from
aspirin.
Although aspirin inhibits COX-1 and also COX-2 (table 1), certain COX-2-mediated
reactions can still occur after aspirin has been given, such as the conversion of
arachidonic acid to the fatty acid 15-hydroxyeicosatetraenoic acid (15-HETE); 15-HETE
is then converted to 15-epi lipoxin A4 by another enzyme, 5-lipoxygenase (5-LOX).
Studies in animals have demonstrated enhanced production of 15-epi lipoxin A4
following aspirin exposure [25]. Furthermore, this lipoxin minimizes gastric damage by
aspirin (ie, it is cytoprotective). This protective effect of 15-epi lipoxin A4 can be
abolished (with more gastric damage resulting) by administering a selective COX-2
inhibitor [25]. Therefore, the combination of low-dose aspirin and a COX-2 selective
inhibitor may lead to more gastrointestinal damage than low-dose aspirin alone.
In contrast to aspirin, which acetylates COX irreversibly, most NSAIDs inhibit COX-1
and COX–2 reversibly [3]. Nevertheless, even transient COX inhibition in the gastric
mucosa by an NSAID is sufficient to predispose the stomach to injury. That this injury is
due to loss of PG-mediated cytoprotection is supported by the observation that NSAID-
related gastric damage is prevented by PGE analogs such as misoprostol [26].
Misoprostol does not primarily protect the stomach by inhibiting its ability to produce
hydrochloric acid because drugs that inhibit gastric acid secretion to the same or slightly
greater extent as misoprostol, such as histamine-2 receptor antagonists (H2RAs), have
little or no protective effect against NSAID-induced gastric damage. On the other hand,
the relative failure of H2RAs to protect the stomach from damage by NSAIDs may be
overcome by using higher H2RA doses or by using more potent acid-inhibitory
compounds such as the proton pump inhibitors. (See "NSAIDs (including aspirin):
Primary prevention of gastroduodenal toxicity".)
Duodenal damage — Aspirin doses as low as 325 mg every other day increase the risk
of duodenal ulcers [27]. In contrast to the stomach, damage to the duodenal mucosa by
aspirin and NSAIDs seems to depend highly upon gastric acid. Thus, not
only misoprostol but also by histamine-2 blockers with their modest acid inhibition can
largely prevent endoscopic evidence of duodenal mucosal injury by NSAIDs. PPIs are
also highly effective. These observations related to the pathogenesis of gastric and
duodenal injury lay the groundwork for using either a PPI (or misoprostol) in the
prevention of NSAID-induced gastroduodenal ulcers. (See "NSAIDs (including aspirin):
Primary prevention of gastroduodenal toxicity".)

233
NSAID-INDUCED GASTRIC TOXICITY NOT MEDIATED THROUGH
COX-1 INHIBITIONWhile inhibition of COX-1 is the major mechanism by which
nonsteroidal anti-inflammatory drugs (NSAIDs) produce gastric injury, mediators
besides prostaglandins (PGs) (and nitric oxide [NO]) may also be involved. As an
example, interference with factors that mediate restitution or adaptive protection may
contribute to mucosal injury.
Adaptive protection refers to the observation that mild gastric irritants induce enhanced
cytoprotection. COX-1, COX-2, and NO as well as various growth factors, such as
transforming growth factor (TGF)-beta and TGF-alpha, appear to participate in these
adaptive processes [28]. Endogenous trefoil factor proteins are also associated with
mucosal protection in indomethacin-induced injury in rats [29]. NSAIDs probably
interfere with growth factors and other mediators responsible for restitution and adaptive
protection, further contributing to their toxicity.
ROLE OF HELICOBACTER PYLORI INFECTIONThe role of H.
pylori infection in nonsteroidal anti-inflammatory drug (NSAID)-induced gastritis or ulcer
formation is complex [30-38].
Multiple studies have evaluated the relationship between NSAIDs and H. pylori in the
development of peptic ulcer disease (PUD) [30,31,33,37,38]. At least two meta-
analyses concluded that NSAID use and H. pylori infection represent independent and
synergistic risk factors for uncomplicated and bleeding peptic ulcer disease [39,40]. In a
systematic review with a total of 21 studies [40]:
●The risk of uncomplicated peptic ulcer disease among NSAID users was
significantly higher among H. pylori positive compared with H. pylori negative
individuals (OR 1.81).
●Ulcers were more common in H. pylori positive compared with H. pylori negative
patients irrespective of NSAID use (OR 4.03) and in NSAID users compared with
nonusers irrespective of H. pylori status (OR 3.10).
A related question is whether eradication of H. pylori prior to NSAID treatment can
reduce the risk of PUD. At least four controlled trials have addressed this issue, but they
used endoscopic rather than clinically apparent ulcers as the endpoint and the study
design and results differed somewhat [34-36,41].
●In one study, 92 patients with musculoskeletal pain, who were also infected
with H. pylori, were randomized to naproxen alone (750 mg per day) for eight
weeks or to a one-week course of triple therapy for H. pylori followed by the same
dose of naproxen [34]. After eight weeks, H. pylori was still present in all 47
patients in the naproxen-only group compared to only 5 of 45 patients (11 percent)
in the triple-therapy group. Development of a peptic ulcer (defined as a mucosal
break of 5 mm or more in diameter with a well-defined ulcer crater by endoscopy)
was more common in the patients treated only with naproxen (26 percent versus 7
percent in those treated with anti-H. pylori therapy); two of the three patients in the
latter group had failure of H. pylori eradication.
●A second trial involved 102 arthritis patients who were randomly assigned to
eradication therapy or placebo eradication therapy followed by a six-month course
of diclofenac (100 mg daily) [35]. Inclusion criteria required that patients had a
positive urea breath test indicating H. pylori infection, required long-term NSAID
treatment and had not previously been treated with a NSAID, and had dyspepsia
234
 or an ulcer history. During follow-up, the probability of endoscopically-detected
ulcers was significantly higher in the placebo-treated group (34 versus 12
percent). The probability of complicated ulcers (defined as "symptomatic or
bleeding") was also higher in the placebo group (27 versus 4 percent).
●A third trial randomly assigned 660 H. pylori-positive patients who were having
musculoskeletal complaints that would require a NSAID for at least five weeks to
receive diclofenac for five weeks, diclofenac plus omeprazole for five weeks,
diclofenac for five weeks with triple therapy for the first week, or diclofenac for five
weeks with triple therapy for the first week followed by omeprazole for the next
four weeks [41]. All three ulcer prophylactic strategies (H. pylori eradication,
proton pump inhibitor therapy, or their combination) were equally effective in
preventing endoscopically-detected NSAID ulcers and in reducing NSAID-related
dyspeptic symptoms.
●Another trial included 285 patients already requiring therapy with NSAIDs, who
were infected with H. pylori and had current or previous PUD, dyspepsia, or both
[36]. Patients were continued on their NSAID and randomized to omeprazole plus
antibiotics for one week or omeprazole plus placebo for one week. Patients in both
groups were equally likely to remain ulcer free at six months (56 versus 53
percent). Furthermore, fewer baseline gastric ulcers (but not duodenal ulcers)
healed among patients who underwent H. pylori eradication. The failure of H.
pylori eradication to enhance ulcer healing at six months was unexpected and
unexplained.
Recommendations regarding H. pylori testing in patients taking requiring NSAIDs are
presented separately. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity".)
RISK OF GASTROINTESTINAL ULCER COMPLICATIONS The risk for
the development of significant nonsteroidal anti-inflammatory drug (NSAID)-induced
gastrointestinal bleeding or perforation due to a peptic ulcer has been evaluated in
multiple studies [27,42-57]. An important determinant is the duration of therapy.
Administration of NSAIDs for a short period of time (less than one week) in healthy
people is unlikely to result in any clinically significant gastroduodenal toxicity. Longer
duration of therapy is associated with an increased risk of developing ulcer
complications. On the other hand, gastroduodenal complications are most common
within the first three months after the initiation of therapy. Gastroduodenal toxicity may
develop even in patients taking low doses of aspirin [57], which, can be associated with
a significant decrease in gastric mucosal prostaglandin concentrations [20].
(See "Aspirin in the primary prevention of cardiovascular disease and cancer", section
on 'Potential risks'.)
Besides duration of exposure, a number of other factors are associated with an
increased risk of gastroduodenal toxicity and complications from NSAIDs. These include
increasing age, higher NSAID dose(s), a past history of gastroduodenal toxicity from
NSAIDs, a past history of peptic ulcer disease, and concurrent use of glucocorticoids,
anticoagulants, and clopidogrel (and probably bisphosphonates and selective serotonin
reuptake inhibitors [SSRIs] as well) [27,42,54,56,58].
Studies have suggested that selective serotonin reuptake inhibitors (SSRIs) are
associated with an increased risk of upper gastrointestinal bleeding, particularly in
235
patients taking NSAIDs [59-64]. A possible mechanism is platelet serotonin depletion,
which may adversely influence the hemostatic response to vascular injury
(see "Overview of hemostasis") [54]. In a meta-analysis that included 15 case-control
studies and 4 cohort studies, the use of SSRIs was associated with an increased risk of
upper gastrointestinal bleeding compared with not using an SSRI (odds ratio [OR] for
case-control studies 1.7, 95% CI 1.4-1.92; OR for cohort studies 1.7, 95% CI 1.1-2.5)
[65]. When the analysis was confined to patients receiving NSAIDs, the risk was even
higher (OR 4.3, 95% CI 2.8-6.4). The risk of toxicity may not be uniform among the
NSAIDs. Many studies of varying design have described gastrointestinal toxicity of
different NSAIDs. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity", section on 'Nonselective NSAIDs'.)
Genetic predisposition to gastroduodenal injury due to polymorphism of cytochrome
P450 2C9 may delay the metabolism of several NSAIDs, with a prolonged duration of
drug exposure enhancing the ulcerogenic effect [66]. Whether these variations can be
used to modify care in individual patients has not yet been determined.
American College of Gastroenterology recommendations — The American College
of Gastroenterology issued guidelines for risk stratifying patients with regard to NSAID-
related gastrointestinal toxicity. These guidelines are discussed elsewhere.
(See "NSAIDs (including aspirin): Primary prevention of gastroduodenal toxicity",
section on 'Risk factors'.)
RELATIONSHIP BETWEEN DYSPEPTIC SYMPTOMS AND
GASTRODUODENAL MUCOSAL INJURYThe majority of the biochemical,
microscopic and macroscopic gastric lesions caused by nonsteroidal anti-inflammatory
drugs (NSAIDs) produce very little in the way of symptoms. On the other
hand, aspirin and nonacetylated salicylates, many NSAIDs, and
even acetaminophen [67], may produce dyspeptic symptoms (epigastric discomfort,
upper abdominal pain, nausea, epigastric fullness or bloating), especially as the dose of
the drug is increased.
There is little correlation between the dyspeptic symptoms sometimes seen with these
drugs and the presence or absence of erosive/ulcerative lesions in the stomach and
duodenum. Two examples illustrate this point. First, patients may present with severe
ulcer complications from an NSAID with no preceding dyspeptic "warning" symptoms.
Second, acetaminophen and salicylsalicylic acid (salsalate) commonly produce
dyspepsia, especially when used in higher doses, but neither blocks gastric COX
activity over a wide range of doses or causes damage to the gastric mucosa (table 1)
[8,67,68].

These observations emphasize that studies evaluating gastric and duodenal toxicity of
NSAIDs should not include dyspepsia as a part of a combined gastrointestinal endpoint.
In most instances the mechanism for the drug-related dyspepsia has not been
determined. Dyspepsia and ulcer formation should be viewed as separate issues when
using analgesic drugs.

INFORMATION FOR PATIENTSUpToDate offers two types of patient education

materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
236
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Beyond the Basics topics (see "Patient education: Nonsteroidal antiinflammatory

drugs (NSAIDs) (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Mucosal damage by aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs)
is primarily a consequence of inhibition of COX-1 in the upper gastrointestinal
tract, although COX-2 inhibition may also play a role. COX inhibition reduces
mucosal generation of protective prostaglandins (PG) such as PGE2. (See 'The
central protective role of prostaglandins' above and 'Mechanisms of
gastroduodenal protection by endogenous prostaglandins' above.)
●The duration and dose of NSAIDs, increasing age, a past history of
gastroduodenal toxicity from NSAIDs, peptic ulcer disease, and concurrent use of
glucocorticoids, anticoagulants, clopidogrel and, possibly, bisphosphonates and
selective serotonin reuptake inhibitors are associated with an increased risk of
gastroduodenal toxicity and complications from NSAIDs. (See 'Risk of
gastrointestinal ulcer complications' above.)
●Strategies to avoid mucosal damage include using anti-inflammatory or analgesic
drugs that have minimal effects on COX-1 at usual doses, such
as acetaminophen, or non-acetylated salicylates, or a selective COX-2 inhibitor,
and prescribing either a potent inhibitor of gastric acid production such as a proton
pump inhibitor or a prostaglandin E analog such as misoprostol together with the
NSAID. (See 'Risk of gastrointestinal ulcer complications' above.)
These strategies, though effective, are expensive and may not be cost-effective
and in the case of selective COX-2 inhibitors may impart an increased risk of
cardiovascular disease. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity".)
Use of UpToDate is subject to the Subscription and License Agreement.

237
NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity
Authors:
Mark Feldman, MD, MACP, AGAF, FACG
Shounak Das, MD
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: May 04, 2021.
INTRODUCTIONNonsteroidal anti-inflammatory drugs (NSAIDs), including aspirin,
cause considerable morbidity and mortality related to gastric and duodenal mucosal
injury [1]. Thus, prevention of NSAID-induced GI toxicity is an important clinical issue.
The prostaglandin E analog misoprostol, H2 receptor antagonists, and proton pump
inhibitors (PPIs) have been evaluated as prophylactic therapies for patients taking
NSAIDs. In addition, the selective COX-2 inhibitors (coxibs) is another strategy for the
reduction of NSAID-related gastroduodenal toxicity. However, concerns about
cardiovascular toxicity has limited development and marketing of coxibs.
(See "Overview of COX-2 selective NSAIDs" and "NSAIDs: Adverse cardiovascular
effects".)
Strategies for the primary prevention of gastroduodenal toxicity due to NSAIDs and low-
dose aspirin will be reviewed here. Emphasis will be placed upon studies that used
clinically relevant end points (symptomatic ulcers and complicated ulcers, including
bleeding, perforating, and obstructing ulcers). Studies using endoscopic detection of
ulcers as the end point will be cited only when data on more meaningful clinical end
points are sparse or lacking. The pathogenesis, treatment, and secondary prevention of
NSAID-induced gastroduodenal injury are discussed separately. (See "NSAIDs
(including aspirin): Pathogenesis of gastroduodenal toxicity" and "NSAIDs (including
aspirin): Treatment of gastroduodenal toxicity" and "NSAIDs (including aspirin):
Secondary prevention of gastroduodenal toxicity".)
RISK FACTORSStudies have evaluated risk factors for gastroduodenal toxicity from
nonsteroidal anti-inflammatory drugs (NSAIDs) and low-dose aspirin and assessment of
these factors is recommended for identifying patients who should be considered for
primary prophylaxis if it is felt that an NSAID or low-dose aspirin must be given [2-5].
(See "NSAIDs (including aspirin): Pathogenesis of gastroduodenal toxicity".)
One prospective study of 34,701 osteoarthritis and rheumatoid arthritis patients ages 50
and above randomized to either a coxib or diclofenac use found the following factors to
be significant predictors for gastrointestinal (GI) toxicity including bleeding, perforation,
obstruction, or uncomplicated ulcer: age >64, a history of prior adverse GI events, or
concurrent use of low-dose aspirin [6].
In the ASPREE primary prevention of cardiovascular disease trial of 19,114 persons
aged 70 years or above, low-dose aspirin therapy increased upper GI bleeding (hazard

238
ratio [HR] 1.87; 95% CI 1.32-2.66), without leading to significant cardiovascular benefit
[5].
According to 2008 American College of Cardiology Foundation/American College of
Gastroenterology (ACG)/American Heart Association guidelines, patients with the
following factors are considered to be at high risk for GI toxicity from NSAIDs [4]:
●History of ulcer disease or ulcer complication
●On dual antiplatelet therapy
●On anticoagulant therapy
●Have two or three of the following:
•Age ≥60 years
•Glucocorticoid use
•Dyspepsia or gastroesophageal reflux disease symptoms
In separate guidelines from the ACG in 2009, patients taking NSAIDs were classified as
being at high, moderate, or low risk for gastroduodenal toxicity [7]:
●High risk was defined as a history of a complicated ulcer or ≥3 risk factors
●Moderate risk was defined as the presence of one or two risk factors
●Low risk was defined as none of the four risk factors

The four risk factors are:

●History of an uncomplicated ulcer

●Age >65 years
●High-dose NSAID therapy
●Concurrent use of aspirin (including low dose), glucocorticoids, or anticoagulants
Use of selective serotonin reuptake inhibitors (SSRIs) has also been associated with an
increased risk of GI bleeding [8,9]. A meta-analysis of observational studies showed an
odds ratio of 2.36 (1.44-3.85) for SSRI associated upper GI hemorrhage. The odds ratio
increased to 6.33 (3.40-11.8) for concurrent SSRI and NSAID use [8]. (See "NSAIDs
(including aspirin): Pathogenesis of gastroduodenal toxicity", section on 'Risk of
gastrointestinal ulcer complications'.)
A large case series analysis involving over 100,000 patients with upper GI bleeding
demonstrated that monotherapy with SSRIs, glucocorticoids, aldosterone antagonists,
nitrates, and calcium channel blockers was associated with an increased risk of
bleeding [9]. Furthermore, glucocorticoids and aldosterone antagonists added to the risk
of bleeding when used together with nonselective NSAIDs [9].
NONSELECTIVE NSAIDsIn a 2013 meta-analysis that utilized data from over
300,000 participants in over 750 trials, all NSAID regimens examined increased upper
gastrointestinal complications including upper gastrointestinal perforation, obstruction,
or bleeding (adjusted rate ratio [ARR] for diclofenac 1.89, 95% CI 1.2-
3.1; ibuprofen 3.97, 95% CI 2.2-7.1; and naproxen 4.2, 95% CI 2.7-6.6) when compared
to placebo [10].
In another meta-analysis of controlled trials involving some of the most commonly
prescribed NSAIDs, the following conclusions were reached [11]:
●The risk of gastrointestinal complications was highest with indomethacin (relative
risk [RR] 2.25), followed by naproxen (RR 1.83), diclofenac (RR

239
 1.73), piroxicam (RR 1.66), tenoxicam (RR 1.43), ibuprofen (RR 1.19),
and meloxicam (RR 1.24).
●The risk was related to the duration of treatment. The average duration of
treatment before observing a significant risk of GI effects was 84 days. However,
an increased risk was apparent as early as seven days with indomethacin.
Similar conclusions were reached in an earlier meta-analysis, which ranked the risk of
various NSAIDs relative to ibuprofen, which at the time was considered to have the
lowest risk [12]. A ranking of increasing risk was found in the following order:
ibuprofen, fenoprofen, aspirin, diclofenac, sulindac, diflunisal, naproxen, indomethacin, t
olmetin, piroxicam, ketoprofen, and azapropazone.
A cohort study conducted in United States veterans suggests that the
NSAID etodolac caused significantly fewer clinically significant upper gastrointestinal
events than naproxen, an effect that was eliminated by concurrent treatment with low-
dose aspirin [13].
Ulcer risk also increased with higher doses of the NSAIDs in the aforementioned meta-
analysis [11]. The following relative risks were observed in a subset of studies that had
described GI toxicity relative to the dose of the NSAID:
●Low-dose ibuprofen (RR 1.6, 95% CI 0.8-3.2)
●High-dose ibuprofen (RR 4.2, 95% CI 1.8-9.8)
●Low-dose naproxen (RR 3.7, 95% CI 1.7-7.7)
●High-dose naproxen (RR 6.0, 95% CI 3.0-12.2)
●Low-dose indomethacin (RR 3.0, 95% CI 2.2-4.2)
●High-dose indomethacin (RR 7.0, 95% CI 4.4-11.2)
While the dosing cutoffs defining "high" versus "low" were somewhat arbitrary and
varied across studies, they nevertheless support the observation that the risk
associated with NSAIDs is related not only to the duration of therapy but also to the
dose of treatment. (See "Aspirin in the primary prevention of cardiovascular disease and
cancer", section on 'Bleeding'.)
Ketorolac, which was not included in the above meta-analyses, is also associated with a
high risk of GI toxicity, particularly when used in higher doses, in older patients, and for
more than five days [14]. One study found that ketorolac was 5.5 times more likely to
cause GI toxicity than other NSAIDs [15]. Because of the risks associated with ketorolac
(both gastrointestinal and renal), its use should be restricted to short-term pain
treatment [16,17]. (See "Nonselective NSAIDs: Overview of adverse effects".)
ENTERIC-COATED AND BUFFERED ASPIRINIt has been proposed that a
way to reduce gastrointestinal (GI) toxicity from aspirin is the use of enteric-coated or
buffered aspirin. Enteric-coated aspirin is designed to resist disintegration in the
stomach, dissolving in the more neutral-to-alkaline environment of the duodenum.
Although enteric-coated aspirin diminishes endoscopic signs of gastroduodenal injury, it
does not protect against the clinically relevant end point of gastrointestinal bleeding [18-
22]. In a placebo-controlled Australian/United States study in healthy older adults,
enteric-coated aspirin 100 mg per day increased the risk of upper GI bleeding (HR 1.87;
95% CI, 1.32-2.66) without significantly reducing cardiovascular events [22]. These
findings are not surprising, since injury severe enough to induce bleeding is thought to
reflect the systemic rather than the topical effects of aspirin [23]. The systemic effect of
aspirin on the stomach and duodenum also probably explains why buffered aspirin is no
240
more effective than plain aspirin in preventing ulcer bleeding [21]. (See "NSAIDs
(including aspirin): Pathogenesis of gastroduodenal toxicity", section on 'Systemic
versus topical effects'.)
Low-dose aspirin for cardiovascular protection — Issues related to bleeding risk of
low-dose aspirin for primary cardiovascular protection are also discussed separately.
(See "Aspirin in the primary prevention of cardiovascular disease and cancer", section
on 'Bleeding'.)
Dual antiplatelet therapy (eg, aspirin and clopidogrel) — Issues related to GI
bleeding with dual antiplatelet therapy in patients with cardiovascular disease are
discussed separately. (See "Coronary artery disease patients requiring combined
anticoagulant and antiplatelet therapy", section on 'Bleeding'.)
ROLE OF HELICOBACTER PYLORIMultiple studies have evaluated the
relationship between Helicobacter pylori and the risk of peptic ulcer disease (PUD) in
nonsteroidal anti-inflammatory drug (NSAID) users. (See "NSAIDs (including aspirin):
Pathogenesis of gastroduodenal toxicity", section on 'Role of Helicobacter pylori
infection'.)

Based upon the available evidence, the following is a reasonable approach:

●Patients with a history of uncomplicated or complicated peptic ulcers (gastric,

duodenal) should be tested for H. pylori prior to beginning a course of NSAID or
low-dose aspirin therapy. If present, H. pylori should be treated with appropriate
therapy, even if it is believed that the prior ulcer was due to NSAIDs.
●In asymptomatic patients with no history of ulcer and not currently taking an
NSAID, physicians can consider H. pylori testing prior to beginning long-term
therapy with a NSAID. A review of this topic suggested that eradication of H.
pylori was beneficial in patients who were naïve to NSAIDs, while little benefit was
observed in patients already taking and tolerating NSAIDs [24]. This "test-and-
treat" approach may be more useful in populations with a relatively high
prevalence of H. pylori infection.
SELECTIVE COX-2 INHIBITORS (COXIBS)The primary effect of nonsteroidal
anti-inflammatory drugs (NSAIDs) is to inhibit cyclooxygenase (COX), thereby impairing
the transformation of arachidonic acid to prostaglandins, prostacyclin, and
thromboxanes. Two isoforms of COX exist: COX-1 and COX-2. COX-1 is constitutive
and involved in gastric and duodenal cytoprotection [25], while COX-2 is inducible and
involved in inflammation and perhaps healing of gastroduodenal lesions. It has been
proposed that the ideal NSAID would inhibit the inducible COX-2 isoform (thereby
decreasing tissue inflammation) without having any effect on the COX-l isoform (thereby
minimizing gastrointestinal [GI] toxicity) [25,26]. (See "NSAIDs: Pharmacology and
mechanism of action" and "Overview of COX-2 selective NSAIDs".)
Clinical trials and meta-analyses have evaluated the gastroduodenal toxicity of coxibs
when compared to nonselective NSAIDs [10]. Detailed information regarding these trials
is presented elsewhere. These data suggest that coxibs are associated with a reduced
risk of gastrointestinal bleeding compared with nonselective NSAIDs but that the risk is
increased compared with placebo. Thus, coxibs may be safer than conventional
NSAIDs for reduction in the risk of gastrointestinal bleeding but are still associated with

241
an increased risk. (See "COX-2 inhibitors and gastroduodenal toxicity: Major clinical
trials".)
Any potential gastroduodenal sparing effect with coxibs may be abrogated when they
are used concurrently with low-dose aspirin therapy for prevention of cardiovascular
disease [27]. One population-based case control study found that coxibs were
associated with a modest reduction in the risk of gastrointestinal bleeding compared
with nonselective NSAIDs (RR 0.6, 95% CI 0.4-0.9 among aspirin nonusers) [28].
Concomitant use of aspirin negated the benefit of coxibs.
A case-control study suggested that patients taking warfarin concomitantly with a
nonselective NSAID or a coxib have an increased risk of hospitalization for upper
gastrointestinal bleeding [29]. The magnitude of risk was similar, suggesting that the
coxibs may not be GI-protective in this population.
Coxib usage declined with concerns related to their cardiovascular toxicity. In the
Prospective Randomized Evaluation of Celecoxib Integrated Safety (PRECISION) trial
in patients with arthritis pain, however, celecoxib was found to be non-inferior
to ibuprofen and naproxen with respect to a composite primary cardiovascular endpoint
[30]. (See "NSAIDs: Adverse cardiovascular effects".)
PREVENTION STRATEGIESOptions for reducing the risk of gastroduodenal
toxicity include (a) using a nonselective nonsteroidal anti-inflammatory drug (NSAID)
together with a proton pump inhibitor (PPI) or misoprostol or (b) using a coxib with or
without a PPI. The use of high-dose H2 blockers is reserved for patients who cannot
tolerate PPIs or misoprostol.
Multiple studies have evaluated a variety of strategies for preventing ulcers in patients
requiring NSAIDs. A meta-analysis of 112 randomized controlled trials found no
evidence supporting the effectiveness of H2 receptor antagonists at standard doses
[31]. However, the risk of symptomatic ulcers was significantly reduced by PPIs (relative
risk [RR] 0.09, 95% CI 0.02-0.47) and by misoprostol (RR 0.36, 95% CI 0.02-0.65).
Coxibs in place of a nonselective NSAID also reduced risk (RR 0.49, 95% CI 0.38-0.65).
Proton pump inhibitors — PPIs are useful for the prevention of NSAID-induced and
low-dose aspirin-induced ulcers [32-37]. The Prospective Randomized Evaluation
of Celecoxib Integrated Safety (PRECISION) trial randomized OA or RA patients whose
arthritis pain was resistant to acetaminophen and who were at increased risk for
cardiovascular disease (nearly one-half taking low-dose aspirin), but who were not at
very high risk for ulcer disease (recent ulcer disease was an exclusionary criteria), to
receive celecoxib (100 mg bid), ibuprofen (600 mg tid), or naproxen (375 mg bid) as
initial NSAID doses to control arthritis pain plus esomeprazole (20 or 40 mg per day in
each group). Clinically significant gastrointestinal (GI) events during a median follow up
period of 34 months occurred in only 0.7 percent, 0.9 percent, and 0.7 percent of
patients, respectively [38]. PPIs are better tolerated than misoprostol [30].
Misoprostol — The risk for NSAID-induced gastric or duodenal ulcer can be decreased
with concomitant use of the prostaglandin E analog misoprostol [38]. In the largest trial,
8843 patients with rheumatoid arthritis receiving continuous therapy with any of 10
nonselective NSAIDs were randomly assigned to receive 200 mcg of misoprostol or
placebo four times daily for six months [2]. Serious upper GI complications (bleeding,
perforation, gastric outlet obstruction) were reduced in patients receiving misoprostol
from 0.95 to 0.38 percent, a relative risk reduction of 40 percent, with an absolute risk
242
reduction of 0.57 percent (ie, 175 patients needed to be treated to reduce one
complication). However, more patients receiving misoprostol rather than placebo
withdrew from the study during the first month (20 versus 15 percent), primarily because
of diarrhea and abdominal discomfort. Misoprostol is not tolerated as well as PPIs
[31,33].
Doses of misoprostol lower than 200 mcg four times daily have fewer side effects [38].
These lower doses have not received FDA approval, and have not been shown to
reduce clinically meaningful ulcer endpoints, however.
PPI (lansoprazole) versus misoprostol
●The efficacies of lansoprazole and misoprostol were compared in a multicenter
trial in which 537 patients who were long-term NSAID users were randomly
assigned to placebo, misoprostol (200 mcg four times daily), or one of two doses
of lansoprazole (15 or 30 mg daily) [33]. After 12 weeks, the 49 percent incidence
of endoscopically detected gastroduodenal ulceration with placebo and was
significantly reduced with misoprostol (7 percent) and, to a lesser extent, with both
doses of lansoprazole (20 and 18 percent, respectively). However, patient
withdrawals were more common in the misoprostol group.
PPI (esomeprazole) used with either nonselective NSAIDs or coxibs
Esomeprazole was evaluated for prevention of NSAID-associated ulcers based upon
the results of two multicenter trials involving a total of 1429 patients who were taking
NSAIDs continuously (nonselective or COX-2 selective, with or without low-
dose aspirin).
Patients were at increased risk for peptic ulcer disease (PUD) either because they were
older than 60 or had a history of an ulcer. They were randomly assigned to co-therapy
with a PPI (20 or 40 mg of esomeprazole) or placebo for six months. All patients
were H. pylori negative and were free of active ulcer disease at randomization. The
primary end point was endoscopically-detected ulcer development [39]. The major
findings in both studies were as follows:
●Endoscopic ulcers developed in a similar proportion of patients taking a
nonselective NSAID or coxib without esomeprazole (17.1 versus 16.5 percent).
This observation underscores that selective COX-2 inhibitors may not protect
against endoscopic ulcer formation in these high-risk patients.
●The cumulative proportion of patients developing endoscopic ulcers at six months
was significantly reduced with esomeprazole co-therapy (17 percent with placebo
versus 5.2 and 4.6 percent with the 20 and 40 mg dose of esomeprazole,
respectively).
●The proportion of patients developing endoscopic ulcers who were taking a coxib
plus esomeprazole was not significantly different from the proportion taking a
nonselective NSAID plus esomeprazole (3 versus 6 percent).
Several important study limitations should be considered. First, only 17 percent of
patients had a history of PUD and none had active ulcer disease at randomization and
thus the studies focused mainly on primary prophylaxis in elderly patients. Second,
patients were not randomized to a nonselective NSAID versus a selective coxib. Thus,
patients who were considered to be at higher risk for UD may have been treated with a
coxib, possibly accounting for the high rate of endoscopic ulcer in this group. Third, the
studies analyzed patients receiving low-dose aspirin as having received a nonselective

243
NSAID even if they were taking a coxib. Fourth, the studies focused on endoscopic, not
clinical, ulcer disease.
Coxib (celecoxib) versus naproxen plus lansoprazole, together with low-dose
aspirin
A randomized endoscopic trial in low-dose aspirin users compared celecoxib 200 mg
per day with the combination of naproxen 500 mg twice daily and lansoprazole [40].
There was no difference between these two approaches, with gastroduodenal ulcers
seen in 10 and 9 percent of patients, respectively.
Coxib (etoricoxib) versus diclofenac
Another randomized controlled trial assessed the cardiovascular toxicity of etoricoxib
versus the nonselective NSAID diclofenac [41]. In a pre-specified analysis of GI toxicity,
etoricoxib was associated with significantly less gastrointestinal toxicity regardless of
PPI or ASA co-administration. This benefit held true only for uncomplicated GI events;
there was no difference between the two groups for complicated GI events [41].
Nonselective NSAID with either a PPI or misoprostol versus a coxib
Data are conflicting about whether coxibs provide additional protection compared with
conventional, nonselective NSAIDs that are combined with either a proton pump
inhibitor or misoprostol:
●In a randomized trial of 4484 patients with osteoarthritis or rheumatoid arthritis,
2238 patients were assigned to receive celecoxib and 2246 patients were
assigned to receive diclofenac plus omeprazole [42]. Patients in the celecoxib arm
were significantly less likely to develop GI toxicity compared with patients in the
diclofenac/omeprazole arm (1 versus 4 percent, hazard ratio 4.3).
●A population based study compared 1382 patients with upper gastrointestinal
complications who were taking a conventional NSAID or a coxib with 22,957 age-
and sex-matched controls [43]. Co-therapy with a PPI or misoprostol or use of a
coxib all significantly reduced the risk of upper gastrointestinal complications.
COX-2 inhibitor use was not more likely to lower complications compared with
PPIs but were superior to low-dose misoprostol. The combination of a coxib with a
PPI was associated with a greater reduction in risk (adjusted OR 0.36 (95% CI
0.28-0.47) as compared to an OR of 0.67 (95% CI 0.48-0.95) for a PPI plus a
conventional, nonselective NSAID.
The issue of using a coxib plus a PPI compared to a nonselective NSAID plus a PPI in
patients at increased risk for PUD remains unsettled. The former approach is probably
non-inferior but may not be superior [30]. Furthermore, the PPI rabeprazole was
associated with more celecoxib-related small intestinal mucosal injury (mostly erosions
as opposed to ulcers) than placebo plus celecoxib [44], although the clinical significance
of this observation is unknown.
H2 receptor antagonists — Standard doses of H2 receptor antagonists were not
effective for the prevention of NSAID-induced gastric ulcers in most reports, although
they may prevent duodenal ulcers [45]. Studies that detected a benefit on gastric ulcer
prevention were short-term (12 to 24 weeks) and focused on endoscopic rather than
clinical endpoints [46,47].
High-dose H2 receptor antagonists have also been studied. Again, the trials were short-
term and focused on endoscopic endpoints. In two randomized trials (REDUCE-1 and
REDUCE-2), patients who required daily NSAIDs for at least six months were assigned

244
to receive either ibuprofen 800 mg plus famotidine 26.6 mg or ibuprofen 800 mg three
times per day [48]. The outcome of interest was ulcer development over 24 weeks of
treatment. The pooled analysis of the two studies comparing the drug combination with
ibuprofen alone found that the combination significantly reduced overall upper
gastrointestinal ulcer development (14 versus 24 percent), gastric ulcer development
(13 versus 21 percent), and duodenal ulcer development (2 versus 7 percent).
Aspirin-phosphatidylcholine combination — A study in healthy volunteers between
the ages of 50 and 74 found that combining 325 mg aspirin per day with
phosphatidylcholine reduced a combined endoscopic endpoint (petechiae, erosions,
ulcers) compared with aspirin alone [49].
Potassium-competitive acid blocker — A novel agent, vonoprazan, has been shown
to be non-inferior to the PPI lansoprazole in a double-blinded, randomized controlled
study for the secondary prevention of NSAID-associated peptic ulcers [50]. Vonoprazan
has been approved for use in Japan but is not available worldwide.
MONITORING PATIENTS TAKING NSAIDSMonitoring patients while on
NSAIDs can be difficult since many patients who develop GI toxicity are asymptomatic
until shortly before the adverse event occurs. An NSAID-induced gastrointestinal
complication should be suspected if the patient develops unexplained blood loss
anemia, iron deficiency, significant dyspepsia, or exhibits overt GI bleeding. The
appropriate test to order is upper gastrointestinal endoscopy, unless ulcer perforation is
suspected in which case computed tomography scan of the abdomen should be
ordered immediately.
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: NSAID-related ulcer
complications" and "Society guideline links: Peptic ulcer disease".)
SUMMARY AND RECOMMENDATIONSAs noted above, multiple studies have
evaluated a variety of strategies for preventing ulcers in patients requiring prolonged
nonsteroidal anti-inflammatory drugs (NSAIDs) or low-dose aspirin for cardiovascular
protection. The following represents recommendations based upon the available data
and expert consensus.
●The risk of gastroduodenal toxicity is related to risk factors described above, the
patient's Helicobacter pylori (H. pylori) status, and the specific type and dose of
NSAID. Significant gastrointestinal (GI) toxicity generally does not begin for
several weeks, although it can appear as early as day 7 with some NSAIDs such
as indomethacin or ketorolac. (See 'Risk factors' above.)
●The ability to modify the risk depends upon the clinical setting. As noted above,
patients with a history of uncomplicated or complicated peptic ulcers (gastric,
duodenal) should be tested for H. pylori prior to beginning long-term use of an
NSAID or low-dose aspirin. If present, H. pylori should be treated with appropriate
therapy, even if it is believed that the prior ulcer was due to NSAIDs. (See 'Role of
helicobacter pylori' above.)
●Attempts should be made to use the lowest dose and shortest duration of NSAID
treatment that is feasible. While it is preferable to use NSAIDs with the lowest
potential for GI toxicity, all NSAIDs are associated with an increased GI risk.

245
●In asymptomatic patients with no history of ulcer and not currently taking an
NSAID, physicians can consider H. pylori testing prior to beginning long-term
therapy with a NSAID. It is possible that successfully treating H. pylori infection in
such individuals will reduce the risk of NSAID-related ulcer complications, but
additional studies of this approach are needed. This "test-and-treat" approach may
be more useful in populations with a relatively high prevalence of H.
pylori infection. In patients with a history of peptic ulcer disease or ulcer
complications requiring an NSAID or low-dose aspirin, we recommend testing
for H. pylori and treating for H. pylori if positive (Grade 1A). (See "NSAIDs
(including aspirin): Secondary prevention of gastroduodenal toxicity".)
●For patients who have multiple risk factors for NSAID-related gastroduodenal
toxicity, options include therapy with a coxib or a nonselective NSAID in
combination with a proton pump inhibitor (PPI) or misoprostol. High-dose H2
receptor antagonists are reserved for patients who cannot tolerate PPIs or
misoprostol. (See 'Risk factors' above and 'Selective COX-2 inhibitors
(COXIBS)' above and 'Proton pump inhibitors' above.)
●We recommend co-administration of a PPI in patients who require an NSAID and
are at high or moderate risk for GI toxicity (Grade 1B). (See 'Risk factors' above.)
We prefer PPIs to other preventive approaches because of their convenience and
relatively good safety profile.
●The approved doses of these ulcer prevention drugs in patients taking
nonselective NSAIDs include misoprostol (200 mcg four times
daily), lansoprazole (15 or 30 mg daily), and esomeprazole (20 or 40 mg daily).
Although not all PPIs have received FDA approval, they probably all have similar
effectiveness.
●If a coxib is used, it is important to realize that any GI protective effect may be
eliminated or reduced with the concomitant use of low-dose aspirin for
cardiovascular prophylaxis. (See 'Coxib (celecoxib) versus naproxen plus
lansoprazole, together with low-dose aspirin' above.)
●An NSAID-induced gastrointestinal complication should be suspected if the
patient develops unexplained blood loss anemia, iron deficiency, severe
dyspepsia, or exhibits overt GI bleeding. Patients who develop NSAID-induced
non-ulcer dyspepsia may respond to H2 receptor antagonists or PPIs at standard
doses, but of these choices only PPIs also reliably reduce the risk of NSAID-
induced ulcer or its complications. Development of dyspepsia may signal that an
ulcer has developed at which point an upper endoscopy can be considered.
Switching to a different NSAID or to a coxib may eliminate dyspepsia and obviate
the need for an H2 antagonist or PPI. Misoprostol does not reduce, and may even
increase, dyspeptic symptoms. (See 'Monitoring patients taking NSAIDs' above
and 'Misoprostol' above.)

246
NSAIDs (including aspirin): Secondary prevention of
gastroduodenal toxicity
Author:
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Feb 01, 2021.
INTRODUCTIONNonselective (ie, inhibiting both cyclooxygenase [COX]-1 and
COX-2) nonsteroidal anti-inflammatory drugs (NSAIDs), including low-dose aspirin, can
cause considerable morbidity and mortality related to gastric and duodenal ulcer
disease, particularly gastrointestinal (GI) bleeding [1,2].
The prevention of recurrent gastroduodenal toxicity associated with NSAID, low-
dose aspirin, or both therapies (secondary prevention) will be reviewed here. Primary
prevention and treatment of NSAID-related and/or low-dose aspirin-related
gastroduodenal toxicity are discussed separately, as is the pathogenesis of the
gastroduodenal toxicity. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity" and "NSAIDs (including aspirin): Treatment of gastroduodenal
toxicity" and "NSAIDs (including aspirin): Pathogenesis of gastroduodenal toxicity".)
SECONDARY PREVENTION OF GASTRODUODENAL
TOXICITYPrevention of recurrent ulcer disease becomes critically important in
patients with a history of gastroduodenal toxicity from nonsteroidal anti-inflammatory
drugs (NSAIDs), low-dose aspirin or both therapies, and who require continued therapy.
Only randomized secondary prevention trials using clinical ulcer disease as the
outcome (rather than an endoscopic outcome) are discussed below.
With continued NSAID therapy — There are occasional patients who must resume or
continue NSAID therapy despite prior symptomatic gastroduodenal ulcers or erosions.
In such high risk individuals, a proton pump inhibitor (PPI) is recommended for as long
as the NSAID is used. In one trial of patients with bleeding ulcers who tested positive
for H. pylori and had to remain on naproxen or low-dose aspirin, maintenance therapy
with omeprazole was more effective in preventing recurrent upper gastrointestinal (GI)
bleeding than anti-H. pylori therapy alone [3]. In clinical practice, however, ulcer patients
who are H. pylori positive are generally treated with a PPI along with H.
pylori eradication. (See "Peptic ulcer disease: Treatment and secondary prevention",
section on 'Eradication of Helicobacter pylori (H. pylori)'.)
When first introduced, COX-2 inhibitors held promise as an alternative to traditional
NSAIDs in preventing recurrent peptic ulcer disease in individuals who required ongoing
anti-inflammatory therapy. (See "NSAIDs (including aspirin): Primary prevention of
gastroduodenal toxicity", section on 'Selective COX-2 inhibitors (COXIBS)' and "NSAIDs
(including aspirin): Pathogenesis of gastroduodenal toxicity".) Simply substituting a
COX-2 inhibitor for the prior nonselective NSAID in patients with NSAID-related ulcer
247
disease and ulcer complications should not be chosen over PPI maintenance therapy.
This was demonstrated in a trial of 273 arthritis patients with NSAID-related bleeding
ulcers. Once the ulcers had healed, the authors found that substituting celecoxib 200
mg bid plus esomeprazole (20 mg bid) for a year was significantly more effective in
preventing recurrent bleeding ulcers than simply substituting celecoxib (plus placebo
bid) for the NSAID that the patient had been taking previously [4]. Ulcer rebleeding rates
after a median follow-up period of 13 months were 0 percent in the
celecoxib/esomeprazole group and 8.9 percent in the celecoxib/placebo group
(P=0.004).
With continued low-dose aspirin therapy — Low-dose aspirin (75 to 325 mg/day) is
of proven benefit in the secondary prevention and, in selected patients, the primary
prevention of cardiovascular disease [5]. However, patients treated with low-dose
aspirin have an increased risk of gastrointestinal bleeding, which must be weighed
against a possible decrease in cardiovascular mortality. The risk of recurrent bleeding in
aspirin users may be higher in patients without a past or current H. pylori infection [6].
(See 'Risk of stopping low-dose aspirin' below.)
Randomized trials support the efficacy of proton pump inhibitor (PPI) therapy in
preventing recurrent bleeding in both H. pylori positive and negative patients receiving
low-dose aspirin therapy. Different approaches using a PPI to prevent recurrent GI
bleeding in patients who continue low-dose aspirin therapy have been evaluated in
several randomized trials performed in East Asia [3,7-12].
Placebo-controlled trials showed a significant reduction in recurrent ulcer complications
with PPI therapy among low-dose aspirin users. The trials can be summarized as
follows:
●Among patients infected with H. pylori, the rate of recurrent ulcer complications
was significantly lower at one year with lansoprazole (30 mg daily) compared to
placebo (1.6 versus 14.8 percent). Lansoprazole or placebo was started after the
ulcer had healed and H. pylori had been eradicated [7].
●A similar difference in recurrent bleeding at one year (0.7 versus 8.6 percent)
in H. pylori negative patients was noted with esomeprazole (20 mg twice daily)
plus low-dose aspirin (80 mg daily) compared to an esomeprazole placebo twice
daily plus clopidogrel (75 mg/day) [8]. A similar trial was stopped early because of
a higher bleeding rate in the clopidogrel group [9].
●Other trials in low-dose aspirin users compared different therapies without a
placebo group. In one, a PPI (pantoprazole) was associated with a significantly
lower rate of recurrent symptomatic or bleeding ulcers/erosions than an H2-
blocker (famotidine) (0 versus 20 percent at 48 weeks) [10]. In a second trial,
limited to patients who were H. pylori positive, there was no significant difference
in recurrent bleeding at six months with omeprazole (20 mg/day) compared to
anti-H. pylori therapy (0.9 versus 1.9 percent) [3]. In a third trial, rabeprazole (5 or
10 mg per day) markedly reduced the incidence of recurrent ulcers over 24 weeks
compared to the comparator therapy, teprenone (a mucosal protective agent), with
no bleeding ulcers occurring in the rabeprazole groups but in nearly 5 percent of
the teprenone group [12].
In summary, patients who require low-dose aspirin for cardiovascular prophylaxis and
have had a bleeding ulcer while taking low-dose aspirin should be treated with a PPI

248
(rather than an H2-blocker), along with H. pylori eradication if they test positive for H.
pylori. (See "Peptic ulcer disease: Treatment and secondary prevention", section on
'Eradication of Helicobacter pylori (H. pylori)'.)
There are no compelling data that suggest that any of the available PPIs are more
effective than another. The table lists the available medications and their usual dose
(table 1).
With continued NSAID and low-dose aspirin therapy — Some patients have multiple
risk factors for gastroduodenal toxicity, including use of an NSAID and low-dose aspirin,
and many of these patients are also at increased risk due their elderly status.
(See "NSAIDs (including aspirin): Primary prevention of gastroduodenal toxicity".)
Patients with a history of a complicated, NSAID-related peptic ulcer (especially if recent)
who must resume an NSAID in addition to low-dose aspirin are at very high GI risk. A
major controlled clinical trial was performed in Hong Kong in patients with bleeding
gastroduodenal ulcers related to NSAID and low-dose aspirin use. Once the ulcer had
healed, patients were placed on celecoxib (COX-2 selective)
plus esomeprazole or naproxen (COX nonselective) plus esomeprazole; both groups
were encouraged to continue taking their low-dose aspirin. Although recurrent ulcer
bleeding over the ensuing 18 months was common in both groups, it was significantly
less common in the celecoxib group than in the naproxen group [13].
In another major clinical trial, 24,081 patients whose arthritis pain could not be
controlled with acetaminophen were randomly assigned to receive
either celecoxib, naproxen, or ibuprofen. Nearly half of them were also receiving low-
dose aspirin for cardiovascular prophylaxis. All patients received esomeprazole co-
therapy. Patients at very high GI risk due to recent ulcer complications or uncomplicated
peptic ulcer disease were excluded, although some patients with a more remote history
of ulcer disease may have been included. Thus, this study mainly compared celecoxib
plus PPI with a nonselective NSAID plus PPI for primary prevention of GI toxicity, but in
some patients, the study likely constituted a secondary prevention trial. Clinically
significant GI events were uncommon during a median follow-up period of 34 months
and were similar in the groups (celecoxib 0.7 percent, naproxen 0.7 percent, and
ibuprofen 0.9 percent) [14].
In summary, if an NSAID and low-dose aspirin must be resumed in high GI risk
patients, celecoxib plus a PPI may be preferred over naproxen plus a PPI. However, in
patients at lower GI risk, celecoxib plus a PPI may not have an advantage over
nonselective NSAIDs plus a PPI.
Risk of stopping low-dose aspirin — Benefits and risks of stopping low-
dose aspirin were addressed in a randomized trial of 156 adults with cardiovascular or
cerebrovascular disease who were being treated with low-dose aspirin and developed
peptic ulcer bleeding. After control of bleeding with endoscopic therapy and use of a
PPI, patients were assigned to continue taking aspirin (80 mg per day) or to stop the
aspirin and switch to a placebo for eight weeks. Stopping aspirin therapy was
associated with a nonsignificant decrease in the primary endpoint of recurrent ulcer
bleeding within 30 days (5.4 percent versus 10.3 percent with continued aspirin
therapy). Stopping aspirin was also associated with a significant increase in the
secondary endpoint of overall mortality at eight weeks (12.3 versus 1.3 percent) [15].
Because the magnitude of this mortality difference was much greater than seen in

249
previous large randomized trials of secondary prevention of cardiovascular disease with
much longer follow-up, the authors appropriately concluded that larger trials are needed
to confirm their findings. (See "Aspirin for the secondary prevention of atherosclerotic
cardiovascular disease", section on 'Efficacy'.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: NSAID-related ulcer complications".)
SUMMARY AND RECOMMENDATIONS
●For patients with symptomatic gastroduodenal ulcers or numerous erosions who
must continue nonsteroidal anti-inflammatory drug (NSAID), low-
dose aspirin therapy or both, we recommend treatment with a proton pump
inhibitor for as long as the NSAID and/or aspirin is used (table 1) (Grade 1A).
(See 'Secondary prevention of gastroduodenal toxicity' above.)
●In addition, we recommend that H. pylori infection be eradicated, if known to be
present (Grade 1A). (See "Peptic ulcer disease: Treatment and secondary
prevention", section on 'Eradication of Helicobacter pylori (H.
pylori)' and "Treatment regimens for Helicobacter pylori in adults".)
●Selecting celecoxib (or perhaps another COX-2 selective agent, if available) plus
a proton pump inhibitor for patients who are at very high gastrointestinal (GI) risk
and who are taking low-dose aspirin for cardiovascular prophylaxis may be more
effective than using a nonselective NSAID plus a proton pump inhibitor in reducing
recurrent bleeding ulcers. Nevertheless, recurrent GI bleeding remains a concern
in such patients. (See 'With continued NSAID therapy' above.)

250
NSAIDs (including aspirin): Treatment of
gastroduodenal toxicity
Author:
Mark Feldman, MD, MACP, AGAF, FACG
Section Editor:
J Thomas Lamont, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Apr 13, 2020.
INTRODUCTIONNonsteroidal antiinflammatory drugs (NSAIDs), including aspirin,
cause considerable morbidity and mortality related to gastric and duodenal ulcer
disease, particularly by causing gastrointestinal (GI) bleeding. NSAIDs are also
important causes of GI bleeding in children [1].
The treatment of gastroduodenal toxicity associated with NSAID therapy will be
reviewed here. Modalities used for primary and secondary prevention of gastroduodenal
toxicity, including the role of cyclooxygenase (COX)-2 selective NSAIDs, are discussed
separately as is the pathogenesis of the gastroduodenal toxicity. The prevention of
recurrent gastroduodenal toxicity is also discussed separately. (See "NSAIDs (including
aspirin): Pathogenesis of gastroduodenal toxicity" and "NSAIDs (including aspirin):
Primary prevention of gastroduodenal toxicity" and "NSAIDs (including aspirin):
Secondary prevention of gastroduodenal toxicity".)
TREATMENTIf a patient develops an ulcer while on a nonsteroidal antiinflammatory
drug (NSAID) or low-dose aspirin, the NSAID or aspirin should be stopped if at all
possible and traditional ulcer therapy with a proton pump inhibitor or an H2 antagonist
started. Proton pump inhibitors are preferred because they are associated with superior
ulcer healing [2-4]. (See "Peptic ulcer disease: Treatment and secondary prevention".) If
the patient has been taking low-dose aspirin for cardiovascular prophylaxis, there is no
consensus as to when to resume the aspirin. The indication for the low-dose aspirin
should be reviewed and the severity of the ulcer presentation considered. For most
patients, the author recommends low-dose aspirin, if still indicated, be restarted
approximately five to seven days after initiating therapy with a proton pump inhibitor.
As in all patients with peptic ulcers, the patient's Helicobacter pylori (H. pylori) status
should also be assessed (if not done previously) [5]; if positive, appropriate therapy
for H. pylori should be instituted [6]. However, the sensitivity of H. pylori testing in the
setting of acute GI bleeding is significantly reduced. (See "Treatment regimens for
Helicobacter pylori in adults" and "Peptic ulcer disease: Treatment and secondary
prevention", section on 'Eradication of Helicobacter pylori (H. pylori)' and "Indications
and diagnostic tests for Helicobacter pylori infection in adults", section on 'Patient
undergoing upper endoscopy'.)
For patients who must remain on low-dose aspirin, NSAID therapy, or both, randomized
trials have shown that ulcer healing occurs more rapidly with a proton pump inhibitor
than an H2 antagonist [7,8], misoprostol [9], or sucralfate [10]. Trials comparing proton

251
pump inhibitor with an H2 antagonist were performed with ranitidine which has since
been withdrawn, but had comparable efficacy with other H2 antagonists. (See "Antiulcer
medications: Mechanism of action, pharmacology, and side effects", section on
'Adverse effects'.)
●In one trial, 541 patients who had ulcers or numerous (>10) erosions in either the
stomach or duodenum and who required continuous treatment with NSAIDs were
randomly assigned to treatment with omeprazole (20 or 40 mg PO daily) or the H2
antagonist ranitidine for four or eight weeks [7]. At eight weeks, healing rates were
greater with omeprazole (80 and 79 versus 63 percent with ranitidine).
●In a second trial, 353 patients with benign gastric ulcers who continued to receive
stable doses of NSAIDs were randomly assigned to treatment
with lansoprazole (15 or 30 mg PO daily) or ranitidine for eight weeks [8]. Healing
rates were higher in patients treated with lansoprazole as compared with ranitidine
(69 and 73 percent versus 53 percent, respectively).
●In another trial, 935 patients who had ulcers or numerous (>10) erosions in the
stomach or duodenum (or both) and who required continuous treatment with
NSAIDs were randomly assigned to treatment with omeprazole (20 or 40 mg PO
daily) or misoprostol (200 mcg four times per day). Patients were treated for four
weeks or, in the absence of ulcer healing, eight weeks [9]. The rates of healing of
gastric ulcers (87 versus 73 percent) and duodenal ulcers (93 versus 77 percent)
were significantly higher with omeprazole (20 mg) than with misoprostol. The
healing rates with 40 mg per day of omeprazole were also superior to misoprostol.
●The efficacy of omeprazole (20 mg daily) was compared to sucralfate (2 g twice
daily) in a much smaller study of 98 patients with gastric and duodenal ulcers who
required continued treatment with NSAIDs [10]. At eight weeks, omeprazole was
significantly superior to sucralfate for healing gastric ulcers (87 versus 52 percent);
the differences in the rates of healing of duodenal ulcers (95 versus 73 percent)
did not reach statistical significance.
There are no data that suggest that any of the available proton pump inhibitors are more
efficacious than another in inducing ulcer healing. The table lists the available
medications and their usual dose (table 1).
Duration of PPI therapy and role of endoscopy — Treatment with a PPI is generally
continued for four to eight weeks depending on the ulcer size and the severity of the
initial clinical presentation. In general, we suggest PPI therapy (eg, omeprazole 20 to 40
mg daily or equivalent dose of an alternative PPI) for four to six weeks for ulcers <1 cm
and PPI therapy for six to eight weeks for ulcers ≥1 cm. Endoscopic follow-up to assess
ulcer healing is usually unnecessary for duodenal ulcers, but should be performed to
assess healing of gastric ulcers if there had been any concern that malignancy had not
been excluded at the initial endoscopy.
Maintenance PPI therapy is indicated in patients who must remain on or resume low-
dose aspirin or NSAID or treatment. (See "NSAIDs (including aspirin): Secondary
prevention of gastroduodenal toxicity".) In the author’s opinion, and those of others [11],
the proven benefits of PPI maintenance therapy for secondary prevention of NSAID- or
low dose aspirin-associated ulcers considerably outweigh potential risks of long term
PPI therapy.

252
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: NSAID-related ulcer complications".)
SUMMARY AND RECOMMENDATIONS
●For patients who develop a symptomatic gastric or duodenal ulcer while on a
nonsteroidal antiinflammatory drug (NSAID) and/or low-dose aspirin, the author
recommends initially stopping the NSAID and/or aspirin. (See 'Treatment' above.)
●For patients in whom the NSAID and/or low-dose aspirin can be discontinued, we
recommend treating with a proton pump inhibitor for four to eight weeks (Grade
1A). (See 'Treatment' above.)
●For patients in whom the NSAID or low-dose aspirin cannot be stopped or must
be restarted before the ulcer is healed, we recommend treatment with a proton
pump inhibitor (table 1) (Grade 1A). Maintenance therapy with a PPI is also
warranted in such patients. (See "Peptic ulcer disease: Treatment and secondary
prevention".)
●The patient's H. pylori status should also be assessed; if positive, appropriate
therapy should be instituted. (See "Treatment regimens for Helicobacter pylori in
adults" and "Indications and diagnostic tests for Helicobacter pylori infection in
adults", section on 'Patient undergoing upper endoscopy'.)

253
Overview of complications of peptic ulcer disease
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editors:
Mark Feldman, MD, MACP, AGAF, FACG
David I Soybel, MD
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Sep 01, 2020.
INTRODUCTIONComplications of peptic ulcer disease (PUD) include bleeding,
penetration, perforation, and gastric outlet obstruction. This topic will provide an
overview of the complications of PUD and their general management. The specific
management of complicated PUD, the endoscopic management of peptic ulcer
bleeding, and the surgical approaches to complications of PUD are discussed
separately. (See "Peptic ulcer disease: Clinical manifestations and
diagnosis" and "Peptic ulcer disease: Treatment and secondary
prevention" and "Overview of the treatment of bleeding peptic ulcers" and "Surgical
management of peptic ulcer disease".)
EPIDEMIOLOGY
Incidence — The risk of complications in patients with chronic PUD is 2 to 3 percent
per year. There has been a consistent decrease in the incidence of bleeding and
perforation and hospitalization rates due to complications of PUD, presumably reflecting
the fall in Helicobacter pylori prevalence [1,2]. As an example, in the United States there
has been a 30 to 40 percent fall in hospitalizations for PUD complications between 1993
and 2006 [3]. Similar data have been reported in other countries [4]. (See "Peptic ulcer
disease: Epidemiology, etiology, and pathogenesis", section on 'Epidemiology'.)
In the United States, bleeding is the most common complication of PUD (73 percent),
followed by perforation (9 percent), and obstruction (3 percent) [3]. A large systematic
review estimated that the annual incidence of peptic ulcer hemorrhage ranges from 19
to 57 cases per 100,000 individuals, and that the annual incidence of ulcer perforation
ranges from 4 to 14 cases per 100,000 individuals [5].
Mortality — In patients with bleeding peptic ulcer, the majority of deaths are related to
multi-organ failure or cardiopulmonary causes rather than to bleeding itself [6]. Not
surprisingly, in patients with gastrointestinal bleeding requiring endoscopic therapy,
hypovolemic shock, multiple co-morbidities, and rebleeding in the hospital were
predictive of mortality [7].
In patients undergoing surgery for perforated PUD, the long-term mortality is high, with
one in three patients dying in the follow-up period. Older age, co-morbid illnesses, and
post-operative complications were predictors of mortality [8].
Penetration of an ulcer into adjacent organs is a rare complication of PUD in the era of
proton pump inhibitors [9]. Penetration occurs in descending order of frequency into the
pancreas, lesser omentum, biliary tract, liver, greater omentum, mesocolon, colon, and
vascular structures. Gastric outlet obstruction is the least frequent complication of PUD
254
in developed countries. Most cases are associated with duodenal or pyloric channel
ulceration. (See "Gastric outlet obstruction in adults", section on 'Peptic ulcer disease'.)
Risk factors for ulcer complications — Complications can occur in patients with PUD
due to any etiology. However, H. pylori infection and nonsteroidal anti-inflammatory
drugs (NSAIDs), including low-dose aspirin, are the primary causes of ulcer bleeding
and perforation [10-13]. (See "Peptic ulcer disease: Epidemiology, etiology, and
pathogenesis".)
H. pylori infection — H. pylori infection increases the risk of peptic ulcer bleeding in
NSAID and aspirin users but not in patients on anticoagulants such as warfarin. H.
pylori infection also increases the risk of bleeding in patients on non-aspirin antiplatelet
therapy and combination antiplatelet therapy with aspirin [14]. H. pylori infection has
also been associated with perforated peptic ulcer [15].
NSAID use
Drug- and dose-specific risk — The magnitude of the risk of PUD complications
associated with NSAIDs vary by the specific drug and are also dose dependent [5]. As
an example, in a study of 2777 patients, the overall relative risk (RR) of bleeding
associated with NSAID use was 5.3 (95% CI 4.5-6.2). However, the risk varied by drug
and was lowest for aceclofenac (RR 3.1, 95% CI 2.3-4.2) and was highest
for ketorolac (RR 14.4, 95% CI 5.2-39.9) [16]. The risk was higher in patients taking
high-dose NSAIDs compared with those taking medium- or low-dose NSAIDs (RR 6.8,
95% CI 5.3-8.8 versus RR 4.0, 95% CI 3.2-5.0). The risk was highest in the first 30 days
of NSAID use (RR 7.6, 95% CI 6.0-9.5). The risk remained high between days 31 and
90 days (RR 7.3, 95% CI 4.0-13.2), but dropped after 91 days (RR 2.6, 95% CI 1.6-4.1).
Use of concomitant medications — Concomitant use of anticoagulants, antiplatelet
agents, corticosteroids, and other NSAIDs, including low-dose aspirin, increase the risk
of ulcer complications [14].
Concurrent H. pylori infection — H. pylori infection is both an independent and
synergistic risk factor for bleeding PUD in patients on NSAIDs [17-19]. A meta-analysis
of nine case-control studies that assessed the prevalence of H. pylori infection and
NSAID use in patients with peptic ulcer bleeding suggested that the H. pylori infection
combined with NSAID use increases the risk of bleeding above that associated with
either risk factor alone [20]. The analysis found that individually, the odds ratios for
bleeding peptic ulcers associated with H. pylori and NSAID use were 1.8 and 4.9,
respectively, whereas the odds ratio increased to 6.1 when both H. pylori and NSAID
were present.
Other risk factors — The most important risk factor for ulcer complications in patients
on NSAIDs is a prior history of clinical ulcer disease or ulcer complications. Other risk
factors for NSAID-induced peptic ulcer complications include advanced age (>65 years)
and chronic debilitating disorders, especially cardiovascular disease [21].
Ulcer characteristics — Refractory ulcers, giant ulcers (>2 cm), and pyloric channel
ulcers are associated with higher complications rates. (See "Peptic ulcer disease:
Clinical manifestations and diagnosis" and "Approach to refractory peptic ulcer disease",
section on 'Refractory peptic ulcer'.)
CLINICAL PRESENTATIONWhile most patients may have typical ulcer
symptoms prior to the development of complications, there is a subset of patients with
silent ulcers and are only brought to medical attention due to peptic ulcer complications.
255
Bleeding — Patients with bleeding from peptic ulcers may present with hematemesis
(either red blood or coffee-ground emesis), or melena (black, tarry stool). In rare cases,
patients have massive bleeding and present with hematochezia (red or maroon blood in
the stool) and orthostatic hypotension. (See "Causes of upper gastrointestinal bleeding
in adults", section on 'Peptic ulcer disease'.)
Gastric outlet obstruction — Ulcers located in the pyloric channel or duodenum can
cause gastric outlet obstruction. Symptoms include early satiety, bloating, nausea,
vomiting, epigastric pain shortly after eating, and weight loss. Prolonged vomiting and
poor fluid intake may lead to hypokalemia and hypochloremic metabolic alkalosis.
(See "Gastric outlet obstruction in adults", section on 'Clinical manifestations'.)
Penetration — Patients with penetrating ulcers often present with a change in
symptoms due to symptomatic involvement of adjacent structures.
●Gastrocolic or duodenocolic fistulas can present with halitosis, feculent vomiting,
postprandial diarrhea, dyspepsia, and weight loss.
●Penetration into a surrounding organ can result in a perivisceral abscess.
●Erosion into vascular structures can result in gastrointestinal hemorrhage (eg,
aortoenteric fistula or erosion into the cystic artery) [22].
●Erosion into the biliary tree can result in a choledochoduodenal fistula that can
cause hemobilia and/or extrahepatic biliary obstruction.
●Fistulization into the pancreatic duct [9].
Perforation — The incidence of perforation is lower than previously reported, ranging
from 3 to 6.5 per 100,000 individuals (image 1) [23,24]. Prepyloric gastric ulcerations
account for most perforations followed by duodenal bulb ulcers [25]. Ulcer perforation
should be suspected in patients who suddenly develop severe, diffuse abdominal pain.
The classic triad of sudden onset of abdominal pain, tachycardia, and abdominal rigidity
is the hallmark of peptic ulcer perforation. If the perforation is walled off, if the gastric
fluid is confined by fibrosis, or if the perforation is retroperitoneal, symptoms may be
much less severe as compared with free intraperitoneal perforations. In such cases, the
upper abdominal pain is more insidious, the presentation often delayed, and the
abdominal examination is frequently equivocal. (See "Overview of gastrointestinal tract
perforation", section on 'Clinical features'.)
DIAGNOSISSpecific complications of PUD are often suspected based on a change
in pattern of existing symptoms or development of new symptoms. As examples, the
development of hematemesis, melena, or hematochezia suggest an ulcer bleed;
nausea, vomiting, and epigastric pain shortly after eating suggest a gastric outlet
obstruction; and sudden onset of abdominal pain, tachycardia, and abdominal rigidity
are suggestive of a perforation. The diagnosis is established by endoscopic evaluation
and/or abdominal imaging. (See 'Clinical presentation' above.)
INITIAL MANAGEMENT
General management
Supportive measures — Supportive measures include fluid resuscitation based on the
hemodynamic status, correction of associated electrolyte abnormalities, and blood
transfusions in selected patients with gastrointestinal bleeding. Patients should be kept
fasting in anticipation of endoscopic or surgical intervention. Early surgical consultation
allows for preoperative preparation should urgent surgical intervention become
necessary. (See "Approach to acute upper gastrointestinal bleeding in adults", section
256
on 'General management' and "Approach to acute upper gastrointestinal bleeding in
adults", section on 'Hemodynamically unstable patients'.)
Acid suppressive therapy — Patients with active peptic ulcer bleeding (eg,
hematemesis, hemodynamic instability) should receive initial acid suppressive therapy
with an intravenous (IV) proton pump inhibitor (PPI; eg, esomeprazole 80 mg bolus).
Typically, endoscopy is performed on these patients within 12 hours. However, if
endoscopy is delayed, a second dose of an IV PPI should be given 12 hours later (eg,
esomeprazole 40 mg). For patients who may have stopped bleeding (eg, patients who
are hemodynamically stable with melena), we give an IV PPI every 12 hours (eg,
esomeprazole 40 mg). Subsequent dosing will then depend on presence of high-risk
stigmata of recent hemorrhage on endoscopic evaluation. (See "Overview of the
treatment of bleeding peptic ulcers", section on 'Acid suppression'.)
In patients with other peptic-ulcer-related complications (gastric outlet obstruction,
penetration, perforation), high-dose twice-daily PPI treatment is reasonable to enhance
healing (eg, oral omeprazole 40 mg twice daily), but dosing should generally be reduced
to once daily after four weeks [26-28].  
The duration of treatment is based on the ulcer location and underlying etiology.
(See "Peptic ulcer disease: Treatment and secondary prevention", section on
'Duration'.)
Discontinue NSAIDs — If aspirin or non-aspirin nonsteroidal anti-inflammatory drugs
(NSAIDs) can be discontinued, even complicated ulcers readily heal and uncommonly
recur. If non-aspirin NSAIDs must be continued, the incidence of recurrent PUD can be
decreased by switching to a COX-2 inhibitor with concomitant therapy with a PPI
or misoprostol (for rare patients who are unable to take PPIs or have contraindications
to PPI use). Likewise, when continued low-dose aspirin is justified, concomitant
cotherapy with a PPI is indicated. Strategies for secondary prevention of
gastroduodenal toxicity due to NSAIDs are discussed in detail elsewhere. (See "NSAIDs
(including aspirin): Secondary prevention of gastroduodenal toxicity".)
Evaluation for H. pylori — Patients should be evaluated for H. pylori. Eradication of H.
pylori dramatically reduces recurrent ulcers and complications. H. pylori testing (eg,
biopsy urease test, urea breath test) in the setting of ulcer bleeding or PPI use may
result in false-negative results, so repeat testing is required for patients whose initial
tests are negative [10]. A urea breath test for H. pylori performed as soon as the patient
has resumed oral feedings is a reasonably sensitive predictor of H. pylori infection.
We typically defer treatment of H. pylori with oral antibiotics until patients are tolerating
oral feedings. Interrupted treatment may encourage resistance and should be avoided.
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults", section
on 'Diagnostic tests'.)
Management of specific complications
●Bleeding – Upper endoscopy is the best initial diagnostic and therapeutic
procedure in the management of bleeding peptic ulcers. Surgery and
transcatheter arteriography/intervention are generally reserved for patients with
failed therapeutic endoscopy. The management of patients with bleeding peptic
ulcers is discussed in detail separately. (See "Overview of the treatment of
bleeding peptic ulcers" and "Approach to acute upper gastrointestinal bleeding in
adults".)

257
 ●Penetrating ulcers – Management of penetrating ulcers should follow the
intensive measures outlined for refractory ulcers. (See "Approach to refractory
peptic ulcer disease".)
●Perforation – Many ulcer-related perforations of the stomach and duodenum
require surgical repair (open or laparoscopic). However, nonoperative
management may be used is selective patients. The management of gastric and
duodenal perforations and surgical repair of PUD are discussed in detail
separately. (See "Overview of gastrointestinal tract perforation", section on 'Initial
management' and "Surgical management of peptic ulcer disease", section on
'Perforated gastric ulcer'.)
●Gastric outlet obstruction – Initial management consists of acid suppression
with a parenteral PPI, avoidance of NSAIDs and, if present, eradication of H.
pylori infection. In patients who fail to respond to a brief (three to seven day) trial
of conservative management, endoscopic dilation is indicated. Surgery is reserved
for patients with a complete pyloric obstruction that cannot be safely dilated, or if
the obstruction persists or recurs despite endoscopic management. (See "Gastric
outlet obstruction in adults", section on 'Management'.)
SUBSEQUENT MANAGEMENT
Upper endoscopy — Upper endoscopy is indicated for most patients with PUD who
present with a PUD complication in order to exclude neoplasia, unless there is
confidence that the ulcer site was adequately evaluated and biopsied endoscopically or
surgically at the time of presentation. To allow ulcers to heal, upper endoscopy is
performed after 8 to 12 weeks. (See "Peptic ulcer disease: Treatment and secondary
prevention", section on 'Repeat upper endoscopy in selected patients' and "Gastric
outlet obstruction in adults", section on 'Management'.)
Prevention of recurrence — The natural history of complicated PUD is for the
recurrence of complications unless the underlying cause can be treated. Strategies to
reduce the risk of recurrence include NSAID avoidance and eradication of H. pylori. The
decision to continue maintenance antisecretory therapy varies based on the patient and
ulcer characteristics and risk factors for recurrent PUD. Strategies to decrease the risk
of recurrent PUD are discussed in detail separately. (See "Peptic ulcer disease:
Treatment and secondary prevention", section on 'Prevention of recurrence'.)
SUMMARY AND RECOMMENDATIONS
●The major complications of peptic ulcer disease (PUD) include perforation,
gastric outlet obstruction, penetration, and bleeding. The risk of complications in
patients with chronic PUD has been decreasing due to Helicobacter
pylori eradication and widespread use of proton pump inhibitors (PPIs). There has
been a decrease in the incidence of bleeding and perforation and hospitalization
rates due to complications of PUD, presumably reflecting the fall in H.
pylori prevalence. (See 'Epidemiology' above.)
●H. pylori and nonsteroidal anti-inflammatory drugs (NSAIDs), including low-
dose aspirin, are the primary causes of ulcer bleeding and perforation. Other risk
factors for ulcer complications include refractory, giant (>2 cm), and pyloric
channel ulcers. (See 'Risk factors for ulcer complications' above.)

258
●While most patients may have typical ulcer symptoms prior to the development of
complications, a subset of patients have silent ulcers and are brought to medical
attention due to peptic ulcer complications. (See 'Clinical presentation' above.)
●Specific complications of PUD are often suspected based on a change in pattern
of existing symptoms or development of new symptoms. As examples, the
development of hematemesis, melena, or hematochezia suggest an ulcer bleed;
nausea, vomiting, and epigastric pain shortly after eating suggest a gastric outlet
obstruction; sudden onset of abdominal pain, tachycardia, and abdominal rigidity
are suggestive of a perforation. Endoscopic evaluation and/or abdominal imaging
are needed to establish the diagnosis. (See 'Diagnosis' above.)
●General measures for all patients with peptic ulcer-related complications include
fluid resuscitation based on the hemodynamic status, correction of associated
electrolyte abnormalities, and high-dose intravenous PPI therapy. Blood
transfusions may be required in selected patients with gastrointestinal bleeding.
Additional measures are specific to the type of complication. (See 'Management of
specific complications' above.)
●Upper endoscopy is indicated in most patients with PUD who present with a
complication, except those with perforation, in order to exclude neoplasia. To
allow for ulcer healing, upper endoscopy is typically repeated after 8 to 12 weeks.
(See 'Upper endoscopy' above.)
●The natural history of complicated PUD is for the recurrence of complications
unless the underlying cause can be treated. Strategies to reduce the risk of
recurrence include NSAID avoidance and eradication of H. pylori. The decision to
continue maintenance antisecretory therapy varies based on the patient and ulcer
characteristics and risk factors for recurrent PUD. (See "Peptic ulcer disease:
Treatment and secondary prevention", section on 'Prevention of recurrence'.)

259
Overview of the treatment of bleeding peptic ulcers
Author:
John R Saltzman, MD, FACP, FACG, FASGE, AGAF
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Anne C Travis, MD, MSc, FACG, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Sep 14, 2021.
INTRODUCTIONUpper gastrointestinal (UGI) bleeding secondary to peptic ulcer
disease is a common medical condition that results in high patient morbidity and
medical care costs. While the majority of patients with bleeding peptic ulcers will stop
bleeding spontaneously and not rebleed during hospitalization, a subgroup of patients is
at high risk for recurrent hemorrhage and requires endoscopic therapy to decrease this
risk [1]. If endoscopic therapy fails, interventional angiography or surgery may be
required.
Despite advances in pharmacologic and endoscopic therapy, mortality rates have not
improved. In a Danish study of 13,498 patients with peptic ulcer bleeding studied
between 2004 and 2011, rates for successful endoscopic therapy were higher in 2010
to 2011 than in 2004 to 2006 (94 versus 89 percent) [2]. In addition, rebleeding rates
were lower (13 versus 18 percent). However, 30-day mortality did not improve (11
percent for both groups), though there was a trend toward decreased mortality after
adjusting for potential confounders (adjusted relative risk 0.89, 95% CI 0.78-1.00).
The pharmacologic and endoscopic management of UGI bleeding due to peptic ulcer
disease will be reviewed here. While there is variability among guidelines, the
discussion that follows is generally consistent with a multidisciplinary international
consensus statement updated in 2019, a 2012 guideline issued by the American
Society for Gastrointestinal Endoscopy, and a 2021 guideline issued by the American
College of Gastroenterology, a 2015 guideline issued by the European Society of
Gastrointestinal Endoscopy, and a 2021 update issued by the European Society of
Gastrointestinal Endoscopy [3-7].
A general approach to patients with UGI bleeding, general treatment of patients with
peptic ulcer disease, an overview of the complications of peptic ulcer disease, a detailed
discussion of the tools used for endoscopic hemostasis, and detailed discussions of
angiographic and surgical management of patients with peptic ulcer disease are
discussed separately. (See "Approach to acute upper gastrointestinal bleeding in
adults" and "Peptic ulcer disease: Treatment and secondary prevention" and "Overview
of complications of peptic ulcer disease" and "Contact thermal devices for the treatment
of bleeding peptic ulcers" and "Angiographic control of nonvariceal gastrointestinal
bleeding in adults" and "Surgical management of peptic ulcer disease".)
APPROACH TO THE PATIENTThe initial evaluation of a patient with upper
gastrointestinal (UGI) bleeding starts with assessing hemodynamic stability and
determining the need for fluid resuscitation and/or blood transfusion. This part of the

260
evaluation is discussed in detail elsewhere. (See "Approach to acute upper
gastrointestinal bleeding in adults".)
Patients with clinically significant UGI bleeding (ie, signs of active UGI bleeding
including hematemesis, melena, or hematochezia, with or without hemodynamic
instability or blood transfusion requirement) should be started on an intravenous proton
pump inhibitor while undergoing their initial evaluation. Once the patient is stabilized,
endoscopy is performed to diagnose high-risk lesions (table 1) [8]. Ulcers that are
actively bleeding and most nonbleeding ulcers that are at high risk for recurrent
bleeding based upon the presence of stigmata of recent hemorrhage require
endoscopic therapy. Ulcers that lack high-risk stigmata can be managed acutely with
acid suppression alone.
Stigmata of recent hemorrhage — Certain endoscopic findings, known as stigmata of
recent hemorrhage, are associated with an increased risk of recurrent bleeding (up to
90 percent depending upon the finding) (table 2) [1].
The appearance of ulcers can be described using the Forrest classification [9]:
●Class Ia – Spurting hemorrhage (picture 1 and movie 1)
●Class Ib – Oozing hemorrhage
●Class IIa – Nonbleeding visible vessel (picture 2A-C)
●Class IIb – Adherent clot (picture 3)
●Class IIc – Flat pigmented spot (picture 4B)
●Class III – Clean ulcer base (picture 4A)
Stigmata of recent hemorrhage are present if anything other than a clean ulcer base is
seen. However, only patients with active bleeding (spurting or oozing), a nonbleeding
visible vessel, or an adherent clot are generally considered to be at high risk for
recurrent bleeding. Most patients with high-risk stigmata require endoscopic therapy to
decrease the risk of recurrent bleeding. On the other hand, patients without these high-
risk stigmata are considered low risk and do not require endoscopic therapy.
(See 'Endoscopic therapy' below and 'Inpatient versus outpatient management' below.)
INPATIENT VERSUS OUTPATIENT MANAGEMENTHospitalization is
required for patients at high-risk for recurrent bleeding, patients with evidence of severe
upper gastrointestinal (UGI) bleeding (hemodynamic instability, blood transfusion
requirement), and patients at increased risk for complications should bleeding recur (eg,
significant coronary artery or cerebrovascular disease, age over 65 years, patients
taking antiplatelet or anticoagulant medications). Patients who are otherwise healthy
and who are at low risk for recurrent UGI bleeding may be safely allowed to eat and be
discharged from the hospital on oral antisecretory therapy once the effects of procedural
sedation have worn off, provided that the patient is reliable and can promptly get
medical care should bleeding recur. (See 'Stigmata of recent hemorrhage' above
and "Gastrointestinal endoscopy in adults: Procedural sedation administered by
endoscopists", section on 'Postprocedure care'.)
In addition to findings at the time of endoscopy, risk stratification scores can help
differentiate patients who require hospitalization from those who are appropriate for
outpatient management. It is recommended in the International Consensus
Recommendations that all patients with upper gastrointestinal bleeding undergo risk
stratification using the Blatchford score (calculator 1) [3]. (See "Approach to acute upper
gastrointestinal bleeding in adults", section on 'Risk stratification'.)
261
Early discharge and return to work can significantly reduce both direct costs (eg,
hospitalization cost) and indirect costs (eg, days lost from work) [10,11]. The safety of
early discharge for low-risk patients was demonstrated in a retrospective study with 72
patients with a bleeding duodenal ulcer [10]. All of the patients were under the age of 60
years and presented with stable vital signs. On endoscopy, they had no stigmata of
recent hemorrhage or only flat spots seen. The mean hospital stay was 1.4 days, and
there were no episodes of recurrent bleeding or significant drops in hemoglobin
concentration two weeks after discharge. The investigators then performed a
prospective study with 75 similar patients and discharged them on the same day of
endoscopy. There were no episodes of recurrent bleeding or significant drops in
hemoglobin one week after discharge.
PHARMACOLOGIC THERAPYAll patients with bleeding peptic ulcers should
receive acid suppressive treatment with a proton pump inhibitor. Our approach is to
start a high-dose intravenous (IV) proton pump inhibitor (PPI) for patients with
suspected clinically significant upper gastrointestinal bleeding prior to endoscopy as
part of their initial management (eg, esomeprazole bolus of 80 mg IV followed by a
continuous infusion of 8 mg per hour). If the patient does not have an ulcer with high-
risk stigmata at the time of endoscopy, we subsequently decrease the dose of the PPI.
This differs from the International Consensus Group guideline in that we start an IV PPI
prior to endoscopy, rather than starting a PPI after endoscopy (with the dose and route
of administration determined by the endoscopic findings and treatment provided).
(See "Approach to acute upper gastrointestinal bleeding in adults", section on 'Acid
suppression'.)
Acid suppression — Treatment with PPIs leads to elevation of gastric pH levels, which
stabilizes blood clots and improves clinical outcomes [12-22]. As a result, PPIs are
recommended for all patients with peptic ulcer bleeding. The optimal approach to PPI
administration prior to endoscopy is unclear. Options include giving an IV PPI every 12
hours or starting a continuous infusion. Our approach is to give a high-dose bolus
(eg, esomeprazole 80 mg) to patients with signs of active bleeding (eg, hematemesis,
hemodynamic instability). Typically, we try to perform endoscopy on these patients
within 12 hours. However, if endoscopy is delayed, a second dose of an IV PPI should
be given 12 hours later (eg, esomeprazole 40 mg). For patients who may have stopped
bleeding (eg, patients who are hemodynamically stable with melena), we give an IV PPI
every 12 hours (eg, esomeprazole 40 mg). Subsequent dosing will then depend on the
endoscopic findings. Oral formulations (eg, esomeprazole 40 mg orally twice daily) are
a reasonable alternative if IV formulations are not available. Pantoprazole and
esomeprazole are the only intravenous formulations available in the United States, and
intravenous lansoprazole has been removed from the world market [3,5,23,24].
(See 'High-dose continuous infusions' below and 'Oral versus intravenous
dosing' below.)
In patients with high-risk stigmata of recent hemorrhage, a high-dose continuous
infusion of a PPI (eg, esomeprazole 8 mg per hour) should be started and continued for
72 hours. If the patient did not receive an IV PPI prior to endoscopy, the patient should
receive a high-dose bolus of PPI (eg, esomeprazole 80 mg) prior to starting the
continuous infusion. The high-dose IV PPI may be switched to a high-dose oral PPI
(eg, omeprazole 40 mg twice daily) 72 hours after endoscopy, provided there is no
262
evidence of recurrent bleeding. (See 'Treatment of persistent and recurrent
bleeding' below.) A randomized trial suggested that twice-daily oral dosing may be
preferable in patients at high risk of rebleeding. In the trial, patients with a Rockall score
≥6 (calculator 2) who received a twice–daily oral PPI had a lower rebleeding rate than
those who received a once-daily oral PPI (11 versus 29 percent at 28 days) [25].
It is unlikely that an IV PPI would be of significant benefit in patients who do not have
active bleeding or other high-risk stigmata (such as a visible vessel or adherent clots)
because their risk of recurrent bleeding is low. Such patients may be switched to a
standard-dose oral PPI (eg, omeprazole 20 mg daily) immediately following endoscopy.
The goal of treatment in low-risk patients should be directed at healing the ulcers and
eliminating precipitating factors (such as Helicobacter pylori and nonsteroidal anti-
inflammatory drugs). (See "Peptic ulcer disease: Treatment and secondary prevention",
section on 'Initial management'.)
High-dose oral PPIs in patients with high-risk stigmata of recent hemorrhage are
continued for a total of two weeks and then changed to a once daily dose. (See "Peptic
ulcer disease: Treatment and secondary prevention", section on 'Initial antisecretory
therapy'.)
Studies of H2 receptor antagonists (H2RAs) in bleeding peptic ulcers have produced
mixed but generally disappointing results and are not recommended for patients with
acute upper gastrointestinal bleeding [14,26-29]. A meta-analysis concluded that there
was a possible minor benefit with IV H2RAs in bleeding gastric ulcers but no benefit
with duodenal ulcers [28]. The improved efficacy seen with PPIs may be due to their
superior ability to maintain a gastric pH at a level above 6.0.
Efficacy of proton pump inhibitors — A meta-analysis of 21 randomized trials that
compared PPIs with either placebo or an H2RA for bleeding ulcers (with or without
endoscopic therapy) found a significant and consistent reduction in the risk of recurrent
bleeding (odds ratio [OR] 0.46, 95% CI 0.33-0.64) and the need for surgery (OR 0.59,
95% CI 0.46-0.76) with PPIs [12]. However, there was no effect on mortality.
●An illustrative trial included 240 patients with ulcers that were actively bleeding or
had a visible vessel. The patients were randomly assigned to IV omeprazole or
placebo following endoscopic hemostasis [16]. Recurrent bleeding was observed
significantly less often in patients receiving omeprazole (7 versus 23 percent), a
finding that led to early termination of the trial. A separate analysis based upon
these data suggested that the addition of IV omeprazole was cost-effective
compared with placebo [30].
●A second randomized trial (published after the meta-analysis [12]) included 638
patients admitted with upper gastrointestinal bleeding [21]. Patients were assigned
to either IV omeprazole or placebo at the time of admission. Patients then
underwent endoscopy the following morning. Patients who received omeprazole
were less likely than those who received placebo to have endoscopic signs of
active bleeding (6 versus 15 percent) or to require endoscopic hemostatic therapy
(19 versus 28 percent).
High-dose continuous infusions — PPIs given for the treatment of bleeding peptic
ulcers are often given as high-dose continuous infusions
(eg, esomeprazole or pantoprazole 80 mg bolus followed by 8 mg per hour). High-dose
continuous PPI infusions decrease rebleeding and mortality rates compared with

263
H2RAs or placebo, whereas non-high dose PPIs decrease rebleeding rates compared
with H2RAs or placebo (but have not been proven to decrease mortality rates) [3]. While
direct comparisons of high-dose continuous PPI infusions with non-high dose PPIs have
not shown a difference in the risk of rebleeding or mortality, the International Consensus
Group guideline favors high-dose continuous PPI infusions in patients with bleeding
ulcers with high-risk stigmata who have undergone successful endoscopic therapy,
since indirect comparisons suggest that high-dose continuous PPI infusions may be
superior to non-high dose PPIs when it comes to mortality.
Oral versus intravenous dosing — Oral dosing of PPIs is less expensive than IV
dosing, and some studies suggest it may be an option for the treatment of patients with
peptic ulcer bleeding [31,32]. However, where available, IV dosing is currently
considered standard of care. If oral dosing is being considered, a high-dose of an oral
PPI should probably be used (eg, omeprazole, pantoprazole, or esomeprazole 40 mg
twice per day).

Studies comparing oral dosing with placebo or an H2RA have shown the following:

●A combined analysis of five trials evaluating oral dosing (with or without

endoscopic therapy) found a significant reduction in the risk of recurrent bleeding
and surgery compared with treatment with placebo or an H2RA [12]. However,
oral dosing was not compared with high-dose IV dosing.
●A randomized trial conducted in 220 patients in a setting where therapeutic
endoscopy was not available found that high-dose oral omeprazole (40 mg twice
daily) was associated with a decreased risk of recurrent bleeding compared with
placebo in patients who had ulcers with visible vessels or adherent clots who
did not undergo endoscopic therapy (11 versus 36 percent) [15].
●Another randomized trial included 160 patients who were at high risk for recurrent
bleeding and had been treated with endoscopic injection therapy [33]. The
patients who were assigned to receive oral omeprazole had a significantly lower
rebleeding rate than those who received placebo (12 versus 26 percent).
Whether there is a difference in outcomes with oral and IV PPI dosing was examined in
a meta-analysis of six randomized trials with 615 patients who underwent endoscopic
treatment for peptic ulcer bleeding [32]. Four of the trials used high-dose IV PPIs (bolus
followed by continuous infusion), and two used lower doses (40 mg
of esomeprazole or omeprazole twice daily). There was no difference between those
who received oral PPIs and those who received IV PPIs with regard to recurrent
bleeding, mean volume of blood transfused, need for surgery, or all-cause mortality.
Patients treated with oral PPIs had a shorter duration of hospital stay (-0.7 days). When
the subset of studies that used high-dose IV PPIs was examined, the results were
similar, though the trend toward decreased length of hospital stay with oral PPIs was no
longer statistically significant.
In a subsequent randomized trial with 244 patients who received endoscopic therapy for
bleeding peptic ulcers, patients were randomly assigned to IV esomeprazole (80 mg
bolus followed by 8 mg/hour infusion) or oral esomeprazole (40 mg orally every 12
hours) [34]. Outcomes were similar between those who received IV esomeprazole and
those who received oral esomeprazole with regard to recurrent bleeding within 30 days
(8 versus 6 percent), blood transfusions (median 2 versus 1 units), need for repeat
264
endoscopic therapy (1.7 versus 2.4 percent), and length of hospital stay (median 4.0
days for both).
High doses of PPIs given orally achieve adequate acid suppression more rapidly than
standard doses (eg, omeprazole 20 mg daily), which may take several days to achieve
adequate acid suppression [35]. In addition, high-dose IV PPIs achieve adequate acid
suppression more rapidly than high-dose oral PPIs. This was demonstrated in a
randomized trial that compared intravenous lansoprazole (90 mg bolus followed by a 9
mg/hour infusion) with oral lansoprazole (120 mg followed by 30 mg every three hours)
in patients with bleeding ulcers [36]. Achievement of an intragastric pH >6 was similar
for IV and oral lansoprazole (68 versus 64 percent), but the group that received IV
lansoprazole achieved a more rapid increase in pH, reaching a mean pH of 6
approximately one hour sooner than the group that received an oral PPI (after two to
three hours compared with three to four hours). Whether such frequent dosing of the
oral PPI was necessary to achieve comparable clinical outcomes is unclear, and we
suggest that if a high-dose oral PPI is going to be used, that it be given twice daily.
(See "Proton pump inhibitors: Overview of use and adverse effects in the treatment of
acid related disorders".)
Somatostatin and octreotide — Somatostatin and its long-acting
analogue octreotide (commonly used in the management of variceal bleeding) have a
theoretical benefit in bleeding ulcer disease because they reduce splanchnic blood flow,
inhibit gastric acid secretion, and may have gastric cytoprotective effects [37,38]. A
clinical benefit has been described in ulcer bleeding, but because of the effectiveness of
PPI and endoscopic therapy, its role is generally limited to settings in which endoscopy
is unavailable or as a means to help stabilize patients before definitive therapy can be
performed.
Several studies have described the experience with these agents in patients with acute
nonvariceal upper gastrointestinal bleeding. A meta-analysis of these trials suggested
that somatostatin was associated with a reduced risk of continued bleeding (relative risk
0.53; 95% CI, 0.43 to 0.63) [39]. The risk also appeared to be reduced with octreotide,
although there were fewer studies. Thus, somatostatin or octreotide can be used as
adjunctive therapy before endoscopy, or when endoscopy is unsuccessful,
contraindicated, or unavailable [40]. The doses used in the above studies were variable;
a typical dose of somatostatin was 250 mcg given as a bolus followed by an infusion of
250 mcg per hour for three to seven days, while a typical dose of octreotide was 50 to
100 mcg given as a bolus followed by an infusion of 25 mcg per hour for up to three
days. Of note, the dose of octreotide used is lower than that used for variceal bleeding
(50 mcg bolus followed by an infusion of 50 mcg per hour), so if there is concern prior to
the endoscopy that the bleed could be due to varices, the higher dose of octreotide
should be used. (See "Methods to achieve hemostasis in patients with acute variceal
hemorrhage", section on 'Somatostatin and its analogs'.)
Prokinetic agents — We suggest that erythromycin be used before endoscopy. A
reasonable dose is 250 mg intravenously over 20 to 30 minutes, 30 to 90 minutes prior
to endoscopy. Patients receiving erythromycin need to be monitored for QTc
prolongation. In addition, drug-drug interactions should be evaluated before giving
erythromycin because it is a cytochrome P450 3A inhibitor (table 3). (See "Approach to
acute upper gastrointestinal bleeding in adults", section on 'Prokinetics'.)

265
ENDOSCOPIC THERAPYEndoscopic therapy is indicated for the treatment of
most ulcers with stigmata of recent hemorrhage that increase the risk of recurrent
bleeding. With appropriate treatment, high-risk lesions have recurrent bleeding rates of
5 to 20 percent, depending upon the endoscopic appearance of the ulcer. On the other
hand, ulcers with a clean base or a flat pigmented spot are at low risk of recurrent
bleeding (3 to 5 percent for clean-based ulcers and 7 to 10 percent for ulcers with a flat
spot) and should not be treated endoscopically (picture 4A-B) [1]. (See "Contact thermal
devices for the treatment of bleeding peptic ulcers".)
Adherent clots that are not easily removed endoscopically (eg, with irrigation or gentle
suctioning of the clot away from the ulcer crater to reveal the underlying stigmata) carry
a 20 to 30 percent risk of recurrent bleeding. A traditional dictum has been to leave
these clots in situ and manage patients medically. More recent experience suggests
that removing the clot (eg, using a cold guillotine technique with a polypectomy snare)
and then treating the underlying ulcer stigmata can significantly reduce the risk of
recurrent bleeding [41,42]. However, meta-analyses have reached variable results. One
meta-analysis of six studies with a total of 240 patients concluded that endoscopic
therapy was associated with a significantly lower rate of recurrent hemorrhage in
patients with bleeding peptic ulcers and adherent clots (8 versus 25 percent with
medical therapy alone) [43]. On the other hand, a later meta-analysis found no clear
evidence supporting specific endoscopic interventions in patients with an adherent clot
[44].
Given the significant risk of recurrent bleeding in patients with adherent clots, we
suggest that gentle attempts be made to remove the clot so that endoscopic treatment
can be applied if needed, provided the ulcer is in a location that is amenable to
endoscopic therapy should bleeding start, that additional surgical and/or interventional
radiology support is available, and that the endoscopist performing the procedure is
comfortable with the techniques involved with clot removal. If these conditions are not
met and the patient is not actively bleeding, high dose IV PPI treatment alone is a
reasonable alternative. (See "Contact thermal devices for the treatment of bleeding
peptic ulcers", section on 'Nonbleeding adherent clots'.)
Choice of endoscopic treatment — Although several types of endoscopic treatment
for bleeding peptic ulcers have been described, including injection therapy, thermal
coagulation, hemostatic clips, fibrin sealant (or glue), argon plasma coagulation, and
combination therapy (typically injection of epinephrine combined with another treatment
modality), relatively few prospective comparative trials have been performed. Currently,
most patients are treated with either thermal coagulation therapy or hemostatic clips,
with or without the addition of injection therapy. This approach is based upon results
from meta-analyses of randomized trials comparing different forms of treatment to
control bleeding [44-46].
One analysis included 74 randomized trials that compared endoscopic methods to
control bleeding in patients considered to be at high risk for recurrent ulcer bleeding (ie,
those with active bleeding, a nonbleeding visible vessel, or an adherent clot) [44]. The
primary endpoint was persistent or recurrent bleeding. The following were the major
conclusions:

266
 ●Compared with epinephrine monotherapy, the risk of further bleeding was
significantly lower with other monotherapies, such as thermal coagulation (relative
risk [RR] 0.6) or epinephrine followed by another modality (RR 0.3).
●Endoscopic clips were more effective than epinephrine alone for preventing
recurrent bleeding (RR 0.2) but were not different than other endoscopic
therapies.
●The efficacy of endoscopic therapies for adherent clots was uncertain.
The choice between endoscopic clips and thermal therapy will often depend upon
factors such as the location of the ulcer and the preference of the endoscopist. Detailed
discussions of the endoscopic techniques used for hemostasis are presented
elsewhere. (See "Contact thermal devices for the treatment of bleeding peptic
ulcers" and "Endoscopic clip therapy in the gastrointestinal tract: Bleeding lesions and
beyond".)
Standard approaches — Standard approaches to treatment include thermal
coagulation and endoscopic clip placement. In addition, both of these modalities can be
combined with injection therapy, an approach known as combination therapy (picture 5).
Injection therapy — Injection therapy should be used in conjunction with other forms of
therapy, such as thermal coagulation or endoscopic clip placement. Injection therapy
should not be used as monotherapy because it is associated with higher rates of
recurrent bleeding than treatment with thermal coagulation, endoscopic clip placement,
or combination therapy [44,47,48].
Injection therapy with dilute epinephrine results in local tamponade and vasospasm
[49,50]. The technique is inexpensive and effective for temporary hemostasis [51-55].
Epinephrine diluted with saline to 1:10,000 to 1:20,000 is injected in 0.5 to 2.0 mL
aliquots in four quadrants within 3 mm of the bleeding site. In patients who are at
increased risk of having an adverse event with epinephrine injection, such as those with
significant cardiac disease or those with lesions close to the esophagogastric junction
(epinephrine injected into this area may enter the systemic circulation without a first
pass through the liver), a dilution of 1:100,000 can be used.
A potential advantage of epinephrine injection is that it is easy to administer, can help
slow or stop bleeding, and can reduce bleeding after attempts at mechanical
hemostasis. During active bleeding, it can also produce a cleaner field, permitting
targeted treatment of the bleeding site. Addition of a sclerosant (eg, ethanolamine)
confers no advantage over injection with epinephrine alone and may cause tissue
damage [53,55]. While no longer commonly performed, injection of absolute ethanol is
another treatment option that is inexpensive and easy to perform [52,54,56,57].
Injection of saline also causes local tamponade, which can be effective in achieving
temporary hemostasis. However, as was seen in studies
comparing epinephrine monotherapy with other therapeutic modalities, saline injection
alone is associated with higher recurrent bleeding rates compared with other standard
therapies. In a randomized trial with 100 patients with high-risk bleeding ulcers,
recurrent bleeding occurred more often with saline monotherapy compared with thermal
coagulation (29 versus 12 percent) [58].
Thermal coagulation — Thermal coagulation with contact probes achieves acute
hemostasis and prevents recurrent bleeding by coaptive coagulation of the underlying
artery in the ulcer base (picture 2B and picture 6) [51]. Coaptive coagulation involves

267
applying pressure to the vessel with the probe to compress it while coagulation is
performed. This results in sealing (coaptation) of vessel. A detailed discussion of
thermal coagulation is presented elsewhere. (See "Contact thermal devices for the
treatment of bleeding peptic ulcers", section on 'Application of thermal coagulation'.)
An alternate form of thermal coagulation uses argon plasma coagulation (APC). This
approach has a theoretical disadvantage for the treatment of bleeding ulcers since it is
not coaptive. Nevertheless, at least two randomized trials suggested that APC can be
effective. (See "Argon plasma coagulation in the management of gastrointestinal
hemorrhage".)
Endoscopic clips — The endoscopic application of endoscopic clips is an alternative
to thermal coagulation. Once applied, the clips achieve hemostasis in a manner similar
to surgical ligation. Placement of an endoscopic clip can be of value even if an ulcer is
not amenable to endoscopic therapy, since it may serve as a radiopaque marker for
subsequent interventional angiographic or surgical intervention. The use of endoscopic
clips in the treatment of upper gastrointestinal bleeding is discussed elsewhere.
(See "Endoscopic clip therapy in the gastrointestinal tract: Bleeding lesions and
beyond".)
Alternative approaches — Alternatives to standard endoscopic therapy that are being
studied include the use of tissue adhesives and agents that promote hemostasis.
Fibrin sealant — One approach involves the use of endoscopically injected fibrin
sealant to achieve initial hemostasis and decrease the rate of recurrent bleeding. An
unblinded, multicenter, randomized trial of 854 patients with actively bleeding
gastroduodenal ulcers compared the safety and efficacy of a single application of fibrin
sealant, daily repeated doses of fibrin sealant until the visible vessel disappeared, or a
single application of the sclerosant polidocanol [59]. While the safety profiles of all three
treatment strategies were similar, the patients who received multiple applications of
fibrin sealant had significantly less recurrent bleeding than the polidocanol group (15
versus 23 percent) and had fewer acute treatment failures (8 versus 13 percent).
However, the resources necessary to perform repeated endoscopies are not
inconsequential, and the precise role for fibrin sealant in this setting remains to be
defined. (See "Fibrin sealants".)
Hemostatic sprays — Hemostatic sprays can be used to control active GI bleeding in a
variety of contexts, particularly when traditional endoscopic techniques fail to control
massive GI bleeding. The 2019 consensus guidelines recommend hemostatic sprays as
a temporizing measure to stop bleeding when conventional modalities to control
bleeding fail or are not available [3]. Hemostatic spray may be deployed in patients with
bleeding from a GI malignancy which may be from multiple sites and is difficult to
control with standard techniques [60].

Hemostatic spray may be delivered by physicians with endoscopic skills, but who lack
training in hemostasis, as a temporizing measure to stabilize the patient until additional
expertise is available or the patient is transported to a different facility. Lesions at high-
risk of rebleeding such as actively bleeding peptic ulcers should undergo additional
endoscopic treatment, such as endoscopic clips to prevent rebleeding.

268
Hemostatic sprays include a nanopowder that promotes hemostasis (Hemospray) and a
starch-derived spray composed of absorbable hemostatic polysaccharides (EndoClot
PHS [Polysaccharide Hemostatic System]).

●Hemostatic nanopowder – Hemostatic nanopowder becomes cohesive and

adhesive when it comes into contact with moisture, forming a stable mechanical
barrier at the site of bleeding. It also has been shown to enhance clot formation
and shorten coagulation time [61]. The powder is delivered through a catheter and
sprayed onto the bleeding site under endoscopic guidance, without the need for
direct tissue contact. Potential advantages of this hemostatic approach are that it
is easy to apply, does not require an en face view of the ulcer, and does not
require direct tissue contact. Hemostatic nanopowder has been used to treat a
wide spectrum of gastrointestinal bleeding sources, including ulcers, tumors,
iatrogenic lesions (eg, postpolypectomy bleeding), and varices [62-67]. Reported
success rates for achieving initial hemostasis in patients with nonvariceal upper
gastrointestinal bleeding are 75 to 100 percent, with rebleeding rates of 9 to 49
percent [62-65,68-73].
This treatment modality is available in Canada and Europe and was FDA
approved in the United States in 2018 [74]. In the data submitted to the FDA, there
were 750 patients treated with hemostatic nanopowder. Hemostasis on index
endoscopy was reported as being achieved in 97.8 percent. There was an overall
rebleeding rate of 10.2 percent [75]. In a subsequent prospective multicenter
international registry study that looked at outcomes in 202 patients with bleeding
peptic ulcers, the primary hemostasis rate with hemostatic nanopowder was 88
percent and rebleeding occurred in 17 percent [72]. Primary hemostasis and
rebleeding outcomes were similar among those who received hemostatic
nanopowder as monotherapy, as part of combination therapy, or as rescue
therapy. However, 30-day all-cause mortality was lower in those who received
combination therapy compared with those who received monotherapy (adjusted
RR 0.51, 95% CI 0.28-0.93, p <0.001).
●Starch-derived hemostatic spray – A starch-derived hemostatic spray composed
of absorbable hemostatic polysaccharides (EndoClot PHS [Polysaccharide
Hemostatic System]) was FDA approved in January 2021 for use in the United
States. The adhesive hemostatic polymer forms a gel adhesive matrix that
adheres to a bleeding site and promotes the clotting cascade. Application of the
powder through the endoscope requires a delivery system that consists of a
delivery catheter, a powder/air mixing chamber connected to a powder dispenser
and an air compressor. This hemostatic spray is FDA approved for hemostasis of
nonvariceal gastrointestinal bleeding, excluding Forrest Ia (spurting) classification
of bleeding.
Data on the starch-derived hemostatic spray are more limited than for the
nanopowder spray. In a prospective observational trial that included 58 patients
with acute upper GI bleeding, treatment success was achieved in 64 percent when
used for primary bleeding control [76]. Another trial prospectively compared 40
patients with upper GI bleeding treated with the starch-derived hemostatic spray
with 303 patients treated with conventional therapies using propensity matching
[77]. It found similar effectiveness of the starch-derived hemostatic spray
269
 compared to conventional therapy. There is also a comparative study of
nanopowder (n=102 patients) versus starch-derived (n=25 patients) hemostatic
sprays that found short-term success at achieving primary hemostasis of 82
percent versus 69 percent, respectively [78].
Second-look endoscopy — Second-look endoscopy refers to the practice of
performing a planned follow-up endoscopy, generally within 24 hours of the initial
endoscopy. If there is active bleeding or a nonbleeding visible vessel, endoscopic
therapy is performed. (See 'Treatment of persistent and recurrent bleeding' below.) Data
are conflicting regarding the benefits of second-look endoscopy, and in general,
guidelines and reviews do not recommend it [23,79-81].
One problem with the available studies is that many were performed without optimal
hemostatic approaches (eg, treatment with epinephrine monotherapy rather than
combination therapy) and did not include the use of proton pump inhibitors. A meta-
analysis that included five randomized trials with 998 patients found that second-look
endoscopy with thermal coagulation was associated with a decreased risk of rebleeding
(relative risk [RR] 0.29; 95% CI 0.11-0.73), but second-look endoscopy with injection
therapy was not (RR 0.85; 95% CI 0.63-1.14). [82]. In addition, second-look endoscopy
was not associated with a decreased risk of surgery or mortality.
A subsequent meta-analysis with eight randomized trials with 938 patients found that
compared with standard medical care, routine second-look endoscopy was associated
with a decreased risk of rebleeding (odds ratio [OR] 0.55; 95% CI 0.37-0.81) and
surgery (OR 0.43; 95% CI 0.19-0.96), but not mortality (OR 0.65; 95% CI 0.26-1.62)
[83]. However, only one of the included studies treated patients with a high-dose proton
pump inhibitor, and three of the studies did not use cautery or endoscopic clipping as
part of the endoscopic treatment. Finally, when two trials that included patients at high
risk for rebleeding were removed, the benefit with second-look endoscopy was no
longer seen.
The 2010 International Consensus Recommendations for the management of patients
with nonvariceal upper gastrointestinal bleeding do not recommend routine use of
second-look endoscopy [23]. However, the guidelines also suggest that patients at
particularly high risk for recurrent bleeding may benefit from second-look endoscopy.

Situations that might warrant a second-look endoscopy include:

●If visualization during the initial endoscopy was limited by blood or debris
●If there is concern on the part of the endoscopist that the prior endoscopic
therapy was suboptimal
In addition, repeat endoscopy is indicated if there is recurrent bleeding in order to
exclude previously missed lesions and/or to retreat the bleeding ulcer. (See 'Treatment
of persistent and recurrent bleeding' below.)
Treatment-related complications — Complications can arise prior to, during, or after
emergency endoscopy [84]. Complications that may occur before endoscopy include
aspiration (especially in a sedated, combative, or encephalopathic patient),
hypoventilation (related to oversedation), or hypotension (due to inadequate volume
replacement or transfusions, in addition to sedation with opiates). (See "Overview of
upper gastrointestinal endoscopy (esophagogastroduodenoscopy)".)

270
Complications of endoscopic therapy include perforation and precipitation (or
worsening) of bleeding. In addition, epinephrine can cause tachycardia and arrhythmias.
Aggressive initial treatment or repeated applications of thermal or injection therapy may
improve hemostasis, but also increase the risk of treatment-induced complications.
Thus, predetermined limits (volume of injection solution, total treatment pulses, and total
energy delivered) should be set and not exceeded for each technique, in order to
minimize the risk of complications. We usually do not exceed 20 mL of epinephrine
(1:10,000 dilution) injection or more than six pulses of thermal coagulation. In addition,
in patients at increased risk for cardiac complications from epinephrine, one option is to
further dilute the epinephrine to 1:100,000.
TREATMENT OF PERSISTENT AND RECURRENT BLEEDING Although
the majority of bleeding ulcers can be controlled endoscopically, some patients have
persistent or recurrent bleeding [85,86]. Persistent bleeding refers to active bleeding
that does not stop despite endoscopic therapy or bleeding that develops during
endoscopic therapy of a nonbleeding lesion that cannot be controlled endoscopically.
Recurrent bleeding refers to bleeding that occurs following spontaneous hemostasis or
after successful endoscopic hemostasis.
The following criteria have been proposed for defining recurrent bleeding in studies [87]:
●Hematemesis or bloody nasogastric aspirate more than six hours after
endoscopy
●Melena after normalization of stool color
●Hematochezia after normalization of stool color or after melena
●Development of tachycardia (heart rate ≥110 beats per minute) or hypotension
(systolic blood pressure ≤90 mmHg) after at least one hour of hemodynamic
stability (ie, no tachycardia or hypotension) in the absence of an alternative
explanation for hemodynamic instability, such as sepsis, cardiogenic shock, or
medications (of note, many endoscopists, ourselves included, consider
tachycardia to be present if the heart rate is greater than 100 beats per minute)
●Hemoglobin drop of 2 g/dL or more after two consecutive stable hemoglobin
values (less than a 0.5 g/dL decrease) obtained at least three hours apart
●Tachycardia or hypotension that does not resolve within eight hours after index
endoscopy despite appropriate resuscitation (in the absence of an alternative
explanation), associated with persistent melena or hematochezia
●Persistently dropping hemoglobin of more than 3 g/dL in 24 hours, associated
with persistent melena or hematochezia

Patients with one episode of recurrent bleeding following initially successful endoscopic
therapy are typically treated with a second attempt at endoscopic therapy. Therapy may
consist of the same therapy initially used or a different endoscopic modality (eg, if
thermocoagulation therapy was used initially it may either be repeated or treatment with
a hemostatic clip employed). If the bleeding was initially controlled with endoscopic
clips, treating with thermal coagulation is an option, even if it is needed next to a clip or
if there might be contact with the clip during treatment (however, thermal coagulation
should not be applied to the clip intentionally with the goal of heating it up). Surgery or
angiography-guided intervention may be indicated for patients who fail endoscopic
therapy (persistent bleeding or recurrent bleeding after two therapeutic endoscopies).

271
This approach was supported by a randomized trial that included 92 patients with
recurrent bleeding after endoscopic therapy who were randomly assigned to repeat
endoscopic therapy or surgery [88]. Long-term control of bleeding was achieved in 35 of
48 patients (73 percent) treated with repeated endoscopic therapy. Salvage surgery
was performed in the 13 patients who failed endoscopic therapy (11 because of
continued bleeding and 2 because of perforation resulting from thermocoagulation).
Mortality was similar between the two groups, though complications were less frequent
in the endoscopy group (15 versus 36 percent).
Another treatment option in patients with rebleeding is the use of an over-the-scope clip
(OTSC). In a multicenter randomized trial, 66 patients with recurrent peptic ulcer
bleeding were randomized to treatment with an OTSC or standard therapy [89].
Persistent bleeding was seen in 19 patients (58 percent) who had standard therapy
versus 5 patients (15 percent) treated with an OTSC (p=0.001). In 14 patients
randomized to the conventional therapy arm, crossover to OTSC occurred because of
persistent bleeding (10 patients) or recurrent bleeding (four patients) and was
successful in all patients. The use of an OTSC to treat recurrent peptic ulcer bleeding is
a promising therapeutic option. (See "Endoscopic clip therapy in the gastrointestinal
tract: Bleeding lesions and beyond".)
Risk factors for persistent or recurrent bleeding — Factors associated with
persistent or recurrent bleeding include, ulcer location, the patient's presentation, the
appearance of the ulcer at the time of endoscopy, size of the ulcer, and presence of
comorbid illnesses.
In one report, gastric ulcers along the lesser curvature and duodenal ulcers along the
posterior wall of the duodenal bulb were at greater risk for severe or recurrent bleeding
compared with ulcers in other locations because of their proximity to large underlying
arteries (left gastric and posterior gastroduodenal arteries, respectively) [85]. In addition,
patients who presented with active hemorrhage, shock, and the lowest hemoglobin
concentrations did less well than those without these risk factors. Factors that
did not predict the outcome of endoscopic therapy were a history of nonsteroidal anti-
inflammatory drug or aspirin use, coagulopathy, previous peptic ulceration, and
concomitant cardiorespiratory disease. In another report, severe bleeding, active
bleeding, fresh blood in the stomach, and large ulcers (2 cm in diameter or larger) were
independent risk factors for therapeutic failure after injection of epinephrine plus heater
probe treatment [86]. Large ulcers tend to have larger blood vessels (>2 to 3 mm), and
ulcers in the posterior duodenal bulb also may have a large vessel (gastroduodenal
artery). These large vessels may be difficult or not possible to control with endoscopic
therapies, and we advise the use of endoscopic combination therapies as well as a low
threshold to advance to other therapies, including angiography and surgery.
A meta-analysis of 10 studies found that predictors of recurrent bleeding included active
bleeding at endoscopy, large ulcers (>1 to 2 cm), posterior duodenal ulcers, and gastric
ulcers on the lesser curvature [90].
End-stage kidney disease may also increase the risk of recurrent bleeding. A case-
control study of 150 patients found that patients with end-stage kidney disease requiring
dialysis were more likely to experience recurrent bleeding than were patients with
chronic kidney disease who were not on dialysis (OR 3.8) or normal controls (OR 3.8)
[91].

272
Surgery — Traditionally, surgery was required for patients who failed endoscopic
therapy, though depending upon local resources and expertise, many patients now
undergo an attempt at interventional angiography prior to surgery. Surgical treatments
for peptic ulcer disease include oversewing of the artery with truncal vagotomy and
pyloroplasty, antrectomy with gastrojejunostomy (Billroth II procedure), and highly
selective vagotomy. Emergency surgery for bleeding peptic ulcer disease involves
oversewing of the ulcer (to ligate the bleeding artery) plus truncal vagotomy (to
decrease acid secretion) and pyloroplasty (for gastric drainage). More time consuming
procedures, such as highly selective vagotomy, can be performed either at standard
laparotomy or laparoscopically for non-emergency ulcer surgery [92]. (See "Surgical
management of peptic ulcer disease", section on 'Bleeding duodenal
ulcer' and "Surgical management of peptic ulcer disease", section on 'Bleeding gastric
ulcer'.)

In addition to failure of endoscopic therapy, other indications for surgery for peptic ulcer
hemorrhage include:

●Hemodynamic instability despite vigorous resuscitation (more than a three unit

transfusion)
●Shock associated with recurrent hemorrhage
●Perforation
Secondary or relative indications include rare blood type, difficult crossmatch, refusal of
transfusion, shock on presentation, advanced age, severe comorbid disease, and
chronic gastric ulcer as the origin of hemorrhage. In addition, surgery may be
appropriate for older adult patients who are unlikely to tolerate prolonged attempts at
resuscitation, large volume transfusions, or periods of hypotension [93]. (See "Surgical
management of peptic ulcer disease", section on 'Natural history of peptic ulcer
disease'.)
If performed emergently, surgery is associated with high mortality rates (up to 36
percent) [94]. On the other hand, early elective surgery is associated with a much lower
mortality rate (0 to 7 percent). Recurrent bleeding rates following surgery vary from 3 to
23 percent [94,95].
Interventional angiography — Angiography with transarterial embolization (TAE) for
persistent or recurrent peptic ulcer bleeding is a less invasive alternative to surgery.
Initial success rates for patients with acute peptic ulcer bleeding are between 52 and 98
percent, with recurrent bleeding rates of 10 to 20 percent [96]. (See "Angiographic
control of nonvariceal gastrointestinal bleeding in adults".)
Indications for interventional angiography for acute nonvariceal upper gastrointestinal
bleeding have been suggested in a consensus statement from the American College of
Radiology [97]:
●Endoscopy is the best initial diagnostic and therapeutic procedure.
●Surgery and transcatheter arteriography/intervention are equally effective
following failed therapeutic endoscopy, but transcatheter
arteriography/intervention should be considered, particularly in patients at high risk
for surgery.
●Transcatheter arteriography/intervention is less likely to be successful in patients
with impaired coagulation.
273
 ●Transcatheter arteriography/intervention is the best technique for treatment of
bleeding into the biliary tree or pancreatic duct.

TAE has been compared with surgical therapy for patients with peptic ulcer bleeding
that could not be controlled endoscopically. In a metaanalysis of 13 observational
studies with 1077 patients, 427 underwent TAE and 650 underwent surgery. There was
a trend toward reduced mortality rates with TAE (22 versus 25 percent, odds ratio [OR]
0.77; 95% CI 0.50-1.18) and complication rates were lower (31 versus 50 percent, OR
0.45; 95% CI 0.30-0.47). Rebleeding was more common with TAE (35 versus 18
percent, OR 2.44; 95% CI 1.77-3.36), as was the need for further intervention (33
versus 18 percent, OR 2.13; 95% CI 1.21-3.77). A subsequent study with 282 patients
found a decreased risk of death with TAE (hazard ration [HR] 0.66; 95% CI 0.46-0.96),
an increased risk of rebleeding (HR 2.48; 95% CI 1.33-4.62), and a lower risk of
complications (8.3 versus 32.2 percent, p <0.0001).

Given the less invasive nature of angiography, we suggest that patients who have failed
endoscopic therapy first undergo attempted angiography with TAE, and that surgery be
reserved for those who fail angiographic therapy. However, surgery is a reasonable
alternative if an interventional radiologist with expertise in TAE is not available, if the
lesion is deemed unlikely to respond to angiographic therapy or if the patient has
underlying conditions that may complicate the ability to perform angiography or TAE
(eg, renal insufficiency). (See "Angiographic control of nonvariceal gastrointestinal
bleeding in adults".)
RESUMPTION OF ANTICOAGULANTS AND ANTIPLATELET
AGENTSData are limited with regard to the appropriate timing for resuming
anticoagulation or antiplatelet agents following endoscopic hemostasis. The timing will
depend on the patient's risk of suffering a thromboembolic event while off of the
medication(s). When to resume these agents after hemostasis has been achieved is
discussed elsewhere. (See "Management of anticoagulants in patients undergoing
endoscopic procedures", section on 'Resuming anticoagulants after
hemostasis' and "Management of antiplatelet agents in patients undergoing endoscopic
procedures".)
FOLLOW-UPAll patients with bleeding peptic ulcers need to be evaluated for factors
that predispose to ulcer formation (eg, H. pylori) and treated as appropriate. This issue
and issues related to maintenance proton pump inhibitor use/discontinuation are
discussed in detail elsewhere. (See "Peptic ulcer disease: Treatment and secondary
prevention" and "Peptic ulcer disease: Epidemiology, etiology, and pathogenesis".)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Gastrointestinal bleeding in adults".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
274
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)")

●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)")
SUMMARY AND RECOMMENDATIONS
●The initial evaluation of the patient with upper gastrointestinal (UGI) bleeding
involves an assessment of hemodynamic stability and the necessity for fluid
resuscitation (table 1). (See "Approach to acute upper gastrointestinal bleeding in
adults".)
●We calculate a bleeding risk score such as the Glasgow-Blatchford score as part
of the initial assessment of patients with UGI bleeding. Patients with a GBS of 0-1
may be considered for out-patients management.
●We recommend endoscopy within 24 hours of presentation for the diagnosis and
treatment of active UGI bleeding and for the prevention of recurrent bleeding
rather than waiting more than 24 hours (Grade 1B). (See "Approach to acute
upper gastrointestinal bleeding in adults".)
●The majority of patients with UGI bleeding due to peptic ulcer disease will stop
bleeding spontaneously, and most will not rebleed during hospitalization.
However, a subgroup of patients is at high risk for recurrent hemorrhage. The
major endoscopic predictors of persistent or recurrent bleeding include active
bleeding at endoscopy, a nonbleeding visible vessel, and an adherent clot (table
2). (See 'Stigmata of recent hemorrhage' above.)
●We recommend that patients with actively bleeding peptic ulcers or ulcers with
high-risk stigmata (such as a visible vessel or adherent clot) receive an
intravenous (IV) proton pump inhibitor (PPI) rather than no acid suppression or an
H2 receptor antagonist (Grade 1A). The optimal approach to PPI administration
prior to endoscopy is unclear. Our approach is to give a high-dose bolus
(eg, esomeprazole 80 mg) to patients with signs of active bleeding (eg,
hematemesis, hemodynamic instability). If endoscopy is delayed beyond 12 hours,
a second dose of an IV PPI should be given (eg, esomeprazole 40 mg). For
patients who may have stopped bleeding (eg, patients who are hemodynamically
stable with melena), we give an IV PPI every 12 hours (eg, esomeprazole 40 mg).
Subsequent dosing will then depend on the endoscopic findings. (See 'Acid
suppression' above and "Approach to acute upper gastrointestinal bleeding in
adults", section on 'Acid suppression'.)
●We recommend that an oral PPI be given when intravenous formulations are not
available (Grade 1A). When used for acute GI bleeding, oral PPIs should be given
twice daily at high doses. (See 'Acid suppression' above.)

275
●We suggest that otherwise healthy patients without severe bleeding or high-risk
stigmata be discharged following endoscopy on a standard-dose proton pump
inhibitor rather than being admitted to the hospital for observation (Grade 2C).
Such patients are at low risk for recurrent bleeding. (See 'Inpatient versus
outpatient management' above.)
●We recommend that patients with active bleeding (oozing or spurting) or a visible
vessel receive endoscopic therapy and acid suppressive therapy rather than acid
suppressive therapy alone (Grade 1B). Endoscopic therapy decreases the risk of
recurrent bleeding in patients with these high-risk stigmata. Patients may be
treated with thermal coagulation or endoscopic clips (with or
without epinephrine injection) but should not be treated with epinephrine
monotherapy due to an increased risk of recurrent bleeding. Application of
hemostatic sprays can be used if traditional techniques fail to achieve hemostasis
and over-the-scope clips may be an option for patient with recurrent bleeding.
(See 'Stigmata of recent hemorrhage' above and 'Endoscopic therapy' above.)
Whether to treat patients with adherent clots will depend upon the location of the
ulcer, the availability of surgical or interventional radiologic backup should
bleeding start, and the experience of the endoscopist. Our approach is to make
gentle attempts to remove the clot so that endoscopic treatment can be applied if
needed, provided the ulcer is in a location that is amenable to endoscopic therapy
should bleeding start.
●Patients with active bleeding, a visible vessel, or an adherent clot should be
started on a high-dose continuous infusion of a PPI (eg, esomeprazole 8 mg per
hour) following endoscopy. The infusion should be continued for 72 hours. If the
patient did not receive a high-dose bolus of a PPI (eg, esomeprazole 80 mg) prior
to endoscopy, the patient should receive the high-dose bolus prior to starting the
continuous infusion. If there are no signs of recurrent bleeding, the patient may
then be switched to a high-dose oral PPI (eg, omeprazole 40 mg twice daily for 14
days followed by a once-daily PPI). Patients with low-risk ulcers can be switched
to a standard-dose oral PPI following endoscopy (eg, omeprazole 20 mg daily).
●We suggest that most patients who fail endoscopic therapy first undergo
attempted transarterial angiographic embolization (TAE) rather than surgery
(Grade 2C). We suggest angiography rather than surgery because it is less
invasive and because surgery is still an option in patients who fail angiographic
therapy. However, surgery is a reasonable alternative if an interventional
radiologist with expertise in TAE is not available, if the lesion is deemed unlikely to
respond to angiographic therapy, or if the patient has underlying conditions that
may complicate the ability to perform angiography or TAE (eg, renal insufficiency).
(See 'Treatment of persistent and recurrent bleeding' above.)
●All patients with bleeding peptic ulcers need to be evaluated for factors that
predispose to ulcer formation (eg, H. pylori) and treated as appropriate.
(See "Peptic ulcer disease: Treatment and secondary prevention" and "Peptic
ulcer disease: Epidemiology, etiology, and pathogenesis".)

276
Peptic ulcer disease: Clinical manifestations and
diagnosis
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Jan 11, 2022.
INTRODUCTIONA peptic ulcer is a defect in the gastric or duodenal mucosa that
extends through the muscularis mucosa into the deeper layers of the wall. Peptic ulcers
may present with dyspeptic or other gastrointestinal symptoms, or may be initially
asymptomatic and then present with complications such as hemorrhage or perforation.
This topic will review the clinical manifestations and diagnosis of peptic ulcer disease.
The etiology, complications, and management of peptic ulcer disease are discussed in
detail, separately. (See "Peptic ulcer disease: Epidemiology, etiology, and
pathogenesis" and "Overview of complications of peptic ulcer disease" and "Peptic ulcer
disease: Treatment and secondary prevention" and "Approach to refractory peptic ulcer
disease" and "Surgical management of peptic ulcer disease".)
CLINICAL MANIFESTATIONS
Asymptomatic — Approximately 70 percent of peptic ulcers are asymptomatic [1].
Patients with silent peptic ulcers may later present with ulcer-related complications such
as hemorrhage or perforation. Between 43 and 87 percent of patients with bleeding
peptic ulcers present without antecedent dyspepsia or other heralding gastrointestinal
symptoms [2-4]. Older adults and individuals on nonsteroidal anti-inflammatory drugs
are more likely to be asymptomatic from their ulcers and later present with ulcer
complications [3,5-7]. (See 'Ulcer complications' below.)
Symptomatic
Abdominal pain — Upper abdominal pain or discomfort is the most prominent
symptom in patients with peptic ulcers. Approximately 80 percent of patients with
endoscopically diagnosed ulcers have epigastric pain [5]. Occasionally the discomfort
localizes to the right or left upper quadrants of the hypochondrium [8]. Radiation of pain
to the back may occur, but back pain as the primary symptom is atypical. In untreated
patients, pain can last a few weeks followed by symptom-free periods of weeks or
months. The "classic" pain of duodenal ulcers occurs two to five hours after a meal
when acid is secreted in the absence of a food buffer, and at night (between about 11
PM and 2 AM) when the circadian pattern of acid secretion is maximal [9].
Patients with peptic ulcers, and particularly pyloric channel ulcers, may have food-
provoked symptoms due to visceral sensitization and gastroduodenal dysmotility [2].
These symptoms include epigastric pain that worsens with eating, postprandial belching
and epigastric fullness, early satiety, fatty food intolerance, nausea, and occasional
vomiting [2,5].

277
Associated symptoms — Patients may have associated symptoms of bloating,
abdominal fullness, nausea, and early satiety that may be provoked by eating.
Gastroesophageal reflux may coexist but may or may not be related to the peptic ulcers.
In one systematic review that included 33 studies of patients with endoscopically
diagnosed peptic ulcer disease, the average prevalence of heartburn or acid
regurgitation was 46 percent [5].
Ulcer complications — Complications may be heralded by new ulcer symptoms or a
change in symptoms or may occur in the absence of typical symptoms. (See "Overview
of complications of peptic ulcer disease", section on 'Clinical presentation'.)
Bleeding — Acute upper gastrointestinal hemorrhage is the most common complication
of peptic ulcer disease. Patients with bleeding from a peptic ulcer may present with
nausea, hematemesis (either red blood or coffee-ground emesis), or melena (black,
tarry stool). In rare cases, patients have massive bleeding and present with
hematochezia (red or maroon blood in the stool) and orthostatic hypotension (picture 1).
(See "Causes of upper gastrointestinal bleeding in adults", section on 'Peptic ulcer
disease'.)
Gastric outlet obstruction — Ulcers located in the pyloric channel or duodenum may
cause gastric outlet obstruction. Symptoms include early satiety, bloating, indigestion,
anorexia, nausea, vomiting, epigastric pain shortly after eating, and weight loss.
(See "Gastric outlet obstruction in adults", section on 'Clinical manifestations'.)
Penetration and fistulization — Peptic ulcers may penetrate through the bowel wall
without a free perforation or leakage of luminal contents into the peritoneal cavity.
Pyloric channel or prepyloric peptic ulcers may penetrate directly into the duodenal bulb,
creating a gastroduodenal fistula, resulting in an acquired "double" pylorus.

The pain of penetration typically becomes more intense, of longer duration, and is
frequently referred to the lower thoracic or upper lumbar spine region. Penetrating
posterior ulcers classically present with a shift from the typical vague visceral discomfort
to a more localized and intense pain that is felt in the back and is not relieved by food or
antacids. This change in symptom pattern may be gradual or sudden.

Patients with penetrating ulcers often present with a change in symptoms due to
symptomatic involvement of adjacent structures.

●Gastrocolic or duodenocolic fistulas can present with halitosis, feculent vomiting,

postprandial diarrhea, dyspepsia, and weight loss.
●Penetration into a surrounding organ can result in an abscess.
●Penetration into vascular structures can result in exsanguinating hemorrhage (eg,
aortoenteric fistula or penetration into the cystic artery) [10-12].
●Penetration into the biliary tree can result in a choledochoduodenal fistula,
extrahepatic biliary obstruction, or hemobilia [13,14].
●Fistulization into the pancreatic duct has also been reported with penetrating
duodenal ulcers [15]. Mild hyperamylasemia can develop with posterior
penetration of either a gastric or duodenal ulcer, but clinical pancreatitis is
uncommon.

278
Perforation — Perforations complicate 2 to 10 percent of patients with peptic ulcer
disease [16]. Ulcer perforation should be suspected in patients who suddenly develop
severe, diffuse abdominal pain. A classic triad of sudden onset of abdominal pain,
tachycardia, and abdominal rigidity is the hallmark of peptic ulcer perforation. Prepyloric
gastric ulcerations account for most perforations followed by duodenal bulb ulcers [17].
(See "Overview of gastrointestinal tract perforation", section on 'Presentations'.)
Laboratory findings — Most patients with uncomplicated peptic ulcers have a normal
complete blood count. Patients may have iron deficiency anemia due to recurrent
gastrointestinal blood loss. Patients with acute gastrointestinal perforation may have
leukocytosis.
Imaging findings — The two direct signs of peptic ulcer disease on abdominal
computed tomography scan include focal discontinuity of the mucosal
hyperenhancement, which reflects disease reaching the muscularis mucosa, and
identification of luminal outpouching. Luminal outpouching corresponds to the ulcer
crater which extends through and beyond the gastroduodenal wall [18]. Other non-
specific signs include gastric wall thickening, perigastric or periduodenal inflammation,
and mucosal enhancement in a focal area [19]. Abdominal CT is not sensitive for
uncomplicated peptic ulcer disease and superficial ulcers may be missed [20].  
DIFFERENTIAL DIAGNOSISThe differential diagnosis of peptic ulcer disease
consists of other causes of dyspepsia and includes drug-induced dyspepsia, biliary
disease, gastric malignancy, and less commonly, chronic pancreatitis. These conditions
can be excluded from peptic ulcer disease by upper endoscopy. The diagnostic
evaluation of a patient with upper abdominal pain and dyspepsia are discussed in detail,
separately. (See "Approach to the adult with dyspepsia" and "Evaluation of the adult
with abdominal pain", section on 'Epigastric pain'.)
DIAGNOSISThe diagnosis of peptic ulcer disease is suspected in patients with
dyspepsia, especially in the setting of nonsteroidal anti-inflammatory drug (NSAID) use
or a history of Helicobacter pylori (H. pylori) infection. Occasionally peptic ulcers may be
suspected based on imaging performed for evaluation of abdominal pain. The diagnosis
of peptic ulcer disease is definitively established by direct visualization of the ulcer on
upper endoscopy.
Upper endoscopy — Endoscopy is the most accurate diagnostic test for peptic ulcer
disease (picture 2). The sensitivity of upper endoscopy in the detection of
gastroduodenal lesions is approximately 90 percent but varies based on the location of
the ulcer and the experience of the endoscopist [21,22].
Indications for ulcer biopsy
Malignant appearing ulcers — All ulcers with malignant features should be biopsied.
Endoscopic features that suggest that an ulcer may be malignant include:
●An ulcerated mass protruding into the lumen
●Folds surrounding the ulcer crater that are nodular, clubbed, fused, or stop short
of the ulcer margin
●Overhanging, irregular, or thickened ulcer margins
Selected benign appearing ulcers — On upper endoscopy, benign ulcers have
smooth, regular, rounded edges, with a flat, smooth ulcer base often filled with exudate.
Routine biopsy of benign-appearing duodenal ulcers is not recommended, as they are
unlikely to harbor malignancy.
279
However, in areas with high gastric cancer incidence, all gastric ulcers should be
biopsied. Whether to biopsy benign-appearing gastric ulcers in areas of low gastric
cancer incidence is controversial. While we generally biopsy benign-appearing gastric
ulcers at the index upper endoscopy as they may harbor malignancy; other experts do
not biopsy gastric ulcers if the patient's history and demographic features suggest a low-
risk of gastric cancer (eg, a young patient in the United States with a history of NSAID
use and a shallow, flat antral ulcer) [23]. (See "Clinical features, diagnosis, and staging
of gastric cancer".)
Specific etiology suspected — In patients with some systemic infiltrative or
inflammatory conditions (eg, sarcoidosis, Crohn disease, eosinophilic gastroenteritis),
peptic ulcers may be due to gastric or duodenal involvement by these conditions. In
such cases, biopsies of the peptic ulcer and surrounding mucosa may be helpful to
confirm the suspected etiology. (See "Unusual causes of peptic ulcer disease", section
on 'Inflammatory and infiltrating disease'.)
Ulcer biopsy technique — If biopsies of an ulcer are indicated, we obtain samples
from all four quadrants of the ulcer. If there are endoscopic features suspicious for
malignancy, we obtain biopsies using jumbo forceps with more extensive sampling
along the edges of the ulcer. We do not routinely perform cytology as this adds little to
the diagnostic yield.
Biopsy for H. pylori — Gastric mucosal biopsy from the gastric antrum and body
should be obtained to exclude H. pylori in patients undergoing upper endoscopy for
peptic ulcer disease unless it is impractical or difficult, such as with a blood-filled
stomach. In such cases, testing should be performed when possible to exclude H.
pylori. (See 'Exclude Helicobacter pylori (H. pylori) infection' below and "Indications and
diagnostic tests for Helicobacter pylori infection in adults", section on 'Patient
undergoing upper endoscopy'.)
ESTABLISHING THE ETIOLOGY
Exclude Helicobacter pylori (H. pylori) infection — All patients diagnosed with peptic
ulcer disease should undergo testing for H. pylori infection.
●Patients with evidence of H. pylori on biopsy should receive eradication therapy.
However, in patients with active bleeding, a negative biopsy result does not
exclude H. pylori, and another test (ideally a urea breath test or a stool antigen
test for H. pylori) should be performed to confirm a negative result.
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults",
section on 'Urea breath testing'.)
●If a gastric mucosal biopsy is not obtained at the time of endoscopy, in the
absence of active GI bleeding we perform a urea breath test or stool antigen
assay to diagnose H. pylori. (See "Indications and diagnostic tests for
Helicobacter pylori infection in adults", section on 'Noninvasive testing'.)
●In patients who receive treatment for H. pylori, eradication of infection should be
confirmed four or more weeks after the completion of therapy [24].
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults",
section on 'Confirmation of eradication'.)
Evaluate nonsteroidal anti-inflammatory drug (NSAID) use — A history of NSAID
use should be sought. Urine salicylate levels are not helpful for the detection of low-

280
dose aspirin use or other widely used NSAIDs. Testing for individual agents is possible
but is not practical and is limited by its expense.
Additional evaluation in selected patients — Other causes of peptic ulcer disease
should be considered when H. pylori and use of NSAIDs have been excluded. Unusual
causes for peptic ulcer disease and the evaluation for H. pylori and NSAID negative
ulcers is discussed in detail, separately. (See "Unusual causes of peptic ulcer disease",
section on 'Evaluation of H. pylori and NSAID negative ulcers'.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Peptic ulcer disease".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)" and "Patient

education: H. pylori infection (The Basics)" and "Patient education: Gastritis (The
Basics)")
●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Peptic ulcers are commonly asymptomatic. Symptomatic peptic ulcers most
commonly present with epigastric pain or food-provoked epigastric discomfort and
fullness, early satiety, and nausea. (See 'Clinical manifestations' above.)
●Complications may be heralded by new ulcer symptoms or a change in
symptoms, or may occur in the absence of typical symptoms. Complications of
peptic ulcers include bleeding, gastric outlet obstruction, penetration into a solid
organ or fistulization into a hollow viscus, and free perforation. (See 'Ulcer
complications' above.)
●The two direct signs of peptic ulcer disease on abdominal computed tomography
scan include focal discontinuity of the mucosal hyperenhancement and luminal
outpouching. (See 'Imaging findings' above.)
●The diagnosis of peptic ulcer disease is suspected in patients with dyspepsia,
especially in the setting of nonsteroidal anti-inflammatory drug (NSAID) use or a
history of Helicobacter pylori (H. pylori) infection. Occasionally peptic ulcers may
281
be diagnosed or suspected based on contrast imaging performed for evaluation of
abdominal pain. The diagnosis of peptic ulcer disease is definitively established by
direct visualization of the ulcer on upper endoscopy. On upper endoscopy, benign
ulcers have smooth, regular, rounded edges, with a flat, smooth ulcer base often
filled with exudate (picture 2).
●All ulcers with malignant features should be biopsied. Endoscopic features that
suggest that an ulcer may be malignant include:
•An ulcerated mass protruding into the lumen
•Folds surrounding the ulcer crater that are nodular, clubbed, fused, or stop
short of the ulcer margin
•Overhanging, irregular, or thickened ulcer margins
●Routine biopsy of benign-appearing duodenal ulcers is not recommended as they
are unlikely to be malignant. In areas with high gastric cancer incidence, gastric
ulcers should be biopsied. The decision to biopsy benign-appearing gastric ulcers
in areas of low gastric cancer incidence is controversial. Although we biopsy
benign-appearing gastric ulcers at the index upper endoscopy as they may harbor
malignancy, other experts do not biopsy gastric ulcers if the patient's history and
demographic features suggest a low risk of gastric cancer (eg, young patient with
a history of NSAID use and a shallow flat antral ulcer). (See 'Diagnosis' above.)
●All patients diagnosed with peptic ulcer disease should undergo testing for H.
pylori infection. Additional evaluation to determine the underlying etiology should
be considered when H. pylori and NSAID use have been excluded.
(See 'Establishing the etiology' above.)

282
Peptic ulcer disease: Epidemiology, etiology, and
pathogenesis
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Mar 02, 2020.
INTRODUCTIONPeptic ulcers are defects in the gastric or duodenal mucosa that
extend through the muscularis mucosae (picture 1). They develop and persist as a
function of the acid-peptic activity in gastric juice. Peptic ulcer disease remains an
important cause of morbidity and health care costs [1].
The natural history of peptic ulcer ranges from healing without intervention to the
development of complications with the potential for significant morbidity and mortality,
such as bleeding and perforation. This topic will review the epidemiology, etiology, and
pathogenesis of peptic ulcer disease. The clinical manifestations, diagnosis, and
management of peptic ulcer disease are discussed in detail, separately. (See "Overview
of complications of peptic ulcer disease".)
EPIDEMIOLOGY
Incidence and prevalence — In a systematic review of 31 published studies, the
pooled incidence of uncomplicated peptic ulcer disease (PUD) was approximately one
case per 1000 person-years in the general population, and the incidence of ulcer
complications was approximately 0.7 cases per 1000 person-years [2].
The incidence and prevalence of PUD varies based upon the presence of Helicobacter
pylori (H. pylori). Higher rates are found in countries where H. pylori infection is higher
[3-6]. The incidence of PUD in H. pylori-infected individuals is approximately 1 percent
per year, a rate that is 6- to 10-fold higher than for uninfected individuals [7,8]. A
systematic review of seven studies from developed countries indicated a population-
based one-year prevalence of PUD of 0.1 to 1.5 percent based on physician diagnosis
and 0.1 to 0.19 percent based on hospitalization data [9]. A study in the United States
reported an endoscopic point prevalence for peptic ulcers in asymptomatic, H. pylori-
positive adults of 2 percent [10]. Other studies, in presumably asymptomatic subjects in
whom H. pylori status was unknown, have reported an endoscopic point prevalence
ranging from 1 and 6 percent [5,11-13].
Ulcer incidence increases with age for both duodenal ulcers (DUs) and gastric ulcers
(GUs), but the incidence of uncomplicated PUD reached a plateau with age, whereas
for complicated PUD, the incidence increases with age [14]. DUs occur two decades
earlier than GUs, particularly in males [14].
Time trends — PUD rates have been steadily falling over the past several decades [14-
22]. In several regions, rates of DUs have fallen more dramatically than for GUs [16,20].
Age-adjusted hospitalization rates for peptic ulcer disease have also declined,

283
suggesting that the decline in H. pylori infection has had a major impact on ulcer rates
[23-25]. In developed countries, the prevalence of ulcer disease in young people has
declined, while nonsteroidal anti-inflammatory drug (NSAID)-related ulcers in older
adults have increased due both to increased life expectancy and frequent use
of aspirin and NSAIDs [16,19-22,26-29]. PUD has shifted from being a male-
predominant disease in western countries to one with a nearly equal prevalence in both
sexes [2,9].
ETIOLOGYPeptic ulcer disease (PUD) is associated with two major
factors: Helicobacter pylori (H. pylori) infection and the consumption of nonsteroidal
anti-inflammatory drugs (NSAIDs). NSAID use and H. pylori infection represent
independent and also synergistic risk factors for uncomplicated and bleeding peptic
ulcer disease. There are also a number of other defined mechanisms for peptic ulcer
disease that are much less common but becoming more evident as the prevalence
of H. pylori declines in developed countries (table 1). (See "Unusual causes of peptic
ulcer disease".)
H. pylori — H. pylori affects a number of aspects of gastrointestinal physiology,
including gastric acid secretion, and mucosal defense mechanisms leading to peptic
ulcer disease. The prevalence of H. pylori both in the general population and in peptic
ulcer patients is decreasing rapidly in developed regions, presumably due to improved
hygiene and decreased H. pylori transmission in early childhood. A systematic review
of H. pylori prevalence around the world showed that the highest reported H.
pylori prevalence areas were in Africa (70.1 percent), South America (69.4 percent),
and Western Asia (66.6 percent). Regions with the lowest reported H. pylori prevalence
were Oceania (24.4 percent), Western Europe (34.3 percent), and Northern America
(37.1 percent) [30].
NSAIDs, including aspirin — NSAID use is associated with a fourfold increase in the
risk of peptic ulcer disease [31]. In addition, NSAIDs are associated with an increase in
risk of complications from peptic ulcer disease which include gastrointestinal bleeding,
perforation, and pyloric obstruction. NSAID-induced ulcers may also be more refractory
to conventional therapy [32]. However, up to 40 percent of patients may not report using
NSAIDs [33]. (See "Nonselective NSAIDs: Overview of adverse effects", section on
'Gastrointestinal effects'.)
Other rare causes — Although some NSAID-negative, H. pylori-negative ulcers are
due to specific causes, such as gastrinoma or viral infections, others remain in the
idiopathic category (table 1). Unusual causes of peptic ulcer disease are discussed in
detail separately. (See "Unusual causes of peptic ulcer disease".)
RISK FACTORSThe presence of Helicobacter pylori (H. pylori) or use of
nonsteroidal anti-inflammatory drugs (NSAIDs) alone is not likely to be sufficient for
ulcer formation. Although gastritis is consistently observed with H. pylori infection and
NSAIDs consistently inhibit mucosal prostaglandin production, the yearly incidence of
clinical ulcer disease among subjects at risk is only approximately 1 percent for both
categories of ulcer [7,8,34]. There are a number of risk factors for ulcer disease.
Smoking — Smoking is an independent risk factor for symptomatic and asymptomatic
peptic ulcer disease (PUD) [35]. PUD risk progressively increases with increasing pack-
years of cigarette smoking [36]. A population-based study found that the prevalence of
ulcer disease in current and former smokers was almost twice as high as non-smokers
284
[37]. Smoking more than 15 cigarettes per day compared with never smoking increased
the risk of a perforated ulcer more than threefold [38]. In addition, ulcers in smokers
appear to be more difficult to treat and may be associated with a higher rate of
recurrence [39-42]. The deleterious effects of smoking may be due to its impact on the
balance of mucosal aggressive and protective factors [43]. However, in patients with H.
pylori, smoking does not appear to be a risk factor for ulcer relapse once H. pylori has
been eradicated [44-46].
Alcohol — Alcohol in high concentrations damages the gastric mucosal barrier. In a
large population based study, heavy drinking (more than 42 drinks per week) increased
the risk of a bleeding ulcer fourfold (RR = 4.4; 95% CI2.3–8.3) compared with drinking
less than one drink per week in high concentrations and also stimulates acid secretion
[47]. In addition, contents of alcoholic beverages other than alcohol are also strong
stimulants of acid secretion. Alcohol consumption is related to the development of
complicated and uncomplicated ulcer disease. In a large population based study,
drinking more than 42 drinks per week increased the risk of a bleeding ulcer fourfold
(RR = 4.4; 95% CI2.3–8.3) compared with drinking less than one drink per week. [38].
Genetic factors — Host genetic factors appear to be important in predisposing to H.
pylori infection and to disease outcomes such as duodenal ulcers (DU) and gastric
cancer. Genes that encode for cytokines account for the individual responses to H.
pylori infection. Genetic variations in pro-inflammatory cytokines (eg, IL-1B, IL-6, IL-8,
and tumor necrosis factor (TNF)-alpha) and anti-inflammatory cytokines (IL-10) are
associated with PUD [48]. A genome-wide association study in Japan found that
specific genetic variations in the Prostate Stem Cell Antigen are associated with gastric
ulcer, duodenal ulcer, and gastric cancer [49]. Genetic polymorphism related to synergy
between host (TNF-alpha promoter) and bacterial (induced by contact with epithelium,
or the iceA1 gene) factors have been related to DU in children [50].
There is also evidence for a genetic predisposition to PUD that is independent of any
predisposition to H. pylori infection [51,52]. An illustrative report found high concordance
of a self-reported ulcer history in monozygotic compared with dizygotic twins regardless
of whether they were reared apart or together [53]. Cross-twin, cross-trait correlations
for monozygotic and dizygotic twins indicated that genetic effects for peptic ulcer were
independent of genetic influences for H. pylori. Hyperpepsinogenemia linked to elevated
serum pepsinogen group A has also been noted in patients with familial clustering of
PUD [52]. Genetic predisposition due to a polymorphism of cytochrome P450 2C9 may
delay the metabolism of several NSAIDs, with a prolonged duration of drug effect,
enhancing the ulcerogenic effect [54].
Blood groups O and A, the Lewis phenotype Le (a + b-), and non-secretors of ABH in
particular have been associated with an increased risk of peptic ulcers [55,56].
However, other studies have failed to find any association of blood group O with H.
pylori infection or with peptic ulcer [57-59]. Studies suggest that strains of H. pylori from
different areas of the world may have different binding affinities for gastric epithelium
accounting for these seemingly contradictory results seen with blood group studies from
different countries [60]. (See "Pathophysiology of and immune response to Helicobacter
pylori infection".)
Other

285
●Diet – Dietary factors have been hypothesized to account for some of the
regional variation of ulcer disease, possibly related to toxins generated with
storage of certain foods or from protective effects of certain foods [61,62]. A
prospective study of a cohort of 47,806 men found no association between the
type or amount of fat intake and the risk of DU [63]. However, high consumption of
fruits and vegetables, dietary fiber, and vitamin A were associated with a reduced
risk of ulcer disease.
Evidence to support the use of a bland diet or dietary restrictions to prevent PUD
are lacking. Although certain foods, beverages, and spices cause dyspepsia, there
are no convincing data that such specific foods cause, perpetuate, or reactivate
peptic ulcers. There is no evidence that coffee consumption is a risk factor for
ulcer disease, although increased consumption may be associated with a higher
rate of infection with H. pylori [64-66]. In one series, coffee increased dyspepsia in
subjects with nonulcer dyspepsia, but not in those with DU when compared with
controls [67].
●Psychologic factors – Two lines of evidence suggest an association between
psychosocial factors and ulcer pathogenesis. First, in prospective studies, poorly
tolerated stress or depressive symptoms at baseline increase the risk of ulcer
development over the next 9 to 15 years [68]. Other psychosocial factors, such as
work-related stress, social problems, and post-traumatic stress disorder, are also
predictive of subsequent ulcer disease [69-72]. Second, several studies have
established that peptic ulcer complications become much more prevalent during
periods of natural disaster or societal catastrophe [68,73,74].
The pathophysiological mechanisms accounting for the effects of stress on the
ulcer formation have not been defined. These effects could be mediated by both
altered behaviors and by altered physiology. A population-based prospective
study in Denmark that included 3379 individuals demonstrated that psychological
stress increased the incidence of peptic ulcer in part by influencing health risk
behaviors [69]. Stress had similar effects on ulcers associated with H.
pylori infection and those unrelated to H. pylori or NSAID use. Stress also
increases acid secretion, but the effects are more prominent in patients with DU
compared with controls [75]. As a result, one must consider not only the stressor,
but the individual's physiological and psychosocial response to the stress;
deleterious effects may only be evident in a subset of predisposed individuals.
Stress, anxiety, and depression have also been observed to impair endoscopic
healing and to promote relapse of endoscopically diagnosed ulcers [68,76,77].
The effects of stress seem to be reversible; patients who develop an ulcer
following traumatic life events, but who are psychologically stable, tend to do well
after the stress has resolved [68].
However, the finding a relationship between psychosocial factors and ulcer
disease does not establish causality. One study found that anxiety and
neuroticism were high in a group of DU patients at the outset, but normalized in
relapse-free patients at the end of a 10-year follow-up period [78]. In some cases,
psychological features may be the result and not the cause of the disease
process.

286
PATHOPHYSIOLOGYConsidering the acid-peptic environment of the stomach and
the noxious agents that are ingested, ulcers are surprisingly uncommon, reflecting the
effectiveness of the protective mechanisms that govern gastric mucosal function and
repair. Primary malfunction of these secretory, defense, or repair mechanisms is a very
uncommon cause of ulcer, if it occurs at all. Most ulcers occur when the normal
mechanisms are disrupted by superimposed processes such as Helicobacter
pylori (H. pylori) infection and the ingestion of nonsteroidal anti-inflammatory drugs
(NSAIDs). NSAIDs, via inhibition of prostaglandin synthesis, affect the amount of gastric
acid generated, the integrity of the mucosal barrier, the amount of bicarbonate and
glutathione generated, and the rate of mucosal blood flow. (See "NSAIDs (including
aspirin): Pathogenesis of gastroduodenal toxicity".)
Gastric acid hypersecretion — Gastric ulcers involving the gastric body or distal
antrum, and gastric ulcers associated with concurrent duodenal ulcers (DU) have
normal or even increased levels of acid secretion. Although only a minority of DU
subjects have true acid hypersecretion, high-normal or modestly elevated values appear
to be a defining characteristic of patients with DU, whether or not they are infected
with H. pylori.
The majority of patients infected with H. pylori develop pangastritis, which over time is
associated with a reduction in gastric acid secretion. However, approximately, 10 to 15
percent of patients with the infection develop an antrum predominant gastritis which is
associated with decreased antral somatostatin concentrations and increased basal and
stimulated gastric acid secretion [79-81]. Somatostatin secretion by the antral D cells is
a feedback mechanism by which gastrin secretion is regulated. Inhibition of
somatostatin secretion may involve cytokines induced by H. pylori infection. Host
genetic factors can also influence the cytokine response to H. pylori infection and
predispose to ulcer disease [79,80,82-84].
Increased acid secretion is also an important factor in some patients with ulcer
recurrences following successful H. pylori eradication. Abnormalities in gastrin and
somatostatin secretion and most abnormalities of acid secretion usually normalize
within one year of H. pylori eradication [79,84]. However, a small study found higher
basal and pentagastrin-stimulated acid secretion in patients with recurrent ulcers
following successful H. pylori eradication, as compared with patients without recurrent
ulcers [85]. (See "Association between Helicobacter pylori infection and duodenal
ulcer", section on 'Pathogenesis of ulcer formation'.)
Among H. pylori-negative subjects, a subset have acid hypersecretion without
hypergastrinemia [83,86]. The mechanisms of acid hypersecretion in these patients
have not been defined. One study of six patients with non-H. pylori, non-NSAID DU
found increased gastrin response to a meal (but not fasting hypergastrinemia) and
increased peak gastric acid secretion [87]. Some subjects with non-H. pylori, non-
NSAID hypersecretion may have a component of muscarinic-dependent, vagal
hyperactivity, although this element is difficult to quantify. In the absence of H. pylori or
gastrinoma, fasting hypergastrinemia is only rarely found in hypersecretory DU patients,
sometimes linked to antral G-cell hyperfunction [88].
Impaired duodenal bicarbonate secretion — The majority of DU patients have
impaired duodenal bicarbonate secretion. The combination of increased gastric acid
secretion and reduced duodenal bicarbonate secretion lowers the pH in the duodenum,
287
which promotes the development of gastric metaplasia (ie, the presence of gastric
epithelium in the first portion of the duodenum) [89]. Excess acid secretion in
gastrinoma patients is also associated with marked gastric metaplasia in the duodenum.
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)" and "Patient

education: H. pylori infection (The Basics)" and "Patient education: Gastritis (The
Basics)")
●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Estimates of the annual incidence of peptic ulcer disease (PUD) range from 0.1
to 0.3 percent. PUD incidence in Helicobacter pylori (H. pylori)-infected individuals
is approximately 1 percent per year, a rate that is 6- to 10-fold higher than for
uninfected subjects. The incidence of PUD increases with age for both duodenal
and gastric ulcers. (See 'Incidence and prevalence' above.)
●Peptic ulcer disease is associated with two major factors: H. pylori infection and
the consumption of nonsteroidal anti-inflammatory drugs (NSAIDs). NSAIDs are
also associated with an increase in risk of complications from peptic ulcer disease
which include gastrointestinal bleeding, perforation, and gastric outlet obstruction.
(See 'Etiology' above.)
●Smoking is a risk factor for PUD. Alcohol in high concentrations damages the
gastric mucosal barrier. Host genetic factors appear to be important in
predisposing to H. pylori infection but can also influence the risk of PUD through
other mechanisms. Dietary factors have been hypothesized to account for some of
the regional variation of ulcer disease, possibly related to toxins generated with
storage of certain foods or from protective effects of certain foods. However,
evidence to support the use of a bland diet or dietary restrictions to prevent PUD
are lacking. (See 'Risk factors' above.)
●Most ulcers occur when the normal secretory, defense, or repair mechanisms of
the stomach are disrupted by superimposed processes such as H. pylori infection
and the ingestion of NSAIDs. (See 'Pathophysiology' above.)

288
289
Peptic ulcer disease: Treatment and secondary
prevention
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Apr 01, 2020.
INTRODUCTIONA peptic ulcer is a defect in the gastric or duodenal wall that
extends through the muscularis mucosa into the deeper layers of the wall. The
management of patients with peptic ulcer disease is based on the etiology, ulcer
characteristics, and anticipated natural history. This topic will review the initial
management of peptic ulcer disease. The management of recurrent peptic ulcer
disease, the complications of peptic ulcer disease, surgical management of peptic ulcer
disease, and the clinical manifestations and diagnosis of peptic ulcer disease are
discussed separately. (See "Approach to refractory peptic ulcer disease" and "Overview
of complications of peptic ulcer disease" and "Surgical management of peptic ulcer
disease" and "Peptic ulcer disease: Clinical manifestations and diagnosis".)
INITIAL MANAGEMENT
Treat the underlying etiology
Eradication of Helicobacter pylori (H. pylori) — Patients with peptic ulcers should be
tested for infection with H. pylori and treated accordingly (algorithm 1) [1-4]. Eradication
of H. pylori in patients with peptic ulcer disease is associated with higher healing rates
in patients with duodenal and gastric ulcers. A meta-analysis of 24 randomized trials
that included 2102 patients with peptic ulcer disease revealed that the 12-month ulcer
remission rates for gastric and duodenal ulcers were significantly higher in patients
successfully eradicated of H. pylori infection as compared with those with a persistent
infection (97 and 98 percent versus 61 and 65 percent, respectively) [5]. In addition,
eradication of H. pylori infection is associated with lower ulcer recurrence rates in
patients with gastric and duodenal ulcers who are not placed on maintenance
antisecretory therapy [6]. Treatment regimens for H. pylori are discussed in detail,
separately. (See "Indications and diagnostic tests for Helicobacter pylori infection in
adults" and "Treatment regimens for Helicobacter pylori in adults".)
Discontinue nonsteroidal anti-inflammatory drugs (NSAIDs) — Patients with peptic
ulcers should be advised to avoid NSAIDs. NSAIDs, including aspirin, increase the risk
of peptic ulcer disease and are associated with an increased risk of complications from
a peptic ulcer.
Rare or unclear cause — Rare causes of ulcer disease (eg, infections, Crohn disease,
ischemia) should be addressed and treated. In patients with peptic ulcer disease of
unclear etiology, additional evaluation is needed to exclude other rare causes of peptic
ulcer disease. (See "Unusual causes of peptic ulcer disease".)

290
Initial antisecretory therapy
Choice of therapy — All patients with peptic ulcers should receive antisecretory
therapy with a proton pump inhibitor (PPI) (eg, omeprazole 20 to 40 mg daily or
equivalent) to facilitate ulcer healing (algorithm 1 and table 1). PPI use results in faster
control of peptic ulcer disease symptoms and higher ulcer healing rates as compared
with H2RA as a consequence of stronger acid suppression. PPIs also heal NSAID-
related ulcers more effectively as compared with H2RAs [7]. (See "Antiulcer
medications: Mechanism of action, pharmacology, and side effects", section on
'Indications and comparative efficacy' and "Proton pump inhibitors: Overview of use and
adverse effects in the treatment of acid related disorders", section on 'Pharmacology'.)
Combining PPIs and H2RAs adds to cost without enhancing healing. Although antacids
and sucralfate can heal duodenal ulcers, they are not routinely recommended to treat
peptic ulcers as PPIs heal ulcers more rapidly and to a greater extent.
Duration — The duration of initial antisecretory therapy varies based on the ulcer
characteristics, the underlying etiology (H. pylori, NSAID use) and the presence of ulcer
complications (eg, bleeding, perforation, penetration, or gastric outlet obstruction) [8].
(See "Overview of complications of peptic ulcer disease" and 'Prevention of
recurrence' below.)
H. pylori-positive ulcer
●Uncomplicated ulcer – In patients with uncomplicated ulcers, PPI
(eg, omeprazole 20 mg twice daily) given for 14 days, along with the antibiotic
regimen to treat H. pylori, is usually adequate to induce healing. Additional
antisecretory therapy is not needed in the absence of persistent or recurrent
symptoms following therapy or other indications for maintenance antisecretory
therapy [9,10]. Eradication of H. pylori even without concurrent acid suppression
therapy heals >90 percent of duodenal ulcers [11,12]. (See 'Prevention of
recurrence' below.)
●Complicated ulcer – All patients with complicated peptic ulcers (ulcers with
bleeding, perforation, penetration, or gastric outlet obstruction) should initially
receive acid suppressive therapy with an intravenous PPI. Once patients are
tolerating oral medications, they should be switched to an oral PPI at high-dose
twice daily to enhance healing (eg, omeprazole 40 mg twice daily). Dosing should
generally be reduced to once daily after four weeks. However, in patients with
bleeding, the intravenous PPI can be switched to a lower oral dose (eg, 20 mg
omeprazole once daily) 72 hours after endoscopy, provided there is no evidence
of recurrent bleeding. The duration of treatment depends on the location and
cause of the ulcer, but is typically between 4 and 12 weeks in total.
In patients with complicated duodenal ulcers, we suggest antisecretory treatment for
four to eight weeks. In patients with complicated gastric ulcers, we suggest
antisecretory therapy for a total duration of 8 to 12 weeks. In patients with gastric ulcers,
we discontinue antisecretory therapy only after ulcer healing has been confirmed by
upper endoscopy. (See "Overview of complications of peptic ulcer disease", section on
'Acid suppressive therapy' and 'Gastric ulcers' below.)
NSAID-induced ulcer — Patients with NSAID-associated ulcers should be treated with
a PPI (eg, omeprazole 20 to 40 mg daily) for four to eight to weeks based on the size of
the ulcer (table 1). In patients with peptic ulcers who need to remain on NSAIDs

291
or aspirin, maintenance antisecretory therapy with a PPI (eg, omeprazole 20 mg daily)
can reduce the risk of ulcer complications or recurrence [13]. (See "NSAIDs (including
aspirin): Secondary prevention of gastroduodenal toxicity", section on 'Secondary
prevention of gastroduodenal toxicity' and 'Prevention of recurrence' below.)
Non-H. pylori, non-NSAID ulcer — In patients with H. pylori-negative ulcers that are
not associated with NSAID use, we suggest initial PPI therapy for four weeks for
uncomplicated duodenal ulcers, and eight weeks for a gastric ulcer or any complicated
ulcer before repeat endoscopic evaluation to assess for ulcer healing. Although the
natural history of these ulcers is unclear, it is important to review the patient's history for
the adequacy of H. pylori testing and for a history of NSAID use. Limited data on the
natural history of these ulcers suggest that they may be more difficult to heal and have a
higher rate of recurrence. In the absence of H. pylori and NSAID use, we continue long-
term acid inhibitory therapy with PPIs [14]. (See 'Prevention of recurrence' below
and 'Repeat upper endoscopy in selected patients' below.)
Other general measures — Underlying risk factors for peptic ulcer disease should be
addressed and treated. Given the many benefits of smoking cessation, we advise
patients to stop smoking and advise them to limit alcohol intake to one alcoholic
beverage a day [15]. Evidence to support the use of a bland diet or dietary restrictions in
patients with peptic ulcer disease is lacking. (See "Peptic ulcer disease: Epidemiology,
etiology, and pathogenesis", section on 'Risk factors'.)
SUBSEQUENT MANAGEMENT
Repeat upper endoscopy in selected patients
Duodenal ulcers — Given the low risk of malignancy in patients with duodenal ulcers,
a repeat upper endoscopy is not routinely recommended, but may be performed in
selected patients four weeks after initial treatment unless symptoms persist or recur
(algorithm 1). Endoscopy should be repeated in patients with signs of ongoing bleeding
[16]. The decision to perform a repeat upper endoscopy in patients with duodenal ulcers
of unclear etiology should be individualized. (See "Unusual causes of peptic ulcer
disease".)
Gastric ulcers — We suggest a surveillance endoscopy (with biopsies of the ulcer if
still present) be performed after 8 to 12 weeks in patients with gastric ulcers and any
one of the following (algorithm 1):
●Symptoms persist despite medical therapy.
●Unclear etiology.
●Giant ulcer (>2 cm).
●Biopsies not performed or inadequate sampling on the index upper endoscopy.
To adequately sample a gastric ulcer, we suggest a minimum of four biopsies
should be obtained, with at least one from each of the four quadrants of the ulcer.
Additional biopsies of the edges should be taken with jumbo forceps if there are
endoscopic features of a malignant gastric ulcer.
●Ulcer appears suspicious for malignancy on index upper endoscopy (mass
lesion, elevated irregular ulcer borders, or abnormal adjacent mucosal folds).
Surveillance endoscopy should be considered in patients whose gastric ulcer
appears endoscopically suspicious for malignancy, even if biopsy samples from
the index endoscopy are benign. False-negative biopsy specimen results have
been reported to occur in 2 percent to 5 percent of malignant ulcers.
292
 ●Patients with bleeding ulcers at initial presentation who show signs of continued
bleeding [16].
●Risks factors for gastric cancer (eg, age >50 years, Helicobacter pylori,
immigrants from a region with high prevalence of gastric cancer [eg, Japan,
Korea, Taiwan, Costa Rica], family history of gastric cancer, the presence of
gastric atrophy, adenoma, dysplasia, intestinal metaplasia).
Our approach is not to perform an upper endoscopy for surveillance in patients with
small benign-appearing antral gastric ulcers due to nonsteroidal anti-inflammatory drug
use that have been adequately biopsied and have no evidence of dysplasia or
malignancy if the patient has no risk factors for gastric cancer [17-19].
There are limited prospective outcome data to guide which patients with peptic ulcers
should undergo surveillance endoscopy. Our recommendations are consistent with the
2010 guidelines by the American Society for Gastrointestinal Endoscopy [16]. The
incidence of gastric cancer on follow-up endoscopy of an apparently benign gastric
ulcer ranges from 0.8 to 4.3 percent. The rationale behind endoscopic follow-up of a
patient with a gastric ulcer is that the absence of symptoms does not reliably exclude
malignancy, and surveillance endoscopy may identify patients with gastric cancer at an
early stage [20,21]. However, it is unclear if endoscopic surveillance is cost-effective or
improves survival [20,22].
Refractory ulcers — A refractory peptic ulcer is defined as an endoscopically proven
ulcer greater than 5 mm in diameter that does not heal after 8 to 12 weeks of treatment
with a proton pump inhibitor (PPI). Approximately 5 to 10 percent of ulcers are refractory
to initial PPI therapy. The evaluation and management of refractory ulcers is discussed
elsewhere. (See "Approach to refractory peptic ulcer disease".)
PREVENTION OF RECURRENCE
Aspirin/nonsteroidal anti-inflammatory drug (NSAID) avoidance — The need
for aspirin and NSAIDs should be carefully assessed in patients with a history of peptic
ulcer disease. NSAIDs, including aspirin, increase the risk of bleeding in patients with
prior peptic ulcer disease. For patients who must continue aspirin or NSAIDs, or both,
we recommend treatment with a proton pump inhibitor (PPI) for as long as the NSAID
and/or aspirin is used. (See "NSAIDs (including aspirin): Secondary prevention of
gastroduodenal toxicity".)
Helicobacter pylori (H. pylori) eradication — In addition, H. pylori infection should be
eradicated, if known to be present. For patients with H. pylori infection who cannot
discontinue NSAIDs, rebleeding rates are reduced with eradication of H. pylori infection
in combination with long-term PPI therapy. (See "Indications and diagnostic tests for
Helicobacter pylori infection in adults", section on 'Confirmation of eradication'.)
In patients treated for H. pylori, eradication of infection should be confirmed four or more
weeks after the completion of therapy [4]. The choice of test depends on the need for an
upper endoscopy (eg, follow-up of bleeding peptic ulcer). (See 'Repeat upper
endoscopy in selected patients' above and "Indications and diagnostic tests for
Helicobacter pylori infection in adults".)
PPIs should be held for one to two weeks prior to testing to reduce false-negative
results. Serologic testing should not be performed to confirm eradication as patients
may continue to have antibodies after eradication. Diagnostic evaluation and treatment
of H. pylori are discussed in detail, separately. (See "Indications and diagnostic tests for
293
Helicobacter pylori infection in adults", section on 'Medications that should be
discontinued prior to testing'.)
Maintenance antisecretory therapy — The decision to continue maintenance
antisecretory therapy varies based on the patient and ulcer characteristics and risk
factors for recurrent peptic ulcer disease [8].
We continue maintenance antisecretory therapy with a PPI (eg, omeprazole 20 mg
daily) in the following patients with peptic ulcer disease (table 1) [23-28]:
●Giant (>2 cm) peptic ulcer and age >50 years or multiple co-morbidities
●H. pylori-negative, NSAID-negative ulcer disease
●Failure to eradicate H. pylori (including salvage therapy)
●Frequently recurrent peptic ulcers (>2 documented recurrences a year)
●Continued NSAID use
The prevention of recurrent gastroduodenal toxicity associated with NSAID, low-
dose aspirin or both therapies (secondary prevention) is discussed in detail, separately.
(See "NSAIDs (including aspirin): Secondary prevention of gastroduodenal
toxicity" and "Overview of the treatment of bleeding peptic ulcers", section on
'Resumption of anticoagulants and antiplatelet agents'.)
TREATMENT DURING PREGNANCY AND LACTATIONWhen peptic ulcer
disease is diagnosed in a woman who is pregnant, the focus of treatment is typically
acid suppression with a proton pump inhibitor (PPI) [29]. If Helicobacter pylori (H. pylori)
is present, antimicrobial treatment is typically deferred until after delivery. However,
there is some evidence that H. pylori can cause severe nausea/vomiting in pregnancy,
including hyperemesis gravidarum [30,31]. Thus, if indicated, H. pylori treatment can be
considered in pregnancy. A meta-analysis of safety studies showed no significant
adverse outcomes with PPI use in pregnant women [32-34]. The available evidence
based on clinical studies in pregnancy supports the safety of older proton pump
inhibitors such as omeprazole and pantoprazole with little data available on the newer
proton pump inhibitors. Similarly, limited data with omeprazole and pantoprazole
suggest that excretion in breast milk does occur but the levels are low [35-37].
(See "Medical management of gastroesophageal reflux disease in adults", section on
'Pregnancy and lactation' and "Prenatal care: Patient education, health promotion, and
safety of commonly used drugs", section on 'Antibiotics'.)
DISEASE COURSEApproximately 60 percent of peptic ulcers heal spontaneously
but with eradication of Helicobacter pylori (H. pylori) infection, ulcer healing rates are
>90 percent [5,38-41]. Approximately 5 to 10 percent of ulcers are refractory to
antisecretory therapy with a proton pump inhibitor (PPI). The risk of complications in
patients with chronic peptic ulcer disease is 2 to 3 percent per year. (See "Approach to
refractory peptic ulcer disease".)
Even with continued PPI use, approximately 5 to 30 percent of peptic ulcers recur within
the first year based on whether H. pylori has been successfully eradicated [42,43].
Recurrent ulcers are usually due to H. pylori infection or nonsteroidal anti-inflammatory
drug use but in rare cases may be due to other etiologies (table 2).
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Peptic ulcer disease".)

294
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Peptic ulcers (The Basics)" and "Patient

education: H. pylori infection (The Basics)" and "Patient education: Gastritis (The
Basics)")
●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Patients with peptic ulcer disease should be tested for Helicobacter pylori (H.
pylori). Patients with H. pylori should be treated with a goal of eradication of H.
pylori infection. In patients treated for H. pylori, eradication of infection should be
confirmed four or more weeks after the completion of eradication therapy.
(See "Indications and diagnostic tests for Helicobacter pylori infection in adults".)
●Patients with peptic ulcers should be advised to avoid nonsteroidal anti-
inflammatory drugs (NSAIDs). Contributing factors should be addressed and
treated (eg, treating medical comorbidities, poor nutritional status, ischemia).
(See 'Rare or unclear cause' above and "Unusual causes of peptic ulcer disease",
section on 'Etiology' and "Unusual causes of peptic ulcer disease", section on
'Evaluation of H. pylori and NSAID negative ulcers'.)
●All patients with peptic ulcer disease should receive antisecretory therapy to
facilitate ulcer healing (algorithm 1). The choice and duration of therapy varies
based on the etiology, ulcer location (eg, gastric or duodenal), and the presence of
ulcer complications (eg, bleeding, perforation, penetration, or gastric outlet
obstruction). (See 'Initial antisecretory therapy' above.)
●Patients with duodenal ulcers who have been treated do not need further
endoscopy unless symptoms persist at four weeks or recur. (See 'Duodenal
ulcers' above.)
●In patients with gastric ulcers, the decision to perform surveillance endoscopy
should be individualized. We suggest surveillance endoscopy (with biopsies of the
ulcer if still present) be performed after 8 to 12 weeks of antisecretory therapy in
patients with gastric ulcers and any one of the following (see 'Repeat upper
endoscopy in selected patients' above):

295
 •Symptoms despite medical therapy
•Unclear etiology
•Giant gastric ulcer (>2 cm)
•Biopsies not performed or inadequate sampling on the index upper
endoscopy
•Ulcer appears suspicious for malignancy on index upper endoscopy (mass
lesion, elevated irregular ulcer borders, or abnormal adjacent mucosal folds)
•Patients with bleeding ulcers at initial presentation who show signs of
continued bleeding
•Risks factors for gastric cancer
●Maintenance antisecretory therapy should be limited to high-risk subgroups of
patients with peptic ulcer disease. These include individuals with any one of the
following (see 'Prevention of recurrence' above):
•Refractory peptic ulcer
•H. pylori-negative, NSAID-negative ulcer disease
•Giant (>2 cm) ulcer and age >50 years or multiple comorbidities
•Failure of H. pylori eradication, including rescue therapies
•Frequently recurrent peptic ulcers (>2 documented recurrences a year)
•Continued NSAID use
●Approximately 60 percent of peptic ulcers heal spontaneously but with
eradication of H. pylori infection, ulcer healing rates are >90 percent. Even with
continued proton pump inhibitor (PPI) use, approximately 5 to 30 percent of peptic
ulcers recur within the first year based on whether H. pylori has been successfully
eradicated. Approximately 5 to 10 percent of ulcers are refractory to antisecretory
therapy with a PPI. The risk of complications in patients with chronic peptic ulcer
disease is 2 to 3 percent per year. (See 'Disease course' above and "Approach to
refractory peptic ulcer disease" and "Overview of complications of peptic ulcer
disease".)

296
Postgastrectomy complications
Author:
Stanley W Ashley, MD
Section Editor:
David I Soybel, MD
Deputy Editor:
Wenliang Chen, MD, PhD
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Nov 10, 2020.
INTRODUCTIONVarious forms of gastric resection and reconstruction are used to
manage a variety of benign and malignant conditions of the stomach. Similar to any
other abdominal surgery, gastric surgery can result in postoperative complications.
Although complications are considerably less common today, historical data would
suggest that approximately one in four patients reports significant symptoms after
gastric surgery; in 2 to 5 percent, these symptoms are disabling [1].
The complications specific to gastric surgery will be reviewed in this topic.
Complications common to all abdominal surgeries, such as bleeding, infection, bowel
obstruction, fascia dehiscence, or hernia, are discussed in another topic
(see "Complications of abdominal surgical incisions"). Complications of bariatric surgical
procedures (eg, Roux-en-Y gastric bypass or sleeve gastrectomy) are discussed
elsewhere. (See "Late complications of bariatric surgical operations" and "Bariatric
operations: Perioperative morbidity and mortality".)
POSTGASTRECTOMY ANATOMYThe derangements in gastrointestinal
function that occur following gastric resection depend upon the portion and volume of
gastric tissue removed, and the type of reconstruction.
Gastric resection — Gastric resections include (see "Partial gastrectomy and
gastrointestinal reconstruction" and "Total gastrectomy and gastrointestinal
reconstruction"):
●Partial gastrectomy (proximal or distal)
●Total gastrectomy
Reconstruction — Reconstructive techniques include (figure 1 and figure 2)
(see "Total gastrectomy and gastrointestinal reconstruction" and "Partial gastrectomy
and gastrointestinal reconstruction", section on 'Gastrointestinal reconstruction'):
●Billroth I
●Billroth II
●Roux-en-Y
●Esophagojejunostomy
DIAGNOSING POSTGASTRECTOMY COMPLICATIONS Complications
should be suspected in postgastrectomy patients who complain of severe or persistent
gastrointestinal symptoms such as epigastric pain, nausea, vomiting, early satiety,
bloating, diarrhea, or weight loss.

The initial evaluation for the majority of patients presenting with postoperative
gastrointestinal symptoms is abdominal computed tomography (CT).
297
 ●Acute abdomen – In the early postoperative period, patients could develop
symptoms of an acute abdomen (including peritonitis, fever, tachycardia, or
hypotension) secondary to an anastomotic or duodenal stump leak. Such patients
should undergo abdominal CT. Upper gastrointestinal series with water-soluble
contrast is a second choice but usually unnecessary. Although endoscopic
approaches to management are being employed in some patients, this is a
therapeutic and not a diagnostic maneuver. (See 'Leak' below.)
●Obstructive symptoms – Patients with acute obstructive symptoms (eg,
epigastric pain, nausea, or vomiting) should likewise undergo abdominal CT.
Abdominal CT is diagnostic of complications that can lead to upper
gastrointestinal tract obstruction, such as afferent or efferent loop syndrome,
jejunal intussusception, or internal hernia. Patients who present with chronic or
intermittent obstruction symptoms may also undergo upper gastrointestinal series
with barium contrast, which is diagnostic of anastomotic stricture and can also
identify marginal ulceration, remnant cancer, and alkaline reflux gastritis.
(See 'Obstruction' below.)
●Chronic dysmotility – Patients with symptoms suggestive of a motility problem
(early satiety and bloating, postprandial vomiting or diarrhea, weight loss) typically
first undergo abdominal CT and upper endoscopy to rule out mechanical
complications such as an obstruction. Nuclear medicine studies then follow to
make the definitive diagnoses of functional complications. A solid food gastric
emptying study is diagnostic of gastric stasis and Roux stasis syndrome.
Technetium biliary scan may be used to identify alkaline gastritis. (See 'Slow
transit' below.)
●Dumping syndrome – Patients with dumping syndrome present with
gastrointestinal discomfort, including nausea, vomiting, cramps, and diarrhea, as
well as vasomotor symptoms such as diaphoresis, palpitations, and flushing 15 to
30 minutes after a meal. The diagnosis of dumping syndrome is made primarily on
clinical grounds, supported by tests such as upper gastrointestinal series and
gastric emptying studies. (See 'Dumping syndrome' below.)
COMPLICATIONS RELATED TO ANASTOMOSISAnastomotic
complications in postgastrectomy patients include anastomotic leak, anastomotic
stricture, and anastomotic ulceration. In addition, postgastrectomy obstruction can occur
anatomically at the level of the gastrointestinal anastomosis or a short distance away
from the anastomosis (eg, afferent or efferent loop).
Leak — Postoperative leak can arise from any of the suture or staple lines, including
the jejunojejunal anastomosis of a Roux-en-Y [2]. An anastomotic leak most commonly
occurs within the first 7 to 10 days after surgery. Patients present with fever,
unexplained tachycardia and/or hypotension, abdominal pain, and/or an acute
abdomen. If an anastomotic leak is suspected, computed tomography should be
obtained, which may show indirect evidence of a leak, such as pneumoperitoneum,
extraluminal contrast, inflammatory stranding, fluid collections, and/or abscess. Contrast
studies, such as an upper gastrointestinal series with Gastrografin, may show the leak
directly. (See "Overview of gastrointestinal tract perforation", section on 'Diagnosis'.)
When an anastomotic leak is confirmed, broad-spectrum antibiotics are initiated. Further
management of the leak will depend on the state of the patient and availability of an
298
interventional radiologist to perform a percutaneous drainage procedure. Contained
collections in an otherwise stable patient may be amenable to percutaneous drainage
[3]. The aim of percutaneous drainage is to create a controlled enterocutaneous fistula
to allow the leak to heal over time.
If nonoperative management is unsuccessful at controlling the leak, the patient is
hemodynamically unstable, or diffuse intra-abdominal contamination is suspected, the
patient is taken to the operating room for exploration, drainage, and anastomotic
revision to prevent or control abdominal sepsis. Although there is some experience with
stenting such leaks, most of this has been done after gastric surgery for obesity
(see "Gastrointestinal endoscopy in patients who have undergone bariatric surgery",
section on 'Anastomotic leak and fistula'). Once the leak is definitively controlled,
antibiotic therapy can be tailored according to microbial sensitivities of cultures obtained
at the time of either percutaneous drainage or operative intervention.
Duodenal stump leak — The most feared anastomotic complication following partial
gastrectomy is a breakdown of the duodenal stump closure, or duodenal stump leak,
following a Billroth II or Roux-en-Y type of procedure (figure 1). The most important goal
of treatment of duodenal stump leak is control of sepsis and drainage of the surgical
bed. In addition to carrying out the routine surgical management of an anastomotic leak
(as described above), the surgeon may decide to insert a tube duodenostomy, if
feasible, depending upon the extent of inflammation. (See "Management of duodenal
trauma in adults", section on 'Repair of duodenal injury'.)
Stricture — The Billroth II reconstruction (figure 1) is susceptible to postsurgical
scarring at the gastrojejunostomy site, resulting in gastric outlet obstruction
characterized by chronic or intermittent bloating with nonbilious vomiting [2]. A stricture
is typically diagnosed by upper gastrointestinal series with special attention to the lateral
views. If a stricture is found on fluoroscopy, endoscopic evaluation with biopsies is
warranted to rule out recurrent cancer. Benign strictures can be treated by dilation,
which may need to be repeated on several occasions. Stenting might also be
considered. (See "Gastrointestinal endoscopy in patients who have undergone bariatric
surgery", section on 'Stomal (anastomotic) stenosis'.).
Anastomotic stricture is also seen in 6 to 20 percent of patients who undergo Roux-en-Y
gastric bypass for weight loss (figure 3) [4]. These patients are managed similarly to
patients with strictures after Billroth II reconstruction with endoscopic dilation. (See "Late
complications of bariatric surgical operations", section on 'Stomal stenosis'.)
Obstruction — Postgastrectomy obstruction can occur at or a distance away from the
gastrointestinal anastomosis. Regardless of the level of obstruction, patients typically
present with nausea, vomiting, early satiety, and/or epigastric abdominal pain, with
progression to intolerance of oral intake. The diagnosis may be suspected based upon
presenting clinical features and physical examination and is confirmed by radiologic
evaluation and/or endoscopy. The diagnosis of gastric outlet obstruction is reviewed in
more detail separately. (See "Gastric outlet obstruction in adults", section on
'Evaluation'.)
Afferent and efferent loop syndrome — Afferent and efferent loop syndromes can
develop after Billroth II reconstruction with a gastrojejunostomy (see "Partial
gastrectomy and gastrointestinal reconstruction", section on 'Billroth II'). They are

299
related to mechanical obstruction of the loops by kinking, anastomotic narrowing,
adhesions, or, rarely, anastomotic ulceration (figure 4) [5].
The afferent loop refers to the duodenojejunal loop proximal to the gastrojejunal
anastomosis (figure 5). It is believed that many of these are related to excessive length
of the afferent loop; afferent loop syndrome may be prevented by keeping the distance
from the ligament of Treitz to the gastrojejunostomy <12 to 15 cm. A patient with an
acute afferent loop obstruction presents with acute onset of severe abdominal pain and
vomiting, which requires immediate operation to prevent bowel necrosis or duodenal
blowout [6]. Chronic afferent loop syndrome is typically associated with postprandial
epigastric pain and intermittent projectile bilious vomiting, which leads to resolution of
the pain for a period of up to several days. In patients suspected of having afferent loop
syndrome based upon symptoms (eg, intermittent projectile bilious vomiting), the
detection of a distended afferent loop on abdominal computed tomography is
diagnostic. Surgical revision of the gastrojejunostomy or conversion to a Roux-en-Y
anastomosis is necessary to treat this problem. Alternatively, a Braun's
enteroenterostomy between the afferent and efferent loops may also decompress the
afferent loop (figure 6) [5].
The efferent loop refers to the jejunal segment distal to the gastrojejunostomy that
drains succus entericus away from the stomach (figure 5). Obstruction of the efferent
loop causes gastric outlet obstruction manifested by symptoms of epigastric pain,
distension, and bilious vomiting. When diagnosed by either computed tomography or
upper gastrointestinal series, surgical correction is the treatment of choice for efferent
loop syndrome.
Jejunal intussusception — Although uncommon, the afferent or efferent loop of a
Billroth II reconstruction can intussuscept into the gastric remnant through the
gastrojejunal anastomosis (jejunogastric intussusception) (figure 7). This unusual cause
of gastric outlet obstruction can cause acute-onset bloating and bloody vomiting in
postgastrectomy patients [2].
Intussusception can be seen on abdominal computed tomography, upper
gastrointestinal series, or endoscopy. On upper intestinal series, there is narrowing of
the distal end of the gastric remnant with an unopacified coil-like distention into the
proximal jejunum (image 1). Jejunal intussusception is generally not reducible;
therefore, surgical resection of the intussuscepting small bowel followed by revision of
the gastrojejunostomy or conversion to a Roux-en-Y reconstruction is necessary.
Internal hernia — Internal hernias can cause gastrointestinal obstruction after a Billroth
II or Roux-en-Y gastrectomy. Patients with internal hernias usually present with acute
abdominal pain with or without abdominal distention or vomiting. Diagnosis is made by
computed tomography. Early surgical intervention is necessary to avoid small bowel
infarction as internal hernias often cause closed-loop small bowel obstruction. Internal
hernias can be prevented at the time of gastric surgery by careful closure of all
mesenteric defects. Specifically, the mesocolon should be sutured to the stomach at the
gastrojejunostomy, and the space between the mesentery of the retrocolic jejunal limb
and mesocolon (ie, Peterson's defect) needs to be closed. (See "Late complications of
bariatric surgical operations", section on 'Internal hernias'.)
Marginal ulcer — After gastric surgery, patients can develop ulcers in the remnant
stomach, duodenum, or jejunum [2,5]. The most common ulcers, marginal ulcers, occur

300
in the jejunum distal to the gastrojejunal anastomosis. The diagnosis and treatment of
marginal ulcers after Roux-en-Y gastric bypass surgery for weight loss are discussed
elsewhere. (See "Late complications of bariatric surgical operations", section on
'Marginal ulcers'.)
In patients who have had peptic ulcer surgery, recurrent ulceration of the stomach or
duodenum may signal incomplete vagotomy, retained gastric antrum, Zollinger-Ellison
syndrome (ie, gastrinoma), nonsteroidal anti-inflammatory drug (NSAID)
use, Helicobacter pylori infection, or cancer. The diagnostic evaluation and treatment of
recurrent ulcer disease is discussed separately. (See 'Peptic ulcer' below
and "Approach to refractory peptic ulcer disease", section on
'Management' and "Approach to refractory peptic ulcer disease", section on 'Address
the etiology and risk factors'.)
COMPLICATIONS RELATED TO MOTILITYFollowing partial gastric
resection, alterations in upper gastrointestinal function inevitably occur because
duodenal or jejunal continuity is lost and the ability of the stomach remnant to function
as a reservoir is impaired (table 1). Termed postgastrectomy syndromes, these
alterations occur in characteristic patterns depending upon the extent of gastric
resection and the type of gastric reconstruction. (See "Partial gastrectomy and
gastrointestinal reconstruction", section on 'Gastrointestinal reconstruction'.)
Rapid transit — The most common symptom that postgastrectomy patients with rapid
transit report is diarrhea. In the postgastrectomy period, diarrhea may be due to
dumping syndrome or postvagotomy diarrhea.
Dumping syndrome — Dumping is a phenomenon usually caused by the destruction
or bypass of the pyloric sphincter. Clinically significant dumping symptoms occur in
about 20 percent of patients after pyloroplasty or distal gastrectomy [7]. Although the
precise mechanism of dumping is incompletely understood, the syndrome is frequently
attributed to the rapid emptying of hyperosmolar chyme (particularly carbohydrates) into
the small bowel [8]. The osmotic gradient is believed to draw fluid into the intestine, and
this may release one or more vasoactive hormones, such as serotonin and vasoactive
intestinal polypeptide. (See "Late complications of bariatric surgical operations", section
on 'Dumping syndrome'.)
●Clinical presentation
•Early dumping – About 15 to 30 minutes after a meal, affected patients
develop gastrointestinal discomfort, including nausea, vomiting, cramps, and
diarrhea, as well as vasomotor symptoms such as diaphoresis, palpitations,
and flushing [9,10]. These symptoms are referred to as early dumping, or
simply dumping syndrome.
•Late dumping – Less often, patients complain of the same constellation of
symptoms hours after eating, the so-called late dumping syndrome. This
phenomenon is not strictly due to alterations of osmotic gradients across the
gastrointestinal (GI) tract but rather is thought to result from hypoglycemia
following a postprandial insulin peak. (See "Postprandial (reactive)
hypoglycemia".)
●Diagnosis – A suggestive pattern of symptoms in a patient who has undergone
gastric surgery should raise the possibility of dumping syndrome. The diagnosis of
dumping syndrome is made primarily on clinical grounds [11]. A monitored
301
 glucose challenge, upper GI series, or gastric emptying studies have been used to
support the diagnosis.
●Treatment – Most patients with dumping can be treated conservatively with
dietary changes (frequent small meals that are high in fiber and protein and low in
carbohydrates, separation of liquid from solid during meals) [9,10]. Symptoms
tend to resolve in most patients as they learn to avoid foods that aggravate the
problem (eg, simple sugar).
Octreotide may also help in severe cases of dumping but is rarely required [8]. A
study of 30 patients with dumping treated with either subcutaneous octreotide,
administered three times a day, or its long-acting formulation (Octreotide LAR),
which is given monthly, reported that both significantly reduced dumping
symptoms and improved quality of life [12]. Patients preferred monthly treatment.
The rare patient with intractable dumping symptoms who fails dietary and medical
therapy may require reoperation [9,10]. In patients who had a distal gastrectomy,
conversion from a loop gastrojejunostomy to a Roux-en-Y reconstruction is the
procedure of choice. This operation slows gastric emptying by impairing motility of
the Roux loop. A gastric remnant of no more than 25 percent should be left to
avoid postoperative Roux stasis syndrome. (See 'Roux stasis syndrome' below.)
Postvagotomy diarrhea — Diarrhea develops in approximately 30 percent of patients
after truncal vagotomy [7]. The pathogenesis is unclear, but it may be related to the
rapid passage of unconjugated bile salts from the denervated biliary tree into the colon,
where they stimulate secretion. Most cases are self-limited. Oral cholestyramine, which
binds bile salts, can be effective in persistent cases. If medical therapy fails, the surgical
option used in the past is to place a 10 cm reversed (antiperistaltic) jejunal loop in
continuity 100 cm distal to the ligament of Treitz; this is rarely needed today.
Slow transit — Postgastrectomy patients with slow transit often present with symptoms
of nausea, vomiting (bilious or nonbilious), epigastric pain or bloating, and early satiety,
leading to weight loss over time. The differential diagnosis of these symptoms includes
gastric stasis, alkaline gastritis, and Roux-stasis syndrome.
Gastric stasis — After gastric surgery, impaired gastric emptying may develop as a
result of postsurgical atony or vagal denervation, or from a small gastric remnant [5].
Symptoms consist of epigastric fullness with meals (early satiety), often followed by
emesis of undigested food, abdominal pain, and weight loss.
The evaluation of a patient suspected of postgastrectomy gastric stasis syndrome
begins with an upper GI series with small bowel series to define postsurgical anatomy
and rule out mechanical obstruction. Upper endoscopy is also typically performed to
rule out anastomotic strictures or marginal ulcers, which could cause or exacerbate
gastric stasis. Upper endoscopy can also disimpact food bezoars, which are commonly
found in chronic gastric stasis patients [13] (see "Gastric bezoars", section on 'Gastric
dysmotility'). Impaired gastric emptying is best diagnosed quantitatively with a nuclear
medicine solid food emptying test.
Symptoms attributed to a small gastric remnant will usually improve with small, frequent
feedings and time to allow the remnant stomach to accommodate. Postoperative gastric
atony may respond to prokinetic agents such as metoclopramide and erythromycin.
Although there is some evidence that gastric pacing may improve the symptoms of

302
primary gastroparesis, its widespread clinical use is not yet achieved. (See "Treatment
of gastroparesis" and "Electrical stimulation for gastroparesis".)
If dietary and medical therapies fail, reoperative gastrectomy may be required. Patients
without prior partial gastrectomy should undergo subtotal (75 percent) gastrectomy;
patients with prior partial gastrectomy should undergo near-total (95 percent)
gastrectomy or total gastrectomy with esophagojejunostomy [14]. A Billroth II
reconstruction with a Braun enteroenterostomy may be preferred to a Roux-en-Y
reconstruction because of potential Roux stasis syndrome associated with the latter
(figure 6). (See 'Roux stasis syndrome' below.)
Alkaline gastritis — Reflux of bile into the stomach is common after operations that
remove or bypass the pylorus. In most patients, there are no serious clinical sequelae
[5,15]. However, approximately 2 percent of patients develop alkaline reflux gastritis, a
syndrome of persistent burning epigastric pain and chronic nausea that is aggravated
by meals. The diagnosis is made primarily by excluding other causes of symptoms,
although endoscopy may reveal gastritis, and technetium biliary scan can demonstrate
excessive reflux of bile into the stomach.
A variety of medical therapies for alkaline gastritis have been reported, but none have
proven particularly effective (see "Acute hemorrhagic erosive gastropathy and reactive
gastropathy"). Surgical therapies aim at separating the remnant stomach from duodenal
content by interposing a loop of jejunum between them. Examples include Roux-en-Y
reconstruction (with a 45 to 60 cm Roux loop), Henley loop (interposition of a 40 cm
isoperistaltic jejunal loop between the gastric remnant and the duodenum), Billroth II
reconstruction with Braun enteroenterostomy (positioned 45 to 60 cm from the
gastrojejunal anastomosis). The reoperative procedure is chosen based upon a
patient's existing anatomy and how much remnant stomach is left [5].
Roux stasis syndrome — Roux-en-Y reconstruction is typically performed as the
primary reconstruction after near-total or total gastrectomy. It is also used as
reoperative treatment for patients with intractable dumping syndrome, severe alkaline
gastritis, or afferent loop syndrome. (See 'Dumping syndrome' above and 'Alkaline
gastritis' above and 'Afferent and efferent loop syndrome' above.)
After undergoing Roux-en-Y reconstruction, a subset of patients develop symptoms of
vomiting, epigastric pain, and weight loss (ie, Roux stasis syndrome). Roux stasis
syndrome is thought to be caused by disordered motility of the Roux loop with net
propulsive activity toward, instead of away from, the stomach. Patients suspected of
Roux stasis syndrome should undergo evaluation with upper GI series, upper
endoscopy, and nuclear medicine gastric emptying study [16]. The findings are similar
to those with gastric stasis (see 'Gastric stasis' above). One important exception is the
finding of a dilated, often flaccid Roux loop.
Medical treatment of Roux stasis syndrome consists of prokinetic agents such
as metoclopramide and erythromycin. When medical therapy fails, surgical therapy
consists of resecting the exiting Roux loop and replacing it with a new Roux-en-Y
reconstruction. To prevent recurrence, further resection of the remnant stomach (near-
total, or 95 percent gastrectomy) is also carried out. Because Roux stasis syndrome is
seen more often in patients with a generous (over 50 percent) gastric remnant, as well
as in patients with truncal vagotomy, alternative reconstructive techniques such as

303
Billroth II reconstruction with or without Braun enteroenterostomy should be used to
avoid Roux-stasis syndrome, when feasible (figure 6) [5].
Gallstones — Gallstones can develop after gastrectomy for cancer due to cholestasis,
which is caused by decreased gallbladder contraction from various mechanisms,
including weight loss, vagotomy, lymph node dissection in the hepatogastric ligament,
and nonphysiologic reconstruction [17].
In one trial, the proportion of patients developing gallstones within 12 months after
gastrectomy was reduced with daily ursodeoxycholic acid (5.3 percent in the 300 mg
group, 4.3 percent in the 600 mg group), compared with placebo (16.7 percent) [18].
LONG-TERM COMPLICATIONS INVOLVING REMNANT
STOMACHFollowing partial gastrectomy, the remnant stomach is susceptible to
developing ulcer disease or cancer. Depending upon the indication for the initial
operation, these ulcers or cancer can be de novo or recurrent.
Peptic ulcer — Gastric or duodenal ulcers can recur because of surgical or medical
reasons [19]. In postgastrectomy patients, retained gastric antrum and incomplete
vagotomy are the two main surgical causes of recurrent peptic ulcer disease. Patients
who had Billroth II gastrectomy can have retained gastric antrum left on the duodenal
margin. Because it is out of the acid stream, the retained gastric antrum secretes an
excessive amount of gastrin, which in turn hyperstimulates gastric acid production,
causing ulceration. In patients with incomplete vagotomy, vagal stimulation of gastric
acid production is not completely abolished, which results in hyperacidity and recurrent
peptic ulcer.
To rule out retained gastric antrum, we typically measure a fasting serum gastrin level.
A high level of gastrin can stimulate overproduction of gastric acid. A secretin test is
then required to differentiate between retained gastric antrum and Zollinger-Ellison
syndrome as the cause for hypergastrinemia. Retained gastric antrum is diagnosed by
an elevated fasting serum gastrin level that is suppressed by intravenous secretin
administration. In contrast, serum gastrin level rises further with secretin in Zollinger-
Ellison syndrome patients (see "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis", section on 'Secretin stimulation test'). Treatment for
retained gastric antrum is surgical resection.
If serum gastrin is not elevated, one may measure gastric acid output at baseline and
after a sham feeding in order to rule out incomplete vagotomy. Compared with basal
acid output, a large increase in acid output stimulated by sham feeding suggests
incomplete vagotomy. Patients with incomplete vagotomy can usually be treated with a
proton pump inhibitor, although in the past, thoracoscopic truncal vagotomy was also
used [20]. (See "Vagotomy".)
Postgastrectomy patients who present with peptic ulcer disease also require an
evaluation to assess for medical causes. These include medications (eg, nonsteroidal
anti-inflammatory drugs), H. pylori infection, Zollinger-Ellison syndrome, and gastric
remnant cancer. The evaluation for medical causes of peptic ulcer disease is reviewed
separately. (See "Approach to refractory peptic ulcer disease".)
Remnant cancer — Patients with a previous partial gastrectomy for benign diseases
are at an increased risk for developing gastric cancer [21]. These gastric remnant or
"stump" carcinomas generally involve the distal extent of the gastric remnant near the

304
gastrojejunal anastomosis [22]. Chronic reflux of bile and pancreatic secretions leading
to chronic inflammation is thought to play a role.
Reported frequencies of gastric remnant carcinoma range from 0.8 to 8.9 percent [23-
31]. Endoscopic screening studies have detected gastric cancer in 4 to 6 percent of
partial gastrectomy patients over time. However, this increased risk has not been
uniformly reported [32,33]. Studies that have demonstrated an increased risk of gastric
remnant cancer suggest that the risk appears to increase 15 to 20 years after the initial
surgery [24,26-28,31,34].

Because a benefit has not been established, endoscopic surveillance of

postgastrectomy patients is not mandatory. However, if surveillance endoscopy is
considered, it is typically initiated after an interval of 15 to 20 years from the time of
gastric surgery. At the time of endoscopy, multiple biopsies from the anastomosis and
gastric remnant should be taken. With or without surveillance, postgastrectomy patients
should also undergo prompt endoscopic evaluation for any significant upper
gastrointestinal symptoms (eg, nausea, vomiting, early satiety, abdominal pain).

NUTRITIONAL DEFICIENCIESThe surgeon should be aware of possible

nutritional deficiencies that can develop following partial gastrectomy, including
malabsorption of vitamins or minerals (table 2) [35-38]. Nutritional deficiencies following
gastric resection are similar to those that can manifest following gastric surgery to
manage obesity, which is discussed elsewhere. (See "Bariatric surgery: Postoperative
nutritional management".)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Gastric surgery for cancer".)
SUMMARY AND RECOMMENDATIONS
●Complications unique to postgastrectomy patients can be related to the
anastomosis or motility and can also occur in the remnant stomach. The
derangements in gastrointestinal function that occur following gastric resection
depend upon the portion and volume of gastric tissue removed and the type of
reconstruction. (See 'Postgastrectomy anatomy' above.)
●Complications should be suspected in postgastrectomy patients who complain of
gastrointestinal symptoms of epigastric pain, nausea, vomiting, early satiety,
bloating, or weight loss. Most of these patients are evaluated with abdominal
computed tomography (CT), upper gastrointestinal series, and/or upper
endoscopy. (See 'Diagnosing postgastrectomy complications' above.)
●Anastomotic leak should be suspected in patients who present acutely with
severe abdominal pain, fever, or hemodynamic instability. Once diagnosed by
upper gastrointestinal series or CT scan, urgent surgical intervention or
percutaneous drainage is required to control the leak and drain any intra-
abdominal sepsis. (See 'Leak' above.)
●Anastomotic strictures should be suspected in patients who complain of bloating
and nonbilious vomiting. Once diagnosed by upper gastrointestinal series,
endoscopic dilation is the treatment of choice. (See 'Stricture' above.)

305
●Obstruction can occur because of afferent/efferent loop syndrome, jejunal
intussusception, or internal hernia. Patients typically present with nausea,
vomiting, early satiety, and/or epigastric abdominal pain, with progression to
intolerance of oral intake. Once the etiology is diagnosed by CT scan, the
treatment is usually surgical. (See 'Obstruction' above.)
●Dumping is a phenomenon caused by the destruction or bypass of the pyloric
sphincter. Dumping symptoms include gastrointestinal discomfort (eg, nausea,
vomiting, cramps, and diarrhea) as well as vasomotor symptoms (eg, diaphoresis,
palpitations, and flushing). Dietary (carbohydrate avoidance) therapies work well
for dumping syndrome, with only rare patients requiring medical (octreotide)
therapies or referral for surgery (Roux-en-Y reconstruction). (See 'Dumping
syndrome' above.)
●Gastric stasis, which is the more common slow transit complication, may develop
as a result of postsurgical atony, vagal denervation, or from a small gastric
remnant following surgical resection. Symptoms consist of epigastric fullness with
meals (early satiety), often followed by emesis of undigested food, abdominal
pain, and weight loss. Once diagnosed by upper gastrointestinal series, upper
endoscopy, and gastric emptying study, dietary and medical (prokinetic agents)
therapies are initiated. Intractable patients require completion near-total or total
gastrectomy. Other complications related to slow transit include alkaline gastritis
and Roux stasis syndrome. (See 'Slow transit' above.)
●In patients who develop peptic ulcer following gastric surgery, possible etiologies
include retained gastric antrum and incomplete vagotomy. Retained gastric
antrum is diagnosed by an elevated serum gastrin level that is suppressed
by secretin. Incomplete vagotomy is diagnosed by documenting a significant
increase from baseline in gastric acid output following a sham feeding. Treatment
for both conditions is surgical resection. (See 'Peptic ulcer' above.)
●Patients with a previous partial gastrectomy for benign diseases are at an
increased risk for developing gastric cancer after 15 to 20 years. Endoscopic
surveillance is required to diagnose gastric remnant cancer in otherwise
asymptomatic patients. Treatment is by surgical resection. (See 'Remnant
cancer' above.)
●The surgeon should be aware of possible nutritional deficiencies that can develop
following partial gastrectomy, including malabsorption of vitamins or minerals
(table 2). (See "Bariatric surgery: Postoperative nutritional management".)

306
Proton pump inhibitors: Overview of use and
adverse effects in the treatment of acid related
disorders
Author:
M Michael Wolfe, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Nov 29, 2021.
What's New
PPI use and inflammatory bowel disease (November 2021)
Due to their ability to alter the gut microbiome, proton pump inhibitors (PPI) may impact
the risk of inflammatory bowel disease (IBD). In a study that included >600,000
individuals followed for a median of 12 years, regular PPI use was associated with an
increased risk of IBD compared with nonusers; however, the absolute risk of IBD was
low [1]. We continue to recommend prescribing PPIs at the lowest effective dose and for
the shortest duration appropriate to the condition being treated. (See "Proton pump
inhibitors: Overview of use and adverse effects in the treatment of acid related
disorders", section on 'Inflammatory bowel disease'.)
Read more

INTRODUCTIONProton pump inhibitors (PPIs) effectively block gastric acid

secretion by irreversibly binding to and inhibiting the hydrogen-potassium ATPase pump
that resides on the luminal surface of the parietal cell membrane.
This topic review will provide an overview of the mechanism of action,
pharmacokinetics, administration, and adverse effects of PPIs. The use and efficacy of
PPIs in specific acid-related disorders is presented separately. (See "Antiulcer
medications: Mechanism of action, pharmacology, and side effects".)
INDICATIONS FOR PPI THERAPYProton pump inhibitor (PPI) therapy is
indicated in the following clinical situations:
●Peptic ulcer disease – PPIs are first-line antisecretory therapy in the treatment
of peptic ulcer disease. (See "Peptic ulcer disease: Treatment and secondary
prevention", section on 'Initial antisecretory therapy'.)
●Gastroesophageal reflux disease – PPIs are indicated in patients with
gastroesophageal reflux disease, including for the treatment of erosive esophagitis
and as maintenance therapy in patients with severe erosive esophagitis or
Barrett’s esophagus. (See "Medical management of gastroesophageal reflux
disease in adults", section on 'Severe or frequent symptoms or erosive
esophagitis'.)
●Zollinger-Ellison syndrome – PPIs, often in high doses, are required to control
gastric acid hypersecretion in patients with gastrin-secreting

307
 tumors. (See "Management and prognosis of the Zollinger-Ellison syndrome
(gastrinoma)", section on 'Proton pump inhibitors'.)
●NSAID-associated ulcers – PPIs are indicated in the primary prevention of
gastroduodenal ulcers associated with NSAID use. (See "NSAIDs (including
aspirin): Primary prevention of gastroduodenal toxicity".)
●Eradication of Helicobacter pylori – PPIs are a component of several first-line
and salvage therapy regimens for H. pylori infection. (See "Treatment regimens for
Helicobacter pylori in adults".)
PHARMACOLOGY
●Mechanism of action – Proton pump inhibitors (PPIs) inhibit H-K-ATPase, the
final step of gastric acid secretion by parietal cells.
PPIs are benzimidazole prodrugs which accumulate specifically and selectively in
the secretory canaliculus of the parietal cell [1]. Within that space, they undergo
an acid catalyzed conversion to a reactive species, the thiophilic sulfonamides,
which are permanent cations. The rate of conversion varies among the
compounds and is inversely proportional to the pKa of the benzimidazole
(rabeprazole >omeprazole, esomeprazole, and lansoprazole >pantoprazole) (table
1) [2]. The reactive species interacts with the external surface of the H-K-ATPase
that faces the lumen of the secretory space of the parietal cell, resulting in
disulfide bond formation with cysteine 813 located within the alpha-subunit of the
enzyme; this is the residue that is intimately involved in hydrogen ion transport.
This covalent inhibition of the enzyme results in a specific and long-lasting
impairment of gastric acid secretion. (See "Physiology of gastric acid secretion".)
●Pharmacokinetics – PPIs are similar in structure and mechanism of action, but
PPIs differ in their pKa, bioavailability, peak plasma levels, and route of excretion
(table 1). The magnitude of these differences are small and their clinical relevance
has not been established.
PPIs are most effective when the parietal cell is stimulated to secrete acid
postprandially, a relationship that has important clinical implications for timing of
administration. Because the amount of H-K-ATPase present in the parietal cell is
greatest after a prolonged fast, PPIs should be administered before the first meal
of the day. In most individuals, once-daily dosing is sufficient to produce the
desired level of acid inhibition, and a second dose, which is occasionally
necessary, should be administered before the evening meal [1]. (See 'Dose and
timing of administration' below.)
Once-daily PPI dosing for five days inhibits maximal gastric acid output by
approximately 66 percent. Since PPIs inhibit only activated enzyme present in the
canalicular membrane, the reduction of gastric acid secretion after an initial dose
will probably be suboptimal. As inactive enzyme is recruited into the secretory
canaliculus, acid secretion will resume, albeit at a reduced level. After the second
dose is given on the next day, more H-K-ATPase will have been recruited and
subsequently inhibited, and after the third dose, additional recruitment and further
acid inhibition will probably occur. Thus, the occasional use of a PPI taken on an
"as needed" basis does not reliably provide adequate acid inhibition and does not
produce a consistent or satisfactory clinical response (in contrast to the H2
antagonists, which have a more rapid onset of action) [1].
308
 Restoration of acid secretion after discontinuing PPIs depends upon enzyme
turnover and the biological reversibility of the disulfide bond. Maximal acid
secretory capacity may not be restored for 24 to 48 hours [1]. (See 'Discontinuing
PPIs' below.)
●Metabolism – PPIs are metabolized via hepatic cytochrome P450 enzymes, with
CYP2C19 having the dominant role. However, the dominance of CYP2C19 over
other pathways varies significantly among the PPIs (table 2). The activity of
CYP2C19 is also determined to some extent by genetic polymorphisms. Two
inactivating mutations have been described as a result; the metabolism of PPIs by
this route may be delayed in these individuals [3,4]. Homozygotes for the wild type
gene rapidly metabolize these drugs, while heterozygotes are intermediate
metabolizers.
Plasma levels of PPI correlate with their metabolism, and differences may
contribute to varying dose requirements and clinical efficacy. As an example, one
study examined the effect of variable metabolism of omeprazole when using this
agent to treat H. pylori in 62 Japanese patients [3]. While eradication was
achieved in all individuals homozygous for a CYP2C19 mutation (ie, slow
metabolizers), successful treatment was achieved in only 60 and 29 percent of
heterozygotes and wild type homozygotes, respectively. In another study that
evaluated the efficacy of lansoprazole in the treatment of 65 patients with
gastroesophageal reflux disease (GERD), slow metabolizers were much more
likely to be asymptomatic as compared with heterozygotes and wild type
homozygotes (85 versus 68 and 46 percent, respectively) [5]. The response rate
in wild type homozygotes with severe GERD was only 16 percent. Wild type
homozygotes (rapid metabolizers) also had the lowest plasma lansoprazole
concentrations. (See 'Dose and timing of administration' below.)
PRETREATMENT CONSIDERATIONS AND MONITORING
Drug interactions — Clinically important drug interactions with proton pump inhibitors
(PPIs) are rare. However, PPI metabolism via hepatic cytochrome P450 enzymes may
lead to specific drug interactions in some individuals. The presence of a CYP2C19 gene
mutation results in higher plasma PPI levels in homozygous individuals. However, if this
metabolic pathway becomes saturated, the isoenzyme can become a major target for
interactions with many drugs, including warfarin, diazepam, clopidogrel,
and phenytoin (table 3). CYP3A4-mediated metabolism may also be activated under
such conditions and become the principal route of drug elimination. Furthermore,
induction of CYP1A, another P450 isoenzyme, in CYP2C19 deficient or saturated
individuals, can make them susceptible to interference with theophylline metabolism.
The specific P450 enzymes involved in PPI metabolism and the potential for
interactions among PPIs varies considerably (table 2) [6-12].

Some other important drug interactions with PPIs include the following:

●Clopidogrel - Some data suggest decreased activation of clopidogrel when used

in conjunction with omeprazole due to shared hepatic cytochrome P450-mediated
metabolism. In 2009, the United States Food and Drug Administration concluded
that patients taking clopidogrel should consult with their clinician if they are taking
or considering taking a PPI, including over-the-counter PPI preparations [13,14].
309
 However, the relevance of these data remains highly controversial. The interaction
of clopidogrel and PPIs are discussed in detail separately. (See "Clopidogrel
resistance and clopidogrel treatment failure".)
●HIV protease inhibitors - PPIs may decrease the absorption of certain HIV
protease inhibitors. PPIs are contraindicated in patients being treated
with rilpivirine. Atazanavir should not be used in patients who require a PPI dose
equivalent to >20 mg omeprazole daily. (See "Overview of antiretroviral agents
used to treat HIV", section on 'Protease inhibitors (PIs)'.)
●Methotrexate - Coadministration of PPIs with high dose methotrexate appears to
be correlated with delayed methotrexate elimination and potentially may lead to
methotrexate toxicity if not monitored appropriately.
For additional information on drug interactions, use the Lexicomp drug
interactions program provided by UpToDate. (See "Overview of the non-acute
management of unstable angina and non-ST elevation myocardial infarction", section on
'Gastrointestinal prophylaxis'.)
Laboratory testing — We limit routine laboratory testing to selected patients on PPI
therapy.
●Magnesium – We obtain serum magnesium levels prior to starting a PPI in
patients who are expected to be on long-term (≥1 year) treatment or in patients
who take PPIs in conjunction with other medications associated with
hypomagnesemia (eg, diuretics). In addition, we obtain magnesium levels
periodically in such patients while they are taking a PPI. The frequency of testing
is based on the clinical history and the presence of symptoms of
hypomagnesemia. As an example, in patients with a history of arrhythmias or QT
interval prolongation, we monitor magnesium levels every six months. The
management of hypomagnesemia is discussed in detail separately.
(See "Hypomagnesemia: Evaluation and treatment".)
●Vitamin B12 – We also obtain vitamin B12 levels yearly in patients on long-term
PPIs [15]. However, routinely monitoring vitamin B12 levels is controversial.
(See 'Magnesium malabsorption' below and 'Vitamin B12 malabsorption' below.)
There are insufficient evidence to support routine bone density monitoring or calcium
supplementation due to proton pump inhibitor use alone [16].
ADMINISTRATION
Intravenous regimen — IV PPIs are indicated prior to endoscopic evaluation in
patients with clinically significant upper gastrointestinal bleeding from a suspected
peptic ulcer. Pantoprazole and esomeprazole are the only PPIs available as an IV
formulation in the United States; IV omeprazole is available in other countries. The use
of PPIs in the treatment of bleeding peptic ulcers and the duration of treatment is
discussed in detail separately. (See "Overview of the treatment of bleeding peptic
ulcers", section on 'Oral versus intravenous dosing' and "Approach to acute upper
gastrointestinal bleeding in adults", section on 'Acid suppression'.)
Oral regimen
Selecting a PPI — The choice of a specific oral PPI and whether over-the-counter
(rather than prescription) PPIs are prescribed are often determined by patient
preference and payer coverage. A systematic review of 12 randomized trials examining

310
the relative effectiveness of different PPI doses and dosing regimens found no
consistent difference in symptom resolution and esophagitis healing rates [17].
In patients unable to swallow pills or capsules, options include an oral suspension
of lansoprazole and a powder formulation of omeprazole-sodium bicarbonate for oral
suspension.
Dose and timing of administration — PPIs should be administered 30 to 60 minutes
before breakfast for maximal inhibition of proton pumps. (See 'Pharmacology' above.)
Dose reduction, particularly for maintenance of healing of erosive esophagitis may be
possible in Asian populations. Polymorphisms in the CYP2C19 gene, which encodes
the cytochrome P450 isoenzyme that metabolizes different PPI preparations, are
common in Asian and other populations [18]. Such gene mutations would render an
individual a "slow metabolizer" and prolong the antisecretory effect of PPIs. In contrast,
the duration of acid inhibition would be decreased in a "rapid metabolizer," and
differences in PPI metabolism might account for incomplete inhibition of acid secretion
and a high prevalence of nocturnal breakthrough symptoms in gastroesophageal reflux
disease patients. (See 'Pharmacology' above and "Approach to refractory
gastroesophageal reflux disease in adults", section on 'Differences in PPI metabolism'.)
Avoidance of concurrent antisecretory agents — PPIs should not administered
concomitantly with antisecretory agents including histamine-2 receptor antagonists
(H2RAs), analogues of prostaglandin E (eg, misoprostol), and somatostatin analogues
(eg, octreotide), because of the marked reduction in acid inhibitory effects [1,19].
Antisecretory drugs can be used with a PPI provided that there is a sufficient time
interval between their administration. As an example, an H2RA can be taken before
bedtime or during the night by individuals who report nocturnal breakthrough symptoms
such as heartburn after taking a PPI in the morning or before dinner.
Switching between PPIs — Switching PPIs is a reasonable strategy in patients with
side-effects to an individual PPI and may be necessary due to cost differences.
Although there is significant interindividual and intraindividual variability in intragastric
pH control between PPIs, there are no consistent difference in relation to symptom
resolution and esophagitis healing rates [17]. Switching PPIs in patients with well-
controlled symptoms may also be associated with increased symptom severity and
decreased patient satisfaction [20]. (See "Approach to refractory gastroesophageal
reflux disease in adults", section on 'Subsequent management'.)
Discontinuing PPIs — PPIs should be prescribed at the lowest dose and for the
shortest duration appropriate to the condition being treated. (See 'Indications for PPI
therapy' above.)
We gradually taper PPI therapy in patients treated with PPIs for longer than six months.
For patients on a standard or high-dose PPI (eg, omeprazole 40 mg daily or twice daily),
we decrease the dose by 50 percent every week. For patients on twice daily dosing, the
initial reduction can be accomplished by decreasing the dosing to once in the morning
before breakfast until the patient is on the lowest dose of the medication. Once on the
lowest dose for one week, the patient is instructed to discontinue the PPI. However, no
specific method for discontinuing PPI therapy has been proven effective, and no
approach is universally accepted. [21,22].
Studies have demonstrated rebound gastric acid hypersecretion following
discontinuation of PPIs in patients with long-term use. The reasons are not entirely

311
clear, but appear to be due in part to the suppression of antral somatostatin expression,
resulting in an increase in antral gastrin release and subsequent disruption of normal
pH-related feedback inhibition of acid secretion that occurs after a meal [1].
(See "Physiology of gastric acid secretion", section on 'Tolerance and acid rebound'.)
ADVERSE EFFECTSLong-term PPI use has been associated with several safety
concerns. However, few of these concerns are supported by consistent data
demonstrating a causal relationship. (See "Physiology of gastrin", section on
'Hypergastrinemia'.)
Gastrointestinal effects
Clostridioides difficile and other enteric infections — PPI use has been associated
with an increased risk of C. difficile infection, even in the absence of antibiotic use [23-
33]. Associations with other enteric infections, including salmonellosis and
campylobacteriosis, have also been reported [34-39]. However, the pathophysiologic
mechanism involved in the increased risk of infection is unclear.
A 2017 meta-analysis of 50 observational studies found that PPI use was significantly
associated with an increased risk of C. difficile infection (relative risk [RR] 1.3; 95% CI
1.1-14). The risk of C. difficile infection appears to be greater with PPIs as compared to
H2 receptor antagonists [30,31].
PPI use has also been associated with an increased risk of recurrent C. difficile infection
[31]. In a 2017 meta-analysis of 16 observational studies that included 7703 patients
with C. difficile infection of whom 1525 (20 percent) had recurrent C. difficile infection,
gastric acid suppression was significantly associated with an increased risk of
recurrent C. difficile infection (odds ratio [OR] 1.5; 95% CI 1.2-1.9) [40]. There was
significant heterogeneity among the studies included in the meta-analysis. In adjusted
analysis using data from nine studies, PPI use was associated with an increased risk of
recurrent C. difficile infection after controlling for patient age and other co-morbid
conditions (OR 1.4; 95% CI 1.1-1.8). (See "Clostridioides difficile infection in adults:
Epidemiology, microbiology, and pathophysiology", section on 'Gastric acid
suppression'.)
Microscopic colitis — PPI use has been associated with microscopic colitis, including
lymphocytic and collagenous colitis. In a case-control study that included 95 cases of
microscopic colitis, exposure to PPIs was significantly higher in patients with
microscopic colitis as compared with controls (38 versus 13 percent, OR 4.5, 95% CI
2.0-9.5) [41]. Similar results have been reported in other case-control studies, however,
it is unclear if this association varies by PPI and if there is a dose-response relationship
in either dose or duration of use [42,43]. (See "Microscopic (lymphocytic and
collagenous) colitis: Clinical manifestations, diagnosis, and management", section on
'Medications'.)
Hypergastrinemia — Induction of hypergastrinemia has been associated with gastric
carcinoid tumors in rats. However, these observations are not generalizable to species
with gastrin physiology more analogous to humans [44]. While patients treated
with omeprazole for up to 11 years have shown some enterochromaffin-like cell
hyperplasia, no dysplasia or neoplastic changes have been observed [45]. An increased
risk of colon cancer due to hypergastrinemia has also not been established [46].
(See "Physiology of gastrin".)

312
Atrophic gastritis — Patients on long-term PPI therapy have a propensity to develop
chronic atrophic gastritis. However, the risk of atrophic gastritis is small, and in the rare
patient who develops atrophic gastritis, the clinical consequences are uncertain
[45,47,48]. (See "Risk factors for gastric cancer".)
Intestinal colonization of multi-drug resistant organisms — PPIs may increase the
risk of intestinal colonization with multi-drug resistant organisms. In a meta-analysis of
12 observational studies that included 22,305 patients, after adjusting for potential
confounders, acid suppression increased the odds of intestinal carriage of multi-drug
resistant organisms of the Enterobacterales order (producing extended-spectrum beta-
lactamases, carbapenemases, or plasmid-mediated AmpC beta-lactamases) and of
vancomycin-resistant enterococci (OR 1.74; 95% CI 1.4-2.2) [49]. Possible mechanisms
include an increase in bacteria that survive transit from the stomach to the intestine due
to reduction in gastric acid by PPIs and direct alteration of the composition of intestinal
microbiota, leading to a decrease in mean species diversity.  
Inflammatory bowel disease — In a study that pooled data from three observational
cohorts and included >600,000 individuals followed for a median of 12 years, the risk of
inflammatory bowel disease (IBD) was increased in regular PPI users as compared with
nonusers (hazard ratio [HR] 1.42; 95% CI 1.22–1.65) [50]. However, the absolute risk of
IBD was low, with a number needed to harm of 3770. In addition, absence of data on
PPI dosing precluded assessment of a dose-response relationship between PPI use
and IBD.
Malabsorption of minerals and vitamins
Magnesium malabsorption — PPIs can cause hypomagnesemia due to reduced
intestinal absorption [51]. A meta-analysis of nine observational studies that included a
total of 109,798 patients found that those who took a PPI had a significantly higher risk
(RR 1.43; 95% CI 1.08-1.88) of developing hypomagnesemia as compared with those
who did not [52]. Clinical manifestations of hypomagnesemia include neuromuscular
excitability (eg, tremor, tetany, convulsions), weakness, and apathy. Severe PPI-
induced hypomagnesemia has been associated with QT interval prolongation and
torsades de pointes [53,54]. The risk of hypomagnesemia appears to be mainly in
patients who have been on PPIs long-term (generally longer than one year) but cases
have been reported within one year of starting PPI therapy [53,55]. This potential risk
has led to recommendations to monitor serum magnesium levels in specific patients at
high risk for hypomagnesemia. Monitoring for hypomagnesemia in patients on PPIs is
discussed in detail separately. (See 'Laboratory testing' above and "Hypomagnesemia:
Clinical manifestations of magnesium depletion", section on 'Overview of clinical
manifestations'.)
Calcium and fracture risk — Although hypochlorhydria could theoretically reduce
calcium absorption, the effect appears to be relevant only for the absorption of water
insoluble calcium (eg, calcium carbonate) and can be overcome by ingestion of a
slightly acidic meal [56]. The absorption of water soluble calcium salts or calcium in
dairy products are not impacted by PPI-induced hypochlorhydria. When calcium
supplementation is necessary in patients taking PPIs, we use calcium supplements that
do not require acid for absorption, such as calcium citrate. (See 'Laboratory
testing' above and "Drugs that affect bone metabolism", section on 'Proton pump
inhibitors'.)

313
PPI-induced hypochlorhydria can augment osteoclastic activity, thereby decreasing
bone density [57,58]. Although an association between PPI use and bone fracture is
plausible, causality has not been established [59]. Nonetheless, the FDA has mandated
revised safety information on all PPIs about a possible increased risk of fractures of the
hip, wrist, and spine with the use of these medications [60]. The association between
PPIs and bone metabolism and risk of fracture are discussed in detail separately.
(See "Drugs that affect bone metabolism", section on 'Proton pump inhibitors'.)
Vitamin B12 malabsorption — Long-term therapy with PPIs has been associated with
vitamin B12 malabsorption [61,62]. However, absorption of oral B12 supplements is not
affected. (See 'Laboratory testing' above and "Treatment of vitamin B12 and folate
deficiencies", section on 'Treatment of vitamin B12 deficiency'.)
Iron malabsorption — Gastric acid plays a role in the absorption of nonheme iron, and
the use of PPIs has been associated with decreased iron absorption [63-67]. However,
in most cases the decreased absorption does not appear to be of clinical significance.
One exception may be in patients who require oral iron supplementation [66,68]. Such
patients may need a higher dose or longer duration of supplementation [66].
(See "Treatment of iron deficiency anemia in adults", section on 'Dosing and
administration (oral iron)'.)
Kidney disease — PPIs can cause acute interstitial nephritis (AIN) [69-72]. Similar to
other cases of drug-induced AIN, AIN due to PPI use is not dose-dependent, and
recurrence or exacerbation can occur with a second exposure to the same or a related
drug. (See "Clinical manifestations and diagnosis of acute interstitial nephritis", section
on 'Drugs'.)
PPI use has also been associated with an increased risk of incident chronic kidney
disease (CKD), CKD progression, and end-stage kidney disease [73-76]. However, the
mechanism underlying the association between PPI use and risk of CKD is not known,
and it is possible that the weak association observed in these studies is due to
methodological limitations (residual confounding) [73,74,77]. Further studies are needed
to help better define an etiologic relationship between PPI use and the development and
worsening of CKD.
Drug-induced lupus — In postmarketing safety surveillance, new onset of cutaneous
lupus erythematosus and systemic lupus erythematosus (SLE), and exacerbation of
existing disease have been reported in patients on PPIs [78-80]. Most cases of CLE-
associated with PPI use are subacute and occur within weeks to years after continuous
PPI therapy. PPI-associated SLE usually occurs days to years after initiating PPI
treatment and typically presents with a rash. Most patients improve within 4 to 12 weeks
of discontinuation of PPI therapy. (See "Drug-induced lupus", section on 'Causative
drugs'.)
Other associations of unclear significance
COVID-19 — It is unclear if PPI use is associated with an increased risk of COVID-19.
In a cross-sectional survey of 86,602 individuals, 53,130 reported prior abdominal pain,
acid reflux, heartburn, and regurgitation symptoms and provided data on H2RA and PPI
use. Of these, 3386 individuals (6.4 percent) self-reported a positive COVID-19 test
[81]. In analyses adjusted for socioeconomic, lifestyle, and clinical comorbidities,
patients who reported PPI use were significantly more likely to report a positive COVID-
19 test with a dose-dependent increase in odds of reporting a positive test (PPI once-

314
daily OR 2.15, 95% CI 1.9-2.4; PPI twice-daily OR 3.7, 95% CI 2.9-4.6). It is possible
that this association is due to residual confounding.
Other studies have demonstrated that patients taking PPIs are at increased risk for
severe clinical outcomes of COVID-19 but have not demonstrated an increase in
susceptibility to SARS-CoV-2 infection [82]. These data require further validation.
Dementia — Although some studies have found a significant association between use
of PPIs and incident dementia, others have not found an association between PPI use
and cognitive function [83-88]. The association between PPI use and dementia may
reflect residual confounding by factors related to both use of PPIs and the development
of dementia and is discussed in detail, separately. (See "Epidemiology, pathology, and
pathogenesis of Alzheimer disease", section on 'Medications'.)
Pneumonia — While observational studies suggest an association between PPI use
and pneumonia, the observed association may be due to confounding such that
individuals prescribed PPIs may be more likely to have other unobserved health
characteristics that predispose them to pneumonia as compared with nonusers [89-96].
(See "Epidemiology, pathogenesis, and microbiology of community-acquired pneumonia
in adults", section on 'Predisposing host conditions' and "Risk factors and prevention of
hospital-acquired and ventilator-associated pneumonia in adults", section on 'Role of
gastric pH'.)
Mortality — It is unclear if PPI use is associated with an increase in risk of death. In an
observational cohort study that included 275,977 new PPI users and 73,335 new
histamine-2 receptor antagonist (H2RA) users, over a median follow-up of 5.7 years, the
incident death rate among new PPI users was higher as compared to those receiving
H2RA (4.5 versus 3.3 per 100 person-years) [97]. New PPI users were significantly
older as compared with new users of H2RAs at the time of study entry (61.7 versus 58.5
years). However, after adjusting for potential confounders, PPI use was associated with
an increase in all-cause mortality as compared with H2RA use (HR 1.25, 95% CI 1.23-
1.28). PPI users also had an increase in risk of death as compared with individuals
without any PPI use and individuals without any acid suppression use (HR 1.15, 95% CI
1.14-1.15; HR 1.23, 95% CI 1.22-1.24, respectively). Among new PPI users, the risk of
death increased with the duration of PPI use. Limitations of the study include its
generalizability as the study cohort primarily consisted of older White males, and lack of
data on the cause of mortality. The underlying basis for this apparent increased risk of
death with PPI use are not known and further studies are needed to evaluate whether
this epidemiologic association is due to unmeasured confounding.
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

315
Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Acid reflux and gastroesophageal reflux

disease in adults (The Basics)")
●Beyond the Basics topics (see "Patient education: Gastroesophageal reflux
disease in adults (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Indications for proton pump inhibitors (PPIs) include the treatment of peptic ulcer
disease, gastroesophageal reflux disease, and Zollinger-Ellison syndrome. PPIs
are effective in the prevention of nonsteroidal anti-inflammatory drug-associated
gastroduodenal mucosal injury and are an important component of several
antimicrobial regimens used in the treatment of Helicobacter pylori infection.
(See 'Indications for PPI therapy' above.)
●PPIs effectively block gastric acid secretion by irreversibly binding to and
inhibiting the hydrogen-potassium ATPase pump that resides on the luminal
surface of the parietal cell membrane. Genetically determined variability in the PPI
metabolism can influence their efficacy. The presence of a CYP2C19 gene
mutation can results in higher plasma PPI levels in homozygous individuals. PPI
metabolism via hepatic cytochrome P450 enzymes may lead to specific drug
interactions in some individuals. However, clinically important drug interactions
with PPIs are rare. (See 'Pharmacology' above.)
●The magnitude of pharmacokinetic differences between PPIs is small and the
clinical relevance of these differences has not been established (table 1).
Intravenous PPI therapy is indicated in patients with clinically significant upper
gastrointestinal bleeding prior to endoscopy for treatment of suspected bleeding
peptic ulcers. (See 'Administration' above.)
●Oral PPIs should be administered 30 to 60 minutes before breakfast for maximal
inhibition of proton pumps. PPIs should not be administered concomitantly with
H2-receptor antagonists, and prostaglandins or somatostatin analogues. In
general, PPIs should be prescribed at the lowest dose and for the shortest
duration appropriate to the condition being treated. When discontinuing PPI
therapy, we taper the dose gradually in patients on PPIs for longer than six
months. (See 'Oral regimen' above.)
●PPI use has been associated with an increased risk
of Clostridioides difficile infection, other enteric infections, and microscopic
colitis. C. difficile infection with diarrhea may occur even in the absence of
antibiotic use. (See 'Gastrointestinal effects' above.)
●PPIs can cause hypomagnesemia due to reduced intestinal absorption. Long-
term therapy with PPIs has been associated with vitamin B12 malabsorption. We
obtain serum magnesium levels prior to starting a PPI in patients who are
expected to be on long term (≥1 year) treatment, or in patients who take PPIs in
conjunction with other medications associated with hypomagnesemia. In addition,
we also monitor magnesium and vitamin B12 levels in patients on long-term PPIs.

316
(See 'Pretreatment considerations and monitoring' above and 'Magnesium
malabsorption' above and 'Vitamin B12 malabsorption' above.)
●Although an association between PPIs and bone fracture is plausible, causality
has not been established. PPIs can decrease the absorption of water insoluble
calcium (eg, calcium carbonate). When calcium supplementation is necessary in
patients taking PPIs, we use calcium supplements that do not require acid for
absorption, such as calcium citrate. (See 'Calcium and fracture risk' above.)
●PPIs can cause acute interstitial nephritis. PPI use has also been associated with
an increased risk of incident chronic kidney disease (CKD), CKD progression, and
end-stage renal disease. However, further studies are needed to help better
define an etiologic relationship between PPI use and the development and
worsening of CKD. (See 'Kidney disease' above.)
●There are conflicting data on the association between PPI use and risk of
dementia and pneumonia. It is also unclear if PPI use is associated with an
increased risk of death. It is possible that these associations are due to residual
confounding and more studies are needed. (See 'Kidney disease' above
and 'Dementia' above and 'Mortality' above.)

317
Surgical management of peptic ulcer disease
Authors:
Ashley H Vernon, MD
Stephen J Ferzoco, MD
Stanley W Ashley, MD
Section Editors:
David I Soybel, MD
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Wenliang Chen, MD, PhD
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Sep 09, 2021.
INTRODUCTIONPeptic ulcer disease was once the most common indication for
gastric surgery but now only infrequently requires operation. Over the last several
decades, the development of potent antisecretory agents (H2 blockers and proton pump
inhibitors) and the recognition that treatment for Helicobacter pylori infection can
eliminate most ulcer recurrences have essentially eliminated the need for elective
surgery [1,2]. However, complications related to peptic ulcer disease continue to occur
and include bleeding, perforation, and gastric outlet obstruction. An understanding of
surgical management remains important since surgery is the mainstay of emergency
treatment of these life-threatening complications and for disease that is refractory to
medical management. Also, there remain a significant number of patients who
underwent surgery prior to the development of current standard medical therapies who
continue to have issues related to their original operation.
The indications for surgery, general principles of ulcer surgery, and respective
treatments for duodenal and gastric ulcers will be reviewed here. The technical aspects
of gastrectomy and vagotomy and their complications are reviewed elsewhere.
(See "Partial gastrectomy and gastrointestinal reconstruction" and "Total gastrectomy
and gastrointestinal reconstruction" and "Vagotomy" and "Postgastrectomy
complications".)
NATURAL HISTORY OF PEPTIC ULCER DISEASEThere has been a
significant decline in the incidence of hospitalization for peptic ulcer disease (PUD)
since 1990 [3,4]. In a study of the Nationwide Inpatient Sample (NIS), hospitalizations
for PUD decreased 30 percent from 1993 to 2006, with the decline being greater for
duodenal compared with gastric ulcers (37 versus 20 percent, respectively) [4]. These
declines are attributed to better medical therapy, including proton pump inhibitors and
regimens that eradicate H. pylori. The natural history and treatment of PUD are
discussed elsewhere. (See "Peptic ulcer disease: Treatment and secondary
prevention".)
Because of the decrease in the hospitalization rate for PUD, surgeons-in-training now
have less exposure to the overall management of PUD, including complications, as well
as some of the more technically demanding procedures for treating PUD, such as highly
selective vagotomy (parietal cell vagotomy) [5,6]. (See "Vagotomy", section on 'Highly
selective vagotomy' and "Postgastrectomy duodenal leak".)

318
INDICATIONS FOR SURGERYElective surgery is uncommonly needed for
peptic ulcer disease in current medical practice. Currently accepted indications for
surgery in the management of peptic ulcer disease include bleeding, perforation,
obstruction, intractable disease, and suspected malignancy [2].
Management of peptic ulcer disease complications — Bleeding is the most common
complication of peptic ulcer disease requiring hospitalization. In a study of the National
Inpatient Sample (NIS), bleeding occurred in 73 percent, perforation in 9 percent, and
obstruction in 3 percent [4]. H. pylori, nonsteroidal anti-inflammatory drugs (NSAIDs),
and the use of low-dose aspirin are the most common etiologies of ulcer bleeding [7]
and ulcer perforation [8-10]. In observational studies, duodenal, antral/pyloric, and
gastric body ulcers account for 60, 20, and 20 percent of perforations, respectively [8,9].
Among these complications, perforation had the highest mortality rate, followed by
obstruction, then hemorrhage.
Bleeding peptic ulcer — Upper gastrointestinal bleeding due to peptic ulcer disease is
a common indication for emergency management of peptic ulcer disease [11]. Most
patients with acute bleeding can be managed with fluid resuscitation and transfusion,
acid suppression therapy, and endoscopic intervention and/or angiographic
embolization. For those who fail these efforts, surgery may become necessary. The
specific indications for surgery for patients with bleeding peptic ulcer are reviewed
separately. (See "Overview of complications of peptic ulcer disease", section on
'Bleeding'.)
Perforated peptic ulcer — Ulcer perforation may be suspected in patients with a
history consistent with peptic ulcer disease who develop the sudden onset of severe,
diffuse abdominal pain. Once a diagnosis of perforation is established, surgical
intervention is usually indicated. (See "Overview of gastrointestinal tract perforation",
section on 'Clinical features' and "Overview of complications of peptic ulcer disease",
section on 'Perforation'.)
Gastric outlet obstruction — Gastric outlet obstruction is the least frequent ulcer
complication in developed countries. Most cases are associated with duodenal or
pyloric channel ulceration; gastric ulceration accounts for only 5 percent of cases.
Surgical consultation should be obtained for patients with chronic partial gastric outlet
obstruction that is refractory to medical treatment and those found to have complete
gastric outlet obstruction or those readmitted with gastric outlet obstruction after recent
"successful treatment." Surgery is indicated if the patient fails to respond to
conservative medical management and endoscopic therapy. (See "Gastric outlet
obstruction in adults", section on 'Peptic ulcer disease'.)
Surgical treatment of gastric outlet obstruction is almost never an emergency. Because
most patients present with some degree of malnutrition and frequently have electrolyte
imbalances (ie, hypokalemic, hypochloremic metabolic alkalosis secondary to vomiting
or nasogastric suctioning), it is important to correct any derangements and optimize the
patient's medical status prior to proceeding with surgery. Failure to identify and correct
these issues increases perioperative morbidity. (See "Maintenance and replacement
fluid therapy in adults".)
It may also be reasonable to try to improve the patient's overall nutritional status with
nutritional support as these patients are frequently malnourished and can have
significant postoperative delayed gastric emptying due to their atonic stomach.
319
(See "Overview of perioperative nutrition support", section on 'Preoperative nutrition
support'.)
Many experienced gastric surgeons advocate preoperative nasogastric sump
decompression to decrease gastric dilation and, hopefully, gastric atony. (See "Inpatient
placement and management of nasogastric and nasoenteric tubes in adults".)
Peptic ulcer disease refractory to medical management — Intractability as an
indication for elective gastric surgery is increasingly unusual. The precise time period
after which an ulcer should be considered intractable remains poorly defined.
Severe symptoms, failure to heal on medical therapy (including antimicrobials), and
relapse while on maintenance therapy or after multiple courses of therapy may indicate
the need for surgery. Ongoing symptoms or ulcers may indicate noncompliance with
proton pump inhibitors and/or persistent H. pylori infection, continued smoking, or
nonsteroidal anti-inflammatory use. The management of these issues is discussed
elsewhere. (See "Approach to refractory peptic ulcer disease".)
For patients with peptic ulcer disease refractory to medical management, gastrinoma
(Zollinger-Ellison syndrome) should be excluded before performing elective surgery for
peptic ulcer disease. (See "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis".)
The management of marginal ulceration following gastric bypass surgery is reviewed
separately [12]. (See "Late complications of bariatric surgical operations", section on
'Marginal ulcers'.)
Suspicion of malignancy — Suspicion of malignancy, usually within a gastric ulcer
that has failed to heal after 12 weeks of medical therapy, indicates the need for elective
surgery, even if biopsies are benign. (See "Risk factors for gastric
cancer" and "Association between Helicobacter pylori infection and gastrointestinal
malignancy" and "Clinical features, diagnosis, and staging of gastric cancer".)
Controversial indications — More controversial and infrequent indications for elective
surgery for peptic ulcer disease may include:
●Intolerance or noncompliance with a medical regimen
●High risk for ulcer complications, such as transplant recipients and those who are
steroid dependent or NSAID dependent
●Giant gastric or duodenal ulcer (although most of these can be healed with
medical therapy) [13]
GENERAL PRINCIPLES OF ULCER SURGERYGastric acid secretion from
parietal cells is regulated by redundant, overlapping pathways that include endocrine
(gastrin), neural (acetylcholine release from the vagus nerves), paracrine (locally
delivered histamine and somatostatin), and autocrine (transforming growth factor alpha)
factors (figure 1). (See "Physiology of gastric acid secretion".)

Surgical approaches to reduce acid secretion have included:

●Sectioning the vagus nerves (truncal vagotomy, highly selective vagotomy,

combinations of vagotomy)
●Eliminating hormonal stimulation from the antrum (ie, antrectomy)
●Decreasing the number of acid-producing parietal cells (subtotal gastrectomy)
●A combination of the above approaches (eg, vagotomy and antrectomy)

320
Each of these procedures reduces acid secretion but has varying effects on gastric
physiology.

Surgical options and issues that are important for choosing an appropriate procedure,
including the choice between an open surgical and laparoscopic approach, are
reviewed in this section. Specific surgical management of refractory peptic ulcer disease
and complications of peptic ulcer disease are discussed below. (See 'Duodenal
ulcer' below and 'Gastric ulcer' below.)
General considerations — The fundamental goals of surgical therapy for peptic ulcer
disease are to:
●Treat or prevent ulcer complications
●Reduce acid secretion in patients who cannot tolerate proton pump inhibitors
●Minimize postoperative sequelae related to the operation

No single procedure satisfies all of these goals, nor is any one operation applicable to
all clinical settings. In some respects, surgical management of peptic ulcer disease
represents a compromise, with the morbidity of ulcer disease replaced by the morbidity
of the operation. Thus, the least morbid procedure that will adequately manage the
patient's problem should be used in each instance.

Preoperative evaluation — Most patients presenting with refractory peptic ulcer

disease or related complications typically undergo some type of diagnostic imaging
study during the initial evaluation (eg, abdominal radiography, computed tomography,
endoscopy), which may provide sufficient information to guide surgery, when it is
indicated. (See "Overview of complications of peptic ulcer disease".)
As an example, computed tomography (CT) of the abdomen and pelvis may identify a
mass as a cause of gastric outlet obstruction due to malignancy (eg, stomach, ampulla,
pancreas, gallbladder, bile ducts, or duodenum) or more unusual causes of gastric
outlet obstruction, such as pancreatitis, Crohn disease, or gastric volvulus. These are
important to recognize preoperatively as these findings alter surgical management.
When a patient presents with gastric outlet obstruction, upper gastrointestinal
endoscopy should also be attempted to determine whether the obstruction is pre- or
postpyloric, to attempt balloon dilatation, and to obtain a biopsy. (See "Gastric outlet
obstruction in adults", section on 'Evaluation' and "Gastric outlet obstruction in adults",
section on 'Peptic ulcer disease'.)
Surgical options — Surgical options range from local therapies that only manage
ulcer-related complications (ie, bleeding, perforation, or obstruction) to definitive ulcer
operations.
Definitive ulcer operations (eg, highly selective vagotomy, truncal vagotomy with gastric
drainage, partial gastrectomy) aim to reduce acid secretion and hence decrease ulcer
recurrence rate. However, these procedures add operative time and can be associated
with increased perioperative morbidity and long-term adverse physiologic sequelae, and
presently there are effective antisecretory medications (eg, proton pump inhibitors) that
make vagotomy less imperative in the current management strategies.
(See "Postgastrectomy complications".)

When choosing between local versus definitive operations, the surgeon must consider:
321
 ●The characteristics of the ulcer (location, chronicity, presence of complications)
●The features of any obstruction (anatomic level, inflammation, pyloric scarring)
●The patient (age, nutrition, comorbid illness, condition on presentation)
●The operation (mortality rate, complications, long-term sequelae)
●Their personal experience
Ulcer bed management — Ulcer bed management may include ligation of bleeding
vessels or the placement and fixation of omentum within the ulcer bed to cover the
defect and promote healing (ie, Graham patch). In the stomach, it may be most
appropriate to excise the ulcer completely.
Vagotomy — Vagotomy is a procedure that transects or removes a portion of the vagus
nerves or branches of the vagus nerves to decrease gastric acid secretion. Although
definitive ulcer operations are no longer frequently performed, either a truncal vagotomy
or a highly selective vagotomy is still considered an integral part of such procedures
[14-18]. Vagotomy for the treatment of acute peptic ulcer disease is discussed in detail
separately. (See "Vagotomy".)
Gastric drainage — If a truncal vagotomy is performed, some form of gastric drainage
procedure should be added to avoid gastric stasis. Gastric drainage procedures divide
or bypass the pyloric sphincter mechanism to facilitate gastric emptying (figure 2).
Pyloroplasty is generally preferred to gastrojejunostomy because the normal anatomic
relationship between the stomach and duodenum is preserved.
●Pyloroplasty is a procedure that widens the opening between the antrum and
duodenum to facilitate passage of gastric contents (figure 2).
•The Heineke-Mikulicz pyloroplasty, which divides the sphincter longitudinally
and closes it transversely, is the simplest and most commonly performed.
•The Finney and Jaboulay pyloroplasties are gastroduodenostomies.
●Gastrojejunostomy (figure 3), which anastomoses the dependent portion of the
stomach to the proximal jejunum, is typically reserved for patients with significant
duodenal bulb scarring that precludes safe pyloroplasty.
Gastrectomy and reconstruction — Partial gastrectomy (eg, antrectomy, subtotal
gastrectomy) removes the portion of the stomach containing the ulcer, the gastrin-
producing cells that stimulate acid secretion, and a variable number of acid-producing
parietal cells depending upon the extent of the resection (figure 4). Antrectomy would
not remove any parietal cells, while subtotal gastrectomy would remove some but not all
parietal cells. To remove all of the gastrin-producing antral tissue, it is necessary to
resect at least 35 percent of the distal stomach. The line of resection should be taken
high onto the lesser curve as antral tissue extends to within 2 cm of the
gastroesophageal junction. (See "Partial gastrectomy and gastrointestinal
reconstruction", section on 'Surgical anatomy and physiology of the stomach'.)
Reconstruction of the stomach is necessary following partial gastrectomy to reestablish
gastrointestinal continuity. The Billroth I, Billroth II, and Roux-en-Y reconstruction
techniques are the most common (figure 5). Techniques for partial gastrectomy and
reconstruction are discussed in detail elsewhere. (See "Partial gastrectomy and
gastrointestinal reconstruction".)
Billroth I reconstruction may be difficult to accomplish due to severe inflammation in the
region around the duodenum and pylorus and is generally not advisable. If a Billroth II is
selected, the gastrojejunostomy should be created to allow gravity to aid drainage. This

322
usually requires anastomosis of the efferent limb to the greater curve of the stomach
and placing the anastomosis in a dependent area on the posterior wall of the stomach.
Although there are no convincing data showing any significant differences in outcomes
of a retrocolic versus antecolic anastomosis, angulation of the proximal limb as it joins
the lesser curvature should be minimized since excessive angulation or twisting can
lead to afferent limb obstruction with potential blowout of the duodenal stump.
(See "Postgastrectomy complications", section on 'Duodenal stump leak'.)
Roux-en-Y gastrojejunostomy is another option for gastric reconstruction after
antrectomy/vagotomy for peptic ulcer disease, but it has several disadvantages over a
Billroth II gastrojejunostomy. The most significant issue with a Roux-en-Y
gastrojejunostomy is that it exacerbates problems with gastric emptying in patients with
gastric outlet obstruction, and motility problems tend to develop over time [19]. While
roux limbs divert biliary and pancreatic secretions away from the stomach, this
advantage does not outweigh the disadvantage of poor drainage, and, thus, it should
not be the preferred method of reconstruction for patients undergoing partial
gastrectomy for peptic ulcer disease. (See "Partial gastrectomy and gastrointestinal
reconstruction", section on 'Choice of reconstruction'.)
Open versus laparoscopic approach — Each of the procedures described above can
be performed laparoscopically, and experience with laparoscopic techniques for
managing peptic ulcer disease is increasing [20-25]. Case series demonstrate that
laparoscopic ulcer surgery is feasible, safe, and likely equally effective; however, the
collective experience remains limited.
A laparoscopic approach may be particularly useful for vagotomy as the magnified
laparoscopic view is very helpful for identifying vagal nerve fibers. (See "Vagotomy",
section on 'Laparoscopic truncal vagotomy'.)
Another appropriate use of the laparoscopic approach is the management of perforated
ulcer with a Graham patch (with or without vagotomy) [26-30].
DUODENAL ULCER
Elective surgery for duodenal ulcer — Elective surgery for duodenal ulcer disease
refractory to medical management is based upon techniques that reduce acid secretion.
For the elective management of duodenal ulcer disease, we prefer either laparoscopic
posterior truncal vagotomy and anterior serosal myotomy or laparoscopic posterior
truncal vagotomy and anterior highly selective vagotomy. With these procedures, pyloric
function is preserved, and thus a drainage procedure, with its inherent risk of bile reflux,
is not needed. If laparoscopic expertise is not available, we suggest referral to a facility
with surgeons who have training and experience with these procedures. These
techniques are reviewed in detail separately. (See "Vagotomy", section on
'Duodenal' and "Vagotomy", section on 'Highly selective vagotomy'.)

In the past, partial gastrectomy, either as an antrectomy and truncal vagotomy, or

subtotal gastrectomy was commonly used in the surgical management of duodenal
ulcer disease as well.

●The simultaneous effects of vagotomy and antrectomy eliminate cholinergic and

gastrin stimulation of acid secretion, such that basal acid secretion is virtually
abolished and stimulated secretion is reduced by nearly 80 percent [31].

323
 ●By removing the distal two-thirds of the stomach, subtotal gastrectomy removes
the major portion of the acid-producing parietal cells and also the gastrin-
producing antrum, reducing basal and stimulated acid secretion by 75 and 50
percent, respectively [31].
However, removal of the pylorus with these procedures leads to rapid emptying of
liquids and solids from the stomach and, potentially, to reflux of intestinal contents into
the stomach. Given the high incidence of postgastrectomy complications, these
procedures are not generally needed for the management of duodenal ulcer disease
refractory to medical management. (See "Partial gastrectomy and gastrointestinal
reconstruction" and "Postgastrectomy complications".)
Bleeding duodenal ulcer — The first priority during emergency surgery for a bleeding
ulcer is control of the bleeding site. If endoscopy has failed to precisely identify the
source of hemorrhage, the proximal duodenum should be opened longitudinally first and
inspected for the bleeding vessel. The incision may need to be extended through the
pylorus (ie, pyloroduodenotomy) to inspect the duodenal bulb and gastric antrum.
Once identified, the bleeding vessel can be controlled directly by suture ligation or
ligation of the gastroduodenal artery at the superior and inferior aspect of the ulcer and
control of the transverse pancreatic branch. While a well-placed "U-stitch" may be all
that is needed, usually more than one suture is necessary, and hemostasis should be
assured before closing the duodenum. Closure with a pyloroplasty (figure 2) and a
truncal vagotomy can be performed. A highly selective vagotomy may be chosen for the
hemodynamically stable patient who has had a duodenotomy, but the benefits of highly
selective vagotomy are less clear if a pyloroduodenotomy has been performed [32,33].
(See "Vagotomy", section on 'Duodenal'.)
For patients in whom H. pylori has not yet been treated, control of bleeding alone
without an acid-reducing procedure may be appropriate, although this approach has not
been systematically evaluated. A single retrospective analysis of American College of
Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) data
suggests that vagotomy and drainage was associated with a significantly lower
postoperative mortality rate than simple oversewing when performed for bleeding ulcers
[34].
Perforated duodenal ulcer — Perforated duodenal ulcers can generally be treated by
closure with a piece of omentum (Graham patch) or, for perforated ulcers close to the
pylorus, by truncal vagotomy with pyloroplasty (incorporating the perforation). Both
approaches are simple and expedient, and the results have generally been good [35-
39].
Simple patch closure of the perforation should be considered in the setting of ongoing
shock, delayed presentation, significant medical comorbidities, or significant peritoneal
contamination. Patch closure may also be appropriate for patients who have never been
treated for peptic ulcer disease and who are candidates for proton pump inhibitors and
antibiotic therapy for H. pylori [40,41]. Nonsteroidal anti-inflammatory drug (NSAID)-
related perforation can generally be treated with simple closure as the drug can almost
always be discontinued or switched to a COX-2 inhibitor.

Large or complex perforated duodenal ulcers may require specialized technique to

close. Although the techniques described below have not changed since they were

324
introduced in the 1960s, they may not be familiar to surgeons of the modern era, who
rarely have chances to practice them.

●Bancroft's technique is best suited for cases in which scarring prevents

adequate dissection of the pylorus. It uses a distal muscular cuff of the antrum to
close the duodenal stump, which requires complete removal of the antral mucosa
to avoid the complications of a retained antrum. Furthermore, the Bancroft
procedure requires the right gastric and right gastroepiploic arteries to be
preserved in order to maintain the blood supply to the local tissue (figure 6). Thus,
the decision to use Bancroft's closure must be made before these vessels are
divided [42,43].
●Nissen's technique is particularly useful when a posterior duodenal ulcer has
penetrated into the pancreas and is not amenable to resection. It requires the
duodenum to be dissected from the proximal portion of the ulcer bed, leaving the
ulcer intact with a small rim of tissue for closure. The duodenum is then brought
up and closed over the ulcer [44,45].

When the risk of significant duodenal leak is perceived to be high, it is still possible to
perform a primary ulcer repair, rather than gastrectomy, by adding a pyloric exclusion to
divert the gastric stream away from the duodenum.

Pyloric exclusion involves closing the pylorus with creating a gastrojejunal anastomosis
for gastric drainage [46]. It can be performed in several different ways:
●One common method is to create a gastrotomy along the greater curvature
through which the pylorus can be grasped and sutured closed with a
nonabsorbable suture. The gastrotomy can then either be closed primarily or used
to complete the gastrojejunostomy.
●Another method fires a noncutting stapler across the pylorus, although care must
be taken to avoid misfire across the proximal duodenum.

Pyloric exclusion with a gastrojejunal anastomosis permits continued oral feeding. In

most cases, the pyloric closure will open spontaneously within several weeks, by which
time the ulcer repair would have healed.

Although pyloric exclusion has been used in patients with traumatic or post-endoscopic
retrograde cholangiopancreatography (ERCP) injuries for years, the results have been
mixed in ulcer patients [47-49]. In our practice, we use pyloric exclusion selectively
when the risk of a large duodenal leak is high, when the degree of inflammation is
severe, or when either the overall health of the patient or the quality of local tissue is
poor, rather than in all patients.
Definitive ulcer surgery with an acid-reducing procedure may be desirable in patients
with perforation, especially when the perforation is close to the pylorus, and in patients
for whom the risk of recurrent complications cannot be reliably reduced (such as rare
patients who will require continued use of an NSAID or those who are likely to resume
NSAID use against medical advice). Most large studies of patients who have undergone
definitive ulcer surgery were performed prior to the recognition of the role of H.
pylori treatment [39,41,50-52]. An illustrative report included 159 patients who were

325
followed more than 10 years after vagotomy and pyloroplasty for perforated duodenal
ulcer [51]. The perioperative mortality was 5.5 percent, ulcers recurred in 8.8 percent,
and postoperative digestive sequelae, notably diarrhea and dumping, developed in 16
percent. Nevertheless, the overall results were considered to be good to excellent in
almost 90 percent of survivors. Similar recurrence rates (but with less morbidity) have
been described in patients who underwent highly selective vagotomy [52].
Gastric outlet obstruction — Inflammation from peptic ulcer disease, especially a
pyloric channel or duodenal ulceration, causes edema and duodenal spasm and is the
most common cause of benign gastric outlet obstruction. Acute inflammation may
respond to medical management or progress to chronic inflammation. Endoscopic
balloon dilation may transiently relieve gastric outlet obstruction, but repeated balloon
dilatation is associated with risks such as perforation, which can complicate future
surgical management [53]. Thus, for a subgroup of patients, surgical management of
gastric outlet obstruction may become essential for the long-term relief of symptoms.
(See "Overview of complications of peptic ulcer disease", section on 'Gastric outlet
obstruction'.)
The goals of surgical treatment are to relieve gastric outlet obstruction and provide
definitive control of peptic ulcer disease. Antrectomy/distal gastrectomy relieves the
obstruction and definitively rules out malignancy. Antrectomy/distal gastrectomy has
traditionally been combined with vagotomy to decrease the risk of ulcer recurrence.
With the pylorus removed, alkaline intestinal contents bathing the stomach stimulate
acid secretion, which can be suppressed by either lifelong antisecretory therapy or
vagotomy. Although vagotomy adds minimal additional morbidity to the procedure, it
can cause gastric atony. There are no data to guide whether antrectomy plus vagotomy
versus antrectomy plus lifelong antisecretory therapy is more efficacious or cost
effective. (See "Vagotomy", section on 'Gastric outlet obstruction'.)
Vagotomy and drainage is an alternative to antrectomy/distal gastrectomy. In a
randomized trial that included 90 consecutive patients with gastric outlet obstruction
secondary to duodenal ulcer, there were no differences in the postoperative course or
the reduction in gastric acid secretion in patients assigned to vagotomy and antrectomy
compared with vagotomy and drainage [14].
Surgery can be challenging because of inflammation and scarring in the pyloric region
that may extend into the porta hepatis. Inflammation in this region can result in distorted
anatomy and lead to potential difficulties in securely closing the duodenal stump
following antrectomy. Because the fibrotic process may alter the usual anatomic
relationships, critical vascular structures and the common bile duct may be nearer to the
duodenal stump than anticipated and difficult to identify. For these reasons, antrectomy
may be a less suitable choice.
When vagotomy and drainage are chosen, debate persists as to the optimal drainage
procedure. Patients whose ulcer disease is severe enough to create obstruction rarely
have a duodenum that is pliable enough to permit a pyloroplasty. In those rare
instances where fibrosis is minimal or confined only to a small area, a Finney
pyloroplasty may be an option [54]. However, foreshortening of the first part of the
duodenum may make a Jaboulay pyloroplasty (latero-lateral gastroduodenostomy) a
better alternative (figure 2). The Jaboulay has gained popularity because of its technical
simplicity and because the anastomosis is performed in healthy tissue, away from the

326
ulcer bed [55]. However, the randomized trial mentioned above suggested that
gastrojejunostomy may be better than a Jaboulay [14]. If the duodenal stump is difficult,
a posterior gastrojejunostomy without antrectomy should provide adequate drainage
and should preferably include biopsy of the ulcer. Although gastrojejunostomy may
theoretically reduce some of the advantages of highly selective vagotomy, vagotomy
should nonetheless be considered. (See "Postgastrectomy duodenal leak".)
GASTRIC ULCERFundamental differences between gastric and duodenal ulcers
affect surgical decision making, even though both are peptic lesions. The most
important is that gastric ulcer may harbor malignancy, and, therefore, the ulcer must be
excised or, at a minimum, generously biopsied. In addition, patients with gastric ulcer
tend to be an older and more debilitated population, increasing the risk for perioperative
morbidity and mortality. (See "Surgical management of invasive gastric cancer".)
Elective surgery for gastric ulcer — A classification system (the Johnson
classification) based upon anatomic location and acid-secretory potential provides a
useful basis for considering elective surgery in the treatment of gastric ulcer [56]:
Type I gastric ulcer — Type I gastric ulcers are the most common type of gastric ulcer.
These occur along the lesser curvature near the junction of fundic and antral mucosa
and occur in the setting of acid hyposecretion.
For most patients with type I gastric ulcer, distal gastrectomy with Billroth I or Billroth II
reconstruction is recommended since this approach removes the ulcer and the diseased
antrum. It also treats an occult malignancy. Recurrent ulcer rates are low (0 to 5
percent), and excellent symptomatic relief is usually achieved [57]. Mortality ranges
from 0 to 6 percent [57,58].
Although type I gastric ulcer has classically been considered the consequence of
inadequate gastric mucosal defense, as opposed to increased acid secretion, many
advocate the addition of some form of vagotomy to the gastric resection. In a
retrospective review of 349 cases of gastric ulcer at the Cleveland Clinic from 1950 to
1979, no significant differences were found in recurrence rates for gastric resection with
or without vagotomy [59]. Median follow-up was more than 11 years. This study also
found comparable results for truncal vagotomy and pyloroplasty plus ulcer excision or
biopsy. Thus, vagotomy is not absolutely necessary but not unreasonable for type I
gastric ulcer.
Although most prefer distal gastrectomy, highly selective vagotomy has been used for
type I gastric ulcer. The value of highly selective vagotomy in gastric ulcer may derive
from its ability to decrease acid secretion while maintaining adequate gastric emptying
and minimizing postoperative duodenogastric reflux. The procedure is performed as for
duodenal ulcer, with the addition of a gastrotomy to resect or biopsy the ulcer bed. This
approach has been questioned on theoretical grounds because highly selective
vagotomy may promote a degree of gastric stasis and gastrin hypersecretion, two
factors hypothesized to contribute to gastric ulcer pathogenesis. Despite these
concerns, clinical results have been promising. In one series of 48 patients, the ulcer
recurrence rate was comparable to gastrectomy at 6.5 percent with few adverse effects
[60]. Ulcer diameter and location may make this highly selective vagotomy difficult for
some patients in whom ulcer-induced inflammation, edema, or scarring may obscure
accurate dissection of the vagus nerves along the lesser curvature of the stomach.

327
Type II gastric ulcer — Type II gastric ulcers occur synchronously with scarring or
ulceration in the duodenum or pyloric channel. They tend to be large, deep ulcers with
poorly defined margins. They frequently occur in younger men and are associated with
increased acid secretion.
For type II gastric ulcer, antrectomy and vagotomy are the preferred approaches.
(See "Partial gastrectomy and gastrointestinal reconstruction".)
Type III gastric ulcer — Type III ulcers are prepyloric, although no precise anatomic
definition exists. They occur in the setting of increased acid secretion and are
approached in a manner similar to duodenal ulcer and type II gastric ulcer. Early
surgical referral is advised for resistant ulcers or those that present with obstructive
symptoms.
Curiously, highly selective vagotomy (as well as medical therapy with H2 receptor
antagonists) has been associated with poor results in type III gastric ulcer, with high
recurrence rates ranging from 16 to 44 percent reported in various series [59]. This
finding, plus the observation that these ulcers may harbor an occult gastric malignancy,
makes antrectomy and vagotomy the most prudent approach. (See "Partial gastrectomy
and gastrointestinal reconstruction".)
Type IV gastric ulcer — Type IV gastric ulcer is distinguished by its anatomic location
high along the lesser curvature, close to the gastroesophageal junction. Antral mucosa
may extend to within 1 to 2 cm of the gastroesophageal junction; thus, type IV ulcers
may represent a subset of type I gastric ulcers. (See 'Type I gastric ulcer' above.)
Type IV ulcers are associated with gastric acid hyposecretion and present early with
dysphagia and reflux. Large ulcer diameter, the degree of surrounding inflammation,
and proximity to the gastroesophageal junction render operative management difficult
and potentially dangerous. If the integrity of the distal esophagus can be assured,
subtotal gastric resection (including the ulcer bed) is considered optimal therapy.
Typically, an attempt is made to preserve a margin of unaffected stomach to facilitate
the anastomosis. Alternatives include the Pauchet procedure, which is a distal
gastrectomy extended along the lesser curvature to include the ulcer, or the Kelling-
Madlener procedure, in which distal gastrectomy is performed but the ulcer is left in
place to avoid compromise of the gastroesophageal junction. Although there is no
consensus in the literature, many surgeons feel more comfortable performing an
esophageal anastomosis knowing that the ulcer is removed. When most of the stomach
is removed, a Roux-en-Y reconstruction if often performed. (See "Partial gastrectomy
and gastrointestinal reconstruction" and "Total gastrectomy and gastrointestinal
reconstruction".)
Recurrent gastric ulcer — Finally, in patients who have undergone multiple failed
procedures for refractory ulcer disease, subtotal or total gastrectomy may be used as
the last line of definitive surgical treatment. Such a procedure would remove the portion
of the stomach containing the ulcer, the gastrin-producing cells that stimulate acid
secretion in the antrum, and the acid-producing parietal cells in the body and fundus of
the stomach. However, it should only be performed by gastrointestinal surgeons with
expertise in these procedures. (See "Total gastrectomy and gastrointestinal
reconstruction".)
Bleeding gastric ulcer — For patients with a bleeding gastric ulcer, partial gastrectomy
with Billroth I or II reconstruction (figure 5) is generally indicated because of the risk for

328
malignancy. For patients with medical comorbidities, ulcer excision combined with
truncal vagotomy and pyloroplasty is an option [61,62]. Ulcer excision alone is
associated with rebleeding in as many as 20 percent of patients.
Perforated gastric ulcer — Because patients with perforated gastric ulcer tend to be
older adults and have comorbidities, surgery is associated with high overall mortality
(ranging from 10 to 40 percent) regardless of treatment [26,61-64].
The choice of procedure is usually made during the operation. The preferred approach
is partial gastrectomy to include the ulcer because of the risk of gastric malignancy,
unless the patient is at an unacceptably high risk because of advanced age, comorbid
disease, intraoperative instability, or severe peritoneal soilage [64]. Wedge resection of
the ulcer or patch closure may be performed in older or sicker patients. Patch closure
alone is associated with postoperative gastric obstruction in approximately 15 percent of
cases. When patch closure is performed, biopsy of the ulcer is necessary to rule out
malignancy [64]. In cases of delayed presentation where the perforation has sealed, no
surgical intervention may be necessary.
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Peptic ulcer disease".)
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10 th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Beyond the Basics topics (see "Patient education: Peptic ulcer disease (Beyond
the Basics)" and "Patient education: Helicobacter pylori infection and treatment
(Beyond the Basics)").
SUMMARY AND RECOMMENDATIONS
●The development of potent antisecretory agents and the recognition that
treatment for Helicobacter pylori infection eradicates most ulcer recurrences has
reduced the need for elective surgical treatment of peptic ulcer disease. Surgical
management is reserved for peptic ulcer disease that is refractory to medical
management, for suspicion of a malignancy within an ulcer, or for the
management of complications of peptic ulcer disease (ie, bleeding, perforation,
obstruction). Other indications are controversial. (See 'Indications for
surgery' above.)

329
●Definitive acid-reducing procedures include highly selective vagotomy, truncal
vagotomy with gastric drainage, and distal gastrectomy with reconstruction. The
choice of procedure depends upon the clinical circumstances. (See 'General
principles of ulcer surgery' above and 'Duodenal ulcer' above and 'Gastric
ulcer' above.)
•For patients with duodenal ulcer and indications for elective surgery, we
suggest vagotomy over another definitive acid-reducing procedure (Grade 2C).
Vagotomy reduces the risk of recurrent ulceration while minimizing
postoperative complications and long-term sequelae. We further suggest a
laparoscopic rather than open approach to vagotomy (Grade 2C). Outcomes
are similar for highly selective vagotomy and anterior seromyotomy with
posterior truncal vagotomy (ie, Taylor procedure), so either is appropriate.
(See 'Elective surgery for duodenal ulcer' above.)
•For patients with duodenal bleeding related to peptic ulcer disease who have
received a fully adequate trial of medical management, we suggest performing
a definitive acid-reducing procedure in addition to managing the ulcer bed
(Grade 2C). When choosing to add an acid-reducing procedure, we suggest
truncal vagotomy and pyloroplasty, rather than another acid-reducing
procedure (Grade 2C). Managing the ulcer bed alone is appropriate to control
bleeding in hemodynamically unstable patients and for managing ulcer
complications in those with significant comorbidities that limit life expectancy.
(See 'Bleeding duodenal ulcer' above.)
•Perforated duodenal ulcers can generally be treated by closure with a piece of
omentum (Graham patch) or, for perforated ulcers close to the pylorus, by
truncal vagotomy with pyloroplasty (incorporating the perforation).
(See 'Perforated duodenal ulcer' above.)
•The extent of gastric resection for gastric ulcer, which may harbor malignancy,
depends upon the size and location of the ulcer. For uncomplicated or
complicated gastric ulcer, we suggest partial gastrectomy and reconstruction
rather than simple ulcer excision in good-risk surgical candidates (Grade 2C).
For patients with significant medical comorbidities, alternatives include ulcer
excision for bleeding or patch closure for perforation, possibly combined with
vagotomy and gastric drainage. (See 'Gastric ulcer' above.)
•Gastric outlet obstruction is the least frequent complication of peptic ulcer
disease, and most cases of gastric outlet obstruction are associated with
duodenal or pyloric channel ulceration, but the potential for malignancy must
be taken into account. Thus, for patients with gastric outlet obstruction
complicating peptic ulcer disease and no major medical comorbidities, we
suggest distal gastrectomy rather than another definitive acid-reducing
procedure (Grade 2C). For patients with significant risk factors for
perioperative morbidity and mortality, a drainage procedure, such as
gastrojejunostomy, is an acceptable alternative. (See 'Gastric outlet
obstruction' above and "Surgical management of invasive gastric cancer".)

330
Unusual causes of peptic ulcer disease
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Mar 02, 2020.
INTRODUCTIONPeptic ulcers are defects in the gastrointestinal mucosa that extend
through the muscularis mucosae. They persist as a function of the acid/peptic activity in
gastric juice. Peptic ulcer disease remains an important cause of morbidity and mortality
and health care costs [1,2]. Helicobacter pylori and nonsteroidal anti-inflammatory drugs
(NSAIDs) account for the large majority of cases of peptic ulcer disease (PUD).
(See "Association between Helicobacter pylori infection and duodenal
ulcer" and "NSAIDs (including aspirin): Pathogenesis of gastroduodenal toxicity".)
This topic will review unusual causes for peptic ulcer disease. The epidemiology, risk
factors, clinical manifestations, diagnosis, and management of peptic ulcer disease are
discussed in detail, separately. (See "Peptic ulcer disease: Epidemiology, etiology, and
pathogenesis" and "Peptic ulcer disease: Clinical manifestations and
diagnosis" and "Overview of complications of peptic ulcer disease" and "Peptic ulcer
disease: Treatment and secondary prevention" and "Approach to refractory peptic ulcer
disease".)
EPIDEMIOLOGYIn a retrospective study that included 386 patients conducted over
five years in a large tertiary hospital in the United Kingdom, Helicobacter pylori (H.
pylori)-negative, NSAID-negative ulcers accounted for 12 percent of all ulcers [3].
Several studies in the United States have shown that less than 75 percent of patients
with duodenal ulcers (DUs) not associated with use of NSAIDs are related to H.
pylori infection. In one study, after excluding NSAID use, 61 percent of DUs and 63
percent of gastric ulcers were H. pylori positive [4].
ETIOLOGY
Non-NSAID medications — There are a number of drugs that may cause or
exacerbate peptic ulcer disease and/or upper gastrointestinal (GI) bleeding (table 1).
●Acetaminophen – A role for acetaminophen in development of GI complications
(including bleeding and perforation) has been suggested in both population-based
studies and a randomized controlled trial [5-8]. As a general rule, the risk appears
to be increased in doses of 2 to 3 g per day or higher. The risk also appears to be
increased with the combination of NSAIDs plus high-dose acetaminophen
compared with either alone [5,8]. In a randomized trial of patients with
osteoarthritis, co-administration of acetaminophen and aspirin was not associated
with a significant difference in endoscopic ulcer rates compared with either drug
alone, but there were more endoscopic erosions and ulcers in the combined group
compared with either alone [9].

331
●Bisphosphonates – Damage and acute ulceration have been observed [10,11].
Population-based studies show that use of bisphosphonates such
as alendronate are associated with a higher risk of upper and lower GI bleeding
[12]. The risk appears to be increased substantially with co-administration of
NSAIDs. (See "Risks of bisphosphonate therapy in patients with osteoporosis",
section on 'Gastrointestinal'.)
●Glucocorticoids – The association between corticosteroids and peptic ulcer
disease (PUD) and its complications is controversial with wide variations in
estimates of the risk of PUD. However, the combination of glucocorticoids and
NSAIDs results in a synergistic increase in the incidence of gastrointestinal
events. (See "Major side effects of systemic glucocorticoids", section on
'Gastrointestinal effects'.)
●Clopidogrel – The antiplatelet agent clopidogrel is a significant risk factor for GI
bleeding, particularly in patients with a prior risk of bleeding or on cotherapy with
low-dose aspirin or NSAIDs [13]. The risk may also be increased with ticlopidine,
but data are limited. (See "Acute ST-elevation myocardial infarction: Antiplatelet
therapy", section on 'Possible early CABG'.)
●Sirolimus – Sirolimus has been associated with aggressive ulcer disease in
patients undergoing liver transplantation [14]. Sirolimus is known to inhibit wound
healing and has been associated with small bowel ulceration [15]. This warrants
cautious use of this agent in patients with a history of PUD and aggressive
management of patients on the drug who develop symptoms or complications
suggestive of PUD. (See "Pharmacology of mammalian (mechanistic) target of
rapamycin (mTOR) inhibitors", section on 'Gastrointestinal system'.)
●Spironolactone – Spironolactone has been associated with a nearly threefold
increased risk of upper GI bleeding [16,17]. The risk appears to be increased with
high doses, suggesting dose-related toxicity [17]. The risk was also increased with
advancing age and use of concomitant ulcerogenic medications. It is hypothesized
that spironolactone, an aldosterone receptor antagonist, may cause ulceration by
impairing the healing of gastric or duodenal erosions [17]. Aldosterone promotes
the formation of fibrous tissue by binding to mineralocorticosteroid receptors; the
effect is modulated by 11 beta hydroxysteroid dehydrogenase enzymes which are
expressed in the stomach and to a small extent in the small intestine.
●Selective serotonin reuptake inhibitors – There are increasing data
suggesting an impact of selective serotonin reuptake inhibitors (SSRIs)
and venlafaxine on the risk of peptic ulcer and upper GI bleeding [18-24]. The risk
of both uncomplicated PUD and ulcer-related bleeding is higher in patients taking
SSRIs [25]. A concomitant H. pylori infection appears to increase the risk of
serious upper GI bleeding in patients on SSRIs [25]. The risk of bleeding is also
higher in patients with a history of peptic ulcer, with increasing age and duration of
SSRI use, and when SSRIs are combined with antiplatelet therapy
(aspirin, clopidogrel) or NSAIDs [25-28]. In a population-based cohort study, the
use of SSRIs alone or in combination with NSAIDs was not associated with an
increased 30-day mortality following a peptic ulcer bleed [29]. In one large nested
case-control study, the risk of upper GI bleeding was only apparent in non-users
of acid inhibitors, suggesting that acid suppression attenuates the impact of SSRIs

332
 [26]. Short-time use of SSRIs (14 days) has also been associated with an
increased risk of bleeding [30]. A systematic review suggested there was an
approximate two-fold increase in the risk for upper GI hemorrhage in SSRI users,
and the risk appears to be greatest in those using NSAIDs or antiplatelet drugs
concomitantly [31]. (See "Selective serotonin reuptake inhibitors: Pharmacology,
administration, and side effects", section on 'Upper gastrointestinal bleeding'.)
●Chemotherapy and molecular targeted therapy – Some cancer
chemotherapeutic agents have been associated with peptic disease. As an
example, patients receiving continuous hepatic artery infusion
of fluorouracil frequently complain of abdominal pain. In one series, endoscopy
usually revealed duodenal, gastric, or pyloric ulcerations and erosions [32]. The
mechanisms have not been established; catheter migration with direct gastric
infusion of chemotherapy did not explain most cases. These lesions responded to
discontinuation of chemotherapy; H2 receptor blockers and proton pump inhibitors
(PPIs) appeared to have little efficacy, but have not been formally tested. Thus,
these ulcers may not truly be peptic ulcers.
An example of a targeted agent that can cause ulceration include erlotinib, a
reversible tyrosine kinase inhibitor of the epidermal growth factor receptor used for
cancer treatment. GI bleeding from peptic ulcers and ulcer perforation have been
reported; not surprisingly, the risk appears greater when erlotinib is combined with
NSAIDs, including cyclooxygenase-2 inhibitors, and in patients with a peptic ulcer
history or treated with anticoagulants [33-36].
Infections
Bacteria
●Isolated duodenal H. pylori colonization – In high-prevalence regions for H.
pylori, it is important to consider duodenal colonization in patients who lack gastric
evidence of H. pylori. In an Italian study 42 of 608 patients (6.9 percent) with
duodenal ulcer had no other obvious cause; 18 of these 42 patients had isolated
duodenal colonization with H. pylori detected by biopsy [37]. H. pylori eradication
therapy was given to these 18 subjects; two had ulcer recurrences compared with
14 out of 20 subjects in the "idiopathic" group. Urea breath tests were positive in
only 3 of the 18 patients with isolated duodenal colonization [37].
●Non-pylori helicobacter – Helicobacters other than H. pylori have been
associated with peptic ulcer disease. Helicobacter heilmannii is the most
frequently described organism and is often associated with antral ulcers [38].
Viruses
●Herpes simplex virus type I – The possible involvement of herpes simplex virus
type 1 (HSV-1) in patients with PUD has been suggested by an increased
incidence of anti-HSV-1 antibodies in patients with peptic ulcers [39,40] and the
detection of DNA and protein specific for HSV-1 in the mucosa at the ulcer margin
in a small proportion of patients with seemingly ordinary ulcers [39,41]. A study
using polymerase chain reaction to identify HSV-I and a CLO-test and histology to
identify H. pylori found HSV-I in 30 and 32 percent of duodenal and gastric ulcers,
respectively [42]. HSV-I was found more frequently in the absence of H.
pylori infection. There is some evidence that the pathogenesis may involve

333
 infected enteric neurons [39,43]. (See "Epidemiology, clinical manifestations, and
diagnosis of herpes simplex virus type 1 infection", section on 'HSV esophagitis'.)
●Cytomegalovirus – Cytomegalovirus (CMV) has been associated with ulcers in
immunocompromised patients in whom it can cause multiple, large, shallow
ulcerations of the stomach and esophagus [44]. The diagnosis has been made by
finding intranuclear inclusion bodies or CMV-DNA in the gastric mucosa in biopsy
specimens taken from the ulcerous region. (See "Epidemiology, clinical
manifestations, and treatment of cytomegalovirus infection in immunocompetent
adults", section on 'Gastrointestinal manifestations' and "AIDS-related
cytomegalovirus gastrointestinal disease", section on 'Gastritis' and "Infection in
the solid organ transplant recipient", section on 'CMV and EBV'.)
●Epstein Barr virus (EBV) – Gastric ulcers related to EBV virus infections have
been reported in both immunocompromised and immunocompetent patients
[45,46]. The endoscopic appearance may resemble malignancy.
Fungi — Fungal infections can also cause gastric ulcers in patients who are
immunocompromised or immunocompetent [47]. Candidal infections have been
described in isolated patients with perforated peptic ulcers. Transmural infections with
the organism may be seen, and a compromised immune status may be a risk factor
[48,49]. Mucormycosis can cause extensive ulceration and necrosis of the gastric wall in
immunocompromised patients. Mucormycosis can cause extensive ulceration and
necrosis of the gastric wall in immunocompromised patients [47].
Mechanical
Obstruction — Unusual cases of duodenal ulcers have been linked to a variety of
gastroduodenal abnormalities, including congenital duodenal webs, hypertrophic pyloric
stenosis, annular pancreas (a rare congenital disorder in which pancreatic tissue
encircles the second portion of the duodenum), and a preduodenal portal vein.
Duodenal ulcers in these settings may present in infancy or childhood, but can also
occur in adolescence or adulthood [50]. As an example, although annular pancreas can
present at birth with high-grade duodenal obstruction, it more commonly presents in the
second to seventh decade, with the peak incidence in between the ages of 30 and 40
years [50,51]. Ulcers with annular pancreas are often postbulbar and can be associated
with basal acid hypersecretion, although mechanisms linking these various "obstructing"
lesions of the pylorus and duodenum to acid hypersecretion and ulcer disease remain
undefined. (See "Annular pancreas".)
Foreign body — Peptic ulceration has also been observed in association with foreign
bodies. Ingestion of multiple magnets and button batteries has been associated with GI
ulceration. Microcoils from previous gastroduodenal artery embolization [52] and
surgical clip migration from prior laparoscopic cholecystectomy have been found in ulcer
craters [53]. Duodenal extension of a gastric trichobezoar can cause ulcerations in the
duodenum (Rapunzel syndrome) [54,55]. (See "Ingested foreign bodies and food
impactions in adults", section on 'Disk batteries' and "Ingested foreign bodies and food
impactions in adults", section on 'Magnets'.)
Post-surgical — Stomal (marginal) ulcers involving the small intestinal mucosa occur in
1 to 16 percent of patients 3 to 18 months after Roux-en-Y gastric bypass surgery for
obesity [56-60]. Patients present with pain, bleeding, or perforation. Ulcers occurring in
the immediate postoperative period present with bleeding and pain, and the bleeding

334
site can usually be reached with the standard upper endoscope. Ulcers that occur
weeks or months after surgery typically present with abdominal pain after meals and are
usually due to marginal ulcerations [61]. A number of causes have been suggested for
marginal ulceration, including surgical technique resulting in ischemia at the
anastomosis, H. pylori infection, and drug-induced ulceration caused by NSAIDS
[56,57]. Antral exclusion (retained gastrin antrum) is a rare cause of peptic ulcers. It has
been described in patients after gastric resection when some antral mucosa was
retained in the duodenal pouch [62].
Acid hypersecretory states
●Gastrinoma – The classic tetrad of the Zollinger-Ellison syndrome consists of a
non-beta islet cell tumor secreting gastrin in association with acid hypersecretion
and severe PUD, which occurs in approximately 90 percent of cases.
(See "Zollinger-Ellison syndrome (gastrinoma): Clinical manifestations and
diagnosis".)
The majority of Zollinger-Ellison tumors are sporadic. However, 20 to 25 percent
occur in association with the multiple endocrine neoplasia syndrome type 1
(MEN1). MEN1 is a familial disorder characterized by an autosomal dominant
predisposition to tumors of the parathyroid glands (which occur in nearly all
patients by age 50), anterior pituitary, and pancreatic islet cells (gastrinoma or
insulinoma). (See "Multiple endocrine neoplasia type 1: Clinical manifestations
and diagnosis".)
Although gastrinoma-associated peptic ulcers may be indistinguishable from
ordinary peptic ulcers, several distinguishing features should raise the level of
suspicion (table 2). (See "Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis".)
●Systemic mastocytosis – Systemic mastocytosis is characterized by mast cell
infiltration of many tissues and symptoms of flushing, pruritus, abdominal pain,
and diarrhea. Dyspepsia, duodenal ulcers, and severe duodenitis occur in 30 to 50
percent of cases and can be associated with basal acid hypersecretion, a
presentation sometimes reminiscent of gastrinoma [63]. While fasting, serum
gastrin levels are normal, serum histamine concentrations may be elevated,
particularly in the setting of ulcer disease, suggesting that circulating histamine
contributes to basal acid hypersecretion [63,64]. (See "Mastocytosis (cutaneous
and systemic) in adults: Epidemiology, pathogenesis, clinical manifestations, and
diagnosis".)
●Myeloproliferative disorders – Acid hypersecretion, an increased risk of peptic
ulcer, and elevated serum histamine concentrations occur in rare
myeloproliferative disorders associated with basophilia such as basophilic
leukemia and chronic myeloid leukemia with marked basophilia [65]. Acid
hypersecretion may only occur when cell lysis occurs during chemotherapy [65].
Polycythemia vera has also been associated with peptic disease. The precipitating
cause may be reduced mucosal blood flow due to increased viscosity. A study
suggested that patients with polycythemia vera are at greater risk for H.
pylori infection and have a higher prevalence of ulcers and erosions in the
stomach and duodenum than controls without the disease [66]. (See "Clinical

335
 manifestations and diagnosis of polycythemia vera", section on 'Gastrointestinal
symptoms'.)
●Antral G-cell hyperfunction – In the pre-H. pylori era, an entity called antral G-
cell hyperfunction was reported, characterized by duodenal ulcers, increased
basal- and/or meal-stimulated gastrin release, and usually increased basal or
maximal acid output [67,68]. It now appears probable that antral G-cell
hyperfunction usually reflects an exaggerated response to H. pylori-induced
hypergastrinemia, being part of the spectrum of H. pylori-associated duodenal
ulcers. In H. pylori-positive patients who appear to have this disorder, anti-H.
pylori therapy induces ulcer healing and suppresses serum gastrin levels [68-70].
Although controversial and rare, antral G-cell hyperfunction can probably occur in
the absence of H. pylori [68,70]. The pathogenesis of hypergastrinemia in this
setting is not known.
Ischemic
Arterial/venous diseases — Mucosal blood flow is critical to mucosal integrity.
Although ischemia is rare because of the rich vascularization of the gastroduodenal
mucosa, vascular insufficiency syndromes involving arterial supply rarely present with
peptic ulceration that is unrelated to H. pylori infection or NSAID use [71-74]. Some
patients complain of typical ulcer pain, while others may have symptoms suggestive of
mesenteric angina with pain following meals. Endoscopy may reveal multiple gastric or
duodenal ulcers. Ulcers related to vascular stenosis often fail PPI therapy, but several
reports indicated response to revascularization [71,74]. Gastroduodenal artery
pseudoaneurysm is also a rare cause of ulceration [75].
Intestinal ulceration can occur after vascular injury by abdominal radiotherapy, with the
second portion of the duodenum being especially sensitive to radiation injury [76].
Gastroduodenal ulceration also occurs following chemoembolization used to treat
hepatic tumors with an incidence between 3 and 5 percent; abdominal pain, nausea,
vomiting, and anorexia herald this complication [77-79]. Careful attention to protocol
may reduce the risks of this complication [78]. (See "Overview of gastrointestinal toxicity
of radiation therapy", section on 'Gastritis'.)
Non-occlusive ischemia — Mucosal vasoconstriction may explain the perforation of
gastroduodenal ulcers that have been described with crack cocaine use [80-84].
Enterocolic lymphocytic phlebitis is a rare cause of GI ischemia apparently due to
inflammation. It usually involves the intestine, but one case had both gastric and
duodenal involvement associated with a chronic non-healing antral ulcer [85].
Hypothermia and hypoxia at high altitudes may also predispose to gastric and duodenal
erosions and ulcers [86,87].
Inflammatory and infiltrating disease
●Sarcoidosis – Involvement of the stomach is the most common site of
sarcoidosis in the GI tract, almost always occurring in association with pulmonary
disease [88,89]. Ulceration resembling PUD can occur with or without
enlargement of mucosal folds. (See "Gastrointestinal, hepatic, pancreatic, and
peritoneal sarcoidosis".)
●Crohn disease – Crohn disease of the stomach and/or duodenum is not
infrequently found [90] but only occasionally causes clinical impact. In the large
majority of cases, more distal Crohn disease is also evident. Involvement presents

336
 with obstruction, ulceration, fistula formation, or bleeding [91]. (See "Clinical
manifestations, diagnosis, and prognosis of Crohn disease in adults", section on
'Other gastrointestinal features'.)
●Other gastroenteritides – Case reports indicate that peptic ulcers, including
perforated ulcers, can occur in association with eosinophilic gastroenteritis [92-94]
and hypereosinophilic syndromes [95]. Refractory gastric ulcers have also been
associated with abundant IgG4-positive plasma cell infiltration and granulomatosis
with polyangiitis [96,97]. (See "Eosinophilic gastrointestinal
diseases" and "Granulomatosis with polyangiitis and microscopic polyangiitis:
Clinical manifestations and diagnosis".)
Other
●Idiopathic hypersecretory duodenal ulcer – Approximately 10 percent of
patients with duodenal ulcers have basal acid output above 15 mmol per hour,
normal serum gastrin levels, and ulcer disease unrelated to H. pylori infection [98].
These cases appear to represent a form of hypersecretory duodenal ulcer, the
mechanisms of which remain to be defined. The acid hypersecretion appears to
respond to long-term treatment with a PPI [98,99]. Such patients may require
relatively high doses of a PPI, as in gastrinoma.
●Stress ulcers in the intensive care unit – Estimates of the incidence of overt GI
bleeding from stress-induced ulceration range from 1.5 to 8.5 percent among all
intensive care unit patients, but may be as high as 15 percent among patients who
do not receive stress ulcer prophylaxis. Patient selection for prophylaxis is
discussed elsewhere. (See "Stress ulcers in the intensive care unit: Diagnosis,
management, and prevention".)
EVALUATION OF H. PYLORI AND NSAID NEGATIVE ULCERS Before
considering unusual causes of peptic ulcer disease (PUD), H. pylori and NSAID use
should be excluded since these two risk factors still account for the large majority of
cases of PUD. False-negative testing for H. pylori and a failure to detect NSAID use are
probably the most common causes of apparently H. pylori-negative, NSAID-negative
ulcers. For H. pylori, particular caution is needed because many of the patients being
evaluated will have been treated with a proton pump inhibitor or antibiotics, rendering
tests dependent upon bacterial density less reliable.
History — The initial evaluation for patients with non-H. pylori, non-NSAID related
peptic ulcer disease should include an assessment for associated diseases as well as
other possible comorbidities.

Some of the key elements of the history include the following:

●Medications (including over-the-counter drugs and supplements).

●Cocaine or methamphetamine use.
●Prior gastric surgery including bariatric procedures or radiation.
●Comorbidities including Crohn disease, sarcoidosis, mastocytosis, multiple
endocrine neoplasia type 1 (MEN1).
●Concomitant history of a hematologic disease including chronic myeloid leukemia
and polycythemia vera.
●Associated symptoms, especially diarrhea which may be due to Zollinger-Ellison
syndrome, Crohn disease, or systemic mastocytosis.
337
 ●A history of peptic ulcers that are resistant to medical therapy, recurrent ulcers,
multiple ulcers beyond the duodenal bulb, which are suggestive of Zollinger-
Ellison Syndrome (ZES) (table 2).
●Family history of peptic ulcer disease or MEN1 may indicate the presence of
ZES.

If an unusual cause of a peptic ulcer is suspected based on the history, additional

testing should be performed as needed to establish the etiology.

Laboratory testing in selected patients — Initial evaluation in a patient with

suspected ZES is with measurement of fasting serum gastrin concentration and
measurement of gastric pH (table 2). In patients with elevated serum gastrin levels/low
gastric pH that are not diagnostic for ZES, we perform a secretin stimulation test.
Laboratory evaluation for ZES is discussed in detail separately. (See 'History' above
and 'Acid hypersecretory states' above and "Zollinger-Ellison syndrome (gastrinoma):
Clinical manifestations and diagnosis", section on 'Clinical manifestations'.)
Repeat upper endoscopy in ulcers without a clear etiology — In patients with
gastric or duodenal ulcers without a clear etiology, we perform an upper endoscopy 8 to
12 weeks after initiating medical therapy (with biopsies of the ulcer if still present) in
order to exclude a malignancy. This upper endoscopy allows for additional biopsies of
the ulcer to exclude neoplastic, infiltrative, or infectious causes of ulceration. We biopsy
both the ulcers and the surrounding mucosa. In addition, we biopsy the duodenum to
detect isolated duodenal colonization by H. pylori. (See 'Bacteria' above
and 'Inflammatory and infiltrating disease' above.)
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Peptic ulcer disease".)
SUMMARY AND RECOMMENDATIONS
●Peptic ulcers are defects in the gastrointestinal mucosa that extend through the
muscularis mucosae. They persist as a function of the acid/peptic activity in
gastric juice. Peptic ulcer disease (PUD) is an important cause of morbidity and
health care costs. Helicobacter pylori and NSAIDs account for the large majority of
cases of peptic ulcer disease. (See 'Introduction' above
and 'Epidemiology' above.)
●Several medications have been implicated in the development of peptic ulcers
either alone or synergistically with other ulcerogenic or antithrombotic agents.
Drugs implicated include acetaminophen (in doses greater than 2 or 3 g daily),
bisphosphonates, cocaine, glucocorticoids, selective serotonin reuptake
inhibitors, clopidogrel, and erlotinib. The risk of some agents may only be evident
in patients with an ulcer or bleeding diathesis, such as those with a prior history of
PUD or bleeding complications. (See 'Non-NSAID medications' above.)
●Other causes of peptic ulcer disease include infections (eg, non-pylori
Helicobacter, Cytomegalovirus), mechanical (eg, post-obstructive, foreign body,
surgery), ischemia, inflammatory and infiltrating diseases of the stomach (eg,
Crohn disease), and acid hypersecretory states (eg, Zollinger-Ellison syndrome).
(See 'Etiology' above.)

338
●Before considering unusual causes of peptic ulcer disease, H. pylori and NSAID
use should be excluded since these two risk factors still account for the large
majority of cases of PUD. False-negative testing for H. pylori and a failure to
detect NSAID use are probably the most common causes of apparently H. pylori-
negative, NSAID-negative ulcers. For H. pylori, particular caution is needed
because many of the patients being evaluated will have been treated with a proton
pump inhibitor or antibiotics, rendering tests dependent upon bacterial density less
reliable. (See 'Evaluation of H. pylori and NSAID negative ulcers' above.)
●The initial evaluation for patients with non-H. pylori, non-NSAID related peptic
ulcer disease should include an assessment for associated diseases as well as
other possible comorbidities. Zollinger-Ellison Syndrome should be suspected in
patients with multiple or refractory peptic ulcers; ulcers distal to the duodenum;
peptic ulcer disease and diarrhea, enlarged gastric folds, or multiple endocrine
neoplasia type 1 (MEN1). ZES should also be suspected in patients with peptic
ulcer disease and a family history of peptic ulcer disease or MEN1.
(See 'History' above.)
●In patients with gastric or duodenal ulcers without a clear etiology, we perform an
upper endoscopy 8 to 12 weeks after initiating medical therapy (with biopsies of
the ulcer if still present) in order to exclude a malignancy. This upper endoscopy
allows for additional biopsies of the ulcer to exclude neoplastic, infiltrative, or
infectious causes of ulceration. We biopsy both the ulcers and the surrounding
mucosa. In addition, we biopsy the duodenum to detect isolated duodenal
colonization by H. pylori. (See 'Repeat upper endoscopy in ulcers without a clear
etiology' above.)

339
Zollinger-Ellison syndrome (gastrinoma): Clinical
manifestations and diagnosis
Author:
Emily Bergsland, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Nov 30, 2021.
What's New
Secretin stimulation testing in suspected Zollinger-Ellison syndrome (November
2021)
In patients being evaluated for Zollinger-Ellison syndrome (ZES), discontinuation of
proton pump inhibitors (PPIs) is recommended before a secretin stimulation test (SST).
In a retrospective study of SSTs that compared test performance on versus off PPIs, the
sensitivity, specificity, and positive predictive value were comparable with no false
positive or negative results on PPI therapy [1]. However, the majority of patients had a
high pretest probability of ZES due to MEN1 syndrome. We continue to suggest PPI
cessation for 1 week prior to SST for ZES but administer high-dose H2 receptor
antagonists until 12 to 30 hours prior to the test and only perform testing once active
ulcers have healed, as stopping a PPI in patients with ZES can cause massive acid
hypersecretion and lead to acute ulcer complications. (See "Zollinger-Ellison syndrome
(gastrinoma): Clinical manifestations and diagnosis", section on 'Secretin stimulation
test'.)
Read more

INTRODUCTIONZollinger-Ellison (ZES) syndrome is characterized by gastric acid

hypersecretion resulting in severe acid-related peptic disease and diarrhea [1,2]. The
clinical manifestations and diagnosis of ZES will be reviewed here. The management of
ZES is discussed separately. (See "Management and prognosis of the Zollinger-Ellison
syndrome (gastrinoma)".)
EPIDEMIOLOGYZollinger-Ellison syndrome (ZES) is caused by secretion of gastrin
by duodenal or pancreatic neuroendocrine tumors (gastrinomas). The annual incidence
of gastrinomas is 0.5 to 2 per million population [3-5]. Most patients are diagnosed
between the ages of 20 and 50, with a higher incidence in males as compared with
females [6]. Approximately 80 percent of gastrinomas are sporadic, but 20 to 30 percent
occur in association with multiple endocrine neoplasia type 1 (MEN1) [7,8].
(See "Multiple endocrine neoplasia type 1: Clinical manifestations and diagnosis".)
Although gastrinomas are one of the most common functional pancreatic
neuroendocrine tumors, only 20 to 25 percent of gastrinomas arise in the pancreas
[6,9]. Approximately 50 to 88 percent of patients with sporadic ZES, and 70 to 100
percent of patients with ZES associated with MEN1, have duodenal gastrinomas [8].
Duodenal gastrinomas are predominantly found in the first part of the duodenum. As
340
compared with pancreatic gastrinomas, duodenal gastrinomas are usually small (<1
cm), are often multiple, and are less likely to have metastasized to the liver at diagnosis
(0 to 10 versus 22 to 35 percent) [6,7,10,11]. In 5 to 15 percent of patients, gastrinomas
arise in non-pancreatic, non-duodenal abdominal (stomach, peripancreatic lymph
nodes, liver, bile duct, ovary), and extra-abdominal (heart, small cell lung cancer)
locations [12-14].
CLASSIFICATION, NOMENCLATURE, AND HISTOLOGYThe World
Health Organization (WHO) classifies neuroendocrine tumors (NETs) arising within the
digestive system based upon the extent to which they resemble their normal non-
neoplastic counterparts (table 1). (See "Pathology, classification, and grading of
neuroendocrine neoplasms arising in the digestive system", section on '2010 and 2019
World Health Organization classification'.)
Histologically, most gastrinomas are well-differentiated NETs with few mitoses and a
histologic appearance that is similar to that of other pancreatic NETs. The cells are
arranged in a solid, trabecular, gyriform, or glandular pattern, with fairly uniform nuclei,
salt-and-pepper chromatin, and finely granular cytoplasm. As with other pancreatic
NETs, the degree of malignancy cannot be predicted by morphologic appearance alone.
The cells produce abundant neurosecretory granules, as reflected in the strong and
diffuse immunohistochemical expression of neuroendocrine markers such as
synaptophysin and chromogranin. Gastrin is the predominant peptide within the
secretory granules of gastrinoma cells, but other neuroendocrine peptides such as
vasoactive intestinal peptide and glucagon can sometimes be identified as well. The
designation of the tumor as a gastrinoma is based upon the presence of a clinical
syndrome that results from tumor production and secretion of gastrin, and not by its
morphologic appearance or the presence of gastrin in the secretory granules. Of note,
not all patients with ectopic gastrin secretion have the symptoms associated with
Zollinger-Ellison syndrome, as in many cases, the hormone is not processed to
biologically active gastrin [15]. If a tumor stains for gastrin or secretes gastrin but does
not produce symptoms of Zollinger-Ellison syndrome, it should not be considered a
gastrinoma. (See "Pathology, classification, and grading of neuroendocrine neoplasms
arising in the digestive system", section on 'Functionality and nomenclature'.)
PATHOPHYSIOLOGYExcessive gastrin secretion from a gastrinoma results in
high gastric acid output (usually four- to sixfold, and up to over 10-fold) due to the
trophic action of gastrin on parietal cells and histamine-secreting enterochromaffin-like
(ECL) cells (picture 1) [15]. In addition, gastrin stimulates parietal cells largely via the
release of histamine. (See "Physiology of gastric acid secretion".)
Chronic diarrhea in Zollinger-Ellison syndrome results from the following [6]:
●The high volume of gastric acid secretion that cannot be fully reabsorbed by the
small intestine and colon.
●The rate of gastric acid secretion exceeds the neutralizing capacity of pancreatic
bicarbonate secretion, resulting in an exceptionally low pH of intestinal contents.
The low pH inactivates pancreatic digestive enzymes, interfering with the
emulsification of fat by bile acids, and damaging intestinal epithelial cells and villi.
Maldigestion and malabsorption both result in steatorrhea.

341
 ●Extremely high serum gastrin concentrations inhibit the absorption of sodium and
water by the small intestine, thereby adding a secretory component to the
diarrhea.
CLINICAL MANIFESTATIONS
Clinical presentation — Peptic ulcer disease (73 to 98 percent), heartburn (52 to 55
percent), diarrhea (60 to 75 percent), weight loss (7 to 53 percent), and complications
from acid hypersecretion (bleeding, stricture, fistulization, perforation) are the most
common symptoms in patients with Zollinger-Ellison syndrome (ZES) (figure 1) [15,16].
Sixty to ninety percent of gastrinomas are malignant [8]. The mean time to diagnosis is
six years [16].
Approximately 1 to 10 percent of patients, especially with metastatic disease or multiple
endocrine neoplasia type 1 (MEN1), have symptoms due to a second hormonal
syndrome (eg, VIPoma, somatostatinoma, glucagonoma, ACTH) [17].
(See "Epidemiology and clinical manifestations of Cushing's syndrome", section on
'Clinical manifestations' and "Somatostatinoma: Clinical manifestations, diagnosis, and
management", section on 'Clinical manifestations' and "VIPoma: Clinical manifestations,
diagnosis, and management", section on 'Clinical features' and "Glucagonoma and the
glucagonoma syndrome", section on 'Clinical features'.)
Endoscopic features — Over 90 percent of patients with ZES develop peptic ulcers
[18]. Patients with ZES, like those with sporadic peptic ulcer disease, often present with
solitary ulcers less than 1 cm in diameter. Approximately 75 percent of ulcers are in the
first portion of the duodenum, 14 percent in the distal duodenum, and 11 percent in the
jejunum [19]. Ulcers are more likely to be refractory to proton pump inhibitor therapy and
to recur as compared with patients with sporadic ulcer disease. Furthermore, in ZES,
ulcers often occur in unusual locations (eg, beyond the first or second fold of the
duodenum). Over 90 percent of patients with ZES often have prominent gastric folds.
Patients may also have evidence of reflux esophagitis. However, strictures of the
esophagus, pylorus, or duodenum are present in less than 10 percent of patients [16].
In the setting of MEN1, duodenal gastrinomas are typically multifocal, small (<0.5 cm),
and associated with lymph node involvement in 40 to 60 percent [15].
DIAGNOSISZollinger-Ellison syndrome (ZES) should be suspected in patients with
multiple or refractory peptic ulcers; ulcers distal to the duodenum; peptic ulcer disease
and diarrhea, enlarged gastric folds, or multiple endocrine neoplasia type 1 (MEN1)
(table 2) [20]. ZES should also be suspected in patients with peptic ulcer disease and a
family history of peptic ulcer disease or MEN1, or in patients with diarrhea that is
responsive to proton pump inhibitors (PPIs). The diagnosis is established by
demonstrating an elevated basal or stimulated gastrin concentration (in the setting of a
low gastric pH).
Evaluation — Initial evaluation in a patient with suspected ZES is with measurement of
fasting serum gastrin concentration and measurement of gastric pH. In patients with
elevated gastrin levels/low gastric pH that are not diagnostic for ZES, we perform
a secretin stimulation test. The calcium infusion study (intravenous infusion with calcium
gluconate) is usually reserved for patients with gastric acid hypersecretion in whom
there is a strong clinical suspicion of gastrinoma despite a negative secretin stimulation
test [21]. Importantly, the diagnosis of ZES can be difficult to make, as evidenced by the
fact the average time from onset of symptoms to diagnosis is >5 years. The symptoms
342
can be nonspecific and/or masked by PPI use [22]. Gastric antral/body biopsy for
atrophic gastritis combined with parietal cell and intrinsic factor antibodies can also be
useful for confirming situations of "appropriate" hypergastrinemia.
In patients in whom PPIs cannot safely be stopped, modern criteria for ZES have been
proposed [8]. An elevated gastrin level, combined with a history of current or recent
peptic ulcer disease, and improvement of diarrhea on a PPI supports a diagnosis of
ZES, particularly in the setting of a positive biopsy for a well differentiated
neuroendocrine tumor (in a patient with or without MEN1). The diagnosis is less clear
without a tissue diagnosis and with only a history of positive somatostatin receptor
imaging (with or without MEN1).

In patients with a positive imaging study and fasting hypergastrinemia without peptic
ulcer disease or diarrhea, a lack of atrophic gastritis and the absence of parietal cell and
intrinsic factor antibodies suggests gastrinoma. However, it is important to note that
atrophic gastritis can be missed with this approach as detection depends on biopsy
location and number, and parietal cell and intrinsic factor antibodies are not always
positive.

Furthermore, the use of somatostatin imaging (eg, 68Ga-DOTATATE positron emission

tomography [PET]/computed tomography [CT] or 111In-DTPA-octreotide with single-
photon emission CT [SPECT]/CT) for diagnosis is of limited value, as it can detect any
benign or malignant process overexpressing somatostatin receptors, and there can be
uptake in nonneoplastic conditions like arthritis, infections, thyroid disease,
granulomatous diseases [8].
Serum gastrin concentration — Fasting serum gastrin should be measured in any
patient suspected of having ZES [6]. A serum gastrin value greater than 10 times the
upper limit of normal (1000 pg/mL) in the presence of a gastric pH below 2 is diagnostic
of ZES. Higher levels are more likely with pancreatic (compared with duodenal) tumors,
larger tumor size, and with metastatic disease [6]. Measurement of gastric pH on a
single specimen is important to exclude secondary hypergastrinemia due to
achlorhydria (eg, atrophic gastritis, pangastritis-associated Helicobacter
pylori infections, renal failure, vagotomy, and ingestion of gastric acid antisecretory
drugs [eg, PPIs]). In such cases the serum gastrin level can exceed 1000 pg/mL, but
the gastric pH is >2. Importantly, these "appropriate" hypergastrinemic conditions are
much more common and need to be distinguished from inappropriate causes, such as
ZES [8]. While assessment of gastric pH is classically required for the diagnosis of ZES,
case series suggest that assessment of gastric acidity is underutilized. This is likely due
to the inability to stop PPIs or unfamiliarity with gastric pH testing).
Approximately two-thirds of patients with ZES have serum gastrin concentrations less
than 10 times the upper limit of normal (between 110 and 1000 pg/mL) [6]. This degree
of hypergastrinemia is nonspecific and can also be present in patients with increased
gastric acid secretion (eg, antral G-cell hyperplasia, gastric outlet obstruction, and
retained gastric antrum). Patients receiving PPIs generally have elevated serum gastrin
levels. Fasting serum gastrin levels can fluctuate even within the same patient, and
elevated levels should be rechecked (see 'Differential diagnosis' below). Importantly, no
level of hypergastrinemia can distinguish appropriate from inappropriate
hypergastrinemia; high levels can be seen in the setting of atrophic gastritis or PPI use
343
[8]. However, persistently normal fasting serum gastrin values are exceedingly
uncommon in ZES [8].
Commercial immunoassay kits in current use vary in their accuracy [22,23]. In one study
that evaluated 12 different commercial assay kits, four of the kits returned falsely low
results in 20 to 80 percent of patients [23]. False negatives are most likely to occur at
relatively low gastrin concentrations, with most assays accurately detecting patients with
ZES whose serum gastrin concentrations are greater than 400 pmol/L (845 pg/mL).
Secretin stimulation test — The secretin stimulation test is used to differentiate
patients with gastrinomas from other causes of hypergastrinemia (eg, in the setting of
gastrin, <10-fold upper limit of normal and gastric pH ≤2). Although, some have
questioned the test’s utility since achlorhydria can lead to false positives [8]. Secretin
stimulates the release of gastrin by gastrinoma cells, and patients with ZES tumors
have a dramatic rise in serum gastrin. In contrast, normal gastric G cells are inhibited by
secretin.
The secretin test should not be performed in a patient on PPIs. False-negative
responses have been reported in 6 to 20 percent of patients [24]. However, other
studies suggest that false-negative results caused by PPIs may be low in frequency,
particularly in patients with a high pretest probability of ZES [25]. False-positive results
occurred in 15 to 39 percent of patients with achlorhydria induced by PPIs or due to
chronic atrophic gastritis [26]. However, if PPIs are discontinued abruptly, patients with
ZES are at high risk to develop complications (such as acute bleeding and perforation)
during the interim week [27]. As such, discontinuation of a PPI in a patient suspected of
having ZES should be done by an experienced provider. Before this is attempted, an
endoscopy should be performed to exclude active ulcer disease. If ulceration is present,
suppressive PPI therapy should be used until active disease is healed before stopping
PPI therapy [8]. One week prior to the secretin study we substitute PPIs with high-dose
H2 receptor antagonists (eg, cimetidine 300 to 600 mg every six hours) until 12 to 30
hours prior to the test [8]. Oral antacids are then taken as needed until midnight prior to
the study. Patients should be advised to seek immediate medical attention for
nasogastric aspiration if they develop significant exacerbation of ZES symptoms (eg,
vomiting, pain, diarrhea) during the taper period [28].
The secretin stimulation test is performed by administering 0.4 micrograms/kg by rapid
infusion intravenously over one minute; a baseline fasting serum gastrin is measured
twice before the secretin is administered and 2, 5, and 10 minutes later. Several criteria
have been proposed to define a positive test; an increase in gastrin levels of greater
than 120 pg/mL over basal fasting levels has a sensitivity and a specificity of 94 and
100 percent, respectively (others use an absolute increase greater than 110 or 200
pg/mL, or a 50 percent increase in gastrin levels) [21,22,29]. Serum gastrin levels
usually peak by 10 minutes (figure 2).
The secretin stimulation test should not be performed in patients with active severe
manifestations of ZES. This includes patients with severe abdominal pain, vomiting and
diarrhea to the point of dehydration, or endoscopic findings of thickened gastric folds or
multiple ulcers, as they are at particular risk for life-threatening consequences of
discontinuing acid suppression. In such patients, tumor localization studies should be
performed. (See 'Tumor localization' below.)

344
Other — Other tests that are supportive of the diagnosis of ZES but are less commonly
performed include the following [8]:
●Serum chromogranin A – Serum chromogranin A is elevated in most patients
with gastrinomas, and the level of elevation tends to correlate with tumor volume
[30]. In contrast to other neuroendocrine tumors, very high levels of chromogranin
A can be seen in gastrinomas without liver metastases [31]. Chromogranin A
levels are usually normal or near normal in patients with high gastrin levels
secondary to chronic atrophic gastritis. (See "Clinical characteristics of well-
differentiated neuroendocrine (carcinoid) tumors arising in the gastrointestinal and
genitourinary tracts", section on 'Stomach'.)
However, serum chromogranin A is less sensitive for a gastrinoma as compared
with fasting serum gastrin levels. In addition, elevated chromogranin A is not
specific to gastrinomas and is a general marker for well-differentiated
neuroendocrine tumors [32]. Elevated chromogranin A levels may be seen in a
number of other conditions including PPI use [33,34].
●Gastric acid secretion studies – Gastric acid secretion studies to measure
basal acid output, which were once pivotal in establishing the diagnosis of ZES,
are no longer performed due to their technical difficulty [26]. Basal acid output is
measured by the passage of a nasogastric tube into the dependent portion of the
stomach with aspiration and quantification of gastric juice production over a one-
hour period. A basal acid output of >15 mEq/hour is supportive of the diagnosis of
ZES.
DIFFERENTIAL DIAGNOSIS
●Antral G-cell hyperplasia – Antral G-cell hyperplasia is a rare entity associated
with increase in the number of G cells and characterized by a marked
hypergastrinemia. Peptic ulcers can also be seen in patients with antral G-cell
hyperplasia. However, unlike Zollinger Ellison syndrome (ZES), antral G-cell
hyperplasia is characterized by a poor response to secretin stimulation test, and
absence of gastrinoma on imaging.
●Retained antrum syndrome – Retained antrum syndrome should be suspected
in patients with recurrent peptic ulceration after gastrectomy. In such cases, peptic
ulcer recurrence results from incomplete excision of the gastric antrum from the
duodenum. Gastrin elevation in patients with retained antrum syndrome is only
modest as compared with patients with ZES, and hypergastrinemia is reversible
with excision of the retained antral remnant.
TUMOR LOCALIZATIONAfter the diagnosis of Zollinger-Ellison syndrome (ZES)
is made, the gastrinoma must be located and staged. (See "Management and prognosis
of the Zollinger-Ellison syndrome (gastrinoma)" and "Multiple endocrine neoplasia type
1: Treatment".)
●Tumor localization begins with an upper endoscopy if not already performed,
cross-sectional imaging with helical, contrast-enhanced, triple-phase computed
tomography (CT) or magnetic resonance imaging (MRI), and somatostatin
receptor-based imaging (somatostatin receptor scintigraphy [SRS]) using 111-In
pentetreotide, or integrated PET/CT using Gallium-68-DOTA-0-Phe 1-Tyr3-
Octreotate (Gallium Ga-68 DOTATATE) or Gallium-68-DOTA-0-Phe1-Tyr3-
Octreotide (Gallium Ga-68 DOTATOC) [35,36]. Because of its greater sensitivity,
345
 Ga-68 DOTATATE or Ga-68 DOTATOC PET imaging may be preferable to
conventional SRS with 111-In pentetreotide [37,38]. (See 'Endoscopic
features' above and "Classification, epidemiology, clinical presentation,
localization, and staging of pancreatic neuroendocrine neoplasms", section on
'Somatostatin-receptor-based imaging'.)
●If CT/MRI and somatostatin receptor-based imaging are negative, and surgery is
being considered, an endoscopic ultrasound (EUS) should be performed because
of its greater sensitivity in detecting small tumors. EUS also permits fine-needle
aspiration for histological identification. (See "Classification, epidemiology, clinical
presentation, localization, and staging of pancreatic neuroendocrine neoplasms",
section on 'Endoscopic ultrasonography'.)
●We reserve invasive testing to localize the tumor with angiography or selective
arterial stimulation and venous sampling with secretin injection for patients who
are strongly suspected of having a gastrinoma in whom imaging is negative [39].
The sensitivity for pancreas tumors is 75 to 100 percent, and the specificity is 95
percent; the sensitivity for duodenal tumors is 38 to 63 percent [22].
●However, in some cases tumor localization can only be achieved at laparotomy
by direct palpation, duodenal transillumination, or intraoperative ultrasound.
(See "Classification, epidemiology, clinical presentation, localization, and staging
of pancreatic neuroendocrine neoplasms", section on 'Intraoperative localization
techniques' and 'Evaluation' above.)
STAGING SYSTEMTwo staging systems are available for pancreatic
neuroendocrine tumors such as gastrinomas, one from the combined American Joint
Committee on Cancer/Union for International Cancer Control (UICC) [40], and another
proposed by the European Neuroendocrine Tumor Society (table 3) [40,41]. Both
staging systems are highly prognostic for both relapse-free and overall survival [42-44].
The newest update of the TNM staging classification (8 th edition, 2017) has a staging
system for neuroendocrine tumors of the pancreas (table 4) that is separate from that
used for exocrine pancreatic tumors [45]. (See "Classification, epidemiology, clinical
presentation, localization, and staging of pancreatic neuroendocrine neoplasms",
section on 'Staging system'.)
ADDITIONAL EVALUATIONWe suggest biochemical studies to screen for
multiple endocrine neoplasia type 1 (MEN1) in patients with Zollinger-Ellison syndrome
(ZES), as 20 to 25 percent of patients have MEN1 and because up to 40 percent of
MEN1/ZES patients have no family history [13].
●All patients with ZES should have serum parathormone levels, ionized calcium
levels, and prolactin levels at diagnosis and periodically thereafter [13]. Screening
for MEN1-associated tumors is discussed separately. (See 'Epidemiology' above
and "Multiple endocrine neoplasia type 1: Clinical manifestations and diagnosis",
section on 'Monitoring for MEN1-associated tumors'.)
●Individuals with a family history of MEN1, suspicious clinical or laboratory findings
(eg, renal colic or nephrolithiases, history of hypercalcemia), or multiple MEN1
tumor types (parathyroid gland, anterior pituitary, and enteropancreatic) should
undergo evaluation for MEN1 syndrome. The management of patients with MEN1
is discussed separately. (See "Multiple endocrine neoplasia type 1: Treatment".)

346
 ●All patients with gastrinoma should at least be considered for testing for an
inherited genetic syndrome. Data suggest that 17 percent of patients with
seemingly sporadic pancreatic neuroendocrine tumors harbor germline alterations
in any one of a variety of genes (including MUTYH, CHEK2, and BRCA2, as well
as MEN1 and VHL).
SOCIETY GUIDELINE LINKSLinks to society and government-sponsored
guidelines from selected countries and regions around the world are provided
separately. (See "Society guideline links: Well-differentiated gastroenteropancreatic
neuroendocrine tumors".)
SUMMARY AND RECOMMENDATIONS
●Zollinger-Ellison syndrome (ZES) is caused by secretion of gastrin by duodenal
or pancreatic neuroendocrine tumors (gastrinomas). The annual incidence of
gastrinomas is 0.5 to 2 per million population. Most patients are diagnosed
between the ages of 20 and 50, with a higher incidence in males as compared
with females. While most gastrinomas are sporadic, 20 to 30 percent occur in
association with multiple endocrine neoplasia type 1 (MEN1) syndrome.
Approximately 50 to 88 percent of sporadic ZES patients, and 70 to 100 percent of
ZES and MEN1, have duodenal gastrinomas. (See 'Epidemiology' above
and 'Classification, nomenclature, and histology' above.)
●Excessive gastrin secretion from a gastrinoma results in high gastric acid output.
Chronic diarrhea results from failure of resorption of the increased gastric acid,
inactivation of pancreatic enzymes, damage to intestinal epithelium, and inhibition
of the absorption of sodium and water. (See 'Classification, nomenclature, and
histology' above and 'Pathophysiology' above.)
●Abdominal pain (75 percent) and diarrhea (73 percent) are the most common
symptoms in patients with ZES. Nearly half of patients have heartburn due to
gastroesophageal reflux (figure 1). Other symptoms include weight loss (17
percent) and gastrointestinal bleeding (25 percent). (See 'Clinical
presentation' above.)
●Over 90 percent of patients with ZES develop peptic ulcers. Patients with ZES
often present with solitary ulcers less than 1 cm in diameter. Approximately 75
percent of ulcers are in the first portion of the duodenum, 14 percent in the distal
duodenum, and 11 percent in the jejunum. Ulcers in ZES may be refractory to
proton pump inhibitors and recur much more often as compared with patients with
sporadic ulcer disease. On upper endoscopy, patients with ZES may have
prominent gastric folds and evidence of reflux esophagitis. However, strictures of
the esophagus, pylorus, or duodenum are present in less than 10 percent of
patients. (See 'Endoscopic features' above.)
●ZES should be suspected in patients with multiple or refractory peptic ulcers;
ulcers distal to the duodenum; peptic ulcer disease and diarrhea, enlarged gastric
folds, an endocrinopathy or MEN1 (table 2). ZES should also be suspected in
patients with peptic ulcer disease and a family history of peptic ulcer disease or
MEN1. The diagnosis is established by demonstrating an elevated basal or
stimulated gastrin concentration, ideally assessed off a PPI (if safe) and in the
setting of a gastric pH <2. (See 'Diagnosis' above.)

347
●Tumor localization begins with an upper endoscopy if not already performed,
cross-sectional imaging with helical, contrast-enhanced, triple-phase computed
tomography (CT) or magnetic resonance imaging (MRI), and somatostatin
receptor-based imaging using somatostatin receptor scintigraphy (SRS) with 111-
In pentetreotide or integrated PET/CT using Gallium Ga-68 DOTATATE or Ga-68
DOTATOC. Because of its greater sensitivity, Ga-68 DOTATATE or Ga-68
DOTATOC PET/CT is preferred over conventional SRS with 111-In pentetreotide,
where available. If CT or MRI and somatostatin receptor-based imaging are
negative, and surgery is being considered, endoscopic ultrasound should be
performed to localize the tumor. We reserve invasive testing to localize the tumor
with angiography or selective arterial stimulation and venous sampling
with secretin injection for patients who are strongly suspected of having a
gastrinoma in whom imaging is negative. However, in some cases, tumor
localization can only be achieved at laparotomy, by direct palpation, duodenal
transillumination, or intraoperative ultrasound. (See 'Tumor localization' above.)
●All patients with gastrinoma should at least be considered for testing for inherited
genetic syndromes. Approximately 20 to 30 percent of patients develop ZES in the
setting of MEN1 syndrome. As such, all patients with ZES should be screened for
MEN1 syndrome with serum parathormone levels, ionized calcium levels, and
prolactin levels at diagnosis, and periodically thereafter. Individuals with a family
history of MEN1, suspicious clinical or laboratory findings (eg, renal colic or
nephrolithiases, history of hypercalcemia), or multiple MEN1 tumor types
(parathyroid gland, anterior pituitary, and enteropancreatic) should undergo
evaluation for MEN1 syndrome. (See 'Additional evaluation' above and "Multiple
endocrine neoplasia type 1: Clinical manifestations and diagnosis", section on
'Diagnosis'.)

348
Physiology of gastric acid secretion
Author:
Nimish B Vakil, MD, AGAF, FACP, FACG, FASGE
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Apr 01, 2020.
INTRODUCTIONThe regulation of acid and pepsin secretion reflects an intricate
balance of chemotransmitters delivered to the gastric mucosa by several pathways that
mediate both stimulatory and inhibitory mechanisms [1]. Similarly, several mechanisms
contribute to the remarkable ability of normal gastroduodenal mucosa to defend itself
against injury from the acid/peptic activity in gastric juice and to rapidly repair injury
when it does occur. Secretory, defense, and healing mechanisms are regulated by the
same type of overlapping neural, endocrine, paracrine, and autocrine control pathways.
Gastric acid facilitates the digestion of protein and the absorption of iron, calcium,
vitamin B12, and is necessary for the absorption of some drugs such
as ketoconazole, itraconazole, and thyroid hormone [2]. Gastric acid, by lowering pH,
kills ingested microorganisms and limits bacterial growth in the stomach and prevents
intestinal infections such as Clostridioides difficile. In addition, gastric acid may have a
role in preventing spontaneous bacterial peritonitis [3-5].
FUNCTIONAL ANATOMY OF THE STOMACHThe stomach consists of three
anatomical (fundus, corpus, and antrum) and two functional areas (oxyntic and pyloric
glands). The oxyntic area comprises approximately 80 percent of the stomach and
contains parietal cells that produce gastric acid. Also present in the oxyntic glands are
neuroendocrine cells producing paracrine and hormonal agents that modify parietal cell
activity. The antrum of the stomach contains pyloric glands and their main feature is the
presence of gastrin secreting G cells. Somatostatin secreting D cells are present in the
pyloric and oxyntic glands and modulate gastrin release [1].
PHASES OF ACID SECRETIONThe physiologic stimulation of acid secretion
has classically been divided into three interrelated phases: cephalic, gastric, and
intestinal [6].
●The cephalic phase is activated by the thought, taste, smell, and sight of food,
and swallowing. It is mediated mostly by cholinergic/vagal mechanisms.
●The gastric phase is due to the chemical effects of food and distension of the
stomach. Gastrin appears to be the major mediator since the response to food is
largely inhibited by blocking gastrin action at its receptors.
●The intestinal phase accounts for only a small proportion of the acid secretory
response to a meal; its mediators remain controversial.

The observation that H2 receptor antagonists block the cephalic and gastric phases
underscores the importance of histamine mediation of the stimulatory response, and
illustrates the interdependence of the different phases.
349
SECRETION OF ACID AND PEPSINGastric acid secretion from parietal cells is
regulated by redundant, overlapping pathways, which include endocrine (gastrin),
paracrine (locally delivered histamine and somatostatin), neural (acetylcholine), and
probably autocrine (transforming growth factor-alpha) factors (figure 1) [7,8].
The parietal cell — In the resting state, parietal cells are filled with secretory vesicles
that coalesce with stimulation to form channels (canaliculi) that drain to the apical lumen
[9]. The secretory membrane lining these structures contains the hydrogen-potassium-
ATPase acid-secreting pump. This pump is always active, but exists in a short-circuited
state in resting vesicles because the pathway necessary for transporting potassium to
the apical surface for exchange with hydrogen is not present or active. With stimulation,
this pathway for potassium-chloride cotransport becomes active, allowing hydrogen-
potassium exchange to occur [10-13].
Parietal cell activation involves an increase in cytoplasmic calcium or generation of
cyclic AMP, followed by activation of a cAMP-dependent protein kinase cascade that
triggers translocation of proton pump containing membranes to the apical surface [6].
The cessation of acid secretion is associated with the re-internalization of the hydrogen-
potassium-ATPase pump. This process is mediated by the cytoplasmic tail of the beta
subunit of the pump. Transgenic mice with a mutation at this site constitutively secrete
acid and continuously express hydrogen-potassium-ATPase at the cell surface [9,14].
Gastrin — Gastrin is the major endocrine regulator of the secretory response to a
protein meal. It is released from gastrin-expressing cells (G cells) localized to the
antrum. Gastrin enhances gastric acid secretion from parietal cells primarily by
stimulating the synthesis and release of histamine from oxyntic mucosal
enterochromaffin-like (ECL) cells [15-18]. However, gastrin also has direct actions on
parietal cells [19,20].
Acid secretion is tightly controlled by a second hormone, somatostatin, which is a potent
inhibitor of both gastrin and histamine synthesis and release, and, therefore, of gastric
acid secretion. (See 'Somatostatin' below and "Physiology of gastrin".)
Gastrin is the best identified trophic regulator of parietal cell mass in humans. This
relationship is evidenced by the presence of gastric hypertrophy in gastrinoma patients
who have chronic exposure to elevated gastrin levels (picture 1), and atrophy of the
parietal cell mass with antrectomy, which decreases gastrin levels.
Gastrin receptors — Cholecystokinin and gastrin have an identical pentapeptide
terminal sequence. Gastrin acts via activation of the cholecystokinin CCK2 receptor
(formerly known as the CCK-B or gastrin receptor), which has equal affinity for
cholecystokinin (CCK) and gastrin [21]. These receptors have been localized to parietal
and ECL cells, but it is likely that the ECL cell gastrin receptor is of greater importance
in regulating acid secretion [18-20]. Mutant mice lacking CCK2 receptors have a
reduction in gastric acid secretion that is primarily due to a reduced number of parietal
and ECL cells and decreased expression of the histamine-forming enzyme histidine
decarboxylase [22,23]. Similar changes are seen in gastrin-deficient mice [23-25].
Gastrin "receptors" have also been found on somatostatin-secreting D cells. However,
this receptor is a CCK1 that has much greater affinity for CCK than for gastrin [26]. This
difference in receptor affinity may explain why gastrin is so much more effective as a
stimulant of acid secretion, while CCK induces greater release of the inhibitor
somatostatin [27]. Knockout mice that have been genetically engineered to be deficient
350
in CCK2 receptors have low acid secretion, while double knockout mice (deficient in
both gastrin and CCK receptors) have robust acid secretion in response to vagal
stimulation or exogenous histamine [28]. These findings are consistent with the
conclusion that CCK1 receptors exert inhibitory effects on acid secretion in vivo,
mediated by release of endogenous somatostatin. (See "Physiology of gastrin".)

The localization of the receptors responsible for trophic actions of gastrin remains
uncertain; the primary receptor appears to be on the ECL-like cell, although receptors
may also be present on gastric stem cells.

Regulation of gastrin secretion — The main stimulants of acid secretion are

histamine released from the enterochromaffin cells (paracrine secretion), gastrin
released from G cells (hormonal), and acetyl choline released from postganglionic
enteric neurons. The principal inhibitor of acid secretion is somatostatin, which is
released from oxyntic glands and antral D cells (paracrine). Physiologic factors
enhancing secretion are vagal activation, food (particularly amino acids), and gastric
distention. In contrast, carbohydrates and fat inhibit acid secretion. Intestinal exposure
is required for the carbohydrate effect, but the mechanisms are uncertain. Fat
stimulates the release of CCK, which is a potent inhibitor of acid secretion; fat also
releases other potential mediators and it activates neural responses.

Complex mechanisms control gastrin release from the antral G cells.

Two meal-related factors stimulate gastrin secretion: gastric distention and amino acids.

●The effect of gastric distention varies with the degree of distension [29]. Low
grade distention activates vasoactive intestinal peptide neurons, which stimulates
somatostatin release and therefore inhibits gastrin secretion. Higher grade
distention causes cholinergic activation, which reverses the pattern to one of
increased gastrin and reduced somatostatin secretion.
●Amino acids induce gastrin release; direct actions on the G cell have been
demonstrated but amino acids also activate both cholinergic neurons and
bombesin neurons [30]. The release of bombesin (also called gastrin-releasing
peptide) from mucosal nerves directly stimulates the G cell [31-33].
●Gastrin itself contributes to this process by enhancing the secretion of
somatostatin [30].
●Cholinergic activation after gastric distention or in response to a meal promotes
acid secretion by shifting the balance of stimulatory and inhibitory mechanisms
toward the stimulatory side, directly activating the parietal cell and stimulating
gastrin release while suppressing somatostatin release [1,29,30].
Histamine — Histamine is the major paracrine stimulator of acid secretion. It is
localized both in mucosal mast cells and in endocrine cells, the latter called ECL cells
because of the silver-staining properties of their granules. The ECL cells are localized to
the acid-secreting oxyntic or body mucosa, in direct proximity to the parietal cell.
Gastrin is the primary stimulus to histamine release from ECL cells [15,18,34]. ECL cells
are also directly stimulated by pituitary adenylate cyclase-activating polypeptide
(PACAP) and vasoactive intestinal peptide (VIP) [24,35]. Somatostatin is a major direct
351
inhibitor of histamine release; calcitonin gene-related peptide (CGRP), peptide YY,
prostaglandins, and galanin also inhibit release [1]. Stimulated ECL cells promptly
degranulate, with release of histamine and pancreastatin from the vesicles; this is
followed by an increase in histamine synthesis [17]. Although gastric mast cells
outnumber ECL cells, gastrin has only been demonstrated to release histamine from
ECL cells [36].
Several lines of evidence indicate that ECL cell histamine is the major physiological
mediator of acid secretion [18,24,37,38]. Inhibitors of the histamine-forming enzyme
histidine decarboxylase (HDC) block the acid secretory response to gastrin, but not to
histamine [24]. Furthermore, both H2 receptor deficient mice and HDC-knockout mice
have near normal basal acid secretion, a preserved acid secretory response to
cholinergic agents, an absent acid secretory response to exogenous gastrin, and
hypergastrinemia [23,24]. Mast cells may deliver histamine to parietal cells following
exposure to certain antigens (or in patients with systemic mastocytosis), but there is no
evidence that they have a role in the normal physiology of acid secretion.
The effects of histamine are largely mediated by the H2 receptors, which explain the
efficacy of H2 receptor blockers in the treatment of acid-peptic disease [39].
(See "Peptic ulcer disease: Treatment and secondary prevention" and "Medical
management of gastroesophageal reflux disease in adults".) These drugs inhibit acid
secretion in response to gastrin, meal, and neural stimulation, clearly establishing that
histamine plays a role as a universal mediator or modulator of the acid secretory
response. Histamine may also act at H3 receptors to increase acid secretion via
inhibition of somatostatin release [40].
Ghrelin — Ghrelin is a 28-amino acid peptide present mainly in the oxyntic mucosa of
the stomach. Ghrelin increases food intake and has been reported to stimulate acid
secretion via release of histamine from ECL cells [41].
Somatostatin — Somatostatin is a potent inhibitor of acid secretion [42]. It is released
from D cells, which are present throughout the gastric mucosa. Although somatostatin
has some effects on parietal cells, its major effects are exerted on the inhibition of
histamine release and to a lesser extent on gastrin release [7]. The secretion of
somatostatin is increased by gastric acid and by gastrin itself, suggesting that a major
function of somatostatin is to modulate the feedback inhibition of the acid secretory
response to gastrin [43]. Consistent with this hypothesis is the observation that mice
lacking the somatostatin receptor have a tenfold increase in basal acid output [44]. This
effect can be abolished by gastrin antibody, suggesting that somatostatin suppresses
gastric acid secretion via inhibition of the action of gastrin. There may be some effect on
gastrin release [32], but it is likely that somatostatin primarily acts by suppressing
gastrin-stimulated release of histamine from ECL cells [34].
Somatostatin secretion is also affected by neural inputs. It is suppressed by cholinergic
activation and increased by vasoactive intestinal peptide activation [29,30].
Neural control — Neural input may serve as an important integrator of secretory
function. The mucosal nerves, containing acetylcholine, bombesin, VIP, and PACAP
mediate the response to the cephalic phase of acid secretion and to gastric distention
and amino acids [29,30].
Acetylcholine is the major stimulatory mediator [7]. The major effects of muscarinic
receptor activation are to increase gastrin release [15], stimulate parietal cells [7], and

352
inhibit somatostatin secretion [30]. Bombesin release also stimulates acid secretion [30],
an effect that is mediated at least in part by enhanced gastrin release [31-33].
Muscarinic receptors on the parietal cells are of the M3 type and these receptors can be
stimulated by fermented alcoholic beverages resulting in increased acid secretion [45].
Vasoactive intestinal peptide release has a dual effect: a weak transient increase in acid
secretion, possibly due to direct effects on ECL cells; and a sustained reduction due to
enhanced release of somatostatin [42,46].
Orexin, nitric oxide, and galanin may also contribute to the neural regulation of acid
secretion [47].
Prostaglandins — Prostaglandins are autocrine factors that inhibit acid secretion,
histamine-stimulated parietal cell function [7], and gastrin-stimulated histamine release
[48,49]. The effect on gastrin release is less clear as both inhibitory and stimulatory
mechanisms have been described [50]. They are generated from cells in the epithelium
and lamina propria. Macrophages and capillary endothelial cells appear to be the
primary source [51]. The mechanisms regulating their release in vivo are not well
understood.
Other secretory regulators — Transforming growth factor-alpha (TGF-alpha) is an
autocrine factor that is present in parietal cells and inhibits gastric acid secretion [52,53].
Peptide YY (PYY) is released postprandially from cells in the ileum and colon and
inhibits the cephalic and gastric phases of acid secretion via central and peripheral
effects [54]. PYY binds to receptors on ECL cells and inhibits gastrin-stimulated
histamine release [54]. (See "Pancreatic polypeptide, peptide YY, and neuropeptide Y".)
Pepsin — Despite the fact that ulcers of the stomach and duodenum are referred to as
"peptic," pepsin and peptic activity have received little attention. Acid plus pepsin is
much more ulcerogenic than acid alone [55], leaving little question that the "peptic" label
appropriately reflects the critical role in ulcer formation of the proteolytic activity in
gastric juice. The potentiating effect of pepsin may be due in part to its mucolytic activity
[56].
Peptic activity is closely linked to acid secretion and gastric pH [57]. This relation is
partly due to peptic digests of dietary protein (primarily amino acids), which are potent
stimulants of gastrin release and acid secretion [30]. In addition, pepsinogen is
converted to the active protease pepsin at low gastric pH; on the other hand, pepsin is
inactivated when the pH is increased above 4 [55]. This pH dependence probably
accounts for the requirement for elevating the intraluminal pH above 4 to heal refractory
ulcers. (See "Approach to refractory peptic ulcer disease".)
Pepsinogen secretion is enhanced by acetylcholine and peptides of the CCK/gastrin
family [58]. In addition, agents that raise cyclic AMP, such as secretin and vasoactive
intestinal peptide, increase pepsinogen secretion in vitro.
MEASURING ACID SECRETION IN HUMANSAspiration of gastric contents
via a nasogastric tube is the easiest method of measuring acid secretion, if collections
are complete [8]. Basal acid secretion can also be reliably measured through an
endoscope during a 15-minute collection period. Alternatively, intragastric titration
allows the actual level of acid secretion to be measured by the quantity of base required
to hold the gastric pH at a predetermined level. Placement of a gastric pH probe allows
hydrogen ion concentration to be measured over a 24-hour period, but measuring

353
hydrogen ion concentration provides only an indirect indicator of the rate of acid
secretion.

Basal acid output (BAO) is the level of acid secretion when the subject is unstimulated;
measurements are widely variable among individuals. Values in normal subjects are
less than 10 mEq per hour, with an average of about 2 mEq per hour.

Maximal acid output is usually measured in response to pentagastrin and averages

about 30 mEq per hour. The upper limits of the normal range (95 percent confidence
level) is about 60 mEq per hour, underlining the point that acid secretion between
duodenal ulcer and normal overlaps considerably.

Tolerance and acid rebound — Two interesting and possibly related aspects of

antisecretory therapy have come to light: tolerance to the administration of H2 receptor
antagonists (H2RA) and acid rebound [18,59].
Tolerance to H2RA develops after as few as seven days of therapy, with diminished
effectiveness against nocturnal and pentagastrin-stimulated acid secretion [60,61]. In
one study, for example, intravenous H2RAs reduced pentagastrin-stimulated acid
secretion by 95 percent prior to oral therapy, compared with only 62 percent inhibition
after nine months of oral therapy [61].

The clinical relevance of tolerance has not been established. It may contribute to the
poor clinical response to H2RA in some patients with ulcer disease. Tolerance does not
occur with proton pump inhibitors, probably because they block the final stage of acid
secretion, the hydrogen-potassium-ATPase pump.

Rebound acid hypersecretion occurs after the cessation of one to nine months of H2RA
therapy. Increases have been noted in nocturnal acid secretion and in the acid
secretory response to a meal [62-64]. In one study, for example, cessation of H2RAs
after 25 days of therapy was associated with a significant increase in intragastric acidity
(17 percent at day three and 14 percent at day six) compared with pretreatment values;
rebound was no longer present after day nine [63]. The magnitude of the rebound
appears to reflect the degree and duration of secretory inhibition.
Although initially controversial [63], larger studies have also found rebound following
treatment with proton pump inhibitors [65-67]. Rebound appears to be more common in
patients not infected with Helicobacter pylori (mean 82 percent increase in basal acid
output and 28 percent increase in maximal acid output at day 15 after cessation,
compared with pretreatment) [65].
A double blind, placebo-controlled study with healthy volunteers found that
approximately 40 percent of subjects treated with eight weeks of PPI had heartburn or
dyspepsia in the four weeks after PPI treatment was stopped, compared with 15 percent
in the placebo group [67]. The study did not directly relate symptoms to rebound acid
hypersecretion, but it suggests that rebound after stopping PPI treatment can provoke
symptoms. A meta-analysis showed that rebound results in symptoms in healthy
volunteers treated short term with PPIs but in patients who have been treated
chronically with proton pump inhibitors, the clinical significance of rebound is less clear
[68].
354
Mechanisms of rebound — There are several potential mechanisms of rebound and
tolerance, although their relative importance is uncertain [59].
●Hypertrophy and hyperfunction of histamine-ECL cells presumably in response to
relative hypergastrinemia [18,65,66].
●Hypertrophy and hyperfunction of parietal cells presumably in response to
hypergastrinemia.
●Upregulation of histamine-independent stimulatory mechanisms mediated by
vagal/cholinergic pathways.
●Down-regulation of inhibitory pathways, such as those mediated by endogenous
somatostatin.
The observation that antisecretory therapy raises serum gastrin in proportion to the
magnitude of acid inhibition led to the hypothesis that hypergastrinemia was the primary
factor underlying rebound acid hypersecretion. However, rebound appears to occur on
nighttime maintenance doses of H2RA, which cause only minimal nocturnal and no
daytime hypergastrinemia. Furthermore, although sustained, prominent
hypergastrinemia causes acid hypersecretion, the threshold duration and magnitude of
hypergastrinemia required for a hypersecretory response remains controversial [69].
Gastrin receptor antagonists, when and if they become available, would test this
mechanism and possibly provide an important therapeutic option.
Antimuscarinic agents, such as atropine, dramatically reduce acid secretion in some
patients in whom nocturnal acid secretion was refractory to H2 blockade [70]. This
observation is consistent the hypothesis that muscarinic pathways play a role in H2RA-
refractory secretion.
Effect of Helicobacter pylori infection — Acute infection with H. pylori causes a
transient hypochlorhydria that may help the organism colonize the stomach. Chronic
infection with H. pylori can be associated with increased or decreased acid secretion,
depending on the severity of the gastritis and the anatomic distribution of gastritis.
Chronic pangastritis is associated with reduced acid production, which initially is caused
by a functional inhibition of parietal cells by inflammatory cytokines and products of H
pylori. Over time, chronic H. pylori infection can cause gastric atrophy and irreversible
hypochlorhydria and eventually achlorhydria. This pattern of gastritis is seen in
approximately 85 percent of patients. A smaller proportion of patients (15 percent) have
a pattern of gastritis that involves primarily the antrum (antrum-predominant). This
pattern of gastritis is associated with decreased somatostatin levels, high gastrin levels,
and increased acid production [71,72].
GASTRIC ACID HYPERSECRETIONGastric acid hypersecretion
(characterized by a basal acid output >15 mEq/hour) is observed in approximately 30
percent of patients with duodenal ulcers. H. pylori infection is a contributing factor, but
some patients have acid hypersecretion independent of H. pylori. Other uncommon
conditions associated with acid hypersecretion include the Zollinger-Ellison syndrome
(due to a gastrinoma), mastocytosis, and a retained antrum following partial
gastrectomy.
SUMMARY
●The physiologic stimulation of acid secretion has classically been divided into
three interrelated phases: cephalic, gastric, and intestinal. (See 'Phases of acid
secretion' above.)
355
●Gastric acid secretion from parietal cells is regulated by endocrine (gastrin),
paracrine (locally delivered histamine and somatostatin), neural (acetylcholine),
and autocrine (transforming growth factor-alpha) factors (figure 1). (See 'Secretion
of acid and pepsin' above.)
●Gastrin is the major endocrine regulator of the secretory response to a protein
meal. Gastrin enhances gastric acid secretion from parietal cells largely via the
release of histamine from enterochromaffin-like (ECL) cells. Gastrin also exerts
trophic action on the gastric mucosa by direct effects on histamine-secreting ECL
cells and possibly by effects on gastric stem cells and parietal cells.
(See 'Gastrin' above and 'Histamine' above.)
●Somatostatin, released from D cells in the gastric mucosa, is a potent inhibitor of
acid secretion. Although somatostatin has some effects on parietal cells, its major
effects are exerted on the inhibition of histamine release and to a lesser extent on
gastrin release. (See 'Somatostatin' above.)
●The mucosal nerves, containing acetylcholine, bombesin, vasoactive intestinal
peptide (VIP), and pituitary adenylate cyclase-activating polypeptide (PACAP),
mediate the response to the cephalic phase of acid secretion and to gastric
distention and amino acids. Acetylcholine is the major stimulatory mediator that
increases gastrin release, stimulates parietal cells, and inhibits somatostatin
secretion. (See 'Regulation of gastrin secretion' above and 'Neural control' above.)
●Prostaglandins are autocrine factors that inhibit acid secretion, histamine-
stimulated parietal cell function, and gastrin-stimulated histamine release.
(See 'Prostaglandins' above.)
●Tolerance to H2 receptor antagonists (H2RA) develops after as few as seven
days of therapy, with diminished effectiveness against nocturnal and pentagastrin-
stimulated acid secretion. Tolerance does not occur with proton pump inhibitors
(PPIs), probably because they block the final stage of acid secretion, the
hydrogen-potassium-ATPase pump.
Rebound acid hypersecretion occurs after the cessation of one or more months of
H2RA or PPI therapy. Therefore, long-term antisecretory therapy should probably
not be abruptly discontinued in patients at risk for complications or recurrence.
(See 'Tolerance and acid rebound' above.)
●Basal acid output (BAO), which is the level of acid secretion when the subject is
unstimulated, is less than 10 mEq per hour in normal subjects, with an average of
about 2 mEq per hour.
Gastric acid hypersecretion (characterized by a basal acid output >15 mEq/hour)
may be seen in patients with chronic H. pylori infection, duodenal ulcers
(independent of H. pylori infection), Zollinger-Ellison syndrome (due to a
gastrinoma), mastocytosis, and a retained antrum following partial gastrectomy.
(See 'Measuring acid secretion in humans' above and 'Gastric acid
hypersecretion' above.)

356
Proton pump inhibitors: Overview of use and
adverse effects in the treatment of acid related
disorders
Author:
M Michael Wolfe, MD
Section Editor:
Mark Feldman, MD, MACP, AGAF, FACG
Deputy Editor:
Shilpa Grover, MD, MPH, AGAF
Contributor Disclosures
All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2021. | This topic last updated: Nov 29, 2021.
What's New
PPI use and inflammatory bowel disease (November 2021)
Due to their ability to alter the gut microbiome, proton pump inhibitors (PPI) may impact
the risk of inflammatory bowel disease (IBD). In a study that included >600,000
individuals followed for a median of 12 years, regular PPI use was associated with an
increased risk of IBD compared with nonusers; however, the absolute risk of IBD was
low [1]. We continue to recommend prescribing PPIs at the lowest effective dose and for
the shortest duration appropriate to the condition being treated. (See "Proton pump
inhibitors: Overview of use and adverse effects in the treatment of acid related
disorders", section on 'Inflammatory bowel disease'.)
Read more

INTRODUCTIONProton pump inhibitors (PPIs) effectively block gastric acid

secretion by irreversibly binding to and inhibiting the hydrogen-potassium ATPase pump
that resides on the luminal surface of the parietal cell membrane.
This topic review will provide an overview of the mechanism of action,
pharmacokinetics, administration, and adverse effects of PPIs. The use and efficacy of
PPIs in specific acid-related disorders is presented separately. (See "Antiulcer
medications: Mechanism of action, pharmacology, and side effects".)
INDICATIONS FOR PPI THERAPYProton pump inhibitor (PPI) therapy is
indicated in the following clinical situations:
●Peptic ulcer disease – PPIs are first-line antisecretory therapy in the treatment
of peptic ulcer disease. (See "Peptic ulcer disease: Treatment and secondary
prevention", section on 'Initial antisecretory therapy'.)
●Gastroesophageal reflux disease – PPIs are indicated in patients with
gastroesophageal reflux disease, including for the treatment of erosive esophagitis
and as maintenance therapy in patients with severe erosive esophagitis or
Barrett’s esophagus. (See "Medical management of gastroesophageal reflux
disease in adults", section on 'Severe or frequent symptoms or erosive
esophagitis'.)
●Zollinger-Ellison syndrome – PPIs, often in high doses, are required to control
gastric acid hypersecretion in patients with gastrin-secreting

357
 tumors. (See "Management and prognosis of the Zollinger-Ellison syndrome
(gastrinoma)", section on 'Proton pump inhibitors'.)
●NSAID-associated ulcers – PPIs are indicated in the primary prevention of
gastroduodenal ulcers associated with NSAID use. (See "NSAIDs (including
aspirin): Primary prevention of gastroduodenal toxicity".)
●Eradication of Helicobacter pylori – PPIs are a component of several first-line
and salvage therapy regimens for H. pylori infection. (See "Treatment regimens for
Helicobacter pylori in adults".)
PHARMACOLOGY
●Mechanism of action – Proton pump inhibitors (PPIs) inhibit H-K-ATPase, the
final step of gastric acid secretion by parietal cells.
PPIs are benzimidazole prodrugs which accumulate specifically and selectively in
the secretory canaliculus of the parietal cell [1]. Within that space, they undergo
an acid catalyzed conversion to a reactive species, the thiophilic sulfonamides,
which are permanent cations. The rate of conversion varies among the
compounds and is inversely proportional to the pKa of the benzimidazole
(rabeprazole >omeprazole, esomeprazole, and lansoprazole >pantoprazole) (table
1) [2]. The reactive species interacts with the external surface of the H-K-ATPase
that faces the lumen of the secretory space of the parietal cell, resulting in
disulfide bond formation with cysteine 813 located within the alpha-subunit of the
enzyme; this is the residue that is intimately involved in hydrogen ion transport.
This covalent inhibition of the enzyme results in a specific and long-lasting
impairment of gastric acid secretion. (See "Physiology of gastric acid secretion".)
●Pharmacokinetics – PPIs are similar in structure and mechanism of action, but
PPIs differ in their pKa, bioavailability, peak plasma levels, and route of excretion
(table 1). The magnitude of these differences are small and their clinical relevance
has not been established.
PPIs are most effective when the parietal cell is stimulated to secrete acid
postprandially, a relationship that has important clinical implications for timing of
administration. Because the amount of H-K-ATPase present in the parietal cell is
greatest after a prolonged fast, PPIs should be administered before the first meal
of the day. In most individuals, once-daily dosing is sufficient to produce the
desired level of acid inhibition, and a second dose, which is occasionally
necessary, should be administered before the evening meal [1]. (See 'Dose and
timing of administration' below.)
Once-daily PPI dosing for five days inhibits maximal gastric acid output by
approximately 66 percent. Since PPIs inhibit only activated enzyme present in the
canalicular membrane, the reduction of gastric acid secretion after an initial dose
will probably be suboptimal. As inactive enzyme is recruited into the secretory
canaliculus, acid secretion will resume, albeit at a reduced level. After the second
dose is given on the next day, more H-K-ATPase will have been recruited and
subsequently inhibited, and after the third dose, additional recruitment and further
acid inhibition will probably occur. Thus, the occasional use of a PPI taken on an
"as needed" basis does not reliably provide adequate acid inhibition and does not
produce a consistent or satisfactory clinical response (in contrast to the H2
antagonists, which have a more rapid onset of action) [1].
358
 Restoration of acid secretion after discontinuing PPIs depends upon enzyme
turnover and the biological reversibility of the disulfide bond. Maximal acid
secretory capacity may not be restored for 24 to 48 hours [1]. (See 'Discontinuing
PPIs' below.)
●Metabolism – PPIs are metabolized via hepatic cytochrome P450 enzymes, with
CYP2C19 having the dominant role. However, the dominance of CYP2C19 over
other pathways varies significantly among the PPIs (table 2). The activity of
CYP2C19 is also determined to some extent by genetic polymorphisms. Two
inactivating mutations have been described as a result; the metabolism of PPIs by
this route may be delayed in these individuals [3,4]. Homozygotes for the wild type
gene rapidly metabolize these drugs, while heterozygotes are intermediate
metabolizers.
Plasma levels of PPI correlate with their metabolism, and differences may
contribute to varying dose requirements and clinical efficacy. As an example, one
study examined the effect of variable metabolism of omeprazole when using this
agent to treat H. pylori in 62 Japanese patients [3]. While eradication was
achieved in all individuals homozygous for a CYP2C19 mutation (ie, slow
metabolizers), successful treatment was achieved in only 60 and 29 percent of
heterozygotes and wild type homozygotes, respectively. In another study that
evaluated the efficacy of lansoprazole in the treatment of 65 patients with
gastroesophageal reflux disease (GERD), slow metabolizers were much more
likely to be asymptomatic as compared with heterozygotes and wild type
homozygotes (85 versus 68 and 46 percent, respectively) [5]. The response rate
in wild type homozygotes with severe GERD was only 16 percent. Wild type
homozygotes (rapid metabolizers) also had the lowest plasma lansoprazole
concentrations. (See 'Dose and timing of administration' below.)
PRETREATMENT CONSIDERATIONS AND MONITORING
Drug interactions — Clinically important drug interactions with proton pump inhibitors
(PPIs) are rare. However, PPI metabolism via hepatic cytochrome P450 enzymes may
lead to specific drug interactions in some individuals. The presence of a CYP2C19 gene
mutation results in higher plasma PPI levels in homozygous individuals. However, if this
metabolic pathway becomes saturated, the isoenzyme can become a major target for
interactions with many drugs, including warfarin, diazepam, clopidogrel,
and phenytoin (table 3). CYP3A4-mediated metabolism may also be activated under
such conditions and become the principal route of drug elimination. Furthermore,
induction of CYP1A, another P450 isoenzyme, in CYP2C19 deficient or saturated
individuals, can make them susceptible to interference with theophylline metabolism.
The specific P450 enzymes involved in PPI metabolism and the potential for
interactions among PPIs varies considerably (table 2) [6-12].

Some other important drug interactions with PPIs include the following:

●Clopidogrel - Some data suggest decreased activation of clopidogrel when used

in conjunction with omeprazole due to shared hepatic cytochrome P450-mediated
metabolism. In 2009, the United States Food and Drug Administration concluded
that patients taking clopidogrel should consult with their clinician if they are taking
or considering taking a PPI, including over-the-counter PPI preparations [13,14].
359
 However, the relevance of these data remains highly controversial. The interaction
of clopidogrel and PPIs are discussed in detail separately. (See "Clopidogrel
resistance and clopidogrel treatment failure".)
●HIV protease inhibitors - PPIs may decrease the absorption of certain HIV
protease inhibitors. PPIs are contraindicated in patients being treated
with rilpivirine. Atazanavir should not be used in patients who require a PPI dose
equivalent to >20 mg omeprazole daily. (See "Overview of antiretroviral agents
used to treat HIV", section on 'Protease inhibitors (PIs)'.)
●Methotrexate - Coadministration of PPIs with high dose methotrexate appears to
be correlated with delayed methotrexate elimination and potentially may lead to
methotrexate toxicity if not monitored appropriately.
For additional information on drug interactions, use the Lexicomp drug
interactions program provided by UpToDate. (See "Overview of the non-acute
management of unstable angina and non-ST elevation myocardial infarction", section on
'Gastrointestinal prophylaxis'.)
Laboratory testing — We limit routine laboratory testing to selected patients on PPI
therapy.
●Magnesium – We obtain serum magnesium levels prior to starting a PPI in
patients who are expected to be on long-term (≥1 year) treatment or in patients
who take PPIs in conjunction with other medications associated with
hypomagnesemia (eg, diuretics). In addition, we obtain magnesium levels
periodically in such patients while they are taking a PPI. The frequency of testing
is based on the clinical history and the presence of symptoms of
hypomagnesemia. As an example, in patients with a history of arrhythmias or QT
interval prolongation, we monitor magnesium levels every six months. The
management of hypomagnesemia is discussed in detail separately.
(See "Hypomagnesemia: Evaluation and treatment".)
●Vitamin B12 – We also obtain vitamin B12 levels yearly in patients on long-term
PPIs [15]. However, routinely monitoring vitamin B12 levels is controversial.
(See 'Magnesium malabsorption' below and 'Vitamin B12 malabsorption' below.)
There are insufficient evidence to support routine bone density monitoring or calcium
supplementation due to proton pump inhibitor use alone [16].
ADMINISTRATION
Intravenous regimen — IV PPIs are indicated prior to endoscopic evaluation in
patients with clinically significant upper gastrointestinal bleeding from a suspected
peptic ulcer. Pantoprazole and esomeprazole are the only PPIs available as an IV
formulation in the United States; IV omeprazole is available in other countries. The use
of PPIs in the treatment of bleeding peptic ulcers and the duration of treatment is
discussed in detail separately. (See "Overview of the treatment of bleeding peptic
ulcers", section on 'Oral versus intravenous dosing' and "Approach to acute upper
gastrointestinal bleeding in adults", section on 'Acid suppression'.)
Oral regimen
Selecting a PPI — The choice of a specific oral PPI and whether over-the-counter
(rather than prescription) PPIs are prescribed are often determined by patient
preference and payer coverage. A systematic review of 12 randomized trials examining

360
the relative effectiveness of different PPI doses and dosing regimens found no
consistent difference in symptom resolution and esophagitis healing rates [17].
In patients unable to swallow pills or capsules, options include an oral suspension
of lansoprazole and a powder formulation of omeprazole-sodium bicarbonate for oral
suspension.
Dose and timing of administration — PPIs should be administered 30 to 60 minutes
before breakfast for maximal inhibition of proton pumps. (See 'Pharmacology' above.)
Dose reduction, particularly for maintenance of healing of erosive esophagitis may be
possible in Asian populations. Polymorphisms in the CYP2C19 gene, which encodes
the cytochrome P450 isoenzyme that metabolizes different PPI preparations, are
common in Asian and other populations [18]. Such gene mutations would render an
individual a "slow metabolizer" and prolong the antisecretory effect of PPIs. In contrast,
the duration of acid inhibition would be decreased in a "rapid metabolizer," and
differences in PPI metabolism might account for incomplete inhibition of acid secretion
and a high prevalence of nocturnal breakthrough symptoms in gastroesophageal reflux
disease patients. (See 'Pharmacology' above and "Approach to refractory
gastroesophageal reflux disease in adults", section on 'Differences in PPI metabolism'.)
Avoidance of concurrent antisecretory agents — PPIs should not administered
concomitantly with antisecretory agents including histamine-2 receptor antagonists
(H2RAs), analogues of prostaglandin E (eg, misoprostol), and somatostatin analogues
(eg, octreotide), because of the marked reduction in acid inhibitory effects [1,19].
Antisecretory drugs can be used with a PPI provided that there is a sufficient time
interval between their administration. As an example, an H2RA can be taken before
bedtime or during the night by individuals who report nocturnal breakthrough symptoms
such as heartburn after taking a PPI in the morning or before dinner.
Switching between PPIs — Switching PPIs is a reasonable strategy in patients with
side-effects to an individual PPI and may be necessary due to cost differences.
Although there is significant interindividual and intraindividual variability in intragastric
pH control between PPIs, there are no consistent difference in relation to symptom
resolution and esophagitis healing rates [17]. Switching PPIs in patients with well-
controlled symptoms may also be associated with increased symptom severity and
decreased patient satisfaction [20]. (See "Approach to refractory gastroesophageal
reflux disease in adults", section on 'Subsequent management'.)
Discontinuing PPIs — PPIs should be prescribed at the lowest dose and for the
shortest duration appropriate to the condition being treated. (See 'Indications for PPI
therapy' above.)
We gradually taper PPI therapy in patients treated with PPIs for longer than six months.
For patients on a standard or high-dose PPI (eg, omeprazole 40 mg daily or twice daily),
we decrease the dose by 50 percent every week. For patients on twice daily dosing, the
initial reduction can be accomplished by decreasing the dosing to once in the morning
before breakfast until the patient is on the lowest dose of the medication. Once on the
lowest dose for one week, the patient is instructed to discontinue the PPI. However, no
specific method for discontinuing PPI therapy has been proven effective, and no
approach is universally accepted. [21,22].
Studies have demonstrated rebound gastric acid hypersecretion following
discontinuation of PPIs in patients with long-term use. The reasons are not entirely

361
clear, but appear to be due in part to the suppression of antral somatostatin expression,
resulting in an increase in antral gastrin release and subsequent disruption of normal
pH-related feedback inhibition of acid secretion that occurs after a meal [1].
(See "Physiology of gastric acid secretion", section on 'Tolerance and acid rebound'.)
ADVERSE EFFECTSLong-term PPI use has been associated with several safety
concerns. However, few of these concerns are supported by consistent data
demonstrating a causal relationship. (See "Physiology of gastrin", section on
'Hypergastrinemia'.)
Gastrointestinal effects
Clostridioides difficile and other enteric infections — PPI use has been associated
with an increased risk of C. difficile infection, even in the absence of antibiotic use [23-
33]. Associations with other enteric infections, including salmonellosis and
campylobacteriosis, have also been reported [34-39]. However, the pathophysiologic
mechanism involved in the increased risk of infection is unclear.
A 2017 meta-analysis of 50 observational studies found that PPI use was significantly
associated with an increased risk of C. difficile infection (relative risk [RR] 1.3; 95% CI
1.1-14). The risk of C. difficile infection appears to be greater with PPIs as compared to
H2 receptor antagonists [30,31].
PPI use has also been associated with an increased risk of recurrent C. difficile infection
[31]. In a 2017 meta-analysis of 16 observational studies that included 7703 patients
with C. difficile infection of whom 1525 (20 percent) had recurrent C. difficile infection,
gastric acid suppression was significantly associated with an increased risk of
recurrent C. difficile infection (odds ratio [OR] 1.5; 95% CI 1.2-1.9) [40]. There was
significant heterogeneity among the studies included in the meta-analysis. In adjusted
analysis using data from nine studies, PPI use was associated with an increased risk of
recurrent C. difficile infection after controlling for patient age and other co-morbid
conditions (OR 1.4; 95% CI 1.1-1.8). (See "Clostridioides difficile infection in adults:
Epidemiology, microbiology, and pathophysiology", section on 'Gastric acid
suppression'.)
Microscopic colitis — PPI use has been associated with microscopic colitis, including
lymphocytic and collagenous colitis. In a case-control study that included 95 cases of
microscopic colitis, exposure to PPIs was significantly higher in patients with
microscopic colitis as compared with controls (38 versus 13 percent, OR 4.5, 95% CI
2.0-9.5) [41]. Similar results have been reported in other case-control studies, however,
it is unclear if this association varies by PPI and if there is a dose-response relationship
in either dose or duration of use [42,43]. (See "Microscopic (lymphocytic and
collagenous) colitis: Clinical manifestations, diagnosis, and management", section on
'Medications'.)
Hypergastrinemia — Induction of hypergastrinemia has been associated with gastric
carcinoid tumors in rats. However, these observations are not generalizable to species
with gastrin physiology more analogous to humans [44]. While patients treated
with omeprazole for up to 11 years have shown some enterochromaffin-like cell
hyperplasia, no dysplasia or neoplastic changes have been observed [45]. An increased
risk of colon cancer due to hypergastrinemia has also not been established [46].
(See "Physiology of gastrin".)

362
Atrophic gastritis — Patients on long-term PPI therapy have a propensity to develop
chronic atrophic gastritis. However, the risk of atrophic gastritis is small, and in the rare
patient who develops atrophic gastritis, the clinical consequences are uncertain
[45,47,48]. (See "Risk factors for gastric cancer".)
Intestinal colonization of multi-drug resistant organisms — PPIs may increase the
risk of intestinal colonization with multi-drug resistant organisms. In a meta-analysis of
12 observational studies that included 22,305 patients, after adjusting for potential
confounders, acid suppression increased the odds of intestinal carriage of multi-drug
resistant organisms of the Enterobacterales order (producing extended-spectrum beta-
lactamases, carbapenemases, or plasmid-mediated AmpC beta-lactamases) and of
vancomycin-resistant enterococci (OR 1.74; 95% CI 1.4-2.2) [49]. Possible mechanisms
include an increase in bacteria that survive transit from the stomach to the intestine due
to reduction in gastric acid by PPIs and direct alteration of the composition of intestinal
microbiota, leading to a decrease in mean species diversity.  
Inflammatory bowel disease — In a study that pooled data from three observational
cohorts and included >600,000 individuals followed for a median of 12 years, the risk of
inflammatory bowel disease (IBD) was increased in regular PPI users as compared with
nonusers (hazard ratio [HR] 1.42; 95% CI 1.22–1.65) [50]. However, the absolute risk of
IBD was low, with a number needed to harm of 3770. In addition, absence of data on
PPI dosing precluded assessment of a dose-response relationship between PPI use
and IBD.
Malabsorption of minerals and vitamins
Magnesium malabsorption — PPIs can cause hypomagnesemia due to reduced
intestinal absorption [51]. A meta-analysis of nine observational studies that included a
total of 109,798 patients found that those who took a PPI had a significantly higher risk
(RR 1.43; 95% CI 1.08-1.88) of developing hypomagnesemia as compared with those
who did not [52]. Clinical manifestations of hypomagnesemia include neuromuscular
excitability (eg, tremor, tetany, convulsions), weakness, and apathy. Severe PPI-
induced hypomagnesemia has been associated with QT interval prolongation and
torsades de pointes [53,54]. The risk of hypomagnesemia appears to be mainly in
patients who have been on PPIs long-term (generally longer than one year) but cases
have been reported within one year of starting PPI therapy [53,55]. This potential risk
has led to recommendations to monitor serum magnesium levels in specific patients at
high risk for hypomagnesemia. Monitoring for hypomagnesemia in patients on PPIs is
discussed in detail separately. (See 'Laboratory testing' above and "Hypomagnesemia:
Clinical manifestations of magnesium depletion", section on 'Overview of clinical
manifestations'.)
Calcium and fracture risk — Although hypochlorhydria could theoretically reduce
calcium absorption, the effect appears to be relevant only for the absorption of water
insoluble calcium (eg, calcium carbonate) and can be overcome by ingestion of a
slightly acidic meal [56]. The absorption of water soluble calcium salts or calcium in
dairy products are not impacted by PPI-induced hypochlorhydria. When calcium
supplementation is necessary in patients taking PPIs, we use calcium supplements that
do not require acid for absorption, such as calcium citrate. (See 'Laboratory
testing' above and "Drugs that affect bone metabolism", section on 'Proton pump
inhibitors'.)

363
PPI-induced hypochlorhydria can augment osteoclastic activity, thereby decreasing
bone density [57,58]. Although an association between PPI use and bone fracture is
plausible, causality has not been established [59]. Nonetheless, the FDA has mandated
revised safety information on all PPIs about a possible increased risk of fractures of the
hip, wrist, and spine with the use of these medications [60]. The association between
PPIs and bone metabolism and risk of fracture are discussed in detail separately.
(See "Drugs that affect bone metabolism", section on 'Proton pump inhibitors'.)
Vitamin B12 malabsorption — Long-term therapy with PPIs has been associated with
vitamin B12 malabsorption [61,62]. However, absorption of oral B12 supplements is not
affected. (See 'Laboratory testing' above and "Treatment of vitamin B12 and folate
deficiencies", section on 'Treatment of vitamin B12 deficiency'.)
Iron malabsorption — Gastric acid plays a role in the absorption of nonheme iron, and
the use of PPIs has been associated with decreased iron absorption [63-67]. However,
in most cases the decreased absorption does not appear to be of clinical significance.
One exception may be in patients who require oral iron supplementation [66,68]. Such
patients may need a higher dose or longer duration of supplementation [66].
(See "Treatment of iron deficiency anemia in adults", section on 'Dosing and
administration (oral iron)'.)
Kidney disease — PPIs can cause acute interstitial nephritis (AIN) [69-72]. Similar to
other cases of drug-induced AIN, AIN due to PPI use is not dose-dependent, and
recurrence or exacerbation can occur with a second exposure to the same or a related
drug. (See "Clinical manifestations and diagnosis of acute interstitial nephritis", section
on 'Drugs'.)
PPI use has also been associated with an increased risk of incident chronic kidney
disease (CKD), CKD progression, and end-stage kidney disease [73-76]. However, the
mechanism underlying the association between PPI use and risk of CKD is not known,
and it is possible that the weak association observed in these studies is due to
methodological limitations (residual confounding) [73,74,77]. Further studies are needed
to help better define an etiologic relationship between PPI use and the development and
worsening of CKD.
Drug-induced lupus — In postmarketing safety surveillance, new onset of cutaneous
lupus erythematosus and systemic lupus erythematosus (SLE), and exacerbation of
existing disease have been reported in patients on PPIs [78-80]. Most cases of CLE-
associated with PPI use are subacute and occur within weeks to years after continuous
PPI therapy. PPI-associated SLE usually occurs days to years after initiating PPI
treatment and typically presents with a rash. Most patients improve within 4 to 12 weeks
of discontinuation of PPI therapy. (See "Drug-induced lupus", section on 'Causative
drugs'.)
Other associations of unclear significance
COVID-19 — It is unclear if PPI use is associated with an increased risk of COVID-19.
In a cross-sectional survey of 86,602 individuals, 53,130 reported prior abdominal pain,
acid reflux, heartburn, and regurgitation symptoms and provided data on H2RA and PPI
use. Of these, 3386 individuals (6.4 percent) self-reported a positive COVID-19 test
[81]. In analyses adjusted for socioeconomic, lifestyle, and clinical comorbidities,
patients who reported PPI use were significantly more likely to report a positive COVID-
19 test with a dose-dependent increase in odds of reporting a positive test (PPI once-

364
daily OR 2.15, 95% CI 1.9-2.4; PPI twice-daily OR 3.7, 95% CI 2.9-4.6). It is possible
that this association is due to residual confounding.
Other studies have demonstrated that patients taking PPIs are at increased risk for
severe clinical outcomes of COVID-19 but have not demonstrated an increase in
susceptibility to SARS-CoV-2 infection [82]. These data require further validation.
Dementia — Although some studies have found a significant association between use
of PPIs and incident dementia, others have not found an association between PPI use
and cognitive function [83-88]. The association between PPI use and dementia may
reflect residual confounding by factors related to both use of PPIs and the development
of dementia and is discussed in detail, separately. (See "Epidemiology, pathology, and
pathogenesis of Alzheimer disease", section on 'Medications'.)
Pneumonia — While observational studies suggest an association between PPI use
and pneumonia, the observed association may be due to confounding such that
individuals prescribed PPIs may be more likely to have other unobserved health
characteristics that predispose them to pneumonia as compared with nonusers [89-96].
(See "Epidemiology, pathogenesis, and microbiology of community-acquired pneumonia
in adults", section on 'Predisposing host conditions' and "Risk factors and prevention of
hospital-acquired and ventilator-associated pneumonia in adults", section on 'Role of
gastric pH'.)
Mortality — It is unclear if PPI use is associated with an increase in risk of death. In an
observational cohort study that included 275,977 new PPI users and 73,335 new
histamine-2 receptor antagonist (H2RA) users, over a median follow-up of 5.7 years, the
incident death rate among new PPI users was higher as compared to those receiving
H2RA (4.5 versus 3.3 per 100 person-years) [97]. New PPI users were significantly
older as compared with new users of H2RAs at the time of study entry (61.7 versus 58.5
years). However, after adjusting for potential confounders, PPI use was associated with
an increase in all-cause mortality as compared with H2RA use (HR 1.25, 95% CI 1.23-
1.28). PPI users also had an increase in risk of death as compared with individuals
without any PPI use and individuals without any acid suppression use (HR 1.15, 95% CI
1.14-1.15; HR 1.23, 95% CI 1.22-1.24, respectively). Among new PPI users, the risk of
death increased with the duration of PPI use. Limitations of the study include its
generalizability as the study cohort primarily consisted of older White males, and lack of
data on the cause of mortality. The underlying basis for this apparent increased risk of
death with PPI use are not known and further studies are needed to evaluate whether
this epidemiologic association is due to unmeasured confounding.
INFORMATION FOR PATIENTSUpToDate offers two types of patient education
materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces
are written in plain language, at the 5th to 6th grade reading level, and they answer the
four or five key questions a patient might have about a given condition. These articles
are best for patients who want a general overview and who prefer short, easy-to-read
materials. Beyond the Basics patient education pieces are longer, more sophisticated,
and more detailed. These articles are written at the 10th to 12th grade reading level and
are best for patients who want in-depth information and are comfortable with some
medical jargon.

365
Here are the patient education articles that are relevant to this topic. We encourage you
to print or e-mail these topics to your patients. (You can also locate patient education
articles on a variety of subjects by searching on "patient info" and the keyword(s) of
interest.)

●Basics topics (see "Patient education: Acid reflux and gastroesophageal reflux

disease in adults (The Basics)")
●Beyond the Basics topics (see "Patient education: Gastroesophageal reflux
disease in adults (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Indications for proton pump inhibitors (PPIs) include the treatment of peptic ulcer
disease, gastroesophageal reflux disease, and Zollinger-Ellison syndrome. PPIs
are effective in the prevention of nonsteroidal anti-inflammatory drug-associated
gastroduodenal mucosal injury and are an important component of several
antimicrobial regimens used in the treatment of Helicobacter pylori infection.
(See 'Indications for PPI therapy' above.)
●PPIs effectively block gastric acid secretion by irreversibly binding to and
inhibiting the hydrogen-potassium ATPase pump that resides on the luminal
surface of the parietal cell membrane. Genetically determined variability in the PPI
metabolism can influence their efficacy. The presence of a CYP2C19 gene
mutation can results in higher plasma PPI levels in homozygous individuals. PPI
metabolism via hepatic cytochrome P450 enzymes may lead to specific drug
interactions in some individuals. However, clinically important drug interactions
with PPIs are rare. (See 'Pharmacology' above.)
●The magnitude of pharmacokinetic differences between PPIs is small and the
clinical relevance of these differences has not been established (table 1).
Intravenous PPI therapy is indicated in patients with clinically significant upper
gastrointestinal bleeding prior to endoscopy for treatment of suspected bleeding
peptic ulcers. (See 'Administration' above.)
●Oral PPIs should be administered 30 to 60 minutes before breakfast for maximal
inhibition of proton pumps. PPIs should not be administered concomitantly with
H2-receptor antagonists, and prostaglandins or somatostatin analogues. In
general, PPIs should be prescribed at the lowest dose and for the shortest
duration appropriate to the condition being treated. When discontinuing PPI
therapy, we taper the dose gradually in patients on PPIs for longer than six
months. (See 'Oral regimen' above.)
●PPI use has been associated with an increased risk
of Clostridioides difficile infection, other enteric infections, and microscopic
colitis. C. difficile infection with diarrhea may occur even in the absence of
antibiotic use. (See 'Gastrointestinal effects' above.)
●PPIs can cause hypomagnesemia due to reduced intestinal absorption. Long-
term therapy with PPIs has been associated with vitamin B12 malabsorption. We
obtain serum magnesium levels prior to starting a PPI in patients who are
expected to be on long term (≥1 year) treatment, or in patients who take PPIs in
conjunction with other medications associated with hypomagnesemia. In addition,
we also monitor magnesium and vitamin B12 levels in patients on long-term PPIs.

366
(See 'Pretreatment considerations and monitoring' above and 'Magnesium
malabsorption' above and 'Vitamin B12 malabsorption' above.)
●Although an association between PPIs and bone fracture is plausible, causality
has not been established. PPIs can decrease the absorption of water insoluble
calcium (eg, calcium carbonate). When calcium supplementation is necessary in
patients taking PPIs, we use calcium supplements that do not require acid for
absorption, such as calcium citrate. (See 'Calcium and fracture risk' above.)
●PPIs can cause acute interstitial nephritis. PPI use has also been associated with
an increased risk of incident chronic kidney disease (CKD), CKD progression, and
end-stage renal disease. However, further studies are needed to help better
define an etiologic relationship between PPI use and the development and
worsening of CKD. (See 'Kidney disease' above.)
●There are conflicting data on the association between PPI use and risk of
dementia and pneumonia. It is also unclear if PPI use is associated with an
increased risk of death. It is possible that these associations are due to residual
confounding and more studies are needed. (See 'Kidney disease' above
and 'Dementia' above and 'Mortality' above.)

367