Professional Documents
Culture Documents
GUIDELINE HEMATEMESIS
DISUSUN OLEH:
PENGUJI :
dr. COANA SUKMAGAUTAMA, Sp.PD, M.Kes
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
Joseph JY Sung,1 Philip WY Chiu,1 Francis K L Chan,1 James YW Lau,1 Khean-lee Goh,2
Lawrence HY Ho,3 Hwoon-young Jung,4 Jose D Sollano,5 Takuji Gotoda,6
Nageshwar Reddy,7 Rajvinder Singh,8 Kentaro Sugano,9 Kai-chun Wu,10
Chun-Yin Wu,11 David J Bjorkman,12 Dennis M Jensen,13 Ernst J Kuipers,14
Angel Lanas15
For numbered affiliations see Abstract practising in community or referral hospitals. The
end of article. Non-variceal upper gastrointestinal bleeding remains an Asia-Pacific Working Group decided that there
important emergency condition, leading to significant was no need to repeat guideline statements previ-
Correspondence to ously recommended unless there was a change of
morbidity and mortality. As endoscopic therapy is
Professor Joseph JY Sung,
Institute of Digestive Disease, the ’gold standard’ of management, treatment of view, but only to highlight recommendations based
Prince of Wales Hospital, the these patients can be considered in three stages: pre- on new evidence reported in the past 5–10 years.
Chinese University of Hong endoscopic treatment, endoscopic haemostasis and post- Therefore, the 2011 consensus recommendations
Kong, Hong Kong, China; endoscopic management. Since publication of the Asia- that are not dealt with in this update are considered
jjysung@cuhk.edu.h k
Pacific consensus on non-variceal upper gastrointestinal to be still valid for the management of NVUGIB.
Received 14 February 2018 bleeding (NVUGIB) 7 years ago, there have been The working group continued to use the same
Revised 28 March 2018 significant advancements in the clinical management of modified Delphi process as before1 but chose to
Accepted 29 March 2018 patients in all three stages. These include pre-endoscopy divide the updated consensus into three sections:
risk stratification scores, blood and platelet transfusion, (1) pre-endoscopic management, (2) endoscopic
use of proton pump inhibitors; during endoscopy new management and (3) post-endoscopic management
haemostasis techniques (haemostatic powder spray and of NVUGIB.
over-the-scope clips); and post-endoscopy management
by second-look endoscopy and medication strategies.
Emerging techniques, including capsule endoscopy Methods
and Doppler endoscopic probe in assessing adequacy The Asia-Pacific Working Group of upper gastro-
of endoscopic therapy, and the pre-emptive use of intestinal bleeding comprises key opinion leaders
angiographic embolisation, are attracting new attention. in the region/countries of Asia and Australasia—
An emerging problem is the increasing use of dual namely, Australia, China, Hong Kong, India,
antiplatelet agents and direct oral anticoagulants in Japan, Korea, Malaysia, Philippines, Singapore and
patients with cardiac and cerebrovascular diseases. Taiwan. We also invited international experts from
Guidelines on the discontinuation and then resumption Europe and North America to share new scientific
of these agents in patients presenting with NVUGIB data and discuss the consensus statements. The
are very much needed. The Asia-Pacific Working Group group met during the Asia-Pacific Digestive Week
examined recent evidence and recommends practical 2017 in Hong Kong.
management guidelines in this updated consensus Literature search include Medline, EMBASE,
statement. the Cochrane Central Register of Controlled Trials
and ISI Web of Knowledge with manual searches
of bibliographies of key articles and abstracts of
major gastroenterology conferences held in the past
Background 5 years, 2012–2017 (Asian Pacific Digestive Week
Important advances in the management of non-var- (APDW), Digestive Disease Week (DDW), United
iceal upper gastrointestinal bleeding (NVUGIB) European Gastroenterology Week (UEGW)). Key
have been made. The concept of pre-endoscopic words used included gastrointestinal bleeding,
treatment has changed. New devices in endoscopic peptic ulcer disease and Asia.
haemostasis have been introduced. The increasing The working group members from the 10 coun-
use of antiplatelets and anticoagulants has compli- tries and regions mentioned above were selected
cated the picture. Endoscopic interventions, such as from the scientific committee of the APDW 2017
mucosectomy and endoscopic submucosal dissec- for their expertise in areas of NVUGIB, evidence-
tion, have become standard care. These procedures based medicine and continuing medical education.
are associated with marked risks of bleeding. For The preparation committee in this working group
these reasons, the Asia-Pacific Working Group felt comprised JJYS, PCYC, FKLC and JYWL, who
that it was necessary to update their consensus drafted the initial statements based on the literature.
recommendations for the management of NVUGIB. A modified Delphi process was used, and these
To cite: Sung JJY, Chiu PWY, Similar to the previous Asia-Pacific consensus drafted statements were sent to all group members
Chan FKL, et al. Gut statements published in 2011,1 this update aims for voting before the meeting, together with
2018;67:1757–1768. to produce management guidelines for clinicians evidence-based reviews and other pertinent
Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276 1757
Guidelines
literature. Each statement was assessed on a five-point Likert Europe, North America, Asia and Oceania. It compared the
scale: (1) accept completely, (2) accept with some reservation, pre-endoscopy (admission RS, AIMS65 and GBS) and post-en-
(3) accept with major reservation, (4) reject with reservation, doscopy scores (full RS, and Progetto Nazionale Emorragia
(5) reject completely. Results and comments were collated by Digestiva (PNED)) for their ability to predict clinical outcome.
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
emails. A statement was accepted when supported by ≥80% of GBS was found to be the best of these scoring systems for
the working group (ie, proportion of the working group voting predicting the need for intervention and mortality.2 GBS also
on the 5-point scale for 1 or 2). Statements that did not reach out-performed admission RS and AIMS65 in predicting the need
consensus support during the first-round voting were modified. for endoscopic treatment. GBS ≥7 is the best score to predict
These modified statements were discussed during the meeting the requirement for endoscopic treatment.
in Hong Kong, followed by a second round of voting with elec- Among these four scoring systems, GBS most accurately
tronic keypads. Participants voted anonymously on statements predicts the need for hospital admission and mortality. In
after discussion and provided comments on the wording of the Korea, a study recruiting 523 patients with NVUGIB concluded
statements, which were progressively finalised through two that GBS predicted the requirement of blood transfusion with
separate iterations. If this again failed to reach consensus, the highest accuracy.3 It was comparable to the full RS and AIMS65
statement was rejected. for predicting 30-day mortality and endoscopic intervention.
Each statement was then assessed for level of evidence by the A multicentre cohort study from Korea comparing GBS and
following criteria: (a) high level of evidence; further research RS confirmed that GBS was useful in predicting the need for
is very unlikely to change our confidence in the estimate of hospital-based intervention, and RS was useful for predicting
effect, (b) moderate level of evidence; further research is likely outcome.4
to have an important impact on our confidence in the estimate A smaller-scale study from Europe comparing GBS, AIMS65
of effect and may change the recommendation and (c) low level and RS in 309 patients presenting with UGIB reported that GBS
of evidence; further research is very likely to have an important was the best for predicting the need for transfusion, identical
impact on our confidence in the estimate of effect and is likely to AIMS65 in predicting endoscopic intervention, and compa-
to change the estimate. rable to AIMS65 and RS in predicting inpatient mortality.5
The conference was supported by an unrestrictive grant from In Denmark, 831 patients with UGIB were enrolled to
the GI Research Fund of the Institute of Digestive Disease of compare the accuracy of GBS, the age-extended GBS, the Baylor
the Chinese University of Hong Kong. Industry support was Bleeding Score and the Cedars-Sinai Medical Centre Predictive
not provided to avoid potential influence on the process of Index using the outcome of (1) need for hospital-based inter-
consensus development. Mandatory written disclosure of finan- vention or 30-day mortality, (2) likelihood of rebleeding and (3)
cial conflicts of interest within 24 months before the meeting mortality.6 GBS accurately identified patients most likely to need
was obtained from all voting participants. a hospital-based intervention, but none of these scores accu-
In this updated consensus meeting, it was decided that the rately predicted mortality or rebleeding.
statements should focus on guiding clinical management of Two studies from Australia were reported. From Victoria, a
NVUGIB rather than on clarifying clinical concepts. There study comparing AIMS65, GBS and RS in 424 patients confirmed
should be minimal overlap or repetition between this updated that the three scores were comparable in predicting a composite
statement and the previous Asia-Pacific consensus on non-var- outcome of rebleeding and endoscopic/radiological intervention,
iceal upper gastrointestinal bleeding1 unless new evidence had and mortality.7 GBS was the best for predicting blood transfu-
arisen that might change the recommendation. The panel also sion. The AIMS65 score was better than both the GBS (area
decided to report the statements that had not reached consensus under the receiver operating characteristic curve (AUROC), 0.80
as these are also considered useful in providing guidelines for the vs 0.76, P<0.027) and the pre-endoscopy RS (AUROC 0.74,
clinical management of such patients. Therefore, the first part of P=0.001) and equivalent to the full RS (AUROC 0.78, P=0.18)
this report consists of consensus statements categorised under (1) in predicting inpatient mortality. The AIMS65 score was better
pre-endoscopic management, (2) endoscopic management, (3) than all other scores in predicting the need for intensive care
post-endoscopic management. The second part consists of state- unit (ICU) admission and length of hospital stay. In Adelaide,
ments that were rejected after deliberations and debate among comparing just GBS and RS, investigators confirmed that GBS
panel members. These latter statements point towards new was superior in predicting blood transfusion and surgery, and
management concepts or strategies, but they are not accepted as was equivalent to RS in predicting the need for endoscopic
recommendations because of insufficient evidence in the existing therapy, rebleeding and death.8
literature. A study from Thailand recruiting close to 1000 patients with
UGIB compared GBS, full RS and pre-endoscopic RS and found
that these scores were better in NVUGIB than variceal UGIB.9
Consensus statements In Iran, 200 patients were recruited to compare the prediction
Pre-endoscopic management value of GBS and full RS. GBS was more accurate than full RS
Statement 1: The Glasgow Blatchford Score (GBS) should be used for need for blood transfusion, rebleeding, ICU admission and
in predicting clinical outcome of all patients presenting with upper endoscopic intervention. RS predicted 30-day mortality better
gastrointestinal bleeding (Accept—agreement: 94.5%, level of than GBS in this study.10
evidence: high) A systematic review identified 16 studies evaluating pre-en-
The most widely validated scoring methods for prediction of doscopic risk scores (GBS, RS and AIMS65) on a composite
recurrent bleeding and mortality are the Rockall (RS), GBS and outcome including recurrent bleeding, need for intervention
AIMS65. Many studies in the past 5 years have compared the and 30-day mortality.11 The review concluded that GBS has
performance of various scoring systems as a risk stratification the highest sensitivity and specificity to predict this outcome
tool. compared with the other scores. The results of these studies were
The largest study is a multicentre prospective cohort that very consistent. Most confirmed that GBS was the best score
recruited over 3000 patients with UGIB from six hospitals in to predict which patients require a hospital-based intervention,
1758 Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276
Guidelines
including blood transfusion, endoscopic therapy and surgery. haemodynamics.14 Early blood transfusion appeared to lead
There is still room for improvement for GBS in predicting to more recurrent bleeding in these patients, and the authors
recurrent bleeding and 30-day or long-term mortality. With the suggested that this was due to reversal of hypercoagulable
large amount of clinical data providing consistent results, The response to haemorrhage thereby encouraging rebleeding and
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
working group recommend that GBS is used for predicting clin- hence the need for surgery.
ical outcome of patients with NVUGIB. Patients with a score of Almost three decades later, a study from Spain randomised
0–1 rarely need any clinical intervention, and can thus be safely 921 patients with severe acute UGIB to a restrictive blood
discharged, with elective endoscopy at a later stage. In contrast, transfusion strategy (transfuse only when haemoglobin
a high score (say 10–12) is associated with frequent need for level <7 g/dL) or a liberal transfusion strategy (transfuse
intervention such as transfusion and therapeutic endoscopy. The only when haemoglobin <9 g/dL).15 Significant differences
level of evidence was graded as high. were demonstrated in the overall survival, further bleeding
and adverse events between the two groups in favour of
Statement 2: Patients with a GBS of<1can be treated as an restrictive transfusion. The major difference, however, was
outpatient seen in the group of patients with cirrhosis and Child-Pugh
class A or B diseases.
A more recent study from the UK (TRIGGER Study) took
(Accept-agreement: 94.5%, level of evidence: high)
a cluster randomised feasibility trial approach in six hospi-
Most of the studies are very consistent in showing that GBS is
tals. A slightly different definition of restrictive transfusion
the best tool to predict which patients presenting with UGIB
strategy (when haemoglobin <8 g/dL) versus liberal transfu-
are likely to require a hospital-based intervention. This first
sion strategy (when haemoglobin <10 g/dL) was adopted.16
pertains to blood transfusion and also to the need for endoscopic
Comparison of centres using different transfusion strate-
treatment and surgery. Several studies also point to setting the
gies showed that there was no significant difference in clin-
threshold of GBS <1–3 as an indicator for patients not requiring
ical outcome, including mortality, thromboembolic events,
hospitalisation.2 4–6 9–11
surgical or radiological intervention, therapeutic interven-
The Upper Gastrointestinal Haemorrhage International
tion, length of hospital stay and serious adverse events.
Consortium has examined the subject of threshold of GBS in
In Asia, a study from Korea randomised patients to
managing UGIB. In their study recruiting patients from four
receive restrictive transfusion (haemoglobin <8 g/dL) or
countries (Scotland, England, Denmark and New Zealand), they
liberal transfusion (target haemoglobin >10 g/dL), 17 and
found that the GBS at a cut-off point of <1 and <2 identified
confirmed that restrictive transfusion was associated with
low-risk patients with a higher level of specificity than setting
less recurrent bleeding. Pooling these four studies together,
the cut-off value of GBS at 0 (40–49% vs 22%).12 The GBS at a
a meta-analysis of randomised controlled trials concluded,
cut-off value of <2 had the highest specificity to detect adverse
admitting moderate heterogeneity of studies and over 90%
outcomes but missed 3% of high-risk patients. Therefore, the
of patients from two studies,15 16 that restrictive blood
authors suggested that a GBS cut-off point at <1 was most suit-
transfusion conferred a lower mortality (relative risk 0.65,
able as a guide for outpatient management.12 Some argue that
95% CI 0.44 to 0.97) and lower rebleeding rate (relative
even the GBS has a relatively low specificity to predict adverse
rate 0.58, 95% CI 0.40 to 0.84). 18 There was, however, no
outcomes.11 This comment is primarily based on a systematic
difference in thromboembolic events associated with blood
review. The working group believe that by adopting a GBS
transfusion.
cut-off point <1, most hospitals can reduce the majority of
A retrospective nationwide survey of 5861 hospital admis-
unnecessary hospital admissions without missing high-risk
sions in Denmark showed that the number of units of red
patients. It will naturally translate into a significant reduction
blood cells is a predictor of the need for repeated endoscopy,
in costs of managing UGIB. As we have pointed out in our
surgery and 30-day mortality.19 In Canada, a retrospective
previous consensus statement,1 a low risk for recurrent bleeding
cohort of 1677 patients with UGIB found that transfusion of
is not the same as low risk for mortality. Patients whose bleeding
red blood cells within 24 hours of presentation was signifi-
is controlled might still die owing to non-bleeding related causes
cantly and independently associated with an increased risk
such as cardiac and pulmonary decompensation.13 Therefore,
of recurrent bleeding.20 Another retrospective cohort from
the GBS score and its cut-off threshold cannot be interpreted
Australia, which included 2228 patients with NVUGIB, also
as a score to predict mortality. The working group agreed that
reported that blood transfusion of more than four units is
existing published data support this statement with a level of
associated with increased risk of further bleeding, but not
evidence graded as high.
with higher mortality.21
It should be pointed out that the mechanisms explaining
Statement 3: Bloodtransfusion should be restricted in the why transfusion may lead to recurrent bleeding in NVUGIB
management of upper gastrointestinal bleeding are not known. Based on existing published data, the Asia-Pa-
cific Working Group recommended a restrictive transfusion
(Accept—agreement: 100%, level of evidence: moderate) strategy, without specifying whether the threshold should
It would seem uncontroversial to replace blood loss in UGIB be 7 g/dL or 8 g/dL. Clinical discretion should be exercised
by transfusion of red blood cell and plasma to restore and when transfusion is given. In patients with active cardiovas-
preserve tissue perfusion and blood pressure. The earliest cular disease, the role of transfusion needs to be individual-
study that suggested that blood transfusion might be harmful ised based on assessment of blood loss and the cardiovascular
to patients with gastrointestinal haemorrhage was published status. In patients with massive active bleeding with dropping
over 30 years ago. A small group of patients with severe UGIB blood pressure, a more liberal transfusion strategy might be
were randomised to receiving either a blood transfusion in necessary. There is also an expectation that more data will
cases of a haemoglobin <8 g/dL or shock, or transfusion of at allow a refinement of this recommendation. The level of
least two units of blood regardless of haemoglobin level and evidence was agreed to be moderate.
Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276 1759
Guidelines
Statement 4: Platelet transfusion has no benefit for patients with A nationwide survey from the UK included 4478 patients. It
upper gastrointestinal bleeding taking antiplatelet agents showed that earlier endoscopy (<12 hours) was not associated
with a lower mortality, or need for surgery, compared with
(Accept—agreement: 88.9%, level of evidence: low) endoscopy offered within 24 hours.27
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
Antiplatelet agents, including aspirin and thienopyridines (eg, Two studies from Asia also examined this question. In Singa-
clopidogrel), are increasingly used in patients with various pore, a cohort study showed that in high-risk UGIB patients
cardiovascular and cerebrovascular conditions as primary and with a GBS>12, timing of endoscopy is the most important
secondary prevention. As aspirin and thienopyridines cause factor associated with all-cause in-hospital mortality.28 The
irreversible blockage of platelet function for the lifespan of the cut-off time of endoscopy that improved survival of such
platelet, approximately 8–10 days, their effects are expected to patients was within 13 hours from presentation. In contrast, a
last for days after discontinuation of this medication in patients study from Hong Kong selected high-risk patients (GBS >12)
with acute UGIB. It is common practice to transfuse platelets for randomisation to urgent (within 6 hours of presentation)
in patients with acute UGIB while receiving antiplatelet agents versus early (within 24 hours of presentation) endoscopy.
despite having normal platelet count. This study did not confirm the benefit of very early endos-
A case–control study compared patients with UGIB with or copy.29 With endoscopy within 6 hours, patients had more
without platelet transfusion irrespective of platelet counts. It active bleeding lesions requiring endoscopic haemostasis, but
showed that platelet transfusion did not reduce bleeding but this conferred no benefit in prevention of recurrent bleeding,
probably increased the overall mortality.22 The use of platelet mortality, requirement of blood transfusion and duration of
transfusion was studied in another retrospective observational hospital stay. Urgent endoscopy within 6 hours of presenta-
study comparing a group of patients with UGIB in the ICU tion for all NVUGIB seems unnecessary. The working group
for clinical outcome.23 There was no demonstrable differ- accepts that in some highly selected patients, such as those who
ence in total hospital stay, amount of blood transfusion and present with haemodynamic shock or instability, an urgent
resulting haemoglobin levels between the platelet-transfused endoscopy, say within 12 hours of admission, may benefit the
and non-transfused group. The platelet-transfused patients, patient after initial resuscitation and stabilisation. However,
however, were found to have had a shorter stay in the ICU. In offering urgent endoscopy to all patients who present with
view of the absence of any demonstrable improvement in clin- NVUGIB is not considered necessary. The level of evidence is
ical outcome, the working group did not recommend platelet graded as moderate.
transfusion in UGIB even if patients are taking antiplatelet
agents. The working group considered the evidence for this
recommendation weak and remarked that this statement Statement 6: Endoscopic haemostatic powder spray (such as
applied only to patients with normal platelet counts. Hemospray) is useful as a stop-gap treatment in NVUGIB
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
(Accept—agreement: 94.4%, level of evidence: moderate) noma. Procedure time was not a clear risk factor for post-ESD
In contrast to haemostatic powder spray, the over-the-scope-clip bleeding but the need for endoscopic haemostasis was a factor.
(OTSC), if successfully applied, appears to provide a firm and Experienced practitioners of ESD stated that the management
definite control of bleeding in NVUGIB. of post-ESD bleeding did not differ from that for peptic ulcer
Made from nitinol alloy, the OTSC fits to the tip of the endo- bleeding. Most post-ESD bleeding occurred within 24 hours
scope and can be deployed by tightening the thread with the after the procedure. Endoscopic devices used in peptic ulcer
hand wheel using a mechanism similar to rubber band variceal bleeding can also be used to treat post-ESD bleeding. Most
ligators. After being released from the applicator, the shape- post-ESD bleeds come from a focal bleeding point and are
memory effect and elasticity of the alloy result in firm closure of confined to mucosal and submucosal layers, without pene-
the clip. Compared with conventional clips, the OTSC can take trating through the muscle layer. Hence, haemostasis in such
up much more tissue by grasping deeper layers of the gastroin- circumstances is relatively easy. There is no agreement on
testinal wall, and hence the device can be used to treat bleeding whether proton pump inhibitors are better than histamine-2
and bowel perforation. receptor antagonists after ESD.45 46 Second-look endoscopy
Several case series reported promising results of successful has not been proved to be associated with less postproce-
haemostasis in the range of 70–100%.39–42 Recurrent bleeding dural bleeding,47 Overall, the working group concluded that
within 7 days occurred in 5–33%. This device, however, is tech- post-EMS and post-ESD bleeds can be managed like peptic
nically slightly more demanding than other through-the-scope ulcer bleeds. The level of evidence was graded as moderate.
haemostatic treatments. Deployment of the OTSC requires
accurate positioning and adequate retraction of tissue (either by Post-endoscopic management
suction or retractor) into the cap of the OTSC before the clip can Statement 9: As an adjunct to endoscopic treatment, high-dose oral
be released properly. The retractor or anchor device is used in proton pump inhibitors can be used to prevent rebleeding
hard fibrotic ulcers, especially those located in difficult positions
such as the high lesser curvature of the stomach. It punctures the
base of the lesion and allows tissue to be pulled into the cap. At (Accept—agreement: 88.9%, level of evidence: moderate)
certain locations in the stomach (eg, proximal lesser curve of the The use of intravenous high-dose proton pump inhibitors
stomach) and duodenum (junction of the first and second part of (PPIs) has become standard practice in the management of
the duodenum), this can be technically challenging. upper gastrointestinal bleeding. At least three randomised
A multicentre randomised control trial comparing through- trials, all from South Asia, showed that oral PPIs, given with
the-scope clips (TTSC) with OTSC has recently been reported,43 or without endoscopic therapy, also reduce the risk of recur-
In that study, 32 patients received TTSC and 33 received rent bleeding from peptic ulcer.48–50 New evidence suggests
OTSC. Initial haemostasis was reported in 62.5% of those who that high-dose oral PPIs may have a similar effect to their
received TTSC and 96.8% received OTSC (P=0.002). Recur- action in preventing recurrent bleeding from peptic ulcers.
rent bleeding within 7 days after treatment occurred at the same A study from Hong Kong recruited 118 high-risk patients
rate in both groups (33.3% vs 24.4%). The interim results of this with Forrest I or IIa/b peptic ulcer bleeding to receive
study suggested that OTSC is a better haemostatic device than either IV esomeprazole plus oral placebo or oral esomepra-
haemoclips in the treatment of peptic ulcer bleeding. In view of zole (40 mg every 12 hours) plus IV PPI placebo.51 Recur-
the promising interim results while waiting for a full report, the rent bleeding within 30 days was reported in 7.7% in the
working group recommends the use of OTSC in treating lesions IV esomeprazole group and 6.4% in the oral esomeprazole
refractory to conventional endoscopic therapy, such as through- group. There was no difference in the requirement for blood
the-scope haemoclips, thermal device or endoscopic injection. transfusion, repeated endoscopic therapy and hospital stay
There may be a role for OTSC in primary therapy, especially between the two groups. It was noted that the study was
in peptic ulcer bleeding with large vessels. This device will add stopped prematurely and was not designed as an equivalent
to the armamentarium for NVUGIB, with the level of evidence trial. The trend suggests that the action of high-dose oral
graded as moderate. PPI peptic ulcer bleeding is comparable to that of IV PPI. In
Taiwan, IV esomeprazole was compared with oral lansopra-
zole (30 mg four times a day for 3 days) in patients with peptic
Statement 8: Endoscopic treatment of delayed bleeding after ulcer bleeding.52 There was no difference in all the clinical
endoscopic mucosal resection or endoscopic submucosal dissection outcome parameters, except that those who received oral PPI
is similar to that for bleeding peptic ulcers had a shorter hospital stay. In Korea, when IV omeprazole
was compared with oral rabeprazole (20 mg twice daily), the
(Accept—agreement: 89%, level of evidence: moderate) recurrent bleeding rate, surgical intervention and mortality
As endoscopic mucosectomy (EMS) and endoscopic submu- between the two groups were comparable.53
cosal dissection (ESD) are gaining popularity among tertiary There is no properly powered RCT to confirm that high-
centres worldwide, complications such as delayed bleeding dose oral PPI is as effective as IV PPI. The working group
require more guideline from experts. A meta-analysis which accepted that high-dose oral PPI can be used to prevent recur-
pooled data from over 70 studies (15 RCTs, three prospec- rent bleeding, but emphasised that it has to be used as an
tive trials, five prospective cohort studies, and 48 retrospec- adjunct to endoscopic therapy. Only after endoscopic haemo-
tive cohort and case–control studies) depicted clearly the stasis is achieved, can high-dose oral PPI be recommended
risks associated with EMS and ESD.44 Post-ESD bleeding to prevent recurrent bleeding. The definition of high-dose
occurred in 5.1% (95% CI 4.5% to 5.7%) of patients. Risk oral PPI has been stated as at least 80 mg of esomeprazole (or
Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276 1761
Guidelines
equivalent dosage of other PPIs). The high oral dose should be therapy. The level of evidence was considered moderate as
maintained for at least 3 days, which is the period of highest only one RCT has been carried out, with limitations.
risk of recurrent bleeding. If the patient’s condition remains
stable, standard dose oral PPI can be resumed afterwards. All Statement 11: Routine second-look endoscopy is not recommended
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
the reports were relatively small and underpowered, and thus for patients after gastric endoscopic submucosal dissection
the level of evidence was graded as moderate, in the hope that
future studies might confirm this finding.
(Accept—agreement: 88.9%, level of evidence: high)
Submucosal arterioles are often exposed after gastric ESD
Statement 10: There is no special preference for a particular proton just above the level of muscularis propria. As mentioned
pump inhibitor when used concomitantly with clopidogrel above, the management of post-ESD haemorrhage is similar
to that for bleeding peptic ulcers. However, the pathology
(Accept—agreement: 94%, level of evidence: moderate) is different as the artificial ulcers created by ESD have less
Concern about a potential interaction between PPIs and fibrosis and hence it is easier to stop bleeding using either
clopidogrel (a prodrug that requires CYP450 for metab- a thermal or mechanical device. In this situation, routine
olism) arose from a number of in vivo platelet aggregation second-look endoscopy may not be necessary.
studies.54 55 Different PPIs may have unequal effects on A single-centre randomised trial comparing routine
clopidogrel metabolism, which further intensifies the debate second-look endoscopy with no second-look endoscopy
about a potential interaction between PPIs and clopidogrel. It after gastric ESD in 155 patients demonstrated no differ-
has been suggested that PPIs such as lansoprazole, pantopra- ence in recurrent bleeding and need for transfusion.68 Two
zole and rabeprazole have less interaction with clopidogrel prospective randomised trials including larger (>35 mm) arti-
since they have fewer inhibitory effects on CYP2C19.56–59 ficial ulcers after gastric ESD also showed no difference in
More data suggest that concomitant use of a PPI with clopi- outcome between those with and without second-look endos-
dogrel increases the risk of major adverse cardiovascular copy.69 70 Another large multicentre prospective randomised
events (MACE) in patients with high cardiovascular risk. trial conducted in Japan, including 262 patients, further
A retrospective cohort study from China, including about demonstrated that routine second-look endoscopy confers no
6200 patients who received aspirin plus clopidogrel with protection against recurrent haemorrhage.71 This result was
a PPI, reported that PPI users had a 3% increase in MACE confirmed by a meta-analysis including four randomised trials
compared with non-users.60 Another retrospective cohort and four non-randomised trials. No difference in post-ESD
from Italy reported an alarming 12% increase in MACE in bleeding rate was shown by the pooled data between those
patients receiving a PPI in conjunction with clopidogrel.61 who did and did not receive second-look endoscopy.47 The
PPIs interact with clopidogrel in patients with coronary, panel concurred with high agreement that based on current
cerebrovascular and peripheral artery disease. This interac- evidence, routine second-look endoscopy is not recommended
tion was implicated in the Factores de Riesgo y ENfermedad for patients after gastric ESD, with a level of evidence graded
Arterial (FRENA) Registry as leading to a doubling of the as high.
incidence of myocardial infarction and ischaemic stroke.62
However, a number of studies showed otherwise. A Japa- Statement 12: Among patients with high cardiothrombotic risk
nese study of patients with coronary stenting receiving dual
receiving antiplatelet agents, these agents should be resumed as
antiplatelet agents showed no effect of PPIs on MACE or
soon as haemostasis can be established
mortality.63 The safety of prescribing a PPI with clopidogrel
was also suggested in a case–control study enrolling 23 655
patients from the Netherlands.64 Beside cardiovascular risk, (Accept—agreement: 100%, level of evidence: high)
the safety profile of using a PPI concomitantly with clopido- Patients with NVUGIB taking antiplatelet agents face a
grel was also confirmed in group of 2765 patients who had dilemma of discontinuing these drugs to facilitate controlling
strokes.65 of haemorrhage or to continue taking these drugs to avoid
Finally, the strongest support came from the only RCT, thromboembolic complications. Using the UK primary care
the Clopidogrel and the Optimisation of Gastrointestinal database, a study reported the cardiovascular and UGIB conse-
Events Trial (COGENT).66 Patients receiving dual anti- quences of low-dose acetylsalicylic acid (ASA) in patients aged
platelet therapy (aspirin and clopidogrel) were randomised 50–84 years.72 Based on this dataset, the attributable risks
to receive omeprazole or placebo. Omeprazole significantly associated with ASA discontinuation for non-fatal myocar-
reduced rates of composite gastrointestinal events from 2.7% dial infarction/coronary death and ischaemic stroke were 17
(without omeprazole) to 1.2% (with omeprazole). There was and 11 per 1000 people, respectively. On the other hand, the
no significant excess of cardiovascular events in 3759 high- risk of UGIB with continued ASA was 1.6 per 1000 people.
risk cardiovascular patients recruited in this study. Although This amounts to eight extra cardiovascular events for a reduc-
the trial was terminated prematurely, the follow-up period tion of 0.4 UGIB events per year.
was sufficient to demonstrate the rate of MACE related to The only RCT examining this subject was reported by the
the medication. These data from COGENT, the only large- Hong Kong group. It showed that immediate resumption of
scale RCT evaluating the effects of PPI on clinical endpoints aspirin for high-risk cardiac patients was critical as it did
in patients requiring dual antiplatelet therapy, provided reas- not increase the risk of fatal haemorrhage but significantly
surance about the safety of PPIs in high-risk cardiovascular improved the 30-day survival.73 Subsequently, a retrospec-
subjects.67 Therefore, the working group recommended the tive analysis of 118 patients who received low-dose aspirin
use of a PPI without preference for a particular type, in patients with 40% of cases discontinuing the drug for 2 years showed
requiring gastrointestinal protection against dual antiplatelet a sevenfold increase in risk for a cardiovascular event and
cardiac death.74 A study from Taiwan comparing 89 patients
receiving esomeprazole alone versus 89 patients receiving
1762 Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276
Guidelines
esomeprazole plus aspirin showed that the ulcer healing rate Statement 14: Among direct oral anticoagulant (DOAC) or warfarin
between the groups was almost identical. There is no evidence users with high cardiothrombotic risk who develop ulcer bleeding,
to suggest that aspirin would delay the healing of peptic ulcer DOAC or warfarin should be resumed as soon as haemostasis is
when treated with a PPI.75 It should be noted that these two established
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
studies used a relative low dose of aspirin (80–100 mg daily).
Even at this dose, the cardioprotective effects of aspirin
are retained and haemorrhagic complications are not detri- (Accept—agreement: 83.3%, level of evidence: low)
mental. Similarly, although data are lacking on non-aspirin Clinical evidence is lacking to support strategies for managing
antiplatelet agents, the protective effects against thromboem- patients with NVUGIB and high cardiothrombotic risk who
bolic events in cardiovascular and cerebrovascular diseases receive DOAC or warfarin. For such patients with atrial fibril-
are considered more important than the increased risk of lation and/or valvular heart diseases, management should
gastrointestinal bleeding. depend on the balance between thrombotic risk and bleeding
No study has investigated the optimal timing for the resump- risk. In acute NVUGIB, DOAC or warfarin should be with-
tion of antiplatelet agents, but one can consider resuming held to facilitate achievement of haemostasis.
these agents on day 1 if endoscopy shows a clean-based ulcer. If the patient is taking warfarin, four-factor prothrombin
In patients who received endoscopic therapy for bleeding, complex concentrate (PCC) and vitamin K or fresh frozen
antiplatelet agents can be resumed 72 hours after the treat- plasma can be given for life-threatening GI bleeding. Warfarin
ment—that is, passing the period of highest risk for recur- reversal should be used for life-threatening bleeding irrespec-
rent bleeding. The working group unanimously endorsed this tive of the international normalised ratio (INR). Current
recommendation of resuming antiplatelet agents early when evidence does not show any correlation between INR at
UGIB is under control. The level of evidence of this statement presentation and outcomes of GI bleeding.79 A combination
was considered as high. of PCC and vitamin K is preferred for urgent reversal of
warfarin. PCC has advantages over fresh-frozen plasma such
Statement 13: In patients receiving dual antiplatelet agents, at as faster onset of action and minimal risk of fluid overload.
least one antiplatelet agent should be resumed in cases of upper Endoscopic therapy should not be delayed in patients with
gastrointestinal bleeding serious UGIB. Warfarin and DOAC treatment should be with-
held in patients with ongoing NVUGIB.80
The Food and Drug Administration in October 2015
(Accept—agreement: 94.4%, level of evidence: low)
granted approval for idarucizumab, a potent monoclonal
The most commonly used antiplatelet agents are ASA
antibody against dabigatran, for use in patients with uncon-
(aspirin), a cyclo-oxygenase inhibitor and thienopyridines
trolled bleeding. In a multicentre study of 503 patients with
that bind to P2Y12 component of the adenosine diphos-
uncontrolled bleeding who were about to undergo an urgent
phate receptors. After stopping ASA, 7–9 days are required
procedure, 5 g of idarucizumab reversed the anticoagulant
to regain full platelet function, whereas in the case of thien-
effect of dabigatran within 4 hours in almost all patients.81
opyridines (such as clopidogrel or prasugrel), the minimum
However, its true benefit in NVUGIB is still unclear because
duration to restore platelet function is 5–7 days. Use of dual
of limited clinical data.
antiplatelet agents, often ASA plus a thienopyridine, may
Once the bleeding is controlled, the decision to resume
confer a threefold increase in the risk of UGIB over single-
anticoagulants should be made by a multidisciplinary team
agent antithrombotic therapy.76 There are no data guiding the
with cardiologist, gastroenterologist, intensivist and, if
management of these patients using dual antiplatelet therapy
feasible, with patient’s participation. The CHA2DS2-VASC
when they develop UGIB. We do not recommend withholding
and HAS-BLED score are useful in assessing the need for
both antiplatelet drugs because the median time to coronary
continuing anticoagulation.82–84 In patients with high cardio-
stent thrombosis can be as short as 7 days with both drugs
vascular risk, resumption of anticoagulants should not be
withheld as compared with 122 days with only clopidogrel
delayed. A large prospective cohort study of patients with
withheld.77 Balancing the risk and benefit of discontinuation
atrial fibrillation in Denmark showed that resumption of
of antiplatelet agents, the ASGE recommends that cessation
single anticoagulants was associated with lowest rate of
of all antiplatelet therapy after PCI should be avoided, and
all-cause mortality85 Most data on vitamin K antagonists
furthermore, when only one antiplatelet agent is used, aspirin
show that resuming the oral anticoagulant is associated with
should be continued as it is associated with a lower risk for
a lower mortality despite more frequent major bleeding.86 A
causing recurrent bleeding. On the other hand, for patients
recent study suggested that dabigatran offers similar protec-
with a high risk of thrombosis, such as those with drug-eluting
tion against thromboembolism with less rebleeding than
coronary stents, clopidogrel should not be discontinued for
warfarin after major bleeding.87 The timing for resumption of
more than 5 days.78
warfarin should be assessed on a patient-by-patient basis. Since
Cardiologists’ opinion should be sought for the commence-
the time required for re-anticoagulation will be prolonged
ment of antiplatelet agents. In patients with low risk of recur-
after warfarin reversal with vitamin K, we recommend early
rent bleeding from the gastrointestinal tract, antiplatelet
resumption of warfarin once haemostasis has been achieved,
agents should not be discontinued at all. The working group
especially in patients with high thromboembolic risk. With
consider that this statement is primarily based on pharmaco-
DOAC, the effect should disappear in 1–2 days after stopping
logical characteristics of the antiplatelet agents. Clinical trials
the drug. Therefore, once bleeding is under control, DOAC
to test the safety of this strategy are eagerly awaited. The
can be resumed earlier (within 1–2 days) bearing in mind that
level of evidence was considered low.
anticoagulation is achieved rapidly within hours and there-
fore might increase the risk of rebleeding. If early anticoagu-
lation is indicated in patients with high thromboembolic risk,
a bridge approach with heparin or enoxaparin is useful. The
Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276 1763
Guidelines
working group considered that this recommendation is based tool and would like to await further clinical studies to prove its
primarily on expert opinion without data from RCTs, hence value.
the level of evidence was graded as low.
Statement 2: Pre-endoscopy intravenous proton pump inhibitors are
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
Rejected statements recommended in stable patients awaiting endoscopy
Statement 1: Video capsule endoscopy can be considered as a triage
tool for patients who require early intervention (Reject—agreement: 72.2%)
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
transfusion requirement and mortality. In the per protocol analysis score, the receiver operating characteristic curve to predict need
of 90 patients who received angiographic embolisation compared for second-look endoscopy appeared promising, but outcome data
with 113 patients who did not, there a trend favouring angiography, were lacking. To date, there is still a lack of evidence to suggest
with a significant reduction in mortality. The size of ulcer (>1.5 cm) that any risk stratification method is effective in selecting patients
is the best predictive parameter associated with benefit of angiog- at high risk who would benefit from second-look endoscopy and
raphy. However, this is a single study with a marginal benefit. The pre-emptive treatment.
working group considered it premature to recommend angiography The working group therefore rejected the statement that a risk
to prevent recurrent bleeding from peptic ulcer after endoscopic stratification score may be useful to identify high-risk bleeding
treatment. More evidence from future clinical trials is necessary. ulcers after endoscopic therapy for second-look endoscopy.
Future studies should be conducted to verify the use of a risk
Statement 4: A risk stratification score should be used to identify stratification system.
high-risk bleeding ulcers after endoscopic therapy for second-look
endoscopy Statement 5: A Doppler endoscopic probe should be used to guide
endoscopic therapy in order to ensure adequate haemostasis
(Reject—agreement: 55.6%)
(Reject—agreement: 66%)
Modified: A risk stratification score may be useful to identify The use of a Doppler endoscopic probe is not new. Previous studies
high-risk bleeding ulcers after endoscopic therapy for second-look have shown that the technique requires skill and the results are often
endoscopy irreproducible. A study from the CURE group has recently reported
that the Doppler endoscopic probe (DEP) can demonstrate major
(Reject—agreement: 23%) stigmata of peptic ulcer bleeding (spurting ulcer, blood clot and
Recurrent bleeding occurs in 8–15% of patients with peptic ulcer visible vessel) and had a significantly higher rate of detecting arterial
bleeding and is associated with a two- to fivefold increase in blood flow than oozing ulcer and ulcer with flat pigmented spot.69
mortality. The objective of routine second-look endoscopy, usually The authors proposed that DEP can improve risk stratification in
performed within 24 hours after index endoscopy, is to pre-emp- the management of peptic ulcer bleeding. This DEP is a FDA-ap-
tively treat peptic ulcers with persistent stigmata of recent haemor- proved disposable probe (Vascular Technology, Nashua, New
rhage before they start bleeding again. Hampshire, USA) that can be inserted into the working channel of
Second-look endoscopy was first investigated for its efficacy in the gastroscope. Subsequently, the same group compared the stan-
prevention of peptic ulcer rebleeding by a few prospective randomised dard endoscopic treatment (using a thermal or mechanical device
trials in the 1990s.105–108 The results from these randomised trials with or without injection) with endoscopic treatment guided by
were conflicting. Some showed that second-look endoscopy and DEP.114 Endoscopic haemostasis was applied to peptic ulcers until
repeated endoscopic haemostasis were effective in preventing recur- DEP showed no Doppler signal of residual blood flow. The results
rent haemorrhage107 108 while others demonstrated no efficacy.105 106 showed that with DEP-guided therapy, recurrent bleeding can be
The reasons for these conflicting results included recruitment of reduced from 26.3% to 11.1%. However, because of the small
patients with different levels of rebleeding risk, a non-standardised number of patients with significant events, there was no difference
method of primary haemostasis and variation in the performance of demonstrated in requirement for surgery or angiography to salvage
second-look endoscopy. A meta-analysis based on eight prospective uncontrolled bleeding, and there was no reduction in mortality.
randomised trials concluded that second-look endoscopy reduced Although a remarkable RCT, the working group concluded that it
rebleeding in the absence of high-dose PPI especially in patients at is too early to recommend DEP for the management of UGIB. This
very high risk.109 Analysis of these pooled data also suggested that is based on the argument that only a single study from a renowned
second-look endoscopy reduced the need for surgery but had no referral centre demonstrated the benefit of this skill-demanding
significant effect on mortality. However, after removing two trials technique. The working group opined that more data, from other
that included patients with a high risk of rebleeding, no benefit from centres recruiting more patients, are required to confirm the value
second-look endoscopy was found.109 110 A recent randomised trial of DEP.
compared intravenous PPI infusion with second-look endoscopy in
patients after receiving endosocopic haemostasis for peptic ulcer Conclusion
bleeding.110 It demonstrated no difference in recurrent bleeding, Since the last Asia-Pacific consensus on NVUGIB published 7 years
need for surgery and mortality between the two treatment strat- ago, significant advancement has been made in the clinical manage-
egies. Furthermore, second-look endoscopy did not appear to be ment of patients before endoscopy (risk stratification scores, blood
cost-effective when offered to all patients. and platelet transfusion, use of PPIs) and in the development of
The question remains whether risk stratification with selection of endoscopic haemostasis (haemostatic powder spray and over-the-
high-risk patients may lead to a benefit from second-look endoscopy scope clips). Emerging techniques, such as the use of capsule endos-
on the next day and with repeated treatment in case of persistent copy for patient triage, a Doppler endoscopic probe for assessing
stigmata. The Baylor Bleeding Score attempted to select high-risk adequacy of endoscopic therapy, and the pre-emptive use of angio-
patients to receive second-look endoscopy versus controls and graphic embolisation, look promising but require further evalua-
found a 24% difference in rate of recurrent bleeding.111 A study tion. The use of PPIs has been further clarified, showing that routine
including 699 patients from Korea showed that use of a non-ste- use of IV PPI is unnecessary but high-dose oral PPI after endoscopy
roidal anti-inflammatory drug, large transfusion volume and failure may be beneficial. Experience in managing NVUGIB after EMR and
to perform second-look endoscopy were risk factors for recurrent ESD is accumulating. An emerging problem is the increasing use of
Sung JJY, et al. Gut 2018;67:1757–1768. doi:10.1136/gutjnl-2018-316276 1765
Guidelines
double antiplatelet agents and direct oral anticoagulants in patients 9 Thanapirom K, Ridtitid W, Rerknimitr R, et al. Prospective comparison of three risk
with cardiac and cerebrovascular diseases, and clinical data to guide scoring systems in non-variceal and variceal upper gastrointestinal bleeding. J
Gastroenterol Hepatol 2016;31:761–7.
management are still minimal. Resumption of antiplatelet agents or 10 Mokhtare M, Bozorgi V, Agah S, et al. Comparison of Glasgow-Blatchford score
anticoagulants is important but the timing and choice of drugs to and full Rockall score systems to predict clinical outcomes in patients with upper
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
be resumed require further clinical trials for guidance. gastrointestinal bleeding. Clin Exp Gastroenterol 2016;9:337–43.
11 Ramaekers R, Mukarram M, Smith CA, et al. The predictive value of preendoscopic
risk scores to predict adverse outcomes in emergency department patients
Author affiliations
1 with upper gastrointestinal bleeding: a systematic review. Acad Emerg Med
Institute of Digestive Disease, The Chinese University of Hong Kong, Hong Kong
2 2016;23:1218–27.
Department of Gastroenterology and Hepatology, University of Malaya, Kuala
12 Laursen SB, Dalton HR, Murray IA, et al. Performance of new thresholds of the
Lumpur, Malaysia
3 Glasgow Blatchford score in managing patients with upper gastrointestinal bleeding.
Yong Loo Lin School of Medicine, National University of Singapore, Singapore
4 Clin Gastroenterol Hepatol 2015;13:115–21.
University of Ulsan College of Medicine, Ulsan, South Korea
5 13 Sung JJ, Tsoi KK, Ma TK, et al. Causes of mortality in patients with peptic ulcer
UST Hospital, University of Santo Tomas, Manila, Philippines
6 bleeding: a prospective cohort study of 10,428 cases. Am J Gastroenterol
Division of Gastroenterology and Hepatology, Nihon University School of Medicine, 2010;105:84–9.
Tokyo, Japan 14 Blair SD, Janvrin SB, McCollum CN, et al. Effect of early blood transfusion on
7
Asian Institute of Gastroenterology, Asian Healthcare Foundation, Hyderabad, India gastrointestinal haemorrhage. Br J Surg 1986;73:783–5.
8
Department of Medicine, Lyell McEwin Hospital, University of Adelaide, Adelaide, 15 Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper
South Australia, Australia gastrointestinal bleeding. N Engl J Med Overseas Ed 2013;368:11–21.
9
Department of Medicine, Jichi Medical School, Shimotsuke, Japan 16 Jairath V, Kahan BC, Gray A, et al. Restrictive versus liberal blood transfusion for
10
State Key Laboratory of Cancer Biology, Xijing Hospital of Digestive Diseases, Xi’an, acute upper gastrointestinal bleeding (TRIGGER): a pragmatic, open-label, cluster
China randomised feasibility trial. Lancet 2015;386:137–44.
11
China Medical University, Taichung, Taiwan 17 Lee JM, Chun HJ, Lee JS, et al. Sa1888 target level for hemoglobin correction in
12
University of Utah College of Health, Salt Lake City, Utah, USA patients with acute non-variceal upper gastrointestinal bleeding. Gastroenterology
13
University of California, Los Angeles, California, USA 2014;146:S-321.
14
Department of Gastroenterology and Hepatology, Erasmus University Medical 18 Odutayo A, Desborough MJ, Trivella M, et al. Restrictive versus liberal blood
Center, Rotterdam, The Netherlands transfusion for gastrointestinal bleeding: a systematic review and meta-analysis of
15
Department of Gastroenterology, University Hospital, Zaragoza, Spain randomised controlled trials. Lancet Gastroenterol Hepatol 2017;2:354–60.
19 Fabricius R, Svenningsen P, Hillingsø J, et al. Effect of transfusion strategy in acute
Acknowledgements Moe Kyaw, John Wong, Jessica Ching and Thomas Lam. non-variceal upper gastrointestinal bleeding: a nationwide study of 5861 hospital
Contributors Study design: JJYS; literature review: JJYS, PCYC, FKLC, JYWL, John admissions in Denmark. World J Surg 2016;40:1129–36.
Wong and Moe Kyaw; draft of the manuscript: JJYS; revision of the manuscript: all 20 Restellini S, Kherad O, Jairath V, et al. Red blood cell transfusion is associated with
authors. increased rebleeding in patients with nonvariceal upper gastrointestinal bleeding.
Aliment Pharmacol Ther 2013;37:316–22.
Funding This research received no specific grant from any funding agency in the 21 Subramaniam K, Spilsbury K, Ayonrinde OT, et al. Red blood cell transfusion
public, commercial or not-for-profit sectors. is associated with further bleeding and fresh-frozen plasma with mortality in
Competing interests None declared. nonvariceal upper gastrointestinal bleeding. Transfusion 2016;56:816–26.
22 Zakko L, Rustagi T, Douglas M, et al. No benefit from platelet transfusion for
Patient consent Not required. gastrointestinal bleeding in patients taking antiplatelet agents. Clin Gastroenterol
Provenance and peer review Not commissioned; externally peer reviewed. Hepatol 2017;15:46–52.
23 Victor N, Umakanthan J, Gandhi A, et al. 419: role of platelet transfusion in
Open access This is an open access article distributed in accordance with the
gastrointestinal bleeding in patients on antiplatelet therapy. Crit Care Med
Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which
2014;42:A1461.
permits others to distribute, remix, adapt, build upon this work non-commercially,
24 Tsoi KK, Ma TK, Sung JJ. Endoscopy for upper gastrointestinal bleeding: how urgent
and license their derivative works on different terms, provided the original work
is it? Nat Rev Gastroenterol Hepatol 2009;6:463–9.
is properly cited and the use is non-commercial. See: http://creativecommons.org/ 25 Kumar NL, Cohen AJ, Nayor J, et al. Timing of upper endoscopy influences outcomes
licenses/by-nc/4.0/ in patients with acute nonvariceal upper GI bleeding. Gastrointest Endosc
© Article author(s) (or their employer(s) unless otherwise stated in the text of the 2017;85:945–52.
article) 2018. All rights reserved. No commercial use is permitted unless otherwise 26 Laursen SB, Leontiadis GI, Stanley AJ, et al. Relationship between timing of
expressly granted. endoscopy and mortality in patients with peptic ulcer bleeding: a nationwide
cohort study. Gastrointest Endosc 2017;85:936–44.
27 Jairath V, Kahan BC, Logan RF, et al. Outcomes following acute nonvariceal upper
References gastrointestinal bleeding in relation to time to endoscopy: results from a nationwide
1 Sung JJ, Chan FK, Chen M, et al. Asia-Pacific Working Group consensus on non- study. Endoscopy 2012;44:723–30.
variceal upper gastrointestinal bleeding. Gut 2011;60:1170–7. 28 Lim LG, Ho KY, Chan YH, et al. Urgent endoscopy is associated with lower mortality
2 Stanley AJ, Laine L, Dalton HR, et al. Comparison of risk scoring systems for in high-risk but not low-risk nonvariceal upper gastrointestinal bleeding. Endoscopy
patients presenting with upper gastrointestinal bleeding: international multicentre 2011;43:300–6.
prospective study. BMJ 2017;356:i6432. 29 Wong JC, Lau JY, Tang RS, et al. 785 urgent versus early endoscopy for upper
3 Park SM, Yeum SC, Kim B-W, et al. Comparison of AIMS65 score and other scoring gastrointestinal bleeding with Glasgow-Blatchford score ≥12. Gastroenterology
systems for predicting clinical outcomes in Koreans with nonvariceal upper 2015;148:S-154.
gastrointestinal bleeding. Gut Liver 2016;10:526–31. 30 Sung JJ, Luo D, Wu JC, et al. Early clinical experience of the safety and effectiveness
4 Yang HM, Jeon SW, Jung JT, et al. Comparison of scoring systems for nonvariceal of Hemospray in achieving hemostasis in patients with acute peptic ulcer bleeding.
upper gastrointestinal bleeding: a multicenter prospective cohort study. J Endoscopy 2011;43:291–5.
Gastroenterol Hepatol 2016;31:119–25. 31 Thayalasekaran S, Dixon S, Mundre P, et al. Mo1118 Hemospray use in the
5 Martínez-Cara JG, Jiménez-Rosales R, Úbeda-Muñoz M, et al. Comparison of management of upper gastrointestinal hemorrhage: a 2-year experience across two
AIMS65, Glasgow-Blatchford score, and Rockall score in a European series of teaching hospitals in the North and South of England. Gut 2017;66:A232.
patients with upper gastrointestinal bleeding: performance when predicting in- 32 Haddara S, Jacques J, Lecleire S, et al. A novel hemostatic powder for upper
hospital and delayed mortality. United European Gastroenterol J 2016;4:371–9. gastrointestinal bleeding: a multicenter study (the “GRAPHE” registry). Endoscopy
6 Laursen SB, Hansen JM, Schaffalitzky de Muckadell OB. The Glasgow Blatchford 2016;48:1084–95.
score is the most accurate assessment of patients with upper gastrointestinal 33 Sinha R, Lockman KA, Church NI, et al. The use of hemostatic spray as an adjunct to
hemorrhage. Clin Gastroenterol Hepatol 2012;10:1130–5. conventional hemostatic measures in high-risk nonvariceal upper GI bleeding (with
7 Robertson M, Majumdar A, Boyapati R, et al. Risk stratification in acute upper GI video). Gastrointest Endosc 2016;84:900–6.
bleeding: comparison of the AIMS65 score with the Glasgow-Blatchford and Rockall 34 Chen YI, Barkun A, Nolan S. Hemostatic powder TC-325 in the management of
scoring systems. Gastrointest Endosc 2016;83:1151–60. upper and lower gastrointestinal bleeding: a two-year experience at a single
8 Bryant RV, Kuo P, Williamson K, et al. Performance of the Glasgow-Blatchford score institution. Endoscopy 2015;47:167–71.
in predicting clinical outcomes and intervention in hospitalized patients with upper
GI bleeding. Gastrointest Endosc 2013;78:576–83.
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
37 Smith LA, Stanley AJ, Bergman JJ, et al. Hemospray application in nonvariceal 2012;13:783–9.
upper gastrointestinal bleeding: results of the Survey to Evaluate the Application of 62 Muñoz-Torrero JF, Escudero D, Suárez C, et al. Concomitant use of proton pump
Hemospray in the Luminal Tract. J Clin Gastroenterol 2014;48:e89–92. inhibitors and clopidogrel in patients with coronary, cerebrovascular, or peripheral
38 Holster IL, Kuipers EJ, Tjwa ET. Hemospray in the treatment of upper gastrointestinal artery disease in the factores de Riesgo y ENfermedad Arterial (FRENA) registry. J
hemorrhage in patients on antithrombotic therapy. Endoscopy 2013;45:63–6. Cardiovasc Pharmacol 2011;57:13–19.
39 Chan SM, Chiu PW, Teoh AY, et al. Use of the over-the-scope clip for treatment 63 Chitose T, Hokimoto S, Oshima S, et al. Clinical outcomes following coronary
of refractory upper gastrointestinal bleeding: a case series. Endoscopy stenting in Japanese patients treated with and without proton pump inhibitor. Circ J
2014;46:428–31. 2012;76:71–8.
40 Wedi E, Gonzalez S, Menke D, et al. One hundred and one over-the-scope-clip 64 Valkhoff VE, ’t Jong GW, Van Soest EM, et al. Risk of recurrent myocardial infarction
applications for severe gastrointestinal bleeding, leaks and fistulas. World J with the concomitant use of clopidogrel and proton pump inhibitors. Aliment
Gastroenterol 2016;22:1844–53. Pharmacol Ther 2011;33:77–88.
41 Manta R, Galloro G, Mangiavillano B, et al. Over-the-scope clip (OTSC) represents 65 Juurlink DN, Gomes T, Mamdani MM, et al. The safety of proton pump inhibitors and
an effective endoscopic treatment for acute GI bleeding after failure of conventional clopidogrel in patients after stroke. Stroke 2011;42:128–32.
techniques. Surg Endosc 2013;27:3162–4. 66 Bhatt DL, Cryer BL, Contant CF, et al. Clopidogrel with or without omeprazole in
42 Kirschniak A, Subotova N, Zieker D, et al. The over-the-scope clip (OTSC) for the coronary artery disease. N Engl J Med Overseas Ed 2010;363:1909–17.
treatment of gastrointestinal bleeding, perforations, and fistulas. Surg Endosc 67 Vaduganathan M, Cannon CP, Cryer BL, et al. Efficacy and safety of proton-pump
2011;25:2901–5. inhibitors in high-risk cardiovascular subsets of the COGENT trial. Am J Med
43 Schmidt A, Goelder S, Messmann H, et al. 62 over-the-scope-clips versus 2016;129:1002–5.
standard endoscopic therapy in patients with recurrent peptic ulcer bleeding 68 Ryu HY, Kim JW, Kim HS, et al. Second-look endoscopy is not associated with better
and a prospective randomized, multicenter trial (Sting). Gastrointest Endosc clinical outcomes after gastric endoscopic submucosal dissection: a prospective,
2017;85:AB50. randomized, clinical trial analyzed on an as-treated basis. Gastrointest Endosc
44 Libânio D, Costa MN, Pimentel-Nunes P, et al. Risk factors for bleeding after 2013;78:285–94.
gastric endoscopic submucosal dissection: a systematic review and meta-analysis. 69 Jensen DM, Ohning GV, Kovacs TO, et al. Doppler endoscopic probe as a guide to risk
Gastrointest Endosc 2016;84:572–86. stratification and definitive hemostasis of peptic ulcer bleeding. Gastrointest Endosc
45 Yang Z, Wu Q, Liu Z, et al. Proton pump inhibitors versus histamine-2-receptor 2016;83:129–36.
antagonists for the management of iatrogenic gastric ulcer after endoscopic mucosal 70 Kim JS, Chung MW, Chung CY, et al. The need for second-look endoscopy to prevent
resection or endoscopic submucosal dissection: a meta-analysis of randomized trials. delayed bleeding after endoscopic submucosal dissection for gastric neoplasms: a
Digestion 2011;84:315–20. prospective randomized trial. Gut Liver 2014;8:480–6.
46 Tomita T, Kim Y, Yamasaki T, et al. Prospective randomized controlled trial to compare
71 Mochizuki S, Uedo N, Oda I, et al. Scheduled second-look endoscopy is not
the effects of omeprazole and famotidine in preventing delayed bleeding and
recommended after endoscopic submucosal dissection for gastric neoplasms (the
promoting ulcer healing after endoscopic submucosal dissection. J Gastroenterol
SAFE trial): a multicentre prospective randomised controlled non-inferiority trial. Gut
Hepatol 2012;27:1441–6.
2015;64:397–405.
47 Kim EH, Park SW, Nam E, et al. Role of second-look endoscopy and prophylactic
72 Cea Soriano L, Bueno H, Lanas A, et al. Cardiovascular and upper gastrointestinal
hemostasis after gastric endoscopic submucosal dissection: a systematic review and
bleeding consequences of low-dose acetylsalicylic acid discontinuation. Thromb
meta-analysis. J Gastroenterol Hepatol 2017;32:756–68.
Haemost 2013;110:1298–304.
48 Khuroo MS, Yattoo GN, Javid G, et al. A comparison of omeprazole and placebo for
73 Sung JJ, Lau JY, Ching JY, et al. Continuation of low-dose aspirin therapy in peptic
bleeding peptic ulcer. N Engl J Med Overseas Ed 1997;336:1054–8.
ulcer bleeding: a randomized trial. Ann Intern Med 2010;152:1–9.
49 Javid G, Masoodi I, Zargar SA, et al. Omeprazole as adjuvant therapy to endoscopic
74 Derogar M, Sandblom G, Lundell L, et al. Discontinuation of low-dose aspirin therapy
combination injection sclerotherapy for treating bleeding peptic ulcer. Am J Med
after peptic ulcer bleeding increases risk of death and acute cardiovascular events.
2001;111:280–4.
Clin Gastroenterol Hepatol 2013;11:38–42.
50 Kaviani MJ, Hashemi MR, Kazemifar AR, et al. Effect of oral omeprazole in reducing
75 Liu CP, Chen WC, Lai KH, et al. Esomeprazole alone compared with esomeprazole
re-bleeding in bleeding peptic ulcers: a prospective, double-blind, randomized,
clinical trial. Aliment Pharmacol Ther 2003;17:211–6. plus aspirin for the treatment of aspirin-related peptic ulcers. Am J Gastroenterol
51 Sung JJ, Suen BY, Wu JC, et al. Effects of intravenous and oral esomeprazole in the 2012;107:1022–9.
prevention of recurrent bleeding from peptic ulcers after endoscopic therapy. Am J 76 Abraham NS, Hartman C, Richardson P, et al. Risk of lower and upper
Gastroenterol 2014;109:1005–10. gastrointestinal bleeding, transfusions, and hospitalizations with complex
52 Yen HH, Yang CW, Su WW, et al. Oral versus intravenous proton pump inhibitors antithrombotic therapy in elderly patients. Circulation 2013;128:1869–77.
in preventing re-bleeding for patients with peptic ulcer bleeding after successful 77 Eisenberg MJ, Richard PR, Libersan D, et al. Safety of short-term discontinuation
endoscopic therapy. BMC Gastroenterol 2012;12:66. of antiplatelet therapy in patients with drug-eluting stents. Circulation
53 Kim HK, Kim JS, Kim TH, et al. Effect of high-dose oral rabeprazole on recurrent 2009;119:1634–42.
bleeding after endoscopic treatment of bleeding peptic ulcers. Gastroenterol Res 78 Acosta RD, Abraham NS, Chandrasekhara V, et al. The management of
Pract 2012;2012:1–8. antithrombotic agents for patients undergoing GI endoscopy. Gastrointest Endosc
54 Fernando H, Bassler N, Habersberger J, et al. Randomized double-blind placebo- 2016;83:3–16.
controlled crossover study to determine the effects of esomeprazole on inhibition of 79 Shingina A, Barkun AN, Razzaghi A, et al. Systematic review: the presenting
platelet function by clopidogrel. J Thromb Haemost 2011;9:1582–9. international normalised ratio (INR) as a predictor of outcome in patients
55 Furuta T, Sugimoto M, Kodaira C, et al. Influence of low-dose proton pump inhibitors with upper nonvariceal gastrointestinal bleeding. Aliment Pharmacol Ther
administered concomitantly or separately on the anti-platelet function of clopidogrel. 2011;33:1010–8.
J Thromb Thrombolysis 2017;43:333–42. 80 Halvorsen S, Storey RF, Rocca B, et al. Management of antithrombotic therapy
56 Ohbuchi M, Noguchi K, Kawamura A, et al. Different effects of proton pump after bleeding in patients with coronary artery disease and/or atrial fibrillation:
inhibitors and famotidine on the clopidogrel metabolic activation by recombinant expert consensus paper of the European Society of Cardiology Working Group on
CYP2B6, CYP2C19 and CYP3A4. Xenobiotica 2012;42:633–40. Thrombosis. Eur Heart J 2017;38:1455–62.
57 Zvyaga T, Chang SY, Chen C, et al. Evaluation of six proton pump inhibitors as 81 Pollack CV, Reilly PA, van Ryn J, et al. Idarucizumab for dabigatran reversal — full
inhibitors of various human cytochromes P450: focus on cytochrome P450 2C19. cohort analysis. N Engl J Med Overseas Ed 2017;377:431–41.
Drug Metab Dispos 2012;40:1698–711. 82 Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification
58 Ogilvie BW, Yerino P, Kazmi F, et al. The proton pump inhibitor, omeprazole, schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
but not lansoprazole or pantoprazole, is a metabolism-dependent inhibitor of JAMA 2001;285:2864–70.
CYP2C19: implications for coadministration with clopidogrel. Drug Metab Dispos 83 Lip GY, Frison L, Halperin JL, et al. Comparative validation of a novel risk score
2011;39:2020–33. for predicting bleeding risk in anticoagulated patients with atrial fibrillation: the
59 Furuta T, Iwaki T, Umemura K. Influences of different proton pump inhibitors on the HAS-BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History
anti-platelet function of clopidogrel in relation to CYP2C19 genotypes. Br J Clin or Predisposition, Labile INR, Elderly, Drugs/Alcohol Concomitantly) score. J Am Coll
Pharmacol 2010;70:383–92. Cardiol 2011;57:173–80.
60 Zou JJ, Chen SL, Tan J, et al. Increased risk for developing major adverse 84 Zulkifly H, Lip GYH, Lane DA. Bleeding risk scores in atrial fibrillation and venous
cardiovascular events in stented Chinese patients treated with dual antiplatelet thromboembolism. Am J Cardiol 2017;120:1139–45.
Gut: first published as 10.1136/gutjnl-2018-316276 on 24 April 2018. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
87 Hernandez I, Zhang Y, Brooks MM, et al. Anticoagulation use and clinical outcomes Endosc 2011;73:900–8.
after major bleeding on dabigatran or warfarin in atrial fibrillation. Stroke 103 Laursen SB, Hansen JM, Andersen PE, et al. Supplementary arteriel embolization
2017;48:159–66. an option in high-risk ulcer bleeding--a randomized study. Scand J Gastroenterol
88 Gralnek IM, Ching JY, Maza I, et al. Capsule endoscopy in acute upper 2014;49:75–83.
gastrointestinal hemorrhage: a prospective cohort study. Endoscopy 2013;45:12–19. 104 Lau JY, Wong K, Chiu PW, et al. Early angiographic embolization after endoscopic
89 Sung JJ, Tang RS, Ching JY, et al. Use of capsule endoscopy in the emergency hemostasis to high risk bleeding peptic ulcers improves outcomes. Gastrointest
department as a triage of patients with GI bleeding. Gastrointest Endosc Endosc 2014;79:AB113.
2016;84:907–13. 105 Villanueva C, Balanzó J, Torras X, et al. Value of second-look endoscopy after
90 Gutkin E, Shalomov A, Hussain SA, et al. Pillcam ESO is more accurate than clinical injection therapy for bleeding peptic ulcer: a prospective and randomized trial.
scoring systems in risk stratifying emergency room patients with acute upper Gastrointest Endosc 1994;40:34–9.
gastrointestinal bleeding. Therap Adv Gastroenterol 2013;6:193–8. 106 Messmann H, Schaller P, Andus T, et al. Effect of programmed endoscopic follow-
91 Meltzer AC, Pinchbeck C, Burnett S, et al. Emergency physicians accurately interpret up examinations on the rebleeding rate of gastric or duodenal peptic ulcers
video capsule endoscopy findings in suspected upper gastrointestinal hemorrhage: a treated by injection therapy: a prospective, randomized controlled trial. Endoscopy
video survey. Acad Emerg Med 2013;20:711–5. 1998;30:583–9.
92 Meltzer AC, Ward MJ, Gralnek IM, et al. The cost-effectiveness analysis of 107 Saeed ZA, Cole RA, Ramirez FC, et al. Endoscopic retreatment after successful initial
video capsule endoscopy compared to other strategies to manage acute upper hemostasis prevents ulcer rebleeding: a prospective randomized trial. Endoscopy
gastrointestinal hemorrhage in the ED. Am J Emerg Med 2014;32:823–32. 1996;28:288–94.
93 Chandran S, Testro A, Urquhart P, et al. Risk stratification of upper GI bleeding with 108 Chiu PW, Lam CY, Lee SW, et al. Effect of scheduled second therapeutic endoscopy
an esophageal capsule. Gastrointest Endosc 2013;77:891–8. on peptic ulcer rebleeding: a prospective randomised trial. Gut 2003;52:1403–7.
94 Daneshmend TK, Hawkey CJ, Langman MJ, et al. Omeprazole versus placebo for 109 El Ouali S, Barkun AN, Wyse J, et al. Is routine second-look endoscopy effective after
acute upper gastrointestinal bleeding: randomised double blind controlled trial. BMJ endoscopic hemostasis in acute peptic ulcer bleeding? A meta-analysis. Gastrointest
1992;304:143–7. Endosc 2012;76:283–92.
95 Hawkey GM, Cole AT, McIntyre AS, et al. Drug treatments in upper gastrointestinal 110 Chiu PW, Joeng HK, Choi CL, et al. High-dose omeprazole infusion compared with
bleeding: value of endoscopic findings as surrogate end points. Gut 2001;49:372–9. scheduled second-look endoscopy for prevention of peptic ulcer rebleeding: a
96 Hulagu S, Demirturk L, Gul S, et al. The effect of omprazole or ranitidine intravenous randomized controlled trial. Endoscopy 2016;48:717–22.
on upper gastrointestinal bleeding. Endoskop J 1995;2:35–43. 111 Saeed ZA, Ramirez FC, Hepps KS, et al. Prospective validation of the Baylor bleeding
97 Naumovski-Mihalic S, Katicic M, Colic-Cvrlje V, et al. Intravenous proton score for predicting the likelihood of rebleeding after endoscopic hemostasis of
pump inhibitor in ulcer bleeding in patients admitted to intensive care unit. peptic ulcers. Gastrointest Endosc 1995;41:561–5.
Gastroenterology 2005;128:A641. 112 Kim SB, Lee SH, Kim KO, et al. Risk factors associated with rebleeding in patients
98 Wallner G, Ciechanski A, Wesolowski M, et al. Treatment of acute upper with high risk peptic ulcer bleeding: focusing on the role of second look endoscopy.
gastrointestinal bleeding with intravenous omeprazole or ranitidine. Eur J Med Res Dig Dis Sci 2016;61:517–22.
1996;8:235–43. 113 Cheng HC, Wu CT, Chen WY, et al. Risk factors determining the need for second-look
99 Lau JY, Leung WK, Wu JCY, et al. Omeprazole before endoscopy in patients with endoscopy for peptic ulcer bleeding after endoscopic hemostasis and proton pump
gastrointestinal bleeding. N Engl J Med Overseas Ed 2007;356:1631–40. inhibitor infusion. Endosc Int Open 2016;4:E255–62.
100 Sreedharan A, Martin J, Leontiadis GI, et al. Proton pump inhibitor treatment 114 Jensen DM, Kovacs TOG, Ohning GV, et al. Doppler endoscopic probe monitoring
initiated prior to endoscopic diagnosis in upper gastrointestinal bleeding. Cochrane of blood flow improves risk stratification and outcomes of patients with severe
Database Syst Rev 2010;304:CD005415. nonvariceal upper gastrointestinal hemorrhage. Gastroenterology 2017;152:1310–8.
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
Adrian J Stanley,1 Loren Laine2
1
Department of Gastroenterology,
Glasgow Royal Infirmary, Glasgow A B S T RAC T
G4 OSF, UK
2
Section of Digestive Diseases, Yale
Upper gastrointestinal bleeding (UGIB) is a common medical emergency, with a reported
School of Medicine, New Haven, mortality of 2-10%. Patients identified as being at very low risk of either needing an
and VA Connecticut Healthcare
System, West Haven, Connecticut, intervention or death can be managed as outpatients. For all other patients, intravenous
CT 06520, USA
Correspondence to: A J Stanley fluids as needed for resuscitation and red cell transfusion at a hemoglobin threshold of 70-80
adrian.stanley@ggc.scot.nhs.uk
and adrianstanley99@gmail.com
g/L are recommended. After resuscitation is initiated, proton pump inhibitors (PPIs) and the
Cite this as: BMJ 2019;364:l536 prokinetic agent erythromycin may be administered, with antibiotics and vasoactive drugs
doi: 10.1136/bmj.l536
recommended in patients who have cirrhosis. Endoscopy should be undertaken within 24
Series explanation: State of the
Art Reviews are commissioned hours, with earlier endoscopy considered after resuscitation in patients at high risk, such as
on the basis of their relevance to
academics and specialists in the US
those with hemodynamic instability. Endoscopic treatment is used for variceal bleeding (for
and internationally. For this reason example, ligation for esophageal varices and tissue glue for gastric varices) and for high risk
they are written predominantly by
US authors non-variceal bleeding (for example, injection, thermal probes, or clips for lesions with active
bleeding or non-bleeding visible vessel). Patients who require endoscopic therapy for ulcer
bleeding should receive high dose proton pump inhibitors after endoscopy, whereas those
who have variceal bleeding should continue taking antibiotics and vasoactive drugs. Recurrent
ulcer bleeding is treated with repeat endoscopic therapy, with subsequent bleeding managed
by interventional radiology or surgery. Recurrent variceal bleeding is generally treated with
transjugular intrahepatic portosystemic shunt. In patients who require antithrombotic agents,
outcomes appear to be better when these drugs are reintroduced early
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
were low, in the range of 1.9-2.5%.12‑14 By contrast, large
• Blood transfusion at a hemoglobin threshold of 70-80 g/L; higher threshold if severe observational cohort studies from Europe suggest higher
bleeding with hypotension
fatality rates of around 10%.15 16 The reason for these
• Risk assessment:
differences is unknown but might be partly related to
-If Glasgow-Blatchford score ≤1 consider outpatient endoscopy and management
reliance on coding in database studies and differences
• Erythromycin (as a prokinetic agent) and proton pump inhibitor may be considered
in practice, such as low risk patients being more often
• Patients with cirrhosis should receive vasoactive drugs and antibiotics
managed in outpatient settings in Europe.
Endoscopic
• Endoscopy is generally recommended within 24 hours in patients admitted to hospital
Sources and selection criteria
––If the patient has severe bleeding with hemodynamic instability, urgent endoscopy We searched PubMed, Medline, and Cochrane databases
should be performed after resuscitation
from 2010 to August 2018 using the search terms gastro-
• Ulcers with active bleeding and non-bleeding visible vessels should receive endoscopic
intestinal hemorrhage, peptic ulcer bleeding, and variceal
therapy; endoscopic therapy may also be used for ulcers with adherent clots
bleeding. References were also identified from the inter-
• Injection therapy (eg, epinephrine), thermal probes (eg, bipolar electrocoagulation, heater
probe), or clips should be used national, UK, European, American, and Asia-Pacific
• Epinephrine injection should always be followed by a second modality guidelines on upper gastrointestinal bleeding published
• Recurrent bleeding should be treated with repeat endoscopic therapy but subsequent during this period in addition to relevant review articles.
bleeding by transarterial embolization or surgery We selected systematic reviews, meta-analyses, RCTs, and
• Esophageal variceal bleeding should be treated with ligation and gastric varices with the observational studies (excluding case reports and small
injection of tissue adhesive (<15 cases) case series). We also excluded articles that
• Refractory variceal bleeding should be treated with transjugular portosystemic shunt were not peer reviewed and those not published in Eng-
• For massive refractory esophageal variceal bleeding a removable covered metal stent is lish. Studies were prioritized by design, as noted above,
preferred to balloon tamponade as a temporizing measure and by patient numbers, quality, and publication date.
Post-endoscopic management
• Patients who have ulcers with high risk lesions (active bleeding, visible vessel, adherent clot) Initial resuscitation
should receive high dose proton pump inhibitors for 72 h As with any new patient with a medical emergency, the
• Patients with cirrhosis should continue antibiotics for up to seven days regardless of the initial clinical evaluation of patients presenting with
bleeding source upper gastrointestinal bleeding involves assessment of
• Variceal bleeding should be treated with vasoactive drugs for up to five days the patient’s airway, breathing, and circulation. Many
• When used for secondary prevention, aspirin should be continued or reintroduced soon after patients are hemodynamically stable at presentation,
hemostasis is achieved but for those with major bleeding, early resuscitation is
• Early reintroduction of other antithrombotic drugs is also recommended after hemostasis is essential. In general, two large bore intravenous cannu-
achieved to reduce thrombotic events and death lae are inserted, although central venous access may be
preferred in certain cases. Regular monitoring of pulse,
Overall quality of evidence blood pressure, and oxygen saturations is crucial. Hypo-
Numerous randomized controlled trials (RCTs) and tension is associated with increased mortality; a multi-
meta-analyses have assessed the use of medical and center observational study of 1882 patients reported an
endoscopic therapy, and the optimal blood transfusion odds ratio of 9.8 (95% confidence interval 5.1 to 19) with
strategy in patients with acute upper gastrointestinal systolic blood pressure <90 mm Hg versus ≥90 mm Hg.11
bleeding, thereby providing high quality data to guide Tracheal intubation may be used to protect the airway in
management. Although the evidence regarding resuscita- patients with severe ongoing hematemesis, especially in
tion, risk assessment, timing of endoscopy, and reintro- those at increased risk of aspiration (such as those with
duction of antithrombotic drugs is of lower quality, large an altered mental status or lack of gag reflex).
recent studies in these areas have helped inform patient No RCTs have assessed fluid resuscitation in upper gas-
management. trointestinal bleeding. By contrast, a comparative study
and an RCT in patients with hemorrhagic shock as a result
Incidence of trauma suggest that a more restrictive fluid resuscita-
The incidence of upper gastrointestinal bleeding in the tion may be better (or not worse) than more intensive fluid
United Kingdom in the 1990s was 103-172/100 000 resuscitation.17 18 The choice of intravenous fluid for ini-
adults per year.10 11 Recent reports from the United States tial resuscitation is unclear, with crystalloids or colloids
using nationwide administrative databases indicate that often being used while the need for the transfusion of
the incidence of hospital admission for the condition blood products is assessed. A meta-analysis of 70 trials
was 61-78 per 100 000 persons in 2009-2012.12‑14 Peptic with 22 392 patients found no difference in mortality
ulcers are the most common cause of hospital admission between colloid and crystalloid solutions for fluid resus-
for upper gastrointestinal bleeding, accounting for just citation in critically ill patients: relative risk 1.01 (0.93 to
over half of all cases.12 14 The incidence of hospital admis- 1.10) for albumin or plasma proteins versus crystalloid
sion for the condition has decreased 21-23% during the solutions, and similar negative results when other col-
past 10 years.12 14 This decrease is largely accounted loids were compared with crystalloids.19 An RCT of 15
for by decreases in peptic ulcer bleeding (and bleeding 802 critically ill hospital inpatients found reduced acute
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
-DLUDWK WR
9LOODQXHYD WR
7RWDOO 3 WR
)DYRUVUHVWULFWLYH )DYRUVOLEHUDO
)DYRUVUHVWULFWLYH )DYRUVOLEHUDO
Fig 1 | Blood transfusion meta-analysis: liberal versus restrictive transfusion for (A) mortality and (B) rebleeding.27 Reproduced
with permission from Elsevier. Abbreviations: CI=confidence interval; RR=relative risk.
kidney injury (odds ratio 0.91, 0.84 to 0.99) and a trend old of 70 g/L versus 90 g/L.25 A subsequent six center
towards reduced mortality in hospital (10.3% v 11.1%; cluster randomized feasibility trial in the UK reported no
P=0.08) with balanced crystalloids versus saline.20 benefit from a liberal transfusion policy when hemoglobin
Whether these data can be fully extrapolated to upper thresholds of 80 g/L versus 100 g/L were compared (mor-
gastrointestinal bleeding is uncertain. tality difference −1%, −8% to 6% in 640 patients).26 A
meta-analysis of five RCTs comprising 1965 patients with
Red blood cell transfusion upper gastrointestinal bleeding reported that restrictive
A meta-analysis of 31 RCTs comprising 12 587 patients transfusion was associated with lower mortality (relative
in a variety of populations found that a more restrictive risk 0.65, 0.44 to 0.97) and reduced rebleeding (0.58,
approach to red cell transfusion (variably defined at 0.40 to 0.84) (fig 1).27
hemoglobin threshold 70-90 g/L) does not adversely affect Thus, in most patients with upper gastrointestinal
outcomes; hospital mortality was lower with a restrictive bleeding, red cell transfusion should be withheld until
strategy but 30 day mortality was not significantly differ- a hemoglobin threshold of 70-80 g/L is reached.5 6 28
ent (risk ratio 0.97, 0.81 to 1.16).21 This systematic review Patients with severe bleeding and hemodynamic com-
concluded that a restrictive policy seemed to be safe in promise, who were generally excluded from the trials
patients with underlying cardiovascular disease but no described above, require transfusion at higher thresholds
evidence was available for patients presenting with acute because their hemoglobin will equilibrate to much lower
coronary syndrome. On that basis, current US guidelines levels as their intravascular volume is repleted with fluid.
recommend transfusion at a threshold of hemoglobin of Transfusion thresholds in patients with cardiovascular
70 g/L for hemodynamically stable adult inpatients and disease, especially those with acute coronary syndrome,
80 g/L for those undergoing orthopedic or cardiac surgery are less certain, but thresholds of 80 g/L or higher are vari-
or with pre-existing cardiovascular disease.22 Others have ably recommended.22 23
suggested a threshold in patients with cardiovascular dis-
ease of ≥80 g/L.23 Risk assessment
Importantly, results for the general populations Many risk assessment scores have been developed for
described above may not be applicable to those with upper patients with upper gastrointestinal bleeding, includ-
gastrointestinal bleeding. In such patients a restrictive ing those that can be calculated early after presentation
transfusion approach appears not only to be safe but also (pre-endoscopy) and those that include endoscopic find-
to provide clinical benefit for rebleeding and mortality. A ings.29‑34 They were designed to predict a variety of end-
small study in 1986 first showed reduced rebleeding with points. We believe pre-endoscopy scores are of greater
restrictive transfusion.24 A large high quality Spanish RCT practical use because it is probably most important to
in 921 patients found significantly lower mortality at six predict risk soon after presentation to help direct man-
weeks (hazard ratio 0.55, 0.33 to 0.92) and rebleeding agement. The use of risk scores has been recommended to
(0.68, 0.47 to 0.98) with a transfusion hemoglobin thresh- stratify patients into those at higher or lower risk of poor
no patients who may come to harm are sent home.
A systematic review of 16 studies on the use of pre-
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
endoscopy scores in emergency departments to predict
intervention, rebleeding, or death concluded that a GBS
of zero provided the highest sensitivity (0.99), although
specificity was very low (0.08).39 In 2012, both US and UK
guidelines recommended that a GBS of zero could be used
to identify very low risk patients who could avoid admis-
sion and be offered outpatient endoscopy.4 28
Subsequently two large international comparative
studies of risk assessment scores, with 3012 and 2305
patients, were published.35 40 These studies reported that
GBS ≤1 was the optimum low risk threshold, with a sen-
sitivity of 99% and specificity of 35-40%. The authors
suggested that this threshold could be used to identify
patients who could be safely discharged from the emer-
gency department for outpatient management, thereby
avoiding admission in 19-24% of patients presenting
with upper gastrointestinal bleeding. This approach has
been accepted by recent European and Asia-Pacific guide-
lines.5 6 No high quality interventional trial has assessed
outcomes after the introduction of a risk score, although
a pre-post design study showed the safety of outpatient
management in 84 patients with a GBS score of zero.38
VSHFLƟFLW\
to fresh frozen plasma because of the smaller volume Pre-endoscopic medical therapy
needed, its more rapid onset, lack of the need to check The use of pre-endoscopic intravenous proton pump
the patient’s blood group, and the minimal infectious inhibitors (PPIs) has been assessed in several studies. A
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
risk.5 US guidelines suggest either four factor PCC plus meta-analysis of six RCTs comprising 2223 patients found
vitamin K or fresh frozen plasma.45 that the use of these drugs before endoscopy is associated
Guidelines also suggest that the international normal- with both reduced high risk stigmata of bleeding and the
ized ratio (INR) should be corrected to <2.5 if possible need for endoscopic therapy (odds ratio 0.68, 0.50 to
before undertaking endoscopy, with the potential need 0.93) but has no effect on patient outcomes, including
for endoscopic therapy if the clinical situation allows.5 45 rebleeding, need for surgery, or mortality (1.12, 0.72 to
This suggestion is based on observational studies which 1.73).56 As a result, UK National Institute for Health and
indicate that the outcome after endoscopic therapy is Care Excellence (NICE) guidelines do not support the
similar in patients with an INR of 1.3-2.7 to that in those routine use of these drugs before endoscopy.4 However,
not taking warfarin.48 49 Other studies report that the INR several other international guidelines suggest PPIs may
value does not predict rebleeding.50 have a role before endoscopy, particularly for patients in
Data on DOACs are limited, but because of their short whom endoscopy may be delayed. 3 5 6 28 57
half lives (5-17 h), anticoagulant activity wanes rapidly Prokinetic agents have been assessed for their ability
over one to two days (in the absence of renal disease). to improve gastric emptying, thereby improving visu-
Thus, European guidelines state that “time is the most alization at endoscopy. Erythromycin, usually given as a
important antidote against DOACs.”5 Although PCC may 250 mg infusion 30-120 minutes before endoscopy, has
be of some use in severe bleeding, particularly for Xa been most widely studied. The most recent meta-analysis
inhibitors, neither vitamin K nor fresh frozen plasma of 598 patients in eight RCTs showed improved visuali-
has been shown to be beneficial.5 51 Reversal agents for zation, reduced need for second look endoscopy, and
dabigatran (idarucizumab)52 and the factor Xa inhibitors reduced length of hospital stay (mean difference −1.75
(andexanet alfa)53 are now approved in the US. Their role days, 2.43 to −1.06) after erythromycin infusion before
in patients with upper gastrointestinal bleeding is unclear endoscopy.58
given the uncertain risk of thrombotic events and the Tranexamic acid (TXA) inhibits the fibrinolytic activity
short half lives of DOACS. They would mainly be used in of plasmin. A meta-analysis reported reduced mortality
patients with a severe ongoing bleed, especially if DOAC with TXA in patients with upper gastrointestinal bleed-
ingestion was recent or if renal disease was present. ing, but many studies were of poor quality and had been
undertaken before the widespread use of PPIs and endo-
Coagulopathy and thrombocytopenia in patients with scopic therapy.59 Furthermore, other outcomes such as
cirrhosis bleeding episodes and transfusions were not reduced.
Interpretation of the complex clotting abnormalities seen Therefore, it is difficult to draw firm conclusions from
in cirrhosis can be difficult. Patients with cirrhosis have these data. A meta-analysis of two RCTs comprising
parallel decreases in procoagulant and anticoagulant 40 138 patients with acute severe traumatic or postpar-
factors.54 The prothrombin time measures procoagulant tum bleeding suggested even a short delay in the admin-
activity only; therefore, prothrombin time or INR is not a istration of TXA reduces benefit: immediate treatment
reliable indicator of coagulation status in patients with improved survival (odds ratio 1.72, 1.42 to 2.10), but
cirrhosis.54 Fresh frozen plasma is often given to patients survival fell 10% with every 15 minutes delay, with no
with upper gastrointestinal bleeding, cirrhosis, and benefit beyond three hours.60 A large international study
raised prothrombin time, but it has not been shown to (HALT-IT) will finish recruitment shortly and the results
provide benefit and could have adverse effects. For these should help clarify the role of TXA in upper gastrointes-
reasons, the most recent US guidelines on portal hyper- tinal bleeding.61
tensive bleeding recommend against correcting INR with
fresh frozen plasma or recombinant factor VIIa in patients Pre-endoscopic medical therapy in patients with cirrhosis
with cirrhosis and acute variceal bleeding.9 Vasoactive drugs (terlipressin, somatostatin, or its ana-
Platelets from patients with cirrhosis generate throm- logs octreotide and vapreotide), which cause splanch-
bin in a similar way to those from healthy controls, nic artery vasoconstriction, are used in patients with
and patients with cirrhosis have pro-hemostatic fac- cirrhosis and variceal bleeding. When combined with
tors (increased von-Willebrand factor and decreased endoscopic therapy, the different vasoactive drugs seem
ADAMTS-13, a protease that cleaves von-Willebrand to have similar efficacy.62 Three double blind RCTs exam-
factor).54 55 However, patients with cirrhosis often have ined the use of vasoactive drugs (terlipressin, somatosta-
thrombocytopenia as a result of splenic sequestra- tin, and vapreotide) given before endoscopy in patients
tion. The experimental finding that a platelet count of with cirrhosis and upper gastrointestinal bleeding.63‑65
56×109/L leads to thrombin generation at the 10th centile Two studies reported less active bleeding at endoscopy
of healthy control values54 provides the basis for giving in the active treatment group,64 65 and the third noted
platelet transfusions at around 50×109/L.4 7 However, no significantly more control of bleeding (clear gastric lav-
studies have assessed platelet thresholds or results with age and stable hemoglobin) at 12 hours with vasoactive
platelet transfusions, and some current guidelines state drug therapy.63
that no recommendation can be made regarding platelet Current guidelines recommend starting vasoactive
transfusions.8 9 drugs as soon as variceal hemorrhage is suspected.7‑9
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
in patients with melena. A technetium-99m labelled red
cell scan may also be used in this situation, but computed
tomography angiography seems to be more accurate.70 71
Early use of capsule endoscopy has been reported, with
3 goals including stratifying risk and determining the timing
of endoscopy,72 but more studies are needed to establish
any potential role.
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
ity (such as a thermal or mechanical) or to decrease the
risk of inducing bleeding with the application of a second
modality.
Most data on non-variceal upper gastrointestinal bleed-
ing are from patients with peptic ulcer bleeding. The For-
rest classification of endoscopic stigmata is commonly
used by endoscopists to identify higher risk lesions that
require the application of endoscopic therapy.84 Endo-
scopic therapy significantly decreases further bleeding
and the need for urgent intervention in patients with
ulcers with spurting or oozing blood (Forrest 1a or 1b) or
with non-bleeding visible vessels (Forrest 2a; fig 4A).28
Patients with adherent clots (Forrest 2b) were not shown
to benefit from endoscopic therapy in a meta-analysis of
RCTs, but results of individual trials were very heteroge-
neous.80 This has led to guideline recommendations that
either endoscopic plus medical therapy or medical man-
agement alone may be used for patients with adherent
clots.5 28 Around 25-50% of patients admitted to hospital
with bleeding ulcers have Forrest 1a, 1b, 2a, or 2b stig-
mata.76 85 Endoscopic treatment is not needed for ulcers
with flat pigmented spots or a clean base (Forrest 2c or
3).5 28
Variceal bleeding
Variceal bleeding accounted for 11% of patients admit-
Fig 4 | (A) Endoscopic view of a large posterior duodenal ulcer with intermittent bleeding from
ted to hospital with acute upper gastrointestinal bleed-
a visible vessel. The patient, a middle aged man taking anticoagulant drugs, was admitted
with hematemesis, hemodynamic instability, and a hemoglobin concentration of 55 g/L.
ing in a nationwide UK audit.15 However, the proportion
After resuscitation, transfusion to hemoglobin 70-80 g/L, and correction of coagulopathy, of patients with variceal bleeding varies widely and is
endoscopy was undertaken. (B) Through-the-scope clips were applied after dilute epinephrine related to the proportion of people with liver disease in
was injected into the four quadrants of the ulcer base. The fibrotic base made application of the the population served. Patients with variceal bleeding
clips problematic. (C) There was ongoing intermittent oozing of blood. Given the high risk ulcer, have a higher mortality than those with non-variceal
hemostatic powder spray was then applied to good effect. High dose intravenous proton pump bleeding, and this is largely related to the severity of
inhibitors (PPIs) were given and the patient was managed in the hospital high dependency unit. underlying liver disease.14 86
Because of the clinical situation and the difficulty in providing endoscopic therapy to this large
The optimal endoscopic therapy for esophageal variceal
fibrotic ulcer, the plan for urgent referral for radiological embolization—should early rebleeding
occur—was clearly documented by the endoscopist as a “rebleeding plan.” (D) Fifteen hours bleeding is variceal band ligation, which is associated
later the patient rebled and became hemodynamically unstable. He was again resuscitated with less rebleeding and fewer side effects than sclero-
appropriately, after which an interventional radiologist performed coil embolization of the therapy.4 7 9 87 If gastric varices are found, ligation can be
gastroduodenal artery. The patient had no further bleeding and was restarted on anticoagulants used for gastroesophageal varices type-1, where esopha-
on day 3. When he was discharged from hospital a week later he was still taking oral PPIs twice geal varices extend several centimeters distally along
daily, but when the 14 day course was finished, the dose was reduced to once daily. the gastric lesser curve. Injection of tissue adhesive (eg,
N-butyl-cyanoacrylate) is the recommended endoscopic
of sclerosants such as absolute ethanol, thermal contact approach for all other types of gastric varices, although
devices such as bipolar electrocoagulation probes or thrombin injection can be considered.7 9 88-90 Thrombin
heater probes, and through-the-scope clips.5 28 RCT data injection has been described for gastric variceal bleeding
supporting efficacy in ulcer bleeding are more limited in cohort studies, but to date no RCTs have compared it
for non-contact thermal devices such as argon plasma with other treatments.91‑93
coagulation.28 Vascular ectasias may also be treated
with thermal methods, commonly argon plasma coagu- Post-endoscopic management
lation; radiofrequency ablation is another thermal con- PPI therapy
tact modality sometimes used for gastric antral vascular A meta-analysis of RCTs comparing PPIs to placebo or no
ectasia.81 therapy in high risk patients undergoing successful endo-
Endoscopic injection of epinephrine should not be scopic therapy showed that high dose PPIs, usually given
used as a single modality treatment. Meta-analyses have as an intravenous bolus of 80 mg followed by continu-
reported lower rates of further bleeding with an alterna- ous infusion at 8 mg/h for 72 hours, reduced rebleeding
tive modality compared with epinephrine alone (relative (relative risk 0.40, 0.28 to 0.59) and mortality (0.41, 0.20
risk 0.58, 0.36 to 0.93) and with epinephrine combined to 0.84). Intermittent intravenous or oral PPIs reduced
with a second modality versus epinephrine alone (relative rebleeding (0.53, 0.35 to 0.78) but not mortality.82 A
meta-analysis of 13 studies found that a bolus followed suggest that the benefit is primarily in those with Child-
by intermittent doses of intravenous or oral PPIs was non- Pugh C disease.102 However the evidence for early TIPS
inferior to continuous infusion (further bleeding risk ratio remains relatively limited and the practicalities may be
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
0.72; one sided 95% confidence interval upper boundary challenging for many units.
0.97), although most individual studies were relatively
small and not designed to answer this question.94 No Reintroduction of antithrombotic drugs
conclusions could be made regarding oral versus intra- Several studies suggest a survival benefit from continuing
venous dosing,94 although oral administration provides or reintroducing antithrombotic drugs after upper gastroin-
an antisecretory effect comparable to equivalent doses testinal bleeding.103 104 This is perhaps unsurprising given
of intravenous PPIs.95 Guidelines have recommended an that mortality after presentation with a bleed is more often
intravenous bolus followed by continuous infusion PPI caused by underlying comorbidities, particularly cardio-
therapy,3 5 6 28 although recent guidelines also suggest con- vascular disease, rather than the bleed itself.77 However,
sidering intermittent high doses of oral or intravenous PPI balancing the risks and benefits of reintroducing these
(eg, 80-160 mg daily in divided doses after an initial 80 drugs after a patient presents with upper gastrointestinal
mg bolus), rather than continuous infusion.5 6 bleeding can be challenging. If an antithrombotic drug is
Patients with peptic ulcer bleeding generally receive reintroduced, a PPI is usually also administered.
four to eight weeks of once daily oral PPIs. Those with Aspirin is the most widely studied antithrombotic drug
low risk endoscopic lesions (clean base, flat spot) should in patients with upper gastrointestinal bleeding. Strati-
receive PPIs once a day from the time of diagnosis. Those fication of patients must be based on whether aspirin is
with high risk endoscopic lesions and clinical features given for secondary or primary cardiovascular preven-
should receive high dose PPIs on days one to three as tion. This is because the benefit in secondary prevention
above, followed by twice daily oral PPI on days four to is far greater than that for primary prevention, with a
14. This regimen is based on an RCT of 187 patients that number needed to treat to prevent myocardial infarction,
showed significantly less rebleeding with twice daily ver- stroke, or vascular death of 67 versus 1745.80
sus once daily PPIs (relative risk 0.41, 0.18 to 0.93) during A randomized study of 156 patients with peptic ulcer
this period.96 bleeding who had been taking aspirin for secondary
The benefits of PPIs outweigh potential risks when prevention reported reduced mortality at eight weeks in
used after bleeding from a peptic ulcer. Multiple phar- those who continued aspirin compared with those who
macodynamics studies report that omeprazole reduces discontinued the drug (1.3% v 12.9%; difference 11.6%,
the antiplatelet effect of clopidogrel, but a double blind 3.7% to 19.5%).103 Therefore, current guidelines suggest
placebo controlled trial in 3761 clopidogrel users found continuing aspirin (or reintroducing the drug within three
no evidence that omeprazole increased cardiovascular days for higher risk endoscopic lesions) once hemosta-
events (hazard ratio 0.99, 0.68 to 1.44).97 The US Food sis is achieved.4-6 105 106 It has been suggested that when
and Drug Administration recommends avoiding the use aspirin has been prescribed for primary prophylaxis, it
of omeprazole or esomeprazole in patients who are taking should be stopped in most patients because the bleed-
clopidogrel.98 ing risk probably outweighs the cardiovascular benefit.28
If primary prevention is still clinically required after the
Patients with cirrhosis and variceal bleeding bleed, it can be reintroduced after the ulcer has healed,
As noted above, in patients with cirrhosis and upper gas- or earlier depending on the clinical situation.5
trointestinal bleeding, antibiotics should be continued No randomized studies are available to guide clinicians
for up to seven days,7‑9 regardless of whether varices are on the reintroduction of thienopyridines (eg, clopidogrel)
identified as the source of the bleeding. Patients who have or anticoagulants. Recent guidelines suggest that for
documented variceal bleeding at endoscopy should also patients receiving dual antiplatelet therapy, at least one
have their vasoactive drugs continued for up to five days.7‑9 drug, usually aspirin, should be reintroduced early as
Combined treatment with endoscopic ligation and vasoac- above, with the second drug withheld for up to five days
tive drugs is superior to ligation alone or vasoactive drugs after hemostasis, or the timing discussed with a cardio-
alone in reducing further bleeding in hospital or during vascular specialist.5 6 105
the first seven days after treatment.99 100 Similar to the situation with antiplatelet agents, obser-
Transjugular intrahepatic portosystemic shunt (TIPS) vational studies in patients who develop upper gastro-
may also be used after initial endoscopic therapy in the intestinal bleeding while taking warfarin indicate that
first three days after presentation for the treatment of acute those who restart warfarin have markedly lower rates of
esophageal variceal bleeding in patients with Child-Pugh death and thromboembolic events, without a higher rate
class C cirrhosis (score 10-13)8 9 A multicenter RCT com- of recurrent bleeding, when compared with those whose
pared early (within 72 h) TIPS placement versus standard warfarin is not restarted.104 107 Recent guidelines suggest
treatment with variceal ligation plus drug therapy in 63 restarting warfarin from “as soon as hemostasis is estab-
patients with Child-Pugh C cirrhosis or Child-Pugh B cir- lished”6 to seven to 15 days after the bleeding event.5 The
rhosis with active bleeding. It reported that more patients indication for anticoagulation should be assessed at the
were free from further bleeding at one year with early time of the bleed, with early reintroduction (zero to seven
TIPS (97% v 50%; P<0.001).101 One year survival was days) recommended in patients with a higher thrombo-
also higher with early TIPS (86% v 61%; P<0.001) and embolic risk.5 6 However, robust data on the optimal tim-
encephalopathy was not increased. Subsequent reports ing of reintroduction are not available.
Data on the timing of the reintroduction of DOACs balloon tamponade has been recommended as a tempo-
after bleeding has been controlled are limited, and this rary bridge to definitive therapy.7 8 However, balloon tam-
clearly depends on the balance of risk between rebleeding ponade is associated with serious complications (such as
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
and thromboembolic events. Use of the CHA2DS2-VASC esophageal rupture and aspiration pneumonia) in about
(Congestive heart failure, Hypertension, Age-2, Diabetes, 12% of patients and its use was lethal in 6% of patients
Stroke/TIA-2, Vascular disease) and HAS-BLED (Hyper- in a case series.117 Recently, removable self expanding
tension, Abnormal renal and liver function, Stroke, prior covered metal esophageal stents designed for the treat-
Bleeding, Labile INRs, Elderly (>65 years), Drugs and ment of severe esophageal variceal bleeding have become
alcohol) scores may help in this situation.108‑110 The recent available (although they are not approved in the US). An
Asia-Pacific guidelines suggest reintroducing DOACs in RCT of 28 patients compared these stents with balloon
patients with a high thromboembolic risk as soon as tamponade in patients with esophageal variceal bleed-
hemostasis is achieved, although others have suggested ing refractory to medical and endoscopic therapy and
that patients should have their anticoagulant restarted at showed that the stents led to improved bleeding control
day 7, with possible bridging therapy with low molecu- (85% v 47%; Fisher’s exact test P=0.055), with similar
lar weight heparin from days 2 to 7 in those with a low mortality.118
bleeding risk.6 In general, patients with an increased
thromboembolic risk should have early reintroduction Emerging treatments
of antithrombotic drugs, because the risks and severity Hemostatic powder spray
of thromboembolic events generally outweigh those of Hemostatic powder spray provides high rates of initial
bleeding events. hemostasis for active non-variceal upper gastrointestinal
bleeding, but with relatively high rebleeding rates, sug-
Management of persistent or recurrent bleeding, including gesting a temporary effect.119‑123 A systematic review of
role of interventional radiology and surgery 195 cases reported initial hemostasis in 92% and a seven
Non-variceal bleeding day rebleeding rate of 21%.112 Currently, hemostatic
The results of an RCT comparing repeat endoscopic ther- powder is often used as a temporary rescue treatment
apy with surgery in 92 patients with recurrent peptic for bleeding that cannot be controlled using established
ulcer bleeding indicate that endoscopic therapy should methods, and it may have a role for the initial control
be repeated when bleeding recurs after initial endoscopic of diffuse bleeding from tumors.6 112 More limited data
control.111 This study reported similar mortality, although are available on hemostatic powder spray for variceal
more patients had complications with surgery (36.4% v bleeding, often as a temporizing method until definitive
14.6%; P=0.03). For patients with persistent or refrac- therapy is applied.124
tory bleeding from non-variceal sources despite optimal
standard endoscopic and medical therapy, the addition Over-the-scope clips
of hemostatic powder spray for temporary control (12-24 Treatment with OTSC is another relatively new technique.
h) or over-the-scope clips (OTSC) as a rescue modality is These clips are much larger than standard through-the-
suggested, in parallel with ongoing resuscitation.5 6 112 113 scope clips so may be successful when applied to larger
For peptic ulcer bleeding not controlled by endoscopic fibrotic lesions or larger feeding vessels. A multicenter
therapy, two recent meta-analyses of observational stud- RCT of 66 patients with recurrent ulcer bleeding after ini-
ies that compared surgery with radiological interven- tial hemostasis showed that significantly fewer patients
tion reported lower rebleeding with surgery, but similar treated with OTSC had further bleeding compared with
mortality and need for further interventions, although those on standard therapy (15% v 58%; difference 42%,
patients receiving radiological intervention were older 22% to 63%).125 Further data are awaited, but at present
and in worse general health.114 115 Because many patients OTSC may be considered as a rescue therapy when stand-
with recurrent bleeding are elderly with comorbidities, ard therapies do not achieve permanent hemostasis.
interventional radiology is generally preferred if locally
available. Therefore, if bleeding continues despite opti- Doppler probes
mal endoscopic therapy, transarterial embolization is Endoscopic Doppler probes have also been studied as a
recommended, although surgery should be considered if guide to endoscopic therapy. A recent dual center RCT
radiological therapy is likely to be delayed.4 5 of 148 patients with severe non-variceal upper gastro-
Prophylactic transarterial embolization of high risk intestinal bleeding that compared endoscopic therapy
ulcers after endoscopic therapy is not recommended: it guided by Doppler probe positive signals with standard
did not significantly reduce rebleeding when compared endoscopic therapy guided by endoscopic stigmata found
with standard treatment (10.2% v 11.4%) in an RCT of reduced 30 day rebleeding with Doppler guided therapy
241 patients.116 (11.1% v 26.3%; P=0.02).126 Further data on the clinical
utility and practicalities of this approach are needed, and
Variceal bleeding currently the use of these probes is not recommended by
Patients with rebleeding after initial endoscopic and guidelines.5 6
medical therapy for varices may have repeat endoscopic
therapy performed.8 TIPS is recommended for those with Guidelines
persistent or severe recurrent bleeding.7‑9 For patients As already noted, international, UK (NICE), American,
with severe bleeding refractory to endoscopic therapy, European, and Asia-Pacific guidelines on upper gastro-
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
20103 identify high and low risk groups should not be used delay endoscopy high risk stigmata and have had
routinely successful endoscopic therapy
US, ACG (ulcer Carry out risk assessment to stratify Consider IV IV PPIs may be considered Within 24 h after resuscitation. After successful endoscopic
bleeding only) into higher and lower risk groups. erythromycin Consider within 12 h if high risk hemostasis, give IV PPI bolus
201228 Consider discharge from ED if GBS=0 features (eg, hemodynamic instability, then infusion to those with active
bloody emesis in hospital) bleeding, NBVV, or adherent clot
US, ASGE No specific recommendation but Suggest IV prokinetic IV PPIs recommended Depends on clinical factors but IV PPI bolus then infusion after
201257 notes that GBS=0 identifies a very low if high probability of recommends within 24 h in the endoscopic therapy for ulcers with
risk group fresh blood or clot in presence of cancer, cirrhosis, high risk stigmata
stomach hematemesis, hypovolemia, or Hb
<80 g/L
UK, NICE 20124 Use GBS before endoscopy and Not assessed Do not give PPIs before endoscopy Within 24 h, but immediately after Offer PPIs if stigmata of recent
full Rockall score after endoscopy. resuscitation if unstable and severe bleeding seen at endoscopy
Consider early discharge if GBS=0 UGIB
Europe, ESGE Patients with GBS 0-1 do not require Recommend IV IV bolus then infusion but should Within 24 h of resuscitation, but IV PPI bolus then 72 h infusion for
20155 early endoscopy or admission erythromycin if not delay endoscopy consider within 12 h if high risk patients who receive endoscopic
clinically severe or features (eg, hemodynamic instability hemostasis and those with
ongoing active UGIB despite resuscitation, inpatient bloody adherent clots. Consider giving
emesis, contraindication to stopping PPIs as intermittent IV bolus or
anticoagulants) high dose oral
Asia-Pacific, Use GBS; adopting a cut off at GBS Not assessed IV PPIs recommended if suspected Within 24 h, but urgent (within 12 After endoscopic hemostasis is
20186 ≤1 allows most hospitals to reduce UGIB awaiting endoscopy h) if hemodynamic instability, after achieved high dose oral PPIs can
unnecessary admissions (especially if endoscopy is not resuscitation and stabilization be used for 72 h as an alternative
available within 24 h) to high dose IV PPIs
Abbreviations: ACG=American College of Gastroenterology; ASGE=American Society for Gastrointestinal Endoscopy; ED=emergency department; ESGE= European Society of Gastrointestinal Endoscopy;
GBS=Glasgow Blatchford score; Hb=hemoglobin; NBVV=non-bleeding visible vessel; NICE=National Institute for Health and Care Excellence; PPIs=proton pump inhibitors; IV=intravenous; UGIB=upper
gastrointestinal bleeding.
QUESTIONS FOR FUTURE RESEARCH clinical practice, and these alterations to management are
now recommended by international guidelines. RCTs and
• What is the optimal approach to fluid resuscitation in
patients with acute upper gastrointestinal bleeding? meta-analyses confirm a benefit from pre-endoscopy anti-
• Can risk assessment tools be developed to allow accurate biotics and vasoactive drugs in patients with cirrhosis,
early identification of high risk patients with upper and post-endoscopy high dose PPIs for high risk peptic
gastrointestinal bleeding, such as those who require ulcer bleeding.
endoscopic therapy or those with high mortality? Endoscopic therapy has advanced dramatically over the
• What is the optimal timing of endoscopy after upper past decades, with recent additions to the endoscopist’s
gastrointestinal bleeding? “toolkit,” including hemostatic powder spray, over-the-
• What is the exact role of hemostatic powder spray, scope clips, and Doppler probes. These join the estab-
over-the-scope clips, and Doppler ultrasound probes in lished and widely studied injection therapies, thermal
the endoscopic management of upper gastrointestinal probes, and clips used for non-variceal bleeding, and
bleeding?
endoscopic band ligation and tissue adhesive injection
• When is the best time to reintroduce antithrombotic drugs
for variceal bleeding. However, the newer modalities
after upper gastrointestinal bleeding?
require further study to clarify their exact role in endo-
scopic management. Technical improvements and more
intestinal bleeding (non-variceal and variceal) have been widely available services for interventional radiology have
published over the past eight years. We have therefore led to it being the most commonly used rescue therapy
referenced them as appropriate throughout. The most for persistent or recurrent upper gastrointestinal bleeding
recent ones—the European (2015) and Asia-Pacific that is refractory to endoscopic treatment. Surgery is now
(2018) guidelines on non-variceal bleeding, and the typically reserved for situations in which interventional
UK, international, and US guidelines on variceal bleed- radiology is unavailable, delayed, or unsuccessful.
ing—differ slightly from earlier ones, largely because they The more widespread use of antiplatelet and anti-
assessed more recently published studies (table). These coagulants drugs has led to uncertainty in managing
guidelines have generally been written by experts in this patients taking these medications who develop upper
field, although methodology has varied. gastrointestinal bleeding. However, recent data suggest
that relatively early reintroduction of these drugs once
Conclusions hemostasis has been achieved is the best approach in
Upper gastrointestinal bleeding remains a common cause those with appropriate cardiovascular indications. New
of presentation to hospitals worldwide, and many recent approaches under investigation for managing upper
studies have assessed the management of patients with gastrointestinal bleeding include the early use of TXA
this condition. The evidence of improved outcomes from and novel endoscopic techniques to reduce rebleeding.
a relatively restrictive approach to blood transfusion and These and other developments will hopefully continue
the ability to identify patients who are at very low risk and to improve management and outcomes for patients with
suitable for outpatient management have recently altered upper gastrointestinal bleeding.
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
saline in critically ill adults. N Engl J Med 2018;378:829-39. 10.1056/
related charities, three patients kindly volunteered to review NEJMoa1711584 pmid:29485925.
21 Carson JL, Stanworth SJ, Roubinian N, et al. Transfusion thresholds
the manuscript. In response to their comments, the sections and other strategies for guiding allogeneic red blood cell transfusion.
on blood transfusion and timing of endoscopy were revised. Cochrane Database Syst Rev 2016;10:CD002042.pmid:27731885.
22 Carson JL, Guyatt G, Heddle NM, et al. Clinical practice guidelines from
Thanks to Deepak Vishwanath for help with the radiological image in fig 4. the AABB: red blood cell transfusion thresholds and storage. JAMA
2016;316:2025-35. 10.1001/jama.2016.9185 pmid:27732721.
Contributors: Both authors contributed equally to the planning and writing
23 Docherty AB, O’Donnell R, Brunskill S, et al. Effect of restrictive versus
of the article and both are responsible for the overall content as guarantors. liberal transfusion strategies on outcomes in patients with cardiovascular
Competing interests: We have read and understood BMJ policy on disease in a non-cardiac surgery setting: systematic review and meta-
declaration of interests and declare the following interests: AJS, none; LL analysis. BMJ 2016;352:i1351. 10.1136/bmj.i1351 pmid:27026510.
has had single consultancies to Takeda and to Bayer and his full competing 24 Blair SD, Janvrin SB, McCollum CN, Greenhalgh RM. Effect of early blood
interests are available from her on request. transfusion on gastrointestinal haemorrhage. Br J Surg 1986;73:783-5.
1 Laine L, Laursen SB, Zakko L, et al. Severity and outcomes of 10.1002/bjs.1800731007 pmid:3533203.
upper gastrointestinal bleeding with bloody vs coffee-grounds 25 Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute
hematemesis. Am J Gastroenterol 2018;113:358-66. 10.1038/ upper gastrointestinal bleeding. N Engl J Med 2013;368:11-21.
ajg.2018.5. pmid:29380820. 10.1056/NEJMoa1211801 pmid:23281973.
2 Lanas A. Editorial: Upper GI bleeding-associated mortality: challenges 26 Jairath V, Kahan BC, Gray A, et al. Restrictive versus liberal blood
to improving a resistant outcome. Am J Gastroenterol 2010;105:90-2. transfusion for acute upper gastrointestinal bleeding (TRIGGER):
10.1038/ajg.2009.517 pmid:20054306. a pragmatic, open-label, cluster randomised feasibility trial.
3 Barkun AN, Bardou M, Kuipers EJ, et al. International Consensus Upper Lancet 2015;386:137-44. 10.1016/S0140-6736(14)61999-
Gastrointestinal Bleeding Conference Group. International consensus 1 pmid:25956718.
recommendations on the management of patients with nonvariceal 27 Odutayo A, Desborough MJ, Trivella M, et al. Restrictive versus liberal
upper gastrointestinal bleeding. Ann Intern Med 2010;152:101-13. blood transfusion for gastrointestinal bleeding: a systematic review and
10.7326/0003-4819-152-2-201001190-00009 pmid:20083829. meta-analysis of randomised controlled trials. Lancet Gastroenterol
4 National Institute for Health and Care Excellence. Acute upper Hepatol 2017;2:354-60. 10.1016/S2468-1253(17)30054-
gastrointestinal bleeding in over 16s: management. CG141. 2012. https:// 7. pmid:28397699.
www.nice.org.uk/guidance/cg141. 28 Laine L, Jensen DM. Management of patients with ulcer bleeding.
5 Gralnek IM, Dumonceau JM, Kuipers EJ, et al. Diagnosis and management Am J Gastroenterol 2012;107:345-60, quiz 361. 10.1038/
of nonvariceal upper gastrointestinal hemorrhage: European Society of ajg.2011.480 pmid:22310222.
Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2015;47:a1-46. 29 Rockall TA, Logan RF, Devlin HB, Northfield TC. Risk assessment after acute
10.1055/s-0034-1393172 pmid:26417980. upper gastrointestinal haemorrhage. Gut 1996;38:316-21. 10.1136/
6 Sung JJ, Chiu PW, Chan FKL, et al. Asia-Pacific working group consensus gut.38.3.316 pmid:8675081.
on non-variceal upper gastrointestinal bleeding: an update 2018. Gut 30 Saeed ZA, Winchester CB, Michaletz PA, Woods KL, Graham DY. A scoring
2018;67:1757-68. 10.1136/gutjnl-2018-316276. pmid:29691276. system to predict rebleeding after endoscopic therapy of nonvariceal
7 Tripathi D, Stanley AJ, Hayes PC, et al. Clinical Services and Standards upper gastrointestinal hemorrhage, with a comparison of heat probe and
Committee of the British Society of Gastroenterology. U.K. guidelines ethanol injection. Am J Gastroenterol 1993;88:1842-9.pmid:8237930.
on the management of variceal haemorrhage in cirrhotic patients. Gut 31 Hay JA, Maldonado L, Weingarten SR, Ellrodt AG. Prospective evaluation
2015;64:1680-704. 10.1136/gutjnl-2015-309262 pmid:25887380. of a clinical guideline recommending hospital length of stay in upper
8 de Franchis R. Baveno VI Faculty. Expanding consensus in portal gastrointestinal tract hemorrhage. JAMA 1997;278:2151-6. 10.1001/
hypertension: report of the Baveno VI Consensus Workshop: Stratifying risk jama.1997.03550240041031 pmid:9417008.
and individualizing care for portal hypertension. J Hepatol 2015;63:743- 32 Blatchford O, Murray WR, Blatchford M. A risk score to predict need for
52. 10.1016/j.jhep.2015.05.022 pmid:26047908. treatment for upper-gastrointestinal haemorrhage. Lancet 2000;356:1318-
9 Garcia-Tsao G, Abraldes JG, Berzigotti A, Bosch J. Portal hypertensive 21. 10.1016/S0140-6736(00)02816-6 pmid:11073021.
bleeding in cirrhosis: Risk stratification, diagnosis, and management: 2016 33 Marmo R, Koch M, Cipolletta L, et al. Italian registry on upper
practice guidance by the American Association for the study of liver diseases. gastrointestinal bleeding (Progetto Nazionale Emorragie Digestive--PNED
Hepatology 2017;65:310-35. 10.1002/hep.28906 pmid:27786365. 2). Predicting mortality in non-variceal upper gastrointestinal bleeders:
10 Rockall TA, Logan RFA, Devlin HB, Northfield TC. Incidence of and validation of the Italian PNED Score and Prospective Comparison with
mortality from acute upper gastrointestinal haemorrhage in the UK. BMJ the Rockall Score. Am J Gastroenterol 2010;105:1284-91. 10.1038/
1995;311:222-6. 10.1136/bmj.311.6999.222 pmid:7627034. ajg.2009.687 pmid:20051943.
11 Blatchford O, Davidson LA, Murray WR, Blatchford M, Pell J. 34 Saltzman JR, Tabak YP, Hyett BH, Sun X, Travis AC, Johannes RS. A simple
Acute upper gastrointestinal haemorrhage in west of Scotland: risk score accurately predicts in-hospital mortality, length of stay, and
case ascertainment study. BMJ 1997;315:510-4. 10.1136/ cost in acute upper GI bleeding. Gastrointest Endosc 2011;74:1215-24.
bmj.315.7107.510 pmid:9329304. 10.1016/j.gie.2011.06.024 pmid:21907980.
12 Laine L, Yang H, Chang SC, Datto C. Trends for incidence of hospitalization 35 Stanley AJ, Laine L, Dalton HR, et al. International Gastrointestinal
and death due to GI complications in the United States from 2001 to Bleeding Consortium. Comparison of risk scoring systems for patients
2009. Am J Gastroenterol 2012;107:1190-5, quiz 1196. 10.1038/ presenting with upper gastrointestinal bleeding: international
ajg.2012.168 pmid:22688850. multicentre prospective study. BMJ 2017;356:i6432. 10.1136/bmj.
13 Abougergi MS, Travis AC, Saltzman JR. The in-hospital mortality rate for i6432 pmid:28053181.
upper GI hemorrhage has decreased over 2 decades in the United States: a 36 Laursen SB, Hansen JM, Schaffalitzky de Muckadell OB. The Glasgow
nationwide analysis. Gastrointest Endosc 2015;81:882-8.e1. 10.1016/j. Blatchford score is the most accurate assessment of patients with upper
gie.2014.09.027 pmid:25484324. gastrointestinal hemorrhage. Clin Gastroenterol Hepatol 2012;10:1130-
14 Wuerth BA, Rockey DC. Changing epidemiology of upper gastrointestinal 1135.e1. 10.1016/j.cgh.2012.06.022 pmid:22801061.
hemorrhage in the last decade: a nationwide analysis. Dig Dis Sci 37 Yang HM, Jeon SW, Jung JT, et al. Daegu-Gyeongbuk Gastrointestinal
2018;63:1286-93. 10.1007/s10620-017-4882-6 pmid:29282637. Study Group (DGSG). Comparison of scoring systems for nonvariceal
15 Hearnshaw SA, Logan RFA, Lowe D, Travis SP, Murphy MF, Palmer KR. Acute upper gastrointestinal bleeding: a multicenter prospective cohort
upper gastrointestinal bleeding in the UK: patient characteristics, diagnoses study. J Gastroenterol Hepatol 2016;31:119-25. 10.1111/
and outcomes in the 2007 UK audit. Gut 2011;60:1327-35. 10.1136/ jgh.13057 pmid:26211939.
gut.2010.228437 pmid:21490373. 38 Stanley AJ, Ashley D, Dalton HR, et al. Outpatient management of patients
16 Nahon S, Hagège H, Latrive JP, et al. Groupe des Hémorragies Digestives with low-risk upper-gastrointestinal haemorrhage: multicentre validation
Hautes de l’ANGH. Epidemiological and prognostic factors involved and prospective evaluation. Lancet 2009;373:42-7. 10.1016/S0140-
in upper gastrointestinal bleeding: results of a French prospective 6736(08)61769-9 pmid:19091393.
multicenter study. Endoscopy 2012;44:998-1008. 10.1055/s-0032- 39 Ramaekers R, Mukarram M, Smith CA, Thiruganasambandamoorthy V. The
1310006 pmid:23108771. predictive value of preendoscopic risk scores to predict adverse outcomes
17 Bickell WH, Wall MJ Jr, , Pepe PE, et al. Immediate versus delayed in emergency department patients with upper gastrointestinal bleeding:
fluid resuscitation for hypotensive patients with penetrating a systematic review. Acad Emerg Med 2016;23:1218-27. 10.1111/
torso injuries. N Engl J Med 1994;331:1105-9. 10.1056/ acem.13101 pmid:27640399.
NEJM199410273311701 pmid:7935634. 40 Laursen SB, Dalton HR, Murray IA, et al. Upper Gastrointestinal
18 Carrick MM, Morrison CA, Tapia NM, et al. Intraoperative hypotensive Hemorrhage International Consortium Upper Gastrointestinal
resuscitation for patients undergoing laparotomy or thoracotomy Hemorrhage International Consortium. Performance of new thresholds
for trauma: Early termination of a randomized prospective clinical of the Glasgow Blatchford score in managing patients with upper
trial. J Trauma Acute Care Surg 2016;80:886-96. 10.1097/ gastrointestinal bleeding. Clin Gastroenterol Hepatol 2015;13:115-21.
TA.0000000000001044 pmid:27015578. e2. 10.1016/j.cgh.2014.07.023 pmid:25058843.
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
42 Chang HY, Zhou M, Tang W, Alexander GC, Singh S. Risk of gastrointestinal 64 Avgerinos A, Nevens F, Raptis S, Fevery J. Early administration of
bleeding associated with oral anticoagulants: population based somatostatin and efficacy of sclerotherapy in acute oesophageal variceal
retrospective cohort study. BMJ 2015;350:h1585. 10.1136/bmj. bleeds: the European Acute Bleeding Oesophageal Variceal Episodes
h1585 pmid:25911526. (ABOVE) randomised trial. Lancet 1997;350:1495-9. 10.1016/S0140-
43 Huang ES, Strate LL, Ho WW, Lee SS, Chan AT. Long-term use of aspirin 6736(97)05099-X pmid:9388396.
and the risk of gastrointestinal bleeding. Am J Med 2011;124:426-33. 65 Calès P, Masliah C, Bernard B, et al. French Club for the Study of Portal
10.1016/j.amjmed.2010.12.022 pmid:21531232. Hypertension. Early administration of vapreotide for variceal bleeding
44 Ahsberg K, Höglund P, Kim WH, von Holstein CS. Impact of in patients with cirrhosis. N Engl J Med 2001;344:23-8. 10.1056/
aspirin, NSAIDs, warfarin, corticosteroids and SSRIs on the site NEJM200101043440104 pmid:11136956.
and outcome of non-variceal upper and lower gastrointestinal 66 Avgerinos A, Sgouros S, Viazis N, et al. Somatostatin inhibits
bleeding. Scand J Gastroenterol 2010;45:1404-15. gastric acid secretion more effectively than pantoprazole in
10.3109/00365521.2010.510567 pmid:20695720. patients with peptic ulcer bleeding: a prospective, randomized,
45 Acosta RD, Abraham NS, Chandrasekhara V, et al. ASGE Standards of placebo-controlled trial. Scand J Gastroenterol 2005;40:515-22.
Practice Committee. The management of antithrombotic agents for 10.1080/00365520510015458 pmid:16036503.
patients undergoing GI endoscopy. Gastrointest Endosc 2016;83:3-16. 67 Whitehouse I, Beglinger C, Fried M, Gyr K. The effect of an octapeptide
10.1016/j.gie.2015.09.035 pmid:26621548. somatostatin analogue (SMS 201-995) and somatostatin-14 (SST-14)
46 Zakko L, Rustagi T, Douglas M, Laine L. No benefit from platelet on pentagastrin-stimulated gastric acid secretion: a comparative study in
transfusion for gastrointestinal bleeding in patients taking antiplatelet man. Hepatogastroenterology 1984;31:227-9.pmid:6150888.
agents. Clin Gastroenterol Hepatol 2017;15:46-52. 10.1016/j. 68 Chavez-Tapia NC, Barrientos-Gutierrez T, Tellez-Avila F, et al.
cgh.2016.07.017 pmid:27464591. Meta-analysis: antibiotic prophylaxis for cirrhotic patients with
47 Razzaghi A, Barkun AN. Platelet transfusion threshold in upper gastrointestinal bleeding—an updated Cochrane review.
patients with upper gastrointestinal bleeding: a systematic Aliment Pharmacol Ther 2011;34:509-18. 10.1111/j.1365-
review. J Clin Gastroenterol 2012;46:482-6. 10.1097/ 2036.2011.04746.x pmid:21707680.
MCG.0b013e31823d33e3 pmid:22688143. 69 Fernández J, Ruiz del Arbol L, Gómez C, et al. Norfloxacin vs ceftriaxone
48 Choudari CP, Rajgopal C, Palmer KR. Acute gastrointestinal haemorrhage in the prophylaxis of infections in patients with advanced cirrhosis
in anticoagulated patients: diagnoses and response to endoscopic and hemorrhage. Gastroenterology 2006;131:1049-56, quiz 1285.
treatment. Gut 1994;35:464-6. 10.1136/gut.35.4.464 pmid:8174982. 10.1053/j.gastro.2006.07.010 pmid:17030175.
49 Wolf AT, Wasan SK, Saltzman JR. Impact of anticoagulation on rebleeding 70 Strate LL, Gralnek IM. ACG clinical guideline: management of patients with
following endoscopic therapy for nonvariceal upper gastrointestinal acute lower gastrointestinal bleeding. Am J Gastroenterol 2016;111:459-
hemorrhage. Am J Gastroenterol 2007;102:290-6. 10.1111/j.1572- 74. 10.1038/ajg.2016.41. pmid:26925883.
0241.2006.00969.x pmid:17100959. 71 Pasha SF, Shergill A, Acosta RD, et al. ASGE Standards of Practice
50 Shingina A, Barkun AN, Razzaghi A, Martel M, Bardou M, Gralnek I. RUGBE Committee. The role of endoscopy in the patient with lower GI
Investigators. Systematic review: the presenting international normalised bleeding. Gastrointest Endosc 2014;79:875-85. 10.1016/j.
ratio (INR) as a predictor of outcome in patients with upper nonvariceal gie.2013.10.039 pmid:24703084.
gastrointestinal bleeding. Aliment Pharmacol Ther 2011;33:1010-8. 72 Nadler M, Eliakim R. The role of capsule endoscopy in acute
10.1111/j.1365-2036.2011.04618.x pmid:21385193. gastrointestinal bleeding. Therap Adv Gastroenterol 2014;7:87-92.
51 Veitch AM, Vanbiervliet G, Gershlick AH, et al. Endoscopy in patients on 10.1177/1756283X13504727 pmid:24587821.
antiplatelet or anticoagulant therapy, including direct oral anticoagulants: 73 Lim LG, Ho KY, Chan YH, et al. Urgent endoscopy is associated with
British Society of Gastroenterology (BSG) and European Society of lower mortality in high-risk but not low-risk nonvariceal upper
Gastrointestinal Endoscopy (ESGE) guidelines. Gut 2016;65:374-89. gastrointestinal bleeding. Endoscopy 2011;43:300-6. 10.1055/s-0030-
10.1136/gutjnl-2015-311110 pmid:26873868. 1256110 pmid:21360421.
52 Pollack CV Jr, , Reilly PA, van Ryn J, et al. Idarucizumab for dabigatran 74 Cho SH, Lee YS, Kim YJ, et al. Outcomes and role of urgent endoscopy
reversal—full cohort analysis. N Engl J Med 2017;377:431-41. 10.1056/ in high-risk patients with acute nonvariceal gastrointestinal
NEJMoa1707278 pmid:28693366. bleeding. Clin Gastroenterol Hepatol 2018;16:370-7. 10.1016/j.
53 Connolly SJ, Milling TJ Jr, , Eikelboom JW, et al. ANNEXA-4 Investigators. cgh.2017.06.029 pmid:28634135.
Andexanet alfa for acute major bleeding associated with factor 75 Lin HJ, Wang K, Perng CL, et al. Early or delayed endoscopy for patients
Xa inhibitors. N Engl J Med 2016;375:1131-41. 10.1056/ with peptic ulcer bleeding. A prospective randomized study. J Clin
NEJMoa1607887 pmid:27573206. Gastroenterol 1996;22:267-71. 10.1097/00004836-199606000-
54 Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl 00005 pmid:8771420.
J Med 2011;365:147-56. 10.1056/NEJMra1011170 pmid:21751907. 76 Laursen SB, Leontiadis GI, Stanley AJ, Møller MH, Hansen JM, Schaffalitzky
55 Tripodi A, Primignani M, Chantarangkul V, et al. Thrombin generation in de Muckadell OB. Relationship between timing of endoscopy
patients with cirrhosis: the role of platelets. Hepatology 2006;44:440-5. and mortality in patients with peptic ulcer bleeding: a nationwide
10.1002/hep.21266 pmid:16871542. cohort study. Gastrointest Endosc 2017;85:936-944.e3. 10.1016/j.
56 Sreedharan A, Martin J, Leontiadis GI, et al. Proton pump inhibitor gie.2016.08.049 pmid:27623102.
treatment initiated prior to endoscopic diagnosis in upper gastrointestinal 77 Sung JJ, Tsoi KK, Ma TK, Yung MY, Lau JY, Chiu PW. Causes of mortality
bleeding. Cochrane Database Syst Rev 2010;7:CD005415.pm in patients with peptic ulcer bleeding: a prospective cohort study
id:20614440. of 10,428 cases. Am J Gastroenterol 2010;105:84-9. 10.1038/
57 Hwang JH, Fisher DA, Ben-Menachem T, et al. Standards of Practice ajg.2009.507 pmid:19755976.
Committee of the American Society for Gastrointestinal Endoscopy. 78 Lee JG, Turnipseed S, Romano PS, et al. Endoscopy-based triage
The role of endoscopy in the management of acute non-variceal significantly reduces hospitalization rates and costs of treating
upper GI bleeding. Gastrointest Endosc 2012;75:1132-8. 10.1016/j. upper GI bleeding: a randomized controlled trial. Gastrointest
gie.2012.02.033 pmid:22624808. Endosc 1999;50:755-61. 10.1016/S0016-5107(99)70154-
58 Rahman R, Nguyen DL, Sohail U, et al. Pre-endoscopic erythromycin 9 pmid:10570332.
administration in upper gastrointestinal bleeding: an updated meta- 79 Bjorkman DJ, Zaman A, Fennerty MB, Lieberman D, Disario JA, Guest-
analysis and systematic review. Ann Gastroenterol 2016;29:312- Warnick G. Urgent vs. elective endoscopy for acute non-variceal upper-GI
7.pmid:27366031. bleeding: an effectiveness study. Gastrointest Endosc 2004;60:1-8.
59 Gluud LL, Klingenberg SL, Langholz E. Tranexamic acid for upper 10.1016/S0016-5107(04)01287-8 pmid:15229417.
gastrointestinal bleeding. Cochrane Database Syst Rev 2012;1:CD00664 80 Laine L. Upper gastrointestinal bleeding due to a peptic ulcer. N Engl J
0.pmid:22258969. Med 2016;374:2367-76. 10.1056/NEJMcp1514257 pmid:27305194.
60 Gayet-Ageron A, Prieto- 81 Maida M, Camilleri S, Manganaro M, Garufi S, Scarpulla G. Radiofrequency
Merino D, Ker K, Shakur H, Ageron FX, Roberts I. Antifibrinolytic Trials ablation for treatment of refractory gastric antral vascular ectasia:
Collaboration. Effect of treatment delay on the effectiveness and safety a systematic review of the literature. Gastroenterol Res Pract
of antifibrinolytics in acute severe haemorrhage: a meta-analysis 2017;2017:5609647. 10.1155/2017/5609647 pmid:28835751.
of individual patient-level data from 40 138 bleeding patients. 82 Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an
Lancet 2018;391:125-32. 10.1016/S0140-6736(17)32455- evidence-based approach based on meta-analyses of randomized
8. pmid:29126600. controlled trials. Clin Gastroenterol Hepatol 2009;7:33-47, quiz 1-2.
61 Roberts I, Coats T, Edwards P, et al. HALT-IT--tranexamic acid for the 10.1016/j.cgh.2008.08.016 pmid:18986845.
treatment of gastrointestinal bleeding: study protocol for a randomised 83 Vergara M, Bennett C, Calvet X, Gisbert JP. Epinephrine injection versus
controlled trial. Trials 2014;15:450. 10.1186/1745-6215-15- epinephrine injection and a second endoscopic method in high-risk
450. pmid:25409738. bleeding ulcers. Cochrane Database Syst Rev 2014;10:CD005584.pm
62 Seo YS, Park SY, Kim MY, et al. Lack of difference among terlipressin, id:25308912.
somatostatin, and octreotide in the control of acute gastroesophageal 84 Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in
variceal hemorrhage. Hepatology 2014;60:954-63. 10.1002/ gastrointestinal bleeding. Lancet 1974;2:394-7. 10.1016/S0140-
hep.27006 pmid:24415445. 6736(74)91770-X pmid:4136718.
BMJ: first published as 10.1136/bmj.l536 on 25 March 2019. Downloaded from http://www.bmj.com/ on 3 June 2020 by guest. Protected by copyright.
United Kingdom: patient characteristics, management and outcomes Validation of clinical classification schemes for predicting stroke: results
in a nationwide audit. Dig Liver Dis 2014;46:419-26. 10.1016/j. from the National Registry of Atrial Fibrillation. JAMA 2001;285:2864-70.
dld.2013.12.010 pmid:24433997. 10.1001/jama.285.22.2864 pmid:11401607.
87 Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for 109 Lip GY, Frison L, Halperin JL, Lane DA. Comparative validation of a novel
treatment of esophageal variceal bleeding. A meta-analysis. Ann Intern risk score for predicting bleeding risk in anticoagulated patients with
Med 1995;123:280-7. 10.7326/0003-4819-123-4-199508150- atrial fibrillation: the HAS-BLED (Hypertension, Abnormal Renal/Liver
00007 pmid:7611595. Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly,
88 Lo GH, Lai KH, Cheng JS, Chen MH, Chiang HT. A prospective, randomized Drugs/Alcohol Concomitantly) score. J Am Coll Cardiol 2011;57:173-80.
trial of butyl cyanoacrylate injection versus band ligation in the management 10.1016/j.jacc.2010.09.024 pmid:21111555.
of bleeding gastric varices. Hepatology 2001;33:1060-4. 10.1053/ 110 Zulkifly H, Lip GYH, Lane DA. Bleeding risk scores in atrial fibrillation and
jhep.2001.24116 pmid:11343232. venous thromboembolism. Am J Cardiol 2017;120:1139-45. 10.1016/j.
89 Tan PC, Hou MC, Lin HC, et al. A randomized trial of endoscopic amjcard.2017.06.058 pmid:28800833.
treatment of acute gastric variceal hemorrhage: N-butyl-2-cyanoacrylate 111 Lau JYW, Sung JJY, Lam YH, et al. Endoscopic retreatment compared
injection versus band ligation. Hepatology 2006;43:690-7. 10.1002/ with surgery in patients with recurrent bleeding after initial endoscopic
hep.21145 pmid:16557539. control of bleeding ulcers. N Engl J Med 1999;340:751-6. 10.1056/
90 Cheng LF, Wang ZQ, Li CZ, Lin W, Yeo AE, Jin B. Low incidence of NEJM199903113401002 pmid:10072409.
complications from endoscopic gastric variceal obturation with butyl 112 Chen Y-I, Barkun AN. Hemostatic powders in gastrointestinal bleeding:
cyanoacrylate. Clin Gastroenterol Hepatol 2010;8:760-6. 10.1016/j. a systematic review. Gastrointest Endosc Clin N Am 2015;25:535-52.
cgh.2010.05.019 pmid:20621678. 10.1016/j.giec.2015.02.008 pmid:26142037.
91 Ramesh J, Limdi JK, Sharma V, Makin AJ. The use of thrombin injections 113 Troland D, Stanley A. Endotherapy for peptic ulcer bleeding.
in the management of bleeding gastric varices: a single-center Gastrointest Endosc Clin N Am 2018;28:277-89. 10.1016/j.
experience. Gastrointest Endosc 2008;68:877-82. 10.1016/j. giec.2018.02.002 pmid:29933775.
gie.2008.02.065 pmid:18534583. 114 Kyaw M, Tse Y, Ang D, Ang TL, Lau J. Embolization versus surgery
92 McAvoy NC, Plevris JN, Hayes PC. Human thrombin for the treatment of gastric for peptic ulcer bleeding after failed endoscopic hemostasis: a
and ectopic varices. World J Gastroenterol 2012;18:5912-7. 10.3748/wjg. meta-analysis. Endosc Int Open 2014;2:E6-14. 10.1055/s-0034-
v18.i41.5912 pmid:23139607. 1365235 pmid:26134614.
93 Smith MR, Tidswell R, Tripathi D. Outcomes of endoscopic human thrombin 115 Beggs AD, Dilworth MP, Powell SL, Atherton H, Griffiths EA. A
injection in the management of gastric varices. Eur J Gastroenterol Hepatol systematic review of transarterial embolization versus emergency
2014;26:846-52. 10.1097/MEG.0000000000000119 pmid:24892515. surgery in treatment of major nonvariceal upper gastrointestinal
94 Sachar H, Vaidya K, Laine L. Intermittent vs continuous proton pump bleeding. Clin Exp Gastroenterol 2014;7:93-104. 10.2147/CEG.
inhibitor therapy for high-risk bleeding ulcers: a systematic review S56725 pmid:24790465.
and meta-analysis. JAMA Intern Med 2014;174:1755-62. 10.1001/ 116 Lau JYW, Pittayanon R, Wong KT, et al. Prophylactic angiographic
jamainternmed.2014.4056 pmid:25201154. embolisation after endoscopic control of bleeding to high-risk peptic
95 Freston JW, Pilmer BL, Chiu YL, et al. Evaluation of the pharmacokinetics ulcers: a randomised controlled trial. Gut 2018;gutjnl-2018-316074.
and pharmacodynamics of intravenous lansoprazole. Aliment 10.1136/gutjnl-2018-316074. pmid:29802172.
Pharmacol Ther 2004;19:1111-22. 10.1111/j.1365- 117 Cook D, Laine L. Indications, technique, and complications of balloon
2036.2004.01942.x pmid:15142201. tamponade for variceal gastrointestinal bleeding. J Intensive Care Med
96 Cheng HC, Wu CT, Chang WL, Cheng WC, Chen WY, Sheu BS. Double 1992;7:212-8. 10.1177/088506669200700408 pmid:10147943.
oral esomeprazole after a 3-day intravenous esomeprazole infusion 118 Escorsell À, Pavel O, Cárdenas A, et al. Variceal Bleeding Study
reduces recurrent peptic ulcer bleeding in high-risk patients: a Group. Esophageal balloon tamponade versus esophageal stent
randomised controlled study. Gut 2014;63:1864-72. 10.1136/ in controlling acute refractory variceal bleeding: A multicenter
gutjnl-2013-306531 pmid:24658598. randomized, controlled trial. Hepatology 2016;63:1957-67. 10.1002/
97 Bhatt DL, Cryer BL, Contant CF, et al. COGENT Investigators. Clopidogrel hep.28360 pmid:26600191.
with or without omeprazole in coronary artery disease. N Engl J Med 119 Sung JJ, Luo D, Wu JC, et al. Early clinical experience of the safety and
2010;363:1909-17. 10.1056/NEJMoa1007964 pmid:20925534. effectiveness of Hemospray in achieving hemostasis in patients with
98 Sanofi; Bristol-Myers Squibb. Plavix. Highlights of prescribing acute peptic ulcer bleeding. Endoscopy 2011;43:291-5. 10.1055/s-
information. 2018. https://www.accessdata.fda.gov/drugsatfda_docs/ 0030-1256311 pmid:21455870.
label/2018/020839s069lbl.pdf. 120 Smith LA, Stanley AJ, Bergman JJ, et al. Hemospray application in
99 Lo GH, Chen WC, Wang HM, et al. Low-dose terlipressin plus banding nonvariceal upper gastrointestinal bleeding: results of the Survey
ligation versus low-dose terlipressin alone in the prevention of very early to Evaluate the Application of Hemospray in the Luminal Tract. J Clin
rebleeding of oesophageal varices. Gut 2009;58:1275-80. 10.1136/ Gastroenterol 2014;48:e89-92. 10.1097/MCG.054.
gut.2008.165910 pmid:19386609. pmid:24326829.
100 Sung JJ, Chung SC, Yung MY, et al. Prospective randomised study of effect 121 Sulz MC, Frei R, Meyenberger C, Bauerfeind P, Semadeni GM, Gubler C.
of octreotide on rebleeding from oesophageal varices after endoscopic Routine use of Hemospray for gastrointestinal bleeding: prospective
ligation. Lancet 1995;346:1666-9. 10.1016/S0140-6736(95)92840- two-center experience in Switzerland. Endoscopy 2014;46:619-24.
5 pmid:8551824. 10.1055/s-0034-1365505 pmid:24770964.
101 García-Pagán JC, Caca K, Bureau C, et al. Early TIPS (Transjugular Intrahepatic 122 Changela K, Papafragkakis H, Ofori E, et al. Hemostatic
Portosystemic Shunt) Cooperative Study Group. Early use of TIPS in patients powder spray: a new method for managing gastrointestinal
with cirrhosis and variceal bleeding. N Engl J Med 2010;362:2370-9. bleeding. Therap Adv Gastroenterol 2015;8:125-35.
10.1056/NEJMoa0910102 pmid:20573925. 10.1177/1756283X15572587 pmid:26082803.
102 Conejo I, Guardascione MA, Tandon P, et al. Multicenter external 123 Kwek BEA, Ang TL, Ong PLJ, et al. TC-325 versus the conventional
validation of risk stratification criteria for patients with variceal bleeding. combined technique for endoscopic treatment of peptic ulcers with high-
Clin Gastroenterol Hepatol 2018;16:132-139.e8. 10.1016/j. risk bleeding stigmata: A randomized pilot study. J Dig Dis 2017;18:323-
cgh.2017.04.042 pmid:28501536. 9. 10.1111/1751-2980.12481. pmid:28485544.
103 Sung JJ, Lau JY, Ching JY, et al. Continuation of low-dose aspirin therapy in 124 Ibrahim M, El-Mikkawy A, Abdel Hamid M, et al. Early
peptic ulcer bleeding: a randomized trial. Ann Intern Med 2010;152:1-9. application of haemostatic powder added to standard
10.7326/0003-4819-152-1-201001050-00179 pmid:19949136. management for oesophagogastric variceal bleeding: a
104 Qureshi W, Mittal C, Patsias I, et al. Restarting anticoagulation and outcomes randomised trial. Gut 2018;gutjnl-2017-314653. 10.1136/
after major gastrointestinal bleeding in atrial fibrillation. Am J Cardiol gutjnl-2017-314653. pmid:29730601.
2014;113:662-8. 10.1016/j.amjcard.2013.10.044 pmid:24355310. 125 Schmidt A, Gölder S, Goetz M, et al. Over the scope clips are more
105 Chan FKL, Goh KL, Reddy N, et al. Management of patients on antithrombotic effective than standard endoscopic therapy for patients with recurrent
agents undergoing emergency and elective endoscopy: joint Asian Pacific bleeding of peptic ulcers. Gastroenterology 2018;155:674-686.e6.
Association of Gastroenterology (APAGE) and Asian Pacific Society for 10.1053/j.gastro.2018.05.037. pmid:29803838.
Digestive Endoscopy (APSDE) practice guidelines. Gut 2018;67:405-17. 126 Jensen DM, Kovacs TOG, Ohning GV, et al. Doppler endoscopic
10.1136/gutjnl-2017-315131 pmid:29331946. probe monitoring of blood flow improves risk stratification and
106 Scott MJ, Veitch A, Thachil J. Reintroduction of anti-thrombotic therapy after a outcomes of patients with severe nonvariceal upper gastrointestinal
gastrointestinal haemorrhage: if and when?Br J Haematol 2017;177:185-97. hemorrhage. Gastroenterology 2017;152:1310-1318.e1. 10.1053/j.
10.1111/bjh.14599. pmid:28272736. gastro.2017.01.042 pmid:28167214.
14 of 13
STAT E O F T H E A RT R E V I E W
15 of 13
STAT E O F T H E A RT R E V I E W
Authors Ian M. Gralnek1, 2, Jean-Marc Dumonceau3, Ernst J. Kuipers4, Angel Lanas5, David S. Sanders6, Matthew Kurien6,
Gianluca Rotondano7, Tomas Hucl8, Mario Dinis-Ribeiro9, Riccardo Marmo10, Istvan Racz11, Alberto Arezzo12,
Ralf-Thorsten Hoffmann13, Gilles Lesur14, Roberto de Franchis15, Lars Aabakken16, Andrew Veitch17, Franco Radaelli18,
Paulo Salgueiro19, Ricardo Cardoso20, Luís Maia19, Angelo Zullo21, Livio Cipolletta22, Cesare Hassan23
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Institutions Institutions listed at end of article.
Bibliography This Guideline is an official statement of the European Society of Gastrointestinal Endoscopy (ESGE). It
DOI http://dx.doi.org/ addresses the diagnosis and management of nonvariceal upper gastrointestinal hemorrhage (NVUGIH).
10.1055/s-0034-1393172
Published online: 0.0.
Endoscopy 2015; 47: 1–46
Main Recommendations
© Georg Thieme Verlag KG MR1. ESGE recommends immediate assessment of patients with clinically severe or ongoing active
Stuttgart · New York hemodynamic status in patients who present with UGIH. In selected patients, pre-endoscopic infu-
ISSN 0013-726X acute upper gastrointestinal hemorrhage (UGIH), sion of erythromycin significantly improves endo-
with prompt intravascular volume replacement scopic visualization, reduces the need for second-
Corresponding author
Ian M. Gralnek, MD, MSHS
initially using crystalloid fluids if hemodynamic look endoscopy, decreases the number of units of
Institute of Gastroenterology instability exists (strong recommendation, mod- blood transfused, and reduces duration of hospital
and Liver Diseases, Ha'Emek erate quality evidence). stay (strong recommendation, high quality evi-
Medical Center MR2. ESGE recommends a restrictive red blood cell dence).
Rappaport Faculty of Medicine, transfusion strategy that aims for a target hemo- MR7. Following hemodynamic resuscitation, ESGE
Technion-Israel Institute of globin between 7 g/dL and 9 g/dL. A higher target recommends early (≤ 24 hours) upper GI endos-
Technology
hemoglobin should be considered in patients copy. Very early (< 12 hours) upper GI endoscopy
Afula, Israel 18101
Fax: +972-4-6495314
with significant co-morbidity (e. g., ischemic car- may be considered in patients with high risk clini-
ian_gr@clalit.org.il diovascular disease) (strong recommendation, cal features, namely: hemodynamic instability (ta-
moderate quality evidence). chycardia, hypotension) that persists despite on-
MR3. ESGE recommends the use of the Glasgow- going attempts at volume resuscitation; in-hospi-
Blatchford Score (GBS) for pre-endoscopy risk stra- tal bloody emesis/nasogastric aspirate; or contra-
tification. Outpatients determined to be at very indication to the interruption of anticoagulation
low risk, based upon a GBS score of 0 – 1, do not re- (strong recommendation, moderate quality evi-
quire early endoscopy nor hospital admission. Dis- dence).
charged patients should be informed of the risk of MR8. ESGE recommends that peptic ulcers with
recurrent bleeding and be advised to maintain spurting or oozing bleeding (Forrest classification
contact with the discharging hospital (strong re- Ia and Ib, respectively) or with a nonbleeding visi-
commendation, moderate quality evidence). ble vessel (Forrest classification IIa) receive endo-
MR4. ESGE recommends initiating high dose intra- scopic hemostasis because these lesions are at
venous proton pump inhibitors (PPI), intravenous high risk for persistent bleeding or rebleeding
bolus followed by continuous infusion (80 mg (strong recommendation, high quality evidence).
then 8 mg/hour), in patients presenting with acute MR9. ESGE recommends that peptic ulcers with an
UGIH awaiting upper endoscopy. However, PPI in- adherent clot (Forrest classification IIb) be consid-
fusion should not delay the performance of early ered for endoscopic clot removal. Once the clot is
endoscopy (strong recommendation, high quality removed, any identified underlying active bleed-
evidence). ing (Forrest classification Ia or Ib) or nonbleeding
MR5. ESGE does not recommend the routine use of visible vessel (Forrest classification IIa) should re-
nasogastric or orogastric aspiration/lavage in pa- ceive endoscopic hemostasis (weak recommenda-
tients presenting with acute UGIH (strong recom- tion, moderate quality evidence).
mendation, moderate quality evidence). MR10. In patients with peptic ulcers having a flat
MR6. ESGE recommends intravenous erythromy- pigmented spot (Forrest classification IIc) or clean
cin (single dose, 250 mg given 30 – 120 minutes base (Forrest classification III), ESGE does not re-
prior to upper gastrointestinal [GI] endoscopy) in commend endoscopic hemostasis as these stigma-
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a2 Guideline
ta present a low risk of recurrent bleeding. In selected clinical set- sis if indicated. In the case of failure of this second attempt at he-
tings, these patients may be discharged to home on standard PPI mostasis, transcatheter angiographic embolization (TAE) or sur-
therapy, e. g., oral PPI once-daily (strong recommendation, moder- gery should be considered (strong recommendation, high quality
ate quality evidence). evidence).
MR11. ESGE recommends that epinephrine injection therapy not MR14. In patients with NVUGIH secondary to peptic ulcer, ESGE re-
be used as endoscopic monotherapy. If used, it should be combined commends investigating for the presence of Helicobacter pylori in
with a second endoscopic hemostasis modality (strong recom- the acute setting with initiation of appropriate antibiotic therapy
mendation, high quality evidence). when H. pylori is detected. Re-testing for H. pylori should be per-
MR12. ESGE recommends PPI therapy for patients who receive formed in those patients with a negative test in the acute setting.
endoscopic hemostasis and for patients with adherent clot not re- Documentation of successful H. pylori eradication is recommended
ceiving endoscopic hemostasis. PPI therapy should be high dose (strong recommendation, high quality evidence).
and administered as an intravenous bolus followed by continuous MR15. In patients receiving low dose aspirin for secondary cardio-
infusion (80 mg then 8 mg/hour) for 72 hours post endoscopy vascular prophylaxis who develop peptic ulcer bleeding, ESGE re-
(strong recommendation, high quality evidence). commends aspirin be resumed immediately following index
MR13. ESGE does not recommend routine second-look endoscopy endoscopy if the risk of rebleeding is low (e. g., FIIc, FIII). In patients
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
as part of the management of nonvariceal upper gastrointestinal with high risk peptic ulcer (FIa, FIb, FIIa, FIIb), early reintroduction
hemorrhage (NVUGIH). However, in patients with clinical evi- of aspirin by day 3 after index endoscopy is recommended, provid-
dence of rebleeding following successful initial endoscopic hemo- ed that adequate hemostasis has been established (strong recom-
stasis, ESGE recommends repeat upper endoscopy with hemosta- mendation, moderate quality evidence).
Abbreviations Introduction
! !
APC argon plasma coagulation Acute upper gastrointestinal hemorrhage (UGIH) is a common
ASA American Society of Anesthesiologists condition worldwide that has an estimated annual incidence of
DAPT dual antiplatelet therapy 40−150 cases per 100 000 population [1, 2], frequently leads to
CHADS2 congestive heart failure, hypertension, age ≥ 75 years, hospital admission, and has significant associated morbidity and
diabetes mellitus, and previous stroke or transient mortality, especially in the elderly. The most common causes of
ischemic attack [risk score] acute UGIH are nonvariceal [1, 2]. This includes peptic ulcers, 28
CI confidence interval % – 59 % (duodenal ulcer 17 % – 37 % and gastric ulcer 11 % – 24 %);
DOAC direct oral anticoagulant mucosal erosive disease of the esophagus/stomach/duodenum,
ESGE European Society of Gastrointestinal Endoscopy 1 % – 47 %; Mallory – Weiss syndrome, 4 % – 7 %; upper GI tract ma-
FFP fresh frozen plasma lignancy, 2 % – 4 %; other diagnosis, 2 % – 7 %; or no exact cause
GBS Glasgow-Blatchford Score identified, 7 % – 25 % [1, 2]. Moreover, in 16 % – 20 % of acute UGIH
GI gastrointestinal cases, more than one endoscopic diagnosis may be identified as
GRADE Grading of Recommendations Assessment, the cause of bleeding. The aim of this evidence-based consensus
Development and Evaluation guideline is to provide medical caregivers with a comprehensive
HR hazard ratio review and recommendations on the clinical and endoscopic
INR international normalized ratio management of NVUGIH.
NBVV nonbleeding visible vessel
NNT number needed to treat
NOAC non-VKA oral anticoagulant Methods
NVUGIH nonvariceal upper gastrointestinal hemorrhage !
PAR protease-activated receptor The ESGE commissioned this guideline on NVUGIH and appoin-
PCC prothrombin complex concentrate ted a guideline leader (I.M.G.) who in collaboration with the Chair
PICO patients, interventions, controls, outcomes of the ESGE Guidelines Committee (C.H.), invited the listed au-
PPI proton pump inhibitor thors to participate in the guideline development and review.
OR odds ratio Key questions were prepared by the coordinating team (I.M.G.
PUB peptic ulcer bleeding and C.H.) and reviewed and approved by all task force members.
RBC red blood cell The coordinating team formed four task force subgroups, each
RCT randomized controlled trial with its own coordinator, and divided the key topics/questions
RR relative risk or risk ratio amongst these four task force subgroups (see Appendix e1, on-
TAE transcatheter angiographic embolization line-only). Task force members included gastroenterologists/gas-
UGIH upper gastrointestinal hemorrhage trointestinal endoscopists, an interventional radiologist, and a
VCE videocapsule endoscopy surgeon. Clinical questions were formulated using the PICO (pa-
VKA vitamin K antagonist tients, interventions, controls, outcomes) methodology.
Each task force subgroup performed a systematic literature
search to identify the relevant literature that was subsequently
used to prepare evidence-based, well-balanced statements on
each of their assigned key questions. The Ovid MEDLINE, EM-
BASE, Google/Google Scholar, and the Cochrane Database of Sys-
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a3
tematic Reviews were searched for English-language articles in- in the “observation group” (P = 0.04 for both comparisons). How-
cluding at a minimum the following key words: nonvariceal up- ever, there is no evidence from randomized controlled trials
per gastrointestinal (GI) hemorrhage/bleeding, peptic ulcer he- (RCTs), for or against early or large-volume intravenous fluid ad-
morrhage/bleeding, fluid resuscitation, fluid therapy, critical ill- ministration in uncontrolled hemorrhage [6, 7]. Moreover, the se-
ness, crystalloid solutions, colloid solutions, plasma transfusions, lection of resuscitation fluid type in critically ill patients requires
red blood cell transfusion, platelet transfusion, hemoglobin, re- careful consideration based on safety, effects on patient out-
strictive transfusion strategy, liberal transfusion strategy, risk comes, and costs. To date, there is ongoing uncertainty regarding
stratification, mortality, rebleeding, anti-thrombotic agent, anti- the ideal fluid administration strategy in this clinical setting [8,
platelet agent, aspirin, dual anti-platelet therapy (DAPT), anti-co- 9].
agulation/anti-coagulant, direct/new oral anticoagulants
(DOACs), coagulopathy, vitamin K inhibitor/antagonist, prokinet-
ESGE recommends a restrictive red blood cell transfusion strategy that aims
ic agent, erythromycin, fresh frozen plasma, nasogastric tube, or- for a target hemoglobin between 7 g/dL and 9 g/dL. A higher target hemo-
ogastric tube, proton pump inhibitor, prokinetic agent, erythro- globin should be considered in patients with significant co-morbidity (e. g.,
ischemic cardiovascular disease) (strong recommendation, moderate quality
mycin, endoscopic hemostasis, injection therapy, thermal ther-
evidence).
apy (contact, non-contact), mechanical therapy/endoscopic clip-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
ping, topical hemostasis therapy, second-look endoscopy, helico- The use of red blood cell (RBC) transfusions may be lifesaving fol-
bacter pylori, H. pylori, transcatheter angiographic embolization lowing massive UGIH. However, the role of RBC transfusion in
(TAE), and surgery. The hierarchy of studies included as part of less torrential GI bleeding remains controversial, with uncertain-
this evidence-based guideline was, in decreasing order of evi- ty existing regarding the hemoglobin level at which blood trans-
dence level, published systematic reviews/meta-analyses, ran- fusion should be initiated. This uncertainty reflects concerns
domized controlled trials (RCTs), prospective and retrospective from both the critical care and gastroenterology literature sug-
observational studies. All selected articles were graded using the gesting poorer outcomes in patients managed with a liberal RBC
Grading of Recommendations Assessment, Development and transfusion strategy [2, 10, 11]. In a recent RCT that included 921
Evaluation (GRADE) system [3, 4]. patients presenting with all causes of acute UGIH, a restrictive
Each task force subgroup proposed statements for each of their RBC transfusion strategy (target hemoglobin, 7 to 9 g/dL) was
assigned key questions which were discussed and voted on dur- compared with a more liberal transfusion strategy (target hemo-
ing the NVUGIH task force guideline meeting held in Berlin, Ger- globin, 9 to 11 g/dL) [12]. The restrictive RBC transfusion group
many in November 2014. In August 2015, a manuscript draft pre- had significantly improved 6-week survival (95 % vs. 91 %; hazard
pared by I.M.G. was sent to all task force members. After agree- ratio [HR] 0.55, 95 % confidence interval [CI] 0.33 – 0.92) and re-
ment on a final version, the manuscript was reviewed by two duced rebleeding (10 % vs.16 %; HR 0.68, 95 %CI 0.47 – 0.98) [12].
members of the ESGE Governing Board and sent for further com- In the subgroup of patients with NVUGIH (n = 699), there was a
ments to the National Societies and ESGE individual members. statistical trend towards lower mortality in the restrictive vs. lib-
After agreement on a final version, the manuscript was submit- eral RBC transfusion strategy (3.7 % vs. 6.9 %, P = 0.065). Because
ted to the journal Endoscopy for publication. All authors agreed the study was not powered to specifically evaluate NVUGIH,
on the final revised manuscript. these findings should be interpreted with caution. Other limita-
This NVUGIH guideline will be considered for review and updat- tions of this study include the exclusion of patients with massive
ing in 2020, or sooner if new relevant evidence becomes avail- exsanguinating bleeding and defined co-morbidities. Further-
able. Any updates to this guideline in the interim will be noted more, all patients underwent endoscopy within 6 hours of pre-
on the ESGE website: http://www.esge.com/esge-guidelines. sentation, which may not be feasible in everyday clinical practice.
html. Coagulopathy at the time of NVUGIH presentation is another fre-
quent and adverse prognostic factor [13]. Published data for the
management of coagulopathy are limited and inconclusive. One
Statements and recommendations small cohort study using an historical comparison group showed
! that aggressive volume resuscitation, including correction of coa-
See ●
" Table 1. gulopathy (international normalized ratio [INR] < 1.8), led to an
improvement in mortality outcomes [5]. In a systematic review
Initial patient evaluation and hemodynamic that evaluated the relevance of initial INR before correction in pa-
resuscitation tients with NVUGIH, INR did not appear to predict rebleeding, yet
after adjusting for potential confounders, an initial INR > 1.5 pre-
dicted mortality (odds ratio [OR] 1.96, 95 %CI 1.13 – 3.41) [14].
ESGE recommends immediate assessment of hemodynamic status in patients
who present with acute upper gastrointestinal hemorrhage (UGIH), with This may in part reflect the presence of underlying liver disease.
prompt intravascular volume replacement initially using crystalloid fluids if There is however no available evidence to help guide coagulopa-
hemodynamic instability exists (strong recommendation, moderate quality
thy correction in critically ill patients and wide variation in man-
evidence).
agement exists in this area, indicating clinical uncertainty re-
The goals of hemodynamic resuscitation are to correct intravas- garding optimal practice [15]. Platelet count has not been shown
cular hypovolemia, restore adequate tissue perfusion, and pre- to be a predictor of either rebleeding or mortality. Currently,
vent multi-organ failure. Early intensive hemodynamic resuscita- there is no high quality evidence to guide platelet transfusion
tion of patients with acute UGIH has been shown to significantly thresholds, although a platelet transfusion threshold of 50 × 109/
decrease mortality [5]. In an observational study of patients with L has been proposed for most patients, with a target of 10 × 109/L
acute UGIH and hemodynamic instability, patients who received for patients in whom platelet dysfunction is suspected [16].
intensive hemodynamic resuscitation had significantly fewer
myocardial infarctions and lower mortality compared with those
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a4 Guideline
Table 1 Summary of Guideline statements and recommendations. Diagnosis and management of nonvariceal upper gastrointestinal hemorrhage: European
Society of Gastrointestinal Endoscopy (ESGE) Guideline.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Pre-endoscopy management
5 For patients taking vitamin K antagonists (VKAs), ESGE recommends withholding the VKA and correcting coagulopathy while taking into account the
patient's cardiovascular risk in consultation with a cardiologist. In patients with hemodynamic instability, administration of vitamin K, supplemented with
intravenous prothrombin complex concentrate (PCC) or fresh frozen plasma (FFP) if PCC is unavailable, is recommended (strong recommendation, low
quality evidence).
6 If the clinical situation allows, ESGE suggests an international normalized ratio (INR) value < 2.5 before performing endoscopy with or without endo-
scopic hemostasis (weak recommendation, moderate quality evidence).
7 ESGE recommends temporarily withholding new direct oral anticoagulants (DOACs) in patients with suspected acute NVUGIH in coordination/consul-
tation with the local hematologist/cardiologist (strong recommendation, very low quality evidence).
8 For patients using antiplatelet agents, ESGE recommends the management algorithm detailed in●
" Fig. 2 (strong recommendation, moderate quality
evidence).
9 ESGE recommends initiating high dose intravenous proton pump inhibitors (PPI), intravenous bolus followed by continuous infusion (80 mg then 8 mg/
hour), in patients presenting with acute UGIH awaiting upper endoscopy. However, PPI infusion should not delay the performance of early endoscopy
(strong recommendation, high quality evidence).
10 ESGE does not recommend the use of tranexamic acid in patients with NVUGIH (strong recommendation, low quality evidence).
11 ESGE does not recommend the use of somatostatin, or its analogue octreotide, in patients with NVUGIH (strong recommendation, low quality evi-
dence).
12 ESGE recommends intravenous erythromycin (single dose, 250 mg given 30 – 120 minutes prior to upper GI endoscopy) in patients with clinically se-
vere or ongoing active UGIH. In selected patients, pre-endoscopic infusion of erythromycin significantly improves endoscopic visualization, reduces the
need for second-look endoscopy, decreases the number of units of blood transfused, and reduces duration of hospital stay (strong recommendation, high
quality evidence).
13 ESGE does not recommend the routine use of nasogastric or orogastric aspiration/lavage in patients presenting with acute UGIH (strong recommen-
dation, moderate quality evidence).
14 In an effort to protect the patient's airway from potential aspiration of gastric contents, ESGE suggests endotracheal intubation prior to endoscopy in
patients with ongoing active hematemesis, encephalopathy, or agitation (weak recommendation, low quality evidence).
15 ESGE recommends adopting the following definitions regarding the timing of upper GI endoscopy in acute overt UGIH relative to patient presentation:
very early < 12 hours, early ≤ 24 hours, and delayed > 24 hours (strong recommendation, moderate quality evidence).
16 Following hemodynamic resuscitation, ESGE recommends early ( ≤ 24 hours) upper GI endoscopy. Very early ( < 12 hours) upper GI endoscopy may be
considered in patients with high risk clinical features, namely: hemodynamic instability (tachycardia, hypotension) that persists despite ongoing attempts
at volume resuscitation; in-hospital bloody emesis/nasogastric aspirate; or contraindication to the interruption of anticoagulation (strong recommenda-
tion, moderate quality evidence).
17 ESGE recommends the availability of both an on-call GI endoscopist proficient in endoscopic hemostasis and on-call nursing staff with technical ex-
pertise in the use of endoscopic devices to allow performance of endoscopy on a 24 /7 basis (strong recommendation, moderate quality evidence).
Endoscopic therapy (peptic ulcer bleeding)
18 ESGE recommends the Forrest (F) classification be used in all patients with peptic ulcer hemorrhage in order to differentiate low and high risk endo-
scopic stigmata (strong recommendation, high quality evidence).
19 ESGE recommends that peptic ulcers with spurting or oozing bleeding (Forrest classification Ia and Ib respectively), or with a nonbleeding visible vessel
(Forrest classification IIa) receive endoscopic hemostasis because these lesions are at high risk for persistent bleeding or rebleeding (strong recommen-
dation, high quality evidence).
20 ESGE recommends that peptic ulcers with an adherent clot (Forrest classification IIb) be considered for endoscopic clot removal. Once the clot is re-
moved, any identified underlying active bleeding (Forrest classification Ia or Ib) or nonbleeding visible vessel (Forrest classification IIa) should receive
endoscopic hemostasis (weak recommendation, moderate quality evidence).
21 In patients with peptic ulcers having a flat pigmented spot (Forrest classification IIc) or clean base (Forrest classification III), ESGE does not recommend
endoscopic hemostasis as these stigmata present a low risk of recurrent bleeding. In selected clinical settings, these patients may be discharged to home on
standard PPI therapy, e. g., oral PPI once-daily (strong recommendation, moderate quality evidence).
22 ESGE does not recommend the routine use of Doppler ultrasound or magnification endoscopy in the evaluation of endoscopic stigmata of peptic ulcer
bleeding (strong recommendation, low quality evidence).
23 For patients with actively bleeding ulcers (FIa, FIb), ESGE recommends combining epinephrine injection with a second hemostasis modality (contact
thermal, mechanical therapy, or injection of a sclerosing agent). ESGE recommends that epinephrine injection therapy not be used as endoscopic mono-
therapy (strong recommendation, high quality evidence).
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a5
Table 1 (Continuation)
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Transcatheter angiographic embolization (TAE) or surgery should be considered if endoscopic treatment fails or is not technically feasible (strong recom-
mendation, low quality evidence).
29 In patients bleeding from upper GI angioectasias, ESGE recommends endoscopic hemostasis therapy. However, there is currently inadequate evidence
to recommend a specific endoscopic hemostasis modality (strong recommendation, low quality evidence).
30 In patients bleeding from upper GI neoplasia, ESGE recommends considering endoscopic hemostasis in order to avert urgent surgery and reduce blood
transfusion requirements. However, no currently available endoscopic treatment appears to have long-term efficacy (weak recommendation, low quality
evidence).
Post endoscopy/endoscopic hemostasis management
31 ESGE recommends PPI therapy for patients who receive endoscopic hemostasis and for patients with adherent clot not receiving endoscopic hemo-
stasis. PPI therapy should be high dose and administered as an intravenous bolus followed by continuous infusion (80 mg then 8 mg /hour) for 72 hours post
endoscopy (strong recommendation, high quality evidence).
32 ESGE suggests considering PPI therapy as intermittent intravenous bolus dosing (at least twice-daily) for 72 hours post endoscopy for patients who
receive endoscopic hemostasis and for patients with adherent clot not receiving endoscopic hemostasis. If the patient’s condition permits, high dose oral
PPI may also be an option in those able to tolerate oral medications (weak recommendation, moderate quality evidence).
33 In patients with clinical evidence of rebleeding following successful initial endoscopic hemostasis, ESGE recommends repeat upper endoscopy with
hemostasis if indicated. In the case of failure of this second attempt at hemostasis, transcatheter angiographic embolization (TAE) or surgery should be
considered (strong recommendation, high quality evidence).
34 ESGE does not recommend routine second-look endoscopy as part of the management of NVUGIH. However, second-look endoscopy may be consid-
ered in selected patients at high risk for rebleeding (strong recommendation, high quality evidence).
35 In patients with NVUGIH secondary to peptic ulcer, ESGE recommends investigating for the presence of Helicobacter pylori in the acute setting with
initiation of appropriate antibiotic therapy when H. pylori is detected. Re-testing for H. pylori should be performed in those patients with a negative test in
the acute setting. Documentation of successful H. pylori eradication is recommended (strong recommendation, high quality evidence).
36 ESGE recommends restarting anticoagulant therapy following NVUGIH in patients with an indication for long-term anticoagulation. The timing for
resumption of anticoagulation should be assessed on a patient by patient basis. Resuming warfarin between 7 and 15 days following the bleeding event
appears safe and effective in preventing thromboembolic complications for most patients. Earlier resumption, within the first 7 days, may be indicated for
patients at high thrombotic risk (strong recommendation, moderate quality evidence).
37 In patients receiving low dose aspirin for primary cardiovascular prophylaxis who develop peptic ulcer bleeding, ESGE recommends withholding as-
pirin, re-evaluating the risks/benefits of ongoing aspirin use in consultation with a cardiologist, and resuming low dose aspirin following ulcer healing or
earlier if clinically indicated (strong recommendation, low quality evidence).
38 In patients receiving low dose aspirin for secondary cardiovascular prophylaxis who develop peptic ulcer bleeding, ESGE recommends aspirin be re-
sumed immediately following index endoscopy if the risk of rebleeding is low (e. g., FIIc, FIII). In patients with high risk peptic ulcer (FIa, FIb, FIIa, FIIb), early
reintroduction of aspirin by day 3 after index endoscopy is recommended, provided that adequate hemostasis has been established (strong recommen-
dation, moderate quality evidence).
39 In patients receiving dual antiplatelet therapy (DAPT) who develop peptic ulcer bleeding, ESGE recommends continuing low dose aspirin therapy. Early
cardiology consultation should be obtained regarding the timing of resuming the second antiplatelet agent (strong recommendation, low quality evi-
dence).
40 In patients requiring dual antiplatelet therapy (DAPT) and who have had NVUGIH, ESGE recommends the use of a PPI as co-therapy (strong recom-
mendation, moderate quality evidence).
endoscopy risk stratification. Outpatients determined to be at very low risk, most widely evaluated and adopted [17 – 19]. However, no single
based upon a GBS score of 0 – 1, do not require early endoscopy nor hospital
scoring tool has been shown to excel at predicting all relevant
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a6 Guideline
Pre-endoscopy management
Table 2 Glasgow-Blatchford Score (GBS).
Initial management of antithrombotic agents
Points (anticoagulants and antiplatelet agents)
Systolic blood pressure, mmHg
100 – 109 1 For patients taking vitamin K antagonists (VKAs), ESGE recommends with-
90 – 99 2 holding the VKA and correcting coagulopathy while taking into account the
< 90 3 patient’s cardiovascular risk in consultation with a cardiologist. In patients
Blood urea nitrogen, mmol/L with hemodynamic instability, administration of vitamin K, supplemented
with intravenous prothrombin complex concentrate (PCC) or fresh frozen
6.5 – 7.9 2
plasma (FFP) if PCC is unavailable, is recommended (strong recommendation,
8.0 – 9.9 3 low quality evidence).
10.0 – 24.9 4
≥ 25.0 6
Hemoglobin for men, g/dL If the clinical situation allows, ESGE suggests an international normalized ratio
12.0 – 12.9 1 (INR) value < 2.5 before performing endoscopy with or without endoscopic
hemostasis (weak recommendation, moderate quality evidence).
10.0 – 11.9 3
< 10.0 6
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Hemoglobin for women, g/dL GI bleeding represents a serious complication of VKA therapy,
10.0 – 11.9 1 with an incidence of 1 % – 4 % per year [29, 30]. Discontinuation
< 10.0 6 of anticoagulants and correction of coagulopathy before endos-
Other risk variables copy is the “standard of practice” in patients with clinically sig-
Pulse ≥ 100 1 nificant GI bleeding [31 – 33]. Because data are limited, specific
Melena 1 strategies to reverse VKAs in a patient with acute overt UGIH
Syncope 2 vary [34]. Practice guidelines recommend urgent reversal in all
Hepatic disease 2 patients presenting with serious, life-threatening bleeding (i. e.,
Cardiac failure 2
hemodynamic instability or shock), either in the case of thera-
TOTAL GBS __________________ peutic or supratherapeutic INR elevations [32, 35]. For patients
GBS restricted for use only in nonhospitalized, ambulatory patients
Risk variables measured at time of patient presentation
who are not actively bleeding and are hemodynamically stable,
GBS = 0 – 1 denotes “low-risk” intravenous vitamin K administration may be an option. When
more urgent reversal is required, administration of prothrombin
complex concentrates (PCCs) or fresh frozen plasma (FFP) is nec-
outcomes in acute UGIH (e. g., rebleeding, need for intervention, essary, with concomitant intravenous administration of 5 – 10 mg
mortality) [19]. This is not surprising as the most validated risk vitamin K to prevent “rebound coagulopathy” once the trans-
scores were derived to assess a specific UGIH outcome: that for fused factors have been cleared. Prothrombin complex concen-
the Rockall score being mortality and for the GBS being the need trates contain clotting factors prepared from pooled and concen-
for intervention [17, 18]. trated human plasma and are preferred over FFP because of sev-
A recent systematic review evaluating the accuracy of the avail- eral advantages, including no need to check the patient’s blood
able UGIH risk stratification tools demonstrated substantial het- group, less risk for volume overload because of smaller transfu-
erogeneity in predicted outcomes and highlighted that methodo- sion volume, faster onset of action, similar thrombotic risk pro-
logical quality of the prediction scores was less than optimal [19]. file, and minimal risk of infectious transmission, albeit at a higher
Regarding the need for intervention, retrospective and prospec- cost [36 – 40]. A recent prospective, nonrandomized, comparative
tive studies have assessed the prognostic value of the GBS vs. the study of 40 warfarin users who presented with UGIH and an INR
Rockall score. These studies showed that the GBS correctly iden- > 2.1 reported that patients who received PCC had a near normal-
tified 98 % (95 %CI 89 % – 100 %) of those patients who did not re- ized INR at 2 hours following infusion (INR = 1.5) while those who
quire any subsequent intervention while 83 % (95 %CI 71 % – 91 %) received FFP had an INR of 2.4 at 6 hours following infusion [38].
of those patients were identified using the Rockall score. Ran- No patient in the PCC group had active bleeding at endoscopy
domized controlled trials and observational studies consistently compared with 7 in the FFP group (0 vs. 35 %, P < 0.01). The risk
indicate that clinical, endoscopic, and social factors may identify of thrombosis following PCC administration approximates 1 %,
patients who may be safely discharged for outpatient manage- and is similar to that reported with FFP [39, 40].
ment [20 – 28]. The most frequent adverse event reported is re-
bleeding ranging between 0.5 % and 4 %, with no deaths or hospi-
ESGE recommends temporarily withholding new direct oral anticoagulants
tal readmissions for surgery reported. Moreover, studies consis- (DOACs) in patients with suspected acute NVUGIH in coordination/consulta-
tently indicate that outpatient management of appropriately se- tion with the local hematologist/cardiologist (strong recommendation, very
low quality evidence).
lected patients with acute UGIH reduces resource utilization [20,
21, 27]. Emergency department discharge without inpatient As an alternative to heparin and VKAs, the new non-VKA oral an-
endoscopy (i. e., outpatient management) should be considered ticoagulants (NOACs; also referred to as direct oral anticoagulants
for patients if: systolic blood pressure ≥ 110 mmHg, pulse < 100 [DOACs]) are being rapidly adopted worldwide, primarily for
beats/minute, hemoglobin ≥ 13.0 g/dL for men or ≥ 12.0 g/dL for thromboembolic prevention in patients with nonvalvular atrial
women, blood urea nitrogen < 18.2 mg/dL, along with the absence fibrillation and for prophylaxis or treatment of venous throm-
of melena, syncope, hepatic disease, and cardiac failure [18]. (See boembolism [41]. These pharmacological agents do however,
Appendix e2, online-only.) present a risk of significant GI bleeding similar to or greater than
that reported with warfarin [42, 43]. Moreover, DOACs differ in
comparison with heparin and VKA. Specifically, in the absence
of renal or hepatic failure, DOAC clearance and the subsequent
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a7
Acute upper GI hemorrhage in a patient using antiplatelet agent(s) (APA) Fig. 1 Algorithm for the management of patients
with acute upper gastrointestinal hemorrhage who
are using antiplatelet agent(s): European Society of
Upper GI endoscopy demonstrates a nonvariceal source of bleeding (e.g. peptic ulcer bleed) Gastrointestinal Endoscopy (ESGE) Guideline.
High risk endoscopic stigmata identified Low risk endoscopic stigmata identified
(FIa, FIb, FIIa, FIIb) (FIIc, FIII)
APA used for primary prophylaxis APA used for primary prophylaxis
▪ Withhold low dose acetylsalicylic acid ▪ Withhold low dose ASA
(ASA) ▪ Re-evaluate risks and benefits of
▪ Re-evaluate risks and benefits of ongoing low dose ASA use
ongoing low dose ASA use ▪ Resume low dose ASA at hospital
▪ Resume low dose ASA after ulcer discharge if clinically indicated
healing or earlier if clinically indicated
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
APA used for secondary prophylaxis APA used for secondary prophylaxis
(known cardiovascular disease) (known cardiovascular disease)
1 Patients on low dose ASA alone 1 Patients on low dose ASA alone
▪ Resume low dose ASA by day 3 following ▪ Continue low dose ASA without interruption
index endoscopy 2 Patients on dual antiplatelet therapy (DAPT)
▪ Second-look endoscopy at the discretion of
▪ Continue DAPT without interruption
the endoscopist may be considered
2 Patients on dual antiplatelet therapy (DAPT) For patients using a non-ASA APA as monotherapy
▪ Continue low dose ASA without interruption (e.g., thienopyridine alone), low-dose ASA may be
▪ Early cardiology consultation for given as substitute for interval period in patients
recommendation on resumption/ with no contraindication or allergy to ASA.
continuation of second APA Early cardiology consultation should be obtained
▪ Second-look endoscopy at the discretion of for further APA recommendations.
the endoscopist may be considered
loss of anticoagulation effect is rapid and predictable (occurring antagonist that inhibits thrombin. The minimum duration of an-
gradually over 12 – 24 hours), routine laboratory tests are not tiplatelet agent discontinuation that allows for restoration of nor-
sensitive for the quantitative assessment of their anticoagulant mal platelet aggregation is 5 – 7 days [52].
activity, and there is currently no specific reversal agent/antidote Studies have shown that in patients taking low dose aspirin for
for emergency use with any DOAC, although potential agents are secondary cardiovascular prophylaxis, all-cause mortality was
in development and may be commercially available in the next lower if aspirin was not discontinued following peptic ulcer
1 – 2 years [44 – 46]. As there are no published clinical trials ad- bleeding [53, 54]. In an RCT, 156 recipients of low dose aspirin
dressing the management of GI bleeding in patients using DOAC, for secondary prophylaxis who had peptic ulcer bleeding were
current recommendations are based on expert opinion or labora- randomized to receive continuous aspirin or placebo [53]. At 8-
tory end-points [47 – 49]. week follow up, all-cause mortality was lower in the patients
At the time of patient presentation with acute UGIH, DOACs randomized to aspirin compared with placebo (1.3 % vs. 12.9 %,
should be temporarily withheld. Given their relatively short 95 %CI 3.7 % – 19.5 %; hazard ratio [HR] 0.20), with the difference
half-life, time is the most important antidote against DOACs. being attributable to cardiovascular, cerebrovascular, or GI com-
Strategies to accelerate anticoagulation reversal are supported plications. The 30-day ulcer rebleeding rate was not significantly
only by data collected from healthy human volunteers, animal greater in the aspirin group. Patients who required dual antipla-
models, and in vitro studies [50]. Based on those data, vitamin K telet therapy (DAPT) were excluded from this study. In a subse-
or FFP have no place as reversal agents for DOACs. Prothrombin quent retrospective analysis that included 118 low dose aspirin
complex concentrates or activated PCC may be considered in pa- recipients who had been treated for peptic ulcer bleeding and fol-
tients with severe or life-threatening bleeding, and hemodialysis lowed-up for a median of 2 years, 47 (40 %) patients stopped as-
can be used to reduce the blood concentration of dabigatran, but pirin [54]. Patients who discontinued aspirin and those who con-
not that of rivaroxaban and apixaban which are more tightly tinued aspirin had similar mortality rates (31 %). However, in a
bound to plasma proteins [48, 49, 51]. Additional data on the clin- subgroup analysis limited to patients with cardiovascular co-
ical effectiveness of these strategies in acutely bleeding patients morbidities, those patients who discontinued aspirin had an al-
are urgently needed. most fourfold increase in the risk of death or acute cardiovascular
event (P < 0.01) [54]. Randomized controlled trials have shown
that neither aspirin nor clopidogrel use impede ulcer healing
For patients using antiplatelet agents, ESGE recommends the management
algorithm detailed in●" Fig. 1 (strong recommendation, moderate quality promoted by proton pump inhibitors (PPI) [55, 56].
evidence).
Pharmacological therapy
Antiplatelet agents include low dose aspirin and thienopyridines
(e. g., clopidogrel, prasugrel, ticlopidine) that irreversibly inhibit
ESGE recommends initiating high dose intravenous proton pump inhibitors
platelet aggregation, ticagrelor a reversible P2Y12 receptor an- (PPI), intravenous bolus followed by continuous infusion (80 mg then 8 mg/
tagonist, and vorapaxar, a protease-activated receptor (PAR-1) hour), in patients presenting with acute UGIH awaiting upper endoscopy.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a8 Guideline
However, PPI infusion should not delay the performance of early endoscopy mucosa visualization (OR 3.43, 95 %CI 1.81 – 6.50; P < 0.01), and
(strong recommendation, high quality evidence).
decreased the need for second-look endoscopy (OR 0.47, 95 %CI
A Cochrane meta-analysis of 6 RCTs (n = 2223 patients) showed 0.26−0.83, P = 0.01), RBC units transfused (weighted mean differ-
that administering PPIs before endoscopy significantly decreases ence −0.41, 95 %CI −0.82 to −0.01, P = 0.04), and duration of hospi-
the incidence of high risk stigmata of hemorrhage at the time of tal stay (weighted mean difference −1.51 days, 95 %CI −2.45 to
index endoscopy (37.2 % vs. 46.5 %; OR 0.67, 95 %CI 0.54 – 0.84) −0.56, P < 0.01) [68].
and the need for endoscopic hemostasis (8.6 % vs. 11.7 %; OR A single intravenous dose of erythromycin is safe and generally
0.68, 95 %CI 0.50 – 0.93), but has no effect on rebleeding, need for well tolerated, with no adverse events reported in the meta-ana-
surgery, or mortality [57]. lyses. Studies that found a significant improvement in endoscopic
Cost– effectiveness studies suggest that high dose PPI infusion visualization with pre-endoscopic erythromycin infusion includ-
prior to endoscopy for patients with UGIH is more effective and ed patients admitted to the intensive care unit because of UGIH
less costly than placebo [58, 59]. (See Appendix e3, online-only.) with clinical evidence of active bleeding or hematemesis or blood
seen on nasogastric lavage. These patients are most likely to ben-
efit from erythromycin infusion prior to endoscopy. The dose of
ESGE does not recommend the use of tranexamic acid in patients with NVU-
GIH (strong recommendation, low quality evidence). erythromycin most commonly used is 250 mg and is infused 30
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
to 120 minutes prior to upper GI endoscopy. A cost– effectiveness
Tranexamic acid reduces clot breakdown by inhibiting the fibri- study found that pre-endoscopy erythromycin infusion in UGIH
nolytic action of plasmin. A recent RCT demonstrated that tra- was cost-effective, primarily due to a reduction in the need for
nexamic acid significantly reduces bleeding-related and all-cause second-look endoscopies [69]. Contraindications to erythromy-
mortality in trauma patients with significant hemorrhage [60]. A cin administration include sensitivity to macrolide antibiotics
Cochrane meta-analysis evaluating the use of tranexamic acid in and prolonged QT interval.
1654 UGIH patients showed a beneficial effect of tranexamic acid Metoclopramide has been less studied, it has been assigned a
on mortality when compared with placebo (relative risk [RR] “black box warning” by the United States Food and Drug Admin-
0.61, 95 %CI 0.42 – 0.89), but not on other patient outcomes in- istration because of the risk of neurologic side effects, and cau-
cluding bleeding, surgery, or transfusion requirements [61]. tion should therefore be advised with the use of this prokinetic
However, the beneficial effect on mortality did not persist in sub- agent.
group analysis. The studies included in this meta-analysis have (See Appendix e6, online-only.)
important limitations that affect their generalizability including
their methodological quality and the fact that the majority were Role of gastric lavage and prophylactic endotracheal
conducted before the widespread use of therapeutic endoscopy intubation
and PPIs. To date, no controlled trial assessing the role of alterna-
tive antifibrinolytic agents (e. g., aminocaproic acid, aprotinin) in
ESGE does not recommend the routine use of nasogastric or orogastric as-
patients with acute UGIH has been reported. (See Appendix e4, piration/lavage in patients presenting with acute UGIH (strong recommenda-
online-only.) tion, moderate quality evidence).
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a9
pared to those patients not intubated, published data show no without hospital admission facilitates discharge in up to 46 % of
significant difference in patient outcomes (e. g., pulmonary as- patients and reduces costs/resource utilization [20, 85]. Dischar-
piration, in-hospital mortality) [75 – 77]. One study suggested ging low risk suspected NVUGIH patients (GBS = 0) directly from
that aspiration was actually more frequent in those patients the emergency department without undergoing upper GI endos-
who had undergone endotracheal intubation prior to upper GI copy has been proposed as a safe and cost-saving option in multi-
endoscopy [75]. At this time, endotracheal intubation prior to ple studies in various clinical settings [18, 86 – 89]. Some investi-
upper GI endoscopy in patients with UGIH does not seem to gators have suggested that using a GBS ≤ 1 (see ● " Table 2) could
make a difference in patient outcome but published data are lim- double the number of patients eligible for ambulatory manage-
ited with small numbers of subjects and low methodological ment while maintaining safety [89].
quality. There are four published studies, one RCT and three prospective
case series, that have evaluated the test characteristics and accu-
Timing of endoscopy racy parameters of video capsule endoscopy (VCE) in risk stratifi-
cation of patients presenting with acute UGIH [90 – 93]. The over-
all sensitivity, specificity, positive predictive value, and negative
ESGE recommends adopting the following definitions regarding the timing of
upper GI endoscopy in acute overt UGIH relative to patient presentation: very predictive value of VCE for detecting blood in the upper GI tract
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
early < 12 hours, early ≤ 24 hours, and delayed > 24 hours (strong recommen- in patients suspected of acute UGIH are 75 %, 76 %, 67 %, and 82 %
dation, moderate quality evidence).
respectively. Because the data are limited, at this time there is no
role for VCE in the emergency department setting in evaluating
Following hemodynamic resuscitation, ESGE recommends early ( ≤ 24 hours) acute upper GIH. However, additional studies are needed to fur-
upper GI endoscopy. Very early ( < 12 hours) upper GI endoscopy may be ther assess VCE in this patient population since, for low to mod-
considered in patients with high risk clinical features, namely: hemodynamic
erate risk UGIH patients, VCE may be a cost-effective modality if
instability (tachycardia, hypotension) that persists despite ongoing attempts
at volume resuscitation; in-hospital bloody emesis/nasogastric aspirate; or post-VCE low risk patients are discharged directly home from the
contraindication to the interruption of anticoagulation (strong recommenda- emergency department and hospital admission is avoided
tion, moderate quality evidence).
[94, 95].
ESGE recommends the availability of both an on-call GI endoscopist proficient Endoscopic management
in endoscopic hemostasis and on-call nursing staff with technical expertise in Endoscopic diagnosis
the use of endoscopic devices to allow performance of endoscopy on a 24 /7
basis (strong recommendation, moderate quality evidence).
ESGE recommends the Forrest (F) classification be used in all patients with
Performance of upper GI endoscopy within 24 hours of patient peptic ulcer hemorrhage in order to differentiate low and high risk endoscopic
presentation with suspected NVUGIH and no contraindication to stigmata (strong recommendation, high quality evidence).
endoscopy has been proposed as a key quality indicator in the
management of upper GI bleeding [78]. In a large European ob-
ESGE recommends that peptic ulcers with spurting or oozing bleeding (Forr-
servational study that included 123 centers in 7 countries, there est classification Ia and Ib, respectively) or with a nonbleeding visible vessel
was wide variation in practice where anywhere from 70 % to 93 % (Forrest classification IIa) receive endoscopic hemostasis because these le-
sions are at high risk for persistent bleeding or rebleeding (strong recom-
of 2660 unselected patients with UGIH underwent upper endos-
mendation, high quality evidence).
copy within 24 hours of hospital admission [79].
Two systematic reviews evaluating the timing of upper GI endos-
copy demonstrated improved risk assessment and reduction in ESGE recommends that peptic ulcers with an adherent clot (Forrest classifica-
hospital length of stay if endoscopy was performed within 24 tion IIb) be considered for endoscopic clot removal. Once the clot is removed,
any identified underlying active bleeding (Forrest classification Ia or Ib) or
hours of patient presentation, yet the impact on need for surgery nonbleeding visible vessel (Forrest classification IIa) should receive endo-
and in-hospital mortality was variable [80, 81]. More recently, a scopic hemostasis (weak recommendation, moderate quality evidence).
retrospective analysis of risk factors for mortality in more than
400 000 patients with NVUGIH found an increased mortality in
In patients with peptic ulcers having a flat pigmented spot (Forrest classifica-
patients who failed to receive upper endoscopy within 1 day of tion IIc) or clean base (Forrest classification III), ESGE does not recommend
hospital admission (OR 1.32, 95 %CI 1.26 – 1.38) [82]. (See Appen- endoscopic hemostasis as these stigmata present a low risk of recurrent
dix e7, online-only.) bleeding. In selected clinical settings, these patients may be discharged to
home on standard PPI therapy, e. g., oral PPI once-daily (strong recommen-
With respect to very early upper GI endoscopy, an RCT that in- dation, moderate quality evidence).
cluded 325 patients with peptic ulcer bleeding showed that up-
per GI endoscopy performed within 12 hours of admission (as The Forrest (F) classification was developed more than 40 years
compared with 12 – 24 hours) resulted in a significant reduction ago in an attempt to standardize the characterization of peptic
in transfusion requirements in patients with bloody nasogastric ulcers [96]. The Forrest classification is defined as follows: FIa
lavage (P < 0.001). No such reduction was observed in patients spurting hemorrhage, FIb oozing hemorrhage, FIIa nonbleeding
with “coffee grounds” or clear lavage [83]. A retrospective analy- visible vessel, FIIb an adherent clot, FIIc flat pigmented spot, and
sis that included 934 UGIH patients showed that in the subset of FIII clean base ulcer [97 – 99]. This classification has been used in
patients having a GBS ≥ 12 (n = 97, 10.4 %), the time lapse between numerous studies that aimed to identify patients at risk of per-
presentation to endoscopy was the lone independent risk factor sistent ulcer bleeding, rebleeding and mortality. Most of these
associated with all-cause in-hospital mortality [84]. In this study, studies have shown that the presence of an ulcer endoscopically
a cutoff time of 13 hours in delay to endoscopy best discrimina- classified as FIa or FIb is an independent risk factor for persistent
ted between patient survival and nonsurvival. bleeding or rebleeding [100 – 107]. A potential limitation of the
In patients who are hemodynamically stable and without serious Forrest classification is that stigmata recognition and identifica-
co-morbidities, RCTs have shown that performing endoscopy
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a10 Guideline
Perform endoscopic hemostasis Consider performing clot removal followed No endoscopic hemostasis required
by endoscopic hemostasis of underlying In select clinical settings, these patients may
high risk stigmata2 have expedited hospital discharge
OR
Medical management with high dose
intravenous PPI
For FIa and FIb stigmata For FIIa stigmata – Once daily oral PPI
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Combination endoscopic therapy using Contact thermal4, mechanical, or – Start regular diet
dilute epinephrine injection plus injection of a sclerosant can be used – Test for Helicobacter pylori, treat if positive
a second hemostasis modality (contact alone as monotherapy or in combination – Document H. pylori eradication
thermal3, mechanical, or sclerosant4) with dilute epinephrine injection
▪ High dose intravenous PPI given as bolus + continuous infusion; If endoscopic hemostasis performed:
can consider intermittent intravenous bolus dosing (minimum ▪ Dilute epinephrine injection circumferential to base of clot
twice-daily) for 72 hours5 followed by clot removal using cold polyp snare guillotine technique
▪ May start clear liquids soon after endoscopy ▪ If underlying high risk stigmata identified after clot removal,
▪ Test for H. pylori, treat if positive apply endoscopic hemostasis as described for FIa, FIb, FIIa stigmata
▪ Document H. pylori eradication
If clinical evidence of ulcer rebleeding, repeat upper endoscopy with endoscopic hemostasis where indicated
If hemostasis not achieved or recurrent rebleeding following second attempt at endoscopic hemostasis,
▪ Consider endoscopic salvage therapy with topical hemostatic spray/over-the-scope clip
▪ Or refer for transcatheter angiographic embolization (TAE) or surgery
Fig. 2 Algorithm for the endoscopic management of patients with nonvariceal upper gastrointestinal hemorrhage (NVUGIH) secondary to peptic ulcer, stra-
tified by endoscopic stigmata: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. GI, gastrointestinal; PPI, proton pump inhibitor.
1
Use of a large single-channel or double-channel therapeutic upper GI endoscope is recommended.
2
The benefit of endoscopic hemostasis may be greater in patients at higher risk for rebleeding, e. g., older age, co-morbidities, in-hospital UGIH.
3
Large size 10-Fr probe recommended.
4
Absolute alcohol, polidocanol, or ethanolamine injected in limited volumes.
5
High dose oral PPI may be an option in those able to tolerate oral medications.
tion, as well as interobserver agreement, may be less than opti- recurrent ulcer bleeding and need for surgery in patients with
mal, although the data are conflicting [108, 109]. FIIa and FIIb ulcers [113, 114].
In addition to the Forrest classification, there are other endo- The indication for endoscopic treatment of FIIb ulcers (adherent
scopic features of peptic ulcers that can predict adverse outcomes clot) remains controversial because of conflicting data. In evalua-
and/or endoscopic treatment failure. These include large-size ul- tion of the natural history of FIIb ulcers (that did not receive
cer (> 2 cm), large-size nonbleeding visible vessel, presence of endoscopic hemostasis), it was found that 25 % of patients re-
blood in the gastric lumen, and ulcer location on the posterior bled within 30 days of follow-up [115]. In patients with FIIb ul-
duodenal wall or the proximal lesser curvature of the stomach cers, RCTs and a meta-analysis comparing medical therapy alone
[100, 101, 103, 105, 110, 111]. with endoscopic hemostasis demonstrated a significant advan-
A meta-analysis of RCTs that evaluated endoscopic hemostasis vs. tage for endoscopic hemostasis in reducing ulcer rebleeding
no endoscopic hemostasis demonstrated that endoscopic hemo- (8.2 % vs. 24.7 %, P < 0.01, yet there was no difference in need for
stasis was effective in preventing persistent or recurrent bleeding surgery or mortality [116 – 118]. In contrast, in a separate RCT,
in actively bleeding ulcers (FIa, FIb: RR 0.29, 95 %CI 0.20 – 0.43; Sung and colleagues reported no ulcer rebleeding in those pa-
number needed to treat [NNT] 2, 95 %CI 2 – 2) as well as in ulcers tients with adherent clots who received medical therapy alone;
with a nonbleeding visible vessel (FIIa: RR 0.49, 95 %CI 0.40 – however the numbers of such patients in the trial were quite
0.59; NNT 5, 95 %CI 4 – 6) [112]. limited (n = 24) [113]. Moreover, a meta-analysis restricted only
●" Fig. 2 presents an algorithm for the endoscopic management of to RCTs showed no benefit for endoscopic hemostasis in patients
bleeding peptic ulcer, stratified by endoscopic stigmata. with an adherent clot (RR 0.31, 95 %CI 0.06 – 1.77) [112].
With respect to the incremental benefit of acid suppression in In patients with peptic ulcers having a flat pigmented spot (FIIc)
addition to endoscopic hemostasis, an RCT and a subsequent or clean base (FIII), rebleeding is rare and therefore endoscopic
meta-analysis found a clear advantage for endoscopic hemostasis hemostasis does not provide a significant advantage [97 – 99].
combined with PPI therapy over PPI therapy alone in preventing
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a11
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
that FIIa ulcers can be classified as low risk or high risk and that hemostasis (RR 1.00, 95 %CI 0.77 – 1.31) [137]. However, a sepa-
some visible vessels classified as low risk using conventional rate meta-analysis [126] found through-the-scope clips to be sig-
endoscopy can be reclassified as high risk using magnification nificantly more effective than thermal therapy in reducing the
endoscopy [125]. However, the classification used has not been risk of recurrent bleeding (OR 0.24, 95 %CI 0.06 – 0.95). Two small
validated and no clinical benefit of this approach has been dem- studies from Japan compared the efficacy of clips versus hemo-
onstrated. static forceps [143, 144]. The first was an RCT conducted in 96 pa-
tients with high risk bleeding gastric ulcers and showed that use
Endoscopic therapy of monopolar, soft coagulation hemostatic forceps was as effec-
tive as clipping [143]. The second was an observational prospec-
tive cohort study on 50 patients in which use of bipolar hemo-
For patients with actively bleeding ulcers (FIa, FIb), ESGE recommends com-
bining epinephrine injection with a second hemostasis modality (contact static forceps was more effective than endoscopic clipping for
thermal, mechanical therapy, or injection of a sclerosing agent). ESGE re- both initial hemostasis (100 % vs. 78.2 %) and preventing recur-
commends that epinephrine injection therapy not be used as endoscopic
rent bleeding (3.7 % vs. 22.2 %) [144]. Unlike thermal therapies
monotherapy (strong recommendation, high quality evidence).
and sclerosing agents, mechanical therapy using clips has the
theoretical benefit of inducing only limited tissue injury, and
For patients with nonbleeding visible vessel (FIIa), ESGE recommends me- therefore may be preferred in patients on antithrombotic therapy
chanical therapy, thermal therapy, or injection of a sclerosing agent as
and those patients undergoing repeat endoscopic hemostasis for
monotherapy or in combination with epinephrine injection. ESGE recom-
mends that epinephrine injection therapy not be used as endoscopic mono- rebleeding. A multidisciplinary expert panel developed an expli-
therapy (strong recommendation, high quality evidence). cit set of evidence-based quality indicators for NVUGIH [78].
Among them, it was felt that patients with ulcer-related bleeding
with high risk stigmata and elevated INR (> 1.5 – 2.0), should re-
For patients with active NVUGIH bleeding not controlled by standard endo-
scopic hemostasis therapies, ESGE suggests the use of a topical hemostatic ceive endoscopic hemostasis using endoscopic clips or a combi-
spray or over-the-scope clip as salvage endoscopic therapy (weak recom- nation of epinephrine injection plus clips.
mendation, low quality evidence). Meta-analyses have shown that combination endoscopic hemo-
Endoscopic hemostasis can be achieved using injection, thermal, stasis therapy (dilute epinephrine injection combined with a sec-
and mechanical modalities (see Box 1), and any endoscopic ther- ond hemostasis modality including injectable, thermal contact
apy is superior to pharmacotherapy in patients with FIa, FIb and probe, or clips) is superior to injection therapy alone, but not to
FIIa ulcers [112, 126]. Meta-analyses show that thermal devices clips or contact thermal therapy alone [126, 139]. There may be
(contact and noncontact), injectable agents other than epine- practical reasons to pre-inject dilute epinephrine before other
phrine (i. e., sclerosing agents, thrombin/fibrin glue), and clips therapies for high risk endoscopic stigmata. Injection of epine-
are all effective methods for achieving hemostasis, with no single phrine may slow or stop bleeding allowing improved visualiza-
modality being superior [112, 126, 137 – 141]. tion for application of subsequent therapy. Adverse events asso-
Epinephrine injection therapy is effective at achieving primary ciated with combination endoscopic hemostasis are low and in-
hemostasis, but inferior to other endoscopic hemostasis mono- clude induction of bleeding (1.7 %) and perforation (0.6 %) [139].
therapies or combination therapy in preventing ulcer rebleeding Recent international consensus guidelines endorse combination
[112, 126, 139]. In the most recently published meta-analysis (19 therapy (dilute epinephrine injection combined with contact
RCTs, 2033 patients), epinephrine plus any second hemostasis thermal therapy, clips, or injection of a sclerosant [e. g., absolute
modality significantly reduced rebleeding (OR 0.53, 95 %CI ethanol]) as appropriate treatment in patients with peptic ulcer
0.35 – 0.81) and emergency surgery (OR 0.68, 95 %CI 0.50 – 0.93) bleeding with high risk endoscopic stigmata [98, 99, 145].
but not mortality as compared with epinephrine injection mono- New endoscopic hemostasis modalities (topical hemostatic
therapy for high risk peptic ulcers [140]. Therefore, it is recom- sprays and over-the-scope clips) are emerging as possible alter-
mended that if epinephrine is used to treat peptic ulcer bleeding native endotherapies for primary hemostasis when bleeding is
with high risk stigmata, it should only be used in combination refractory or not amenable to standard endoscopic hemostasis
with a second endoscopic hemostasis modality [97 – 99, 141]. therapies [136, 146]. Moreover, several small retrospective stud-
With respect to contact thermal therapy (e. g., bipolar electrocoa- ies have reported that an over-the-scope clip (OVESCO), may
gulation, heater probe), a meta-analysis restricted only to RCTs have a role as rescue hemostasis therapy for severe NVUGIH
found that contact thermal therapy was significantly more effec- when conventional endoscopic treatment modalities fail [133,
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a12 Guideline
134, 147]. An inert nanopowder (Hemospray) that causes im- studies [164, 167, 168]. Mechanical therapy appears to be safe, yet
mediate hemostasis when sprayed onto active bleeding [136, data are insufficient to make a clear recommendation of one he-
148] has recently been used as a primary hemostasis agent or as mostasis modality over another [164, 167, 169, 170].
a second-line salvage therapy. Several prospective uncontrolled The proximal stomach and duodenum are the most common lo-
studies, a large European registry [149 – 154] and a systematic re- cations for Dieulafoy lesions [171]. Endoscopic hemostasis is war-
view of the current limited data suggests that Hemospray is safe ranted if technically feasible. Observational studies have reported
and effective and may be best used in high risk cases as a tempor- the superiority of combined, thermal and mechanical methods
izing measure or a bridge toward more definitive treatment over injection monotherapy, in achieving primary hemostasis,
[136]. Other topical agents, such as the starch-derived polysac- preventing rebleeding, and in reducing the need for rescue ther-
charide hemostatic system (EndoClot) and the Ankaferd blood apy, yet with no proven mortality benefit [172 – 180]. All endo-
stopper are also emerging [136]. However, RCTs directly compar- scopic hemostasis modalities (e. g., band ligation, through-the-
ing topical agents with traditional hemostasis methods are re- scope clips, over-the-scope clips, contact thermal coagulation,
quired to better define their optimal role and safety in the endo- and argon plasma coagulation) appear safe and have similar re-
scopic management of NVUGIH. ported outcomes [171 – 180]. Selective TAE has been described
as an effective rescue therapy if endoscopic hemostasis fails or
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
in patients who are poor surgical candidates [181, 182]. If both
For patients with acid-related causes of NVUGIH different from peptic ulcers
(e. g., erosive esophagitis, gastritis, duodenitis), ESGE recommends treatment endoscopic and angiographic therapies fail, surgery should be
with high dose PPI. Endoscopic hemostasis is usually not required and select- considered.
ed patients may be discharged early (strong recommendation, low quality
Studies on endoscopic hemostasis therapy of angioectasias of the
evidence).
upper GI tract are observational and include only a limited num-
ber of subjects. In two recent meta-analyses, endoscopic hemo-
ESGE recommends that patients with a Mallory – Weiss lesion that is actively stasis therapy (e. g., argon plasma coagulation, heater probe, bi-
bleeding receive endoscopic hemostasis. There is currently inadequate evi-
polar coagulation, monopolar coagulation, band ligation, YAG la-
dence to recommend a specific endoscopic hemostasis modality. Patients
with a Mallory – Weiss lesion and no active bleeding can receive high dose PPI ser) is reported to be initially effective and safe, yet bleeding re-
therapy alone (strong recommendation, moderate quality evidence). currence rates are significant [183, 184]. Given the low quality of
evidence and scarcity of comparative data, a recommendation on
a specific endoscopic hemostasis treatment is not permitted at
ESGE recommends that a Dieulafoy lesion receive endoscopic hemostasis
using thermal, mechanical (hemoclip or band ligation), or combination ther- this time.
apy (dilute epinephrine injection combined with contact thermal or mechan- There are limited published data on the role of endoscopic hemo-
ical therapy) (strong recommendation, moderate quality evidence). Trans- stasis in bleeding due to upper GI tract neoplasia and evidence to
catheter angiographic embolization (TAE) or surgery should be considered if
endoscopic treatment fails or is not technically feasible (strong recommen- support a specific modality is scarce [185 – 188]. Numerous
dation, low quality evidence). endoscopic hemostasis modalities (e. g., injection, thermal, me-
chanical, topical spray/powder) have been reported, generally
with limited impact on primary hemostasis, prevention of re-
In patients bleeding from upper GI angioectasias, ESGE recommends endo-
scopic hemostasis therapy. However, there is currently inadequate evidence
bleeding, or mortality. However, endoscopic treatment may avert
to recommend a specific endoscopic hemostasis modality (strong recom- urgent surgery, reduce transfusion requirements, and may pro-
mendation, low quality evidence). vide a temporary bridge to oncologic therapy and/or selective
embolization [185 – 188].
In patients bleeding from upper GI neoplasia, ESGE recommends considering
endoscopic hemostasis in order to avert urgent surgery and reduce blood Management following endoscopy/
transfusion requirements. However, no currently available endoscopic treat- endoscopic hemostasis
ment appears to have long-term efficacy (weak recommendation, low quality
evidence).
ESGE recommends PPI therapy for patients who receive endoscopic hemo-
Erosive esophagitis, gastritis and duodenitis are common causes stasis and for patients with adherent clot not receiving endoscopic hemosta-
of NVUGIH and generally have a benign course and excellent sis. PPI therapy should be high dose and administered as an intravenous bolus
followed by continuous infusion (80 mg then 8 mg/hour) for 72 hours post
prognosis [2, 64, 155 – 158]. Meta-analyses show that acid sup-
endoscopy (strong recommendation, high quality evidence)
pression therapy is effective, with high dose PPI therapy being
significantly more effective than H2-receptor antagonists and no
observed differences in effectiveness amongst PPIs [159, 160]. ESGE suggests considering PPI therapy as intermittent intravenous bolus
dosing (at least twice-daily) for 72 hours post endoscopy for patients who re-
Endoscopic hemostasis is usually not required in this patient
ceive endoscopic hemostasis and for patients with adherent clot not receiving
population and selected patients are candidates for early hospital endoscopic hemostasis. If the patient’s condition permits, high dose oral PPI
discharge. may also be an option in those able to tolerate oral medications (weak re-
commendation, moderate quality evidence).
Although spontaneous resolution of bleeding is frequent, obser-
vational studies have demonstrated that acute UGIH secondary Based upon previously published meta-analytic data, evidence-
to Mallory – Weiss syndrome has a mortality similar to that of based guidelines on NVUGIH have recommended that PPI ther-
peptic ulcer bleeding [161, 162]. Risk factors for adverse out- apy be given as an 80 mg intravenous bolus followed by 8 mg/
comes include older age, medical co-morbidities, and active hour continuous infusion to reduce rebleeding, surgery, and mor-
bleeding at the time of endoscopy. The latter supports early tality in patients with high risk ulcers that had undergone suc-
endoscopy to stratify risk and to perform endoscopic hemostasis cessful endoscopic hemostasis [98, 99, 189, 190]. More recently
if active bleeding is identified [162 – 166]. Despite suggestions however, a meta-analysis of RCTs of high risk bleeding ulcers
that mechanical methods (clips and band ligation) are more ef- treated with endoscopic hemostasis compared intermittent PPI
fective than epinephrine injection, this has not been found in all dosing (oral or intravenous) with the currently recommended
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a13
post hemostasis PPI regimen of 80 mg intravenous bolus followed and over-the-scope clips may also be considered as rescue/sal-
by 8 mg/hour continuous infusion [191]. In that meta-analysis, vage therapy. Although limited, emerging data suggest that he-
Sachar et al reported that the risk ratio of recurrent ulcer bleed- mostatic powder may be successfully employed as salvage hemo-
ing within 7 days for intermittent infusion of PPI vs. bolus plus stasis therapy [154, 198]. The over-the-scope clip (OTSC) has also
continuous infusion of PPI was 0.72 (upper boundary of one-si- proven an effective and safe therapeutic option for severe acute
ded 95 %CI 0.97), with an absolute risk difference of −2.64 %. Risk GI bleeding when conventional endoscopic treatment modalities
ratios for other outcomes, including radiologic/surgical interven- fail [134, 147].
tion and mortality, showed no differences between infusion regi- (See Appendix e9, online-only.)
mens. These meta-analytic data indicate that intermittent PPI
therapy appears comparable to the currently recommended regi-
ESGE does not recommend routine second-look endoscopy as part of the
men of intravenous bolus plus continuous PPI infusion post management of NVUGIH. However, second-look endoscopy may be consid-
endoscopic hemostasis. It should be noted however, that inter- ered in selected patients at high risk for rebleeding (strong recommendation,
high quality evidence).
mittent PPI bolus dosing is associated with a somewhat higher
risk of rebleeding that in general can be managed endoscopically. Routine second-look endoscopy is defined as a scheduled repeat
Given the pharmacodynamic profile of PPIs, consideration should endoscopic assessment of the previously diagnosed bleeding le-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
be given to use of high dose PPI infusion given at least twice-dai- sion usually performed within 24 hours following the index
ly, and using high dose oral PPIs in patients able to tolerate oral endoscopy [98]. This strategy employs repeat endoscopy regard-
medications [191]. The concept of high dose PPI varies between less of the type of bleeding lesion, perceived rebleeding risk, or
the different studies used in the meta-analysis conducted by Sa- clinical signs of rebleeding. A meta-analysis that evaluated the ef-
char et al. However, it appears that an 80 mg oral PPI dose fol- fectiveness of routine second-look endoscopy in NVUGIH report-
lowed by 40 – 80 mg orally every 12 hours for 72 hours yields an ed a significant reduction in rebleeding (OR 0.55, 95 %CI 0.37 –
intragastric pH similar to that reported with continuous intrave- 0.81) and need for emergency surgery (OR 0.43, 95 %CI 0.19 –
nous PPI infusion following successful endoscopic hemostasis of 0.96), but not mortality (OR 0.65, 95 %CI 0.26 – 1.62) [199]. How-
high risk peptic ulcers [192]. This is but one study, and therefore ever, only one included study in that meta-analysis utilized high
we need more data to confirm these findings before drawing firm dose intravenous PPI, and in that study no benefit for second-
practical conclusions for the post-endoscopy management of pa- look endoscopy was observed, while any protective effect was
tients with NVUGIH. These data are in agreement with an RCT limited only to high risk patients (e. g., those with active bleeding
that randomized patients to high dose continuous infusion of at index endoscopy). Similarly, scheduled second-look endoscopy
esomeprazole vs. 40 mg of oral esomeprazole twice-daily for 72 does not appear to be cost-effective outside the subgroup of pa-
hours (118 vs. 126 patients respectively) [193]. Recurrent bleed- tients thought to be at high risk for recurrent ulcer bleeding
ing at 30 days was reported in 7.7 % and 6.4 % of patients, respec- [200]. Thus, the clinical utility and cost– efficiency of routine sec-
tively (difference −1.3 percentage points, 95 %CI −7.7 to 5.1 per- ond-look endoscopy in unselected patients remains to be proven.
centage points). However, this study was conducted in an Asian
population (e. g., PPI slow metabolizers) and its findings may not
In patients with NVUGIH secondary to peptic ulcer, ESGE recommends inves-
be generalizable to Western NVUGIH populations. Moreover, this tigating for the presence of Helicobacter pylori in the acute setting with initia-
study was stopped prematurely since it was not designed as an tion of appropriate antibiotic therapy when H. pylori is detected. Re-testing for
H. pylori should be performed in those patients with a negative test in the
equivalency trial, and based on the preliminary data, thousands
acute setting. Documentation of successful H. pylori eradication is recom-
of patients would have been required in order to complete the mended (strong recommendation, high quality evidence).
study. (See Appendix e8, online-only.)
Peptic ulcer remains the most frequent cause of acute NVUGIH
with H. pylori infection remaining the primary cause of peptic ul-
In patients with clinical evidence of rebleeding following successful initial
endoscopic hemostasis, ESGE recommends repeat upper endoscopy with he- cer disease [201, 202]. Indeed, when H. pylori is eradicated, the
mostasis if indicated. In the case of failure of this second attempt at hemo- risk of ulcer rebleeding is reported to be extremely low [203,
stasis, transcatheter angiographic embolization (TAE) or surgery should be
204]. However, the false-negative rate of H. pylori diagnostic test-
considered (strong recommendation, high quality evidence).
ing is higher if the test is performed at the time of the acute
An RCT comparing endoscopic therapy with surgery for recurrent bleeding episode as compared to later follow-up [205]. A meta-
peptic ulcer bleeding after successful initial endoscopic control of regression analysis including 8496 bleeding peptic ulcer patients
bleeding showed that 35/48 (73 %) of patients randomized to found an H. pylori prevalence of 72 %, with the infection rate
endoscopic re-treatment had long-term control of their peptic being significantly higher when diagnostic testing was delayed
ulcer bleeding, avoided surgery, and had a lower rate of adverse until at least 4 weeks following the bleeding event (OR 2.08, 95 %
events as compared to the surgery-treated patients [194]. The re- CI 1.10 – 3.93; P = 0.024) [206]. Therefore, it is advisable to re-test
maining 13 patients underwent salvage surgery because of failed at a later time those patients who had a negative H. pylori test in
repeat endoscopic hemostasis (n = 11) or perforation due to con- the acute setting.
tact thermal therapy (n = 2). When H. pylori infection is found, eradication therapy should be
If further bleeding occurs following a second endoscopic treat- initiated and guided by patient and local factors [98, 99]. Docu-
ment, surgery for low risk patients or interventional radiology mentation of successful H. pylori eradication is strongly recom-
for high risk patients should be considered [195]. In recent sys- mended given the high risk of recurrent ulcer bleeding in the
tematic reviews and meta-analyses comparing TAE with surgery presence of persistent H. pylori infection [98, 99]. (See Appendix
for peptic ulcer bleeding after failed endoscopic hemostasis, a e10, online-only.)
higher rebleeding rate was observed following TAE. No signifi-
cant difference in mortality or need for additional interventions
ESGE recommends restarting anticoagulant therapy following NVUGIH in
was shown between treatments [196, 197]. Hemostatic powder patients with an indication for long-term anticoagulation. The timing for
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a14 Guideline
resumption of anticoagulation should be assessed on a patient by patient ba- HR 1.9, 95 %CI 0.6 – 6.0), yet a 10-fold reduced risk of all-cause
sis. Resuming warfarin between 7 and 15 days following the bleeding event
appears safe and effective in preventing thromboembolic complications for mortality at 8 weeks (1.3 % vs. 12.9 %; difference 11.6 percentage
most patients. Earlier resumption, within the first 7 days, may be indicated for points, 95 %CI 3.7 – 19.5 percentage points; HR 0.2 95 %CI 0.06 –
patients at high thrombotic risk (strong recommendation, moderate quality 0.60) and a lower mortality rate related to cardiovascular, cere-
evidence).
brovascular, or gastrointestinal events (1.3 % vs. 10.3 %; difference
Retrospective, observational data have shown that resuming an- 9 percentage points, 95 %CI 1.7 – 16.3 percentage points; HR 0.2,
ticoagulation in patients with GI bleeding is associated with a 95 %CI 0.05 – 0.70), compared with those patients in whom aspir-
lower risk of thrombosis and death [207 – 209]. Restarting war- in was withheld [53]. Patients who required DAPT were excluded
farin therapy within 7 days of the index bleeding event was asso- from this study. The antiplatelet effect of aspirin lasts for approxi-
ciated with an approximately twofold increased risk of rebleed- mately 5 days (although new active platelets increase in number
ing [207, 209]. Conversely, as compared with resuming warfarin each day), and the risk of early recurrent bleeding is high in the
beyond 30 days, resuming warfarin between 7 and 30 days did first 3 days [53]. Therefore, restarting aspirin on day 3 in patients
not increase the risk of rebleeding, but did significantly decrease with high risk endoscopic stigmata is a reasonable trade-off be-
the risk of thromboembolism and improved survival [209]. These tween the risks of rebleeding and thrombosis. In patients with
data appear to support that resumption of anticoagulation after 7 peptic ulcer bleeding with no high risk endoscopic stigmata, as-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
days of interruption is safe and effective in preventing throm- pirin can be resumed immediately as RCTs have shown that nei-
boembolic complications for most patients. However, in patients ther aspirin nor clopidogrel use impede ulcer healing promoted
at high thrombotic risk (e. g., chronic atrial fibrillation with pre- by PPIs [53, 55, 56]. No high level evidence helps guide the timing
vious embolic event, CHADS2 score ≥ 3, mechanical prosthetic for resumption of P2Y12 platelet receptor inhibitors (e. g., clopido-
heart valve, recent [within past 3 months] deep venous thrombo- grel) following NVUGIH. However, in view of its similar antiplate-
sis or pulmonary embolism, and patients with known severe hy- let activity, it seems reasonable to apply a similar management
percoagulable state), for whom early resumption of anticoagula- strategy. Moreover, there is no evidence in the literature to help
tion within the first week following an acute bleeding event guide the management of patients receiving DAPT in the setting
might be appropriate, bridging therapy using unfractionated or of NVUGIH. The overriding principle of balancing bleeding and
low molecular weight heparin may be considered [210]. No data thrombotic event risks requires close collaboration between the
are currently available to guide the management of DOACs fol- gastroenterology and cardiology teams.
lowing NVUGIH. Yet caution in the early resumption of DOACs is
required because of their rapid onset of action and the current
In patients requiring dual antiplatelet therapy (DAPT) and who have had
lack of reversal agents. (See Appendix 11, online-only.) NVUGIH, ESGE recommends the use of a PPI as co-therapy (strong recom-
mendation, moderate quality evidence.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a15
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
including thrombin, fibrin, and cyanoacrylate glues, which are scope and over-the-scope) and band ligation devices. Endo-
used to create a primary seal at the site of bleeding. scopic clips are deployed directly onto a bleeding site and ty-
Endoscopic injection is performed using needles which con- pically slough off within days to weeks after placement [131].
sist of an outer sheath and an inner hollow-core needle (19 – Hemostasis is achieved by mechanical compression of the
25 gauge). The endoscopist or nursing assistant can retract the bleeding site.
needle into the sheath for safe passage through the working Clips are available in a variety of jaw lengths and opening
channel of the endoscope. When the catheter is passed out of widths. The delivery catheter consists of a metal cable within
the working channel and placed near the site of bleeding, the a sheath enclosed within a Teflon catheter. After insertion of
needle is extended out of the sheath and the solution injected the catheter through the working channel of the endoscope,
into the submucosa using a syringe attached to the catheter the clip is extended out of the sheath, positioned over the tar-
handle [126]. get area and opened with the plunger handle. A rotation
mechanism on the handle is available on some commercially
Thermal therapy available clips and this allows the endoscopist to change the
Thermal devices used in the treatment of upper gastrointesti- orientation of the clip at the site of bleeding. The jaws of the
nal (UGI) bleeding are divided into contact and noncontact clip are applied with pressure and closed onto the target tis-
modalities. Contact thermal devices include heater probes sue by using the device handle. Some clips may be opened,
which generate heat directly and bipolar electrocautery closed, and repositioned, whereas others are permanently de-
probes which generate heat indirectly by passage of an electri- ployed and released upon clip closure. Some clips are provid-
cal current through the tissue. Noncontact thermal devices in- ed with a reusable delivery sheath, greatly reducing costs. Si-
clude argon plasma coagulation (APC) tools. Heat generated milarly, some clips are automatically released on deployment,
from these devices leads to edema, coagulation of tissue pro- while others require repositioning of the plunger handle to re-
teins, contraction of vessels, and indirect activation of the co- lease the deployed clip from the catheter [131].
agulation cascade, resulting in a hemostatic bond [126, 127]. The over-the-scope clip device includes an applicator cap, a
Contact thermal probes use local tamponade (mechanical nitinol clip, and a hand wheel [132, 133]. The applicator cap,
pressure of the probe tip directly onto the bleeding site) com- with the mounted nitinol clip, is affixed to the tip of the endo-
bined with heat or electrical current to coagulate blood ves- scope in a manner similar to that of a variceal band ligation
sels, a process known as “coaptive coagulation.” Heater probes device. Caps are available in three sizes to accommodate var-
(available in 7-Fr and 10-Fr sizes) consist of a Teflon-coated ious endoscope diameters: 11 mm, 12 mm, and 14 mm. Caps
hollow aluminum cylinder with an inner heating coil com- are also available in two lengths (3 mm and 6 mm) to allow
bined with a thermocoupling device at the tip of the probe to variation in the amount of tissue grasped. Clips come in three
maintain a constant energy output (measured in joules, com- different shapes of teeth: rounded, pointed and long-pointed.
monly 15 – 30 joules of thermal energy are delivered). An Clips with rounded teeth are used where the goal is tissue
endoscopist-controlled foot pedal activates the heater probe compression to achieve hemostasis. The applicator cap incor-
and provides waterjet irrigation. Multipolar/bipolar electro- porates a clip release thread, which is pulled retrogradely
cautery contact probes (7-Fr and 10-Fr sizes) deliver thermal through the working channel of the endoscope and fixed
energy by completion of an electrical local circuit (no ground- onto a hand wheel mounted on the working-channel access
ing pad required) between two electrodes on the tip of the port of the endoscope. The clip is released by turning the
probe as current flows through nondesiccated tissue. As the hand wheel, in a manner similar to deploying a variceal liga-
targeted tissue desiccates, there is a decrease in electrical con- tion band [134].
ductivity, limiting the maximum temperature, depth, and Last, endoscopic band ligation devices, commonly used in
area of tissue injury. An endoscopist-controlled foot pedal esophageal variceal bleeding, have also been reported for
controls the delivery of the energy [127]. The standard setting treatment of NVUGIH (e. g., for Dieulafoy lesion) and involve
for use in achieving hemostasis in peptic ulcer bleeding is 15 – the placement of elastic bands over tissue to produce mechan-
20 watts, which is delivered in 8 – 10-second applications ical compression and tamponade.
(commonly referred to as tamponade stations) [96].
APC, a noncontact thermal modality, uses high frequency,
monopolar alternating current conducted to the target tissue
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a16 Guideline
Topical therapy Additional topical hemostatic sprays include EndoClot and the
Topical hemostatic sprays have been used in acute NVUGIH Ankaferd Blood Stopper [135, 136]. EndoClot (EndoClot Plus
with promising results, but thus far in a limited number of pa- Inc, Santa Clara, California, USA) is a starch-derived compound
tients and without any comparative data regarding standard that rapidly absorbs water from serum and concentrates pla-
endoscopic hemostasis therapies [135, 136]. Advantages of telets, red blood cells, and coagulation proteins at the bleeding
noncontact, spray catheter delivery of hemostatic agents in- site to accelerate the clotting cascade. Hemostatic sprays de-
clude ease of use, lack of need for precise lesion targeting, ac- rived from plant products/extracts have also been evaluated.
cess to lesions in difficult locations, and the ability to treat a Clinical experience with these agents for endoscopic hemosta-
large surface area. sis is currently limited to the off-label use of the Ankaferd
Topical hemostatic sprays include TC-325, (Hemospray, Cook Blood Stopper (Ankaferd Health Products Ltd, Istanbul, Tur-
Medical Inc, Winston-Salem, North Carolina, USA), which is a key). This topical agent promotes formation of a protein
proprietary, inorganic, absorbent powder that rapidly concen- mesh that acts as an anchor for erythrocyte aggregation with-
trates clotting factors at the bleeding site, forming a coagulum. out significantly altering coagulation factors or platelets and is
Hemospray comes in a hand-held device consisting of a pres- delivered onto the bleeding site via an endoscopic spray cath-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
surized CO2 canister, a through-the-scope delivery catheter, eter until an adherent coagulum is formed. The particles are
and a reservoir for the powder cartridge. The powder is deliv- subsequently cleared from the bleeding site within hours to
ered via pushbutton in 1 – 2-second bursts until hemostasis is days later. The overall efficacy of these topical agents is un-
achieved. The maximum amount of TC-325 that can be safely known in brisk arterial bleeding and may be limited because
administered during a single treatment session has not yet of the rapid “wash-away” effect of the hemostatic agent by on-
been established [135, 136]. The coagulum typically sloughs going blood flow.
within 3 days and is naturally eliminated. Hemospray has re-
ceived regulatory clearance in some countries.
16
Department of Medical Gastroenterology, Rikshospitalet University Hospital,
ESGE guidelines represent a consensus of best practice based on Oslo, Norway
the available evidence at the time of preparation. They may not 17
Department of Gastroenterology, Royal Wolverhampton Hospitals NHS
apply in all situations and should be interpreted in the light of Trust, Wolverhampton, United Kingdom
18
Department of Gastroenterology, Valduce Hospital, Como, Italy
specific clinical situations and resource availability. Further con- 19
Department of Gastroenterology, Centro Hospitalar do Porto, Portugal
trolled clinical studies may be needed to clarify aspects of these 20
Department of Gastroenterology, Centro Hospitalar e Universitário de
statements, and revision may be necessary as new data appear. Coimbra, Portugal
21
Endoscopy Unit, Nuovo Regina Margherita Hospital, Rome, Italy
Clinical consideration may justify a course of action at variance 22
Gastroenterology and Endoscopy Department, Antonio Cardarelli Hospital,
to these recommendations. ESGE guidelines are intended to be Naples, Italy
23
an educational device to provide information that may assist Digestive Endoscopy Unit, Catholic University, Rome, Italy
endoscopists in providing care to patients. They are not rules
and should not be construed as establishing a legal standard of References
care or as encouraging, advocating, requiring, or discouraging 1 van Leerdam ME. Epidemiology of acute upper gastrointestinal bleed-
ing. Best Pract Res Clin Gastroenterol 2008; 22: 209 – 224
any particular treatment.
2 Hearnshaw SA, Logan RF, Lowe D et al. Acute upper gastrointestinal
bleeding in the UK: patient characteristics, diagnoses and outcomes
Competing interests: None in the 2007 UK audit. Gut 2011; 60: 1327 – 1335
3 Guyatt GH, Oxman AD, Vist GE et al. GRADE: an emerging consensus on
rating quality of evidence and strength of recommendations. BMJ
Institutions
1
Institute of Gastroenterology and Liver Diseases, Ha'Emek Medical Center, 2008; 336: 924 – 926
Afula, Israel 4 Dumonceau JM, Hassan C, Riphaus A et al. European Society of Gastro-
2
Rappaport Faculty of Medicine Technion-Israel Institute of Technology, inestinal Endoscopy (ESGE) guideline development policy. Endoscopy
Haifa, Israel 2012; 44: 626 – 629
3
Gedyt Endoscopy Center, Buenos Aires, Argentina 5 Baradarian R, Ramdhaney S, Chapalamadugu R et al. Early intensive re-
4
Departments of Internal Medicine and Gastroenterology and Hepatology, suscitation of patients with upper gastrointestinal bleeding decreases
Erasmus MC University Medical Center, Rotterdam, The Netherlands mortality. Am J Gastroenterol 2004; 99: 619 – 622
5
University of Zaragoza, Aragon Health Research Institute (IIS Aragon), 6 Kwan I, Bunn F, Chinnock P et al. Timing and volume of fluid adminis-
CIBERehd, Spain
6 tration for patients with bleeding. Cochrane Database Syst Rev 2014;
Department of Gastroenterology, Sheffield Teaching Hospitals, United
3: CD002245
Kingdom
7
Division of Gastroenterology and Digestive Endoscopy, Maresca Hospital, 7 Spahn DR, Bouillon B, Cerny V et al. Management of bleeding and coa-
Torre del Greco, Italy gulopathy following major trauma: an updated European guideline.
8
Department of Gastroenterology and Hepatology, Institute for Clinical and Crit Care 2013; 17: R76
Experimental Medicine, Prague, Czech Republic 8 Roberts I, Alderson P, Bunn F et al. Colloids versus crystalloids for fluid
9
Department of Gastroenterology, IPO Porto, Portugal and CINTESIS, Porto resuscitation in critically ill patients. Cochrane Database Syst Rev
Faculty of Medicine, Portgal 2004; 4: CD000567
10
Division of Gastroenterology, Hospital Curto, Polla, Italy 9 Myburgh JA, Finfer S, Bellomo R et al. Hydroxyethyl starch or saline for
11
First Department of Internal Medicine and Gastroenterology, Petz Aladar, fluid resuscitation in intensive care. N Engl J Med 2012; 367: 1901 –
Hospital, Gyor, Hungary
12 1911
Department of Surgical Sciences, University of Torino, Torino, Italy
13 10 Marik PE, Corwin HL. Efficacy of red blood cell transfusion in the criti-
Institute and Polyclinic for Diagnostic Radiology, University Hospital
Dresden-TU, Dresden, Germany cally ill: a systematic review of the literature. Crit Care Med 2008; 36:
14
Department of Gastroenterology, Ambroise Paré Hospital, Boulogne, France 2667 – 2674
15
Department of Biomedical and Clinical Sciences, University of Milan,
Gastroenterology Unit, Luigi Sacco University Hospital, Milan, Italy
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a17
11 Restellini S, Kherad O, Jairath V et al. Red blood cell transfusion is asso- 34 Radaelli F, Paggi S, Terruzzi V et al. Management of warfarin-associated
ciated with increased rebleeding in patients with nonvariceal upper coagulopathy in patients with acute gastrointestinal bleeding: a cross-
gastrointestinal bleeding. Aliment Pharmacol Ther 2013; 37: 316 – 322 sectional physician survey of current practice. Dig Liver Dis 2011; 43:
12 Villanueva C, Colomo A, Bosch A et al. Transfusion strategies for acute 444 – 447
upper gastrointestinal bleeding. N Engl J Med 2013; 368: 11 – 21 35 Patriquin C, Crowther M. Treatment of warfarin-associated coagulopa-
13 Jairath V, Kahan BC, Stanworth SJ et al. Prevalence, management, and thy with vitamin K. Expert Rev Hematol 2011; 4: 657 – 665
outcomes of patients with coagulopathy after acute nonvariceal upper 36 Baron TH, Kamath PS, McBane RD. New anticoagulant and antiplatelet
gastrointestinal bleeding in the United Kingdom. Transfusion 2013; agents: a primer for the gastroenterologist. Clin Gastroenterol Hepatol
53: 1069 – 1076 2014; 12: 187 – 195
14 Shingina A, Barkun AN, Razzaghi A et al. Systematic review: the pre- 37 Leissinger CA, Blatt PM, Hoots WK et al. Role of prothrombin complex
senting international normalized ratio (INR) as a predictor of outcome concentrates in reversing warfarin anticoagulation: a review of the lit-
in patients with upper nonvariceal gastrointestinal bleeding. Aliment erature. Am J Hematol 2008; 83: 137 – 143
Pharmacol Ther 2011; 33: 1010 – 1018 38 Karaca MA, Erbil B, Ozmen MM. Use and effectiveness of prothrombin
15 Karam O, Tucci M, Combescure C et al. Plasma transfusion strategies for complex concentrates vs. fresh frozen plasma in gastrointestinal he-
critically ill patients. Cochrane Database Syst Rev 2013; 12: CD010654 morrhage due to warfarin usage in the ED. Am J Emerg Med 2014; 32:
16 Razzaghi A, Barkun AN. Platelet transfusion threshold in patients with 660 – 664
upper gastrointestinal bleeding: a systematic review. J Clin Gastroen- 39 Dentali F, Marchesi C, Pierfranceschi MG et al. Safety of prothrombin
terol 2012; 46: 482 – 486 complex concentrates for rapid anticoagulation reversal of vitamin K
17 Rockall TA, Logan RF, Devlin HB et al. Risk assessment after acute upper antagonists. A meta-analysis. Thromb Haemost 2011; 106: 429 – 438
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
gastrointestinal haemorrhage. Gut 1996; 38: 316 – 321 40 Hickey M, Gatien M, Taljaard M et al. Outcomes of urgent warfarin re-
18 Blatchford O, Murray WR, Blatchford M. A risk score to predict need for versal with frozen plasma versus prothrombin complex concentrate in
treatment for upper-gastrointestinal haemorrhage. Lancet 2000; 356: the emergency department. Circulation 2013; 128: 360 – 364
1318 – 1321 41 Chai-Adisaksopha C, Crowther M, Isayama T et al. The impact of bleed-
19 de Groot NL, Bosman JH, Siersema PD et al. Prediction scores in gastro- ing complications in patients receiving target-specific oral anticoagu-
intestinal bleeding: a systematic review and quantitative appraisal. lants: a systematic review and meta-analysis. Blood 2014; 124: 2450 –
Endoscopy 2012; 44: 731 – 739 2458
20 Lee JG, Turnipseed S, Romano PS et al. Endoscopy-based triage signifi- 42 Holster IL, Valkhoff VE, Kuipers EJ et al. New oral anticoagulants in-
cantly reduces hospitalization rates and costs of treating upper GI crease risk for gastrointestinal bleeding: a systematic review and
bleeding: a randomized controlled trial. Gastrointest Endosc 1999; meta-analysis. Gastroenterology 2013; 145: 105 – 112
50: 755 – 761 43 Ruff CT, Guigliano RP, Braunwald E et al. Comparison of the efficacy and
21 Cipolletta L, Bianco MA, Rotondano G et al. Outpatient management for safety of new oral anticoagulants with warfarin in patients with atrial
low-risk nonvariceal upper GI bleeding: a randomized controlled trial. fibrillation: a meta-analysis of randomized trials. Lancet 2014; 383:
Gastrointest Endosc 2002; 55: 1 – 5 955 – 962
22 Brullet E, Campo R, Calvet X et al. A randomized study of the safety of 44 Lu G, DeGuzman FR, Hollenbach SJ et al. A specific antidote for reversal
outpatient care for patients with bleeding peptic ulcer treated by of anticoagulation by direct and indirect inhibitors of coagulation fac-
endoscopic injection. Gastrointest Endosc 2004; 60: 15 – 21 tor Xa. Nat Med 2013; 19: 446 – 451
23 Longstreth GF, Feitelberg SP. Outpatient care of selected patients with 45 Ansell JE, Bakhru SH, Laulicht BE et al. Use of PER977 to reverse the an-
acute non-variceal upper gastrointestinal haemorrhage. Lancet 1995; ticoagulant effect of edoxaban. N Engl J Med 2014; 371: 2141 – 2142
345: 108 – 111 46 Pollack CVJr, Reilly PA, Eikelboob J et al. Idarucizumab for dabigatran re-
24 Longstreth GF, Feitelberg SP. Successful outpatient management of versal. N Engl J Med 2015; 373: 511 – 520
acute upper gastrointestinal hemorrhage: use of practice guidelines 47 Abraham NS, Castillo DL. Novel anticoagulants: bleeding risk and man-
in a large patient series. Gastrointest Endosc 1998; 47: 219 – 222 agement strategies. Curr Opin Gastroenterol 2013; 29: 676 – 683
25 Rockall TA, Logan RF, Devlin HB et al. Selection of patients for early dis- 48 Desai J, Kolb JM, Weitz JI et al. Gastrointestinal bleeding with the new
charge or outpatient care after acute upper gastrointestinal haemor- oral anticoagulants – defining the issues and the management strate-
rhage. national audit of acute upper gastrointestinal haemorrhage. gies. Thromb Haemost 2013; 110: 205 – 212
Lancet 1996; 347: 1138 – 1140 49 Makris M, Van Veen JJ, Tait CR et al. British Committee for Standards in
26 Lai KC, Hui WM, Wong BC et al. A retrospective and prospective study Haematology. Guideline on the management of bleeding in patients on
on the safety of discharging selected patients with duodenal ulcer antithrombotic agents. Br J Haematol 2013; 160: 35 – 46
bleeding on the same day as endoscopy. Gastrointest Endosc 1997; 50 Siegal DM, Cuker A. Reversal of novel oral anticoagulants in patients
45: 26 – 30 with major bleeding. J Thromb Thrombolysis 2013; 35: 391 – 398
27 Cebollero-Santamaria F, Smith J, Gioe S et al. Selective outpatient man- 51 Fawole A, Daw HA, Crowther MA. Practical management of bleeding
agement of upper gastrointestinal bleeding in the elderly. Am J Gastro- due to the anticoagulants dabigatran, rivaroxaban, and apixaban. Clev
enterol 1999; 94: 1242 – 1247 Clin J Med 2013; 80: 443 – 451
28 Gralnek IM, Dulai GS. Incremental value of upper endoscopy for triage 52 Boustiere C, Veitch A, Vanbiervliet G et al. Endoscopy and antiplatelet
of patients with acute non-variceal upper GI hemorrhage. Gastrointest agents. European Society of Gastrointestinal Endoscopy (ESGE) guide-
Endosc 2004; 60: 9 – 14 line. Endoscopy 2011; 43: 445 – 461
29 Guerrouij M, Uppal CS, Alklabi A et al. The clinical impact of bleeding 53 Sung JJY, Lau JYW, Ching JYL et al. Continuation of low dose aspirin
during oral anticoagulant therapy: assessment of morbidity, mortality therapy in peptic ulcer bleeding: a randomized trial. Ann Intern Med
and post-bleed anticoagulant management. J Thromb Thrombolysis 2010; 152: 1 – 9
2011; 31: 419 – 423 54 Derogar M, Sandblom G, Lundell L et al. Discontinuation of low dose as-
30 Holbrook A, Schulman S, Witt DM et al. Evidence-based management of pirin therapy after peptic ulcer bleeding increases risk of death and
anticoagulant therapy: Antithrombotic therapy and prevention of acute cardiovascular events. Clin Gastroenterol Hepatol 2013; 11:
thrombosis, 9th ed: American College of Chest Physicians Evidence- 38 – 42
Based Clinical Practice Guidelines. Chest 2012; 141: e152S – 184S DOI 55 Liu C-P, Chen W-C, Lai K-H et al. Esomeprazole alone compared with
10.1378/chest.11–2295 esomeprazole plus aspirin for the treatment of aspirin-related peptic
31 Irwin ST, Ferguson R, Weilert F et al. Supratherapeutic anticoagulation ulcers. Am J Gastroenterol 2012; 107: 1022 – 1029
at presentation is associated with reduced mortality in nonvariceal 56 Ng FH, Wong BCY, Wong SY et al. Clopidogrel plus omeprazole compar-
upper gastrointestinal hemorrhage. Endosc Int Open 2014; 2: E148 – ed with aspirin plus omeprazole for aspirin-induced symptomatic
E152 DOI 10.1055/s-0034–1377287 [Epub 2014 Jul 10] peptic ulcers/erosions with low to moderate bleeding/re-bleeding
32 Tran HA, Chunilal SD, Harper PL et al. An update of consensus guide- risk – a single-blind, randomized controlled study. Aliment Pharmacol
lines for warfarin reversal. Med J Aust 2013; 198: 198 – 199 Ther 2004; 19: 359 – 365
33 Choudari CP, Rajgopal C, Palmer KR. Acute gastrointestinal hemorrhage 57 Sreedharan A, Martin J, Leontiadis GI et al. Proton pump inhibitor treat-
in anticoagulated patients: diagnoses and response to endoscopic ment initiated prior to endoscopic diagnosis in upper gastrointestinal
treatment. Gut 1994; 35: 464 – 466 bleeding. Cochrane Database Syst Rev 2010; 7: CD005415
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a18 Guideline
58 Sabah AlS, Barkun AN, Herba K et al. Cost-effectiveness of proton-pump 83 Lin HJ, Wang K, Perng CL et al. Early or delayed endoscopy for patients
inhibition before endoscopy in upper gastrointestinal bleeding. Clin with peptic ulcer bleeding. A prospective randomized study. J Clin Gas-
Gastroenterol Hepatol 2008; 6: 418 – 425 troenterol 1996; 22: 267 – 271
59 Tsoi KKF, Lau JYW, Sung JJY. Cost-effectiveness analysis of high-dose 84 Lim L, Ho K, Chan Y et al. Urgent endoscopy is associated with lower
omeprazole infusion before endoscopy for patients with upper-GI mortality in high-risk but not low-risk nonvariceal upper gastrointes-
bleeding. Gastrointest Endosc 2008; 67: 1056 – 1063 tinal bleeding. Endoscopy 2011; 43: 300 – 306
60 Shakur H, Roberts I. CRASH-2 trial collaborators. et al. Effects of tra- 85 Bjorkman DJ, Zaman A, Fennerty MB et al. Urgent vs. elective endos-
nexamic acid on death, vascular occlusive events, and blood transfu- copy for acute non-variceal upper-GI bleeding: an effectiveness study.
sion in trauma patients with significant haemorrhage (CRASH-2): a Gastrointest Endosc 2004; 60: 1 – 8
randomized, placebo-controlled trial. Lancet 2010; 376: 23 – 32 86 Stanley AJ, Ashley D, Dalton HR et al. Outpatient management of pa-
61 Gluud LL, Klingenberg SL, Langholz E. Tranexamic acid for upper gastro- tients with low-risk upper-gastrointestinal haemorrhage: multicentre
intestinal bleeding. Cochrane Database Syst Rev 2012; 1: CD006640 validation and prospective evaluation. Lancet 2009; 373: 42 – 47
62 Raptis S, Dollinger HC, von Berger L et al. Effects of somatostatin on gas- 87 McLaughlin C, Vine L, Chapman L et al. The management of low-risk
tric secretion and gastrin release in man. Digestion 1975; 13: 15 – 26 primary upper gastrointestinal haemorrhage in the community. Eur J
63 Hearnshaw SA, Logan RF, Lowe D et al. Use of endoscopy for manage- Gastroenterol Hepatol 2012; 24: 288 – 293
ment of acute upper gastrointestinal bleeding in the UK: results of a 88 Girardin M, Bertolini D, Ditisheim S et al. Use of Glasgow-Blatchford
nationwide audit. Gut 2010; 59: 1022 – 1029 bleeding score reduces hospital stay duration and costs for patients
64 Enestvedt BK, Gralnek IM, Mattek N et al. An evaluation of endoscopic with low-risk upper GI bleeding. Endosc Int Open 2014; 2: E74 – E79
indications and findings related to non-variceal upper-GI hemorrhage DOI 10.1055/s-0034–1365542 Epub 2014 May 7
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
in a large multicenter consortium. Gastrointest Endosc 2008; 67: 89 Laursen SB, Dalton HR, Murray IA et al. Performance of new thresholds
422 – 429 of the Glasgow Blatchford score in managing patients with upper gas-
65 Barkun AN, Bardou M, Martel M et al. Prokinetics in acute upper GI trointestinal bleeding. Clin Gastroenterol Hepatol 2015; 13: 115 – 121
bleeding: a meta-analysis. Gastrointest Endosc 2010; 72: 1138 – 1145 90 Rubin M, Hussain SA, Shalomov A et al. Live view video capsule endos-
66 Szary NM, Gupta R, Choudhary A et al. Erythromycin prior to endos- copy enables risk stratification of patients with acute upper GI bleed-
copy in acute upper gastrointestinal bleeding: a meta-analysis. Scand ing in the emergency room: a pilot study. Dig Dis Sci 2011; 56: 786 –
J Gastroenterol 2011; 46: 920 – 924 791
67 Bai Y, Guo JF, Li ZS. Meta-analysis: erythromycin before endoscopy for 91 Chandran S, Testro A, Urquhart P et al. Risk stratification of upper GI
acute upper gastrointestinal bleeding. Aliment Pharmacol Ther 2011; bleeding with an esophageal capsule. Gastrointest Endosc 2013; 77:
34: 166 – 171 891 – 898
68 Theivanayagam S, Lim RG, Cobell WJ et al. Administration of erythro- 92 Gralnek IM, Ching JYL, Maza I et al. Capsule endoscopy in acute upper
mycin before endoscopy in upper gastrointestinal bleeding: a meta-a- gastrointestinal hemorrhage: a prospective cohort study. Endoscopy
nalysis of randomized controlled trials. Saudi J Gastroenterol 2013; 19: 2013; 45: 12 – 19
205 – 210 93 Meltzer AC, Ali MA, Kresiberg RB et al. Video capsule endoscopy in the
69 Winstead NS, Wilcox CM. Erythromycin prior to endoscopy for acute emergency department: a prospective study of acute upper gastroin-
upper gastrointestinal haemorrhage: a cost–effectiveness analysis. Ali- testinal hemorrhage. Ann Emerg Med 2013; 61: 438 – 443
ment Pharmacol Ther 2007; 26: 1371 – 1377 94 Meltzer AC, Pinchbeck C, Burnett S et al. Emergency physicians accu-
70 Srygley FD, Gerardo CJ, Tran T et al. Does this patient have a severe up- rately interpret video capsule endoscopy findings in suspected upper
per gastrointestinal bleed? JAMA 2012; 307: 1072 – 1079 gastrointestinal hemorrhage: a video survey. Acad Emerg Med 2013;
71 Aljebreen AM, Fallone CA, Barkun AN. Nasogastric aspirate predicts 20: 711 – 715
high-risk endoscopic lesions in patients with acute upper GI bleeding. 95 Meltzer AC, Ward MJ, Gralnek IM et al. The cost–effectiveness analysis
Gastrointest Endosc 2004; 59: 172 – 178 of video capsule endoscopy compared to other strategies to manage
72 Pateron D, Vicaut E, Debuc E et al. Erythromycin infusion or gastric la- acute upper gastrointestinal hemorrhage in the ED. Am J Emerg Med
vage for upper gastrointestinal bleeding: a multicenter randomized 2014; 32: 823 – 832
controlled trial. Ann Emerg Med 2011; 57: 582 – 589 96 Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in gastrointestinal
73 Huang ES, Karsan S, Kanwal F et al. Impact of nasogastric lavage on out- bleeding. Lancet 1974; 2: 394 – 397
comes in acute GI bleeding. Gastrointest Endosc 2011; 74: 971 – 980 97 Gralnek IM, Barkun AN, Bardou M. Management of acute bleeding from
74 Singer AJ, Richman PB, Kowalska A et al. Comparison of patient and a peptic ulcer. N Engl J Med 2008; 359: 928 – 937
practitioner assessments of pain from commonly performed emergen- 98 Barkun AN, Bardou M, Kuipers EJ et al. International consensus recom-
cy department procedures. Ann Emerg Med 1999; 33: 652 – 658 mendations on the management of patients with nonvariceal upper
75 Koch DG, Arguedas MR, Fallon MB. Risk of aspiration pneumonia in sus- gastrointestinal bleeding. Ann Intern Med 2010; 152: 101 – 113
pected variceal hemorrhage: the value of prophylactic endotracheal 99 Laine L, Jensen DM. Management of patients with ulcer bleeding. Am J
intubation prior to endoscopy. Dig Dis Sci 2007; 52: 2225 – 2228 Gastroenterol 2012; 107: 345 – 360
76 Rehman A, Iscimen R, Yilmaz M et al. Prophylactic endotracheal intuba- 100 Chung IK, Kim EJ, Lee MS et al. Endoscopic factors predisposing to re-
tion in critically ill patients undergoing endoscopy for upper GI he- bleeding following endoscopic hemostasis in bleeding peptic ulcers.
morrhage. Gastrointest Endosc 2009; 69: 55 – 59 Endoscopy 2001; 33: 969 – 975
77 Rudolph SJ, Landsverk BK, Freeman ML. Endotracheal intubation for air- 101 Guglielmi A, Ruzzenente A, Sandri M et al. Risk assessment and predic-
way protection during endoscopy for severe upper GI hemorrhage. tion of rebleeding in bleeding gastroduodenal ulcer. Endoscopy 2002;
Gastrointest Endosc 2003; 57: 58 – 61 34: 778 – 786
78 Kanwal F, Barkun A, Gralnek IM et al. Measuring quality of care in pa- 102 Zaragoza AM, Tenías JM, Llorente MJ et al. Prognostic factors in gastro-
tients with nonvariceal upper gastrointestinal hemorrhage: develop- intestinal bleeding due to peptic ulcer: construction of a predictive
ment of an explicit quality indicator set. Am J Gastroenterol 2010; model. J Clin Gastroenterol 2008; 42: 786 – 790
105: 1710 – 1718 103 Elmunzer BJ, Young SD, Inadomi JM et al. Systematic review of the pre-
79 Lanas A, Aabakken L, Fonseca J et al. Variability in the management of dictors of recurrent hemorrhage after endoscopic hemostatic therapy
nonvariceal upper gastrointestinal bleeding in Europe: an observa- for bleeding peptic ulcers. Am J Gastroenterol 2008; 103: 2625 – 2632
tional study. Adv Ther 2012; 29: 1026 – 1036 104 Marmo R, Del Piano M, Rotondano G et al. Mortality from nonulcer
80 Spiegel BM, Vakil NB, Ofman JJ. Endoscopy for acute nonvariceal upper bleeding is similar to that of ulcer bleeding in high-risk patients
gastrointestinal tract hemorrhage: is sooner better? A systematic re- with nonvariceal hemorrhage: a prospective database study in Italy.
view Arch Intern Med 2001; 161: 1393 – 1404 Gastrointest Endosc 2012; 75: 263 – 272
81 Tsoi KKF, Ma TKW, Sung JJY. Endoscopy for upper gastrointestinal 105 Bratanic A, Puljiz Z, Ljubicicz N et al. Predictive factors of rebleeding
bleeding: how urgent is it? Nat Rev Gastroenterol Hepatol 2009; 6: and mortality following endoscopic hemostasis in bleeding peptic ul-
463 – 469 cers. Hepatogastroenterology 2013; 60: 112 – 117
82 Wysocki JD, Srivastav S, Winstead NS. A nationwide analysis of risk fac- 106 Sung JJ, Barkun A, Kuipers EJ et al. Intravenous esomeprazole for pre-
tors for mortality and time to endoscopy in upper gastrointestinal hae- vention of recurrent peptic ulcer bleeding: a randomized trial. Ann
morrhage. Aliment Pharmacol Ther 2012; 36: 30 – 36 Intern Med 2009; 50: 455 – 464
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a19
107 de Groot NL, van Oijen MG, Kessels K et al. Reassessment of the predic- 132 Chuttani R, Barkun A, Carpenter S et al. Endoscopic clip application
tive value of the Forrest classification for peptic ulcer rebleeding and devices. Gastrointest Endosc 2006; 63: 746 – 750
mortality: can classification be simplified? Endoscopy 2014; 46: 46 – 133 Kirschniak A, Kratt T, Stüker D et al. A new endoscopic over-the-scope
52 clip system for treatment of lesions and bleeding in the GI tract: first
108 Lau JY, Sung JJ, Chan AC et al. Stigmata of hemorrhage in bleeding pep- clinical experiences. Gastrointest Endosc 2007; 66: 162 – 167
tic ulcers: an interobserver agreement study among international ex- 134 Kirschniak A, Subotova N, Zieker D et al. The over-the-scope clip
perts. Gastrointest Endosc 1997; 46: 33 – 36 (OTSC) for the treatment of gastrointestinal bleeding, perforations,
109 Mondardini A, Barletti C, Rocca G et al. Non-variceal upper gastroin- and fistulas. Surg Endosc 2011; 25: 2901 – 2905
testinal bleeding and Forrest’s classification: diagnostic agreement 135 Gottlieb KT, Banerjee S, Barth BA. ASGE Technology Committee. et al.
between endoscopists from the same area. Endoscopy 1998; 30: Endoscopic closure devices. Gastrointest Endosc 2012; 76: 244 – 251
508 – 512 136 Barkun AN, Moosavi S, Martel M. Topical hemostatic agents: a sys-
110 Lin HJ, Perng CL, Lee FY et al. Clinical courses and predictors for re- tematic review with particular emphasis on endoscopic application
bleeding in patients with peptic ulcers and non-bleeding visible ves- in GI bleeding. Gastrointest Endosc 2013; 77: 692 – 700
sels: a prospective study. Gut 1994; 35: 1389 – 1393 137 Sung JJ, Tsoi KK, Lai LH et al. Endoscopic clipping versus injection and
111 Cheng CL, Lin CH, Kuo CJ et al. Predictors of rebleeding and mortality thermo-coagulation in the treatment of non-variceal upper gastroin-
in patients with high-risk bleeding peptic ulcers. Dig Dis Sci 2010; 55: testinal bleeding: a meta-analysis. Gut 2007; 56: 1364 – 1373
2577 – 2583 138 Calvet X, Vergara M, Brullet E et al. Addition of a second endoscopic
112 Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an evi- treatment following epinephrine injection improves outcome in
dence-based approach based on meta-analyses of randomized con- high-risk bleeding ulcers. Gastroenterology 2004; 126: 441 – 450
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
trolled trials. Clin Gastroenterol Hepatol 2009; 7: 33 – 47 139 Marmo R, Rotondano G, Piscopo R et al. Dual therapy versus mono-
113 Sung J, Chan F, Lau J et al. The effect of endoscopic therapy in patients therapy in the endoscopic treatment of high-risk bleeding ulcers: a
receiving omeprazole for bleeding ulcers with nonbleeding visible meta-analysis of controlled trials. Am J Gastroenterol 2007; 102:
vessels or adherent clots: a randomized comparison. Ann Intern 279 – 289
Med 2003; 139: 237 – 243 140 Vergara M, Bennett C, Calvet X et al. Epinephrine injection versus
114 Andriulli A, Annese V, Caruso N et al. Proton-pump inhibitors and out- epinephrine injection and a second endoscopic method in high risk
come of endoscopic hemostasis in bleeding peptic ulcers: a series of bleeding ulcers. Cochrane Database Syst Rev 2014; 10: CD005584
meta-analyses. Am J Gastroenterol 2005; 100: 207 – 219 141 Hwang JH, Fisher DA, Ben-Menachem T et al. Standards of Practice
115 Lin JH, Wang K, Perng CL et al. Natural history of bleeding peptic ul- Committee of the American Society for Gastrointestinal Endoscopy.
cers with a tightly adherent blood clot: a prospective observation. The role of endoscopy in the management of acute non-variceal up-
Gastrointest Endosc 1996; 43: 470 – 473 per GI bleeding. Gastrointest Endosc 2012; 75: 1132 – 1138
116 Jensen DM, Kovacs TO, Jutabha R et al. Randomized trial of medical or 142 Yuan Y, Wang C, Hunt RH. Endoscopic clipping for acute nonvariceal
endoscopic therapy to prevent recurrent ulcer hemorrhage in pa- upper-GI bleeding: a meta-analysis and critical appraisal of random-
tients with adherent clots. Gastroenterology 2002; 123: 407 – 413 ized controlled trials. Gastrointest Endosc 2008; 68: 339 – 351
117 Bleau BL, Gostout CJ, Sherman KE et al. Recurrent bleeding from peptic 143 Arima S, Sakata Y, Ogata S et al. Evaluation of hemostasis with soft co-
ulcer associated with adherent clot: a randomized study comparing agulation using endoscopic hemostatic forceps in comparison with
endoscopic treatment with medical therapy. Gastrointest Endosc metallic hemoclips for bleeding gastric ulcers: a prospective, ran-
2002; 56: 1 – 6 domized trial. J Gastroenterol 2010; 45: 501 – 505
118 Kahi CJ, Jensen DM, Sung JJY et al. Endoscopic therapy versus medical 144 Kataoka M, Kawai T, Hayama Y et al. Comparison of hemostasis using
therapy for bleeding peptic ulcer with adherent clot: a meta-analysis. bipolar hemostatic forceps with hemostasis by endoscopic hemoclip-
Gastroenterology 2005; 129: 855 – 862 ping for nonvariceal upper gastrointestinal bleeding in a prospective
119 Wong RC, Chak A, Kobayashi K et al. Role of Doppler US in acute peptic non-randomized trial. Surg Endosc 2013; 27: 3035 – 3038
ulcer hemorrhage: can it predict failure of endoscopic therapy? Gas- 145 Sung JJ, Chan FK, Chen M et al. Asia-Pacific Working Group consensus
trointest Endosc 2000; 52: 315 – 121 on non-variceal upper gastrointestinal bleeding. Gut 2011; 60:
120 Kohler B, Maier M, Benz C et al. Acute ulcer bleeding. A prospective 1170 – 1177
randomized trial to compare Doppler and Forrest classifications in 146 Wong Kee Song LM, Banerjee S, Barth BA et al. Emerging technologies
endoscopic diagnosis and therapy. Dig Dis Sci 1997; 42: 1370 – 1374 for endoscopic hemostasis. Gastrointest Endosc 2012; 75: 933 – 937
121 Fullarton GM, Murray WR. Prediction of rebleeding in peptic ulcers by 147 Manta R, Galloro G, Mangiavillano B et al. Over-the-scope clip (OTSC)
visual stigmata and endoscopic Doppler ultrasound criteria. Endos- represents an effective endoscopic treatment for acute GI bleeding
copy 1990; 22: 68 – 71 after failure of conventional techniques. Surg Endosc 2013; 27:
122 Kohler B, Riemann JF. Endoscopic injection therapy of Forrest II and III 3162 – 3164
gastroduodenal ulcers guided by endoscopic Doppler ultrasound. 148 Giday SA, Kim Y, Krishnamurty DM et al. Long-term randomized con-
Endoscopy 1993; 25: 219 – 223 trolled trial of a novel nanopowder hemostatic agent (TC-325) for
123 van Leerdam ME, Rauws EA, Geraedts AA et al. The role of endoscopic control of severe arterial upper gastrointestinal bleeding in a porcine
Doppler US in patients with peptic ulcer bleeding. Gastrointest En- model. Endoscopy 2011; 43: 296 – 269
dosc 2003; 58: 677 – 684 149 Chen YI, Barkun AN, Soulellis C et al. Use of the endoscopically applied
124 Chen VK, Wong RC. Endoscopic doppler ultrasound versus endoscopic hemostatic powder TC-325 in cancer-related upper GI hemorrhage:
stigmata-directed management of acute peptic ulcer hemorrhage: a preliminary experience. Gastrointest Endosc 2012; 75: 1278 – 1281
multimodel cost analysis. Dig Dis Sci 2007; 52: 149 – 160 150 Leblanc S, Vienne A, Dhooge M et al. Early experience with a novel he-
125 Cipolletta L, Bianco MA, Salerno R et al. Improved characterization of mostatic powder used to treat upper GI bleeding related to malignan-
visible vessels in bleeding ulcers by using magnification endoscopy: cies or after therapeutic interventions. Gastrointest Endosc 2013; 78:
results of a pilot study. Gastrointest Endosc 2010; 72: 413 – 418 169 – 175
126 Barkun AN, Martel M, Toubouti Y et al. Endoscopic hemostasis in pep- 151 Holster IL, Kuipers EJ, Tjwa ET. Hemospray in the treatment of upper
tic ulcer bleeding for patients with high-risk lesions: a series of meta- gastrointestinal hemorrhage in patients on antithrombotic therapy.
analyses. Gastrointest Endosc 2009; 69: 786 – 799 Endoscopy 2013; 45: 63 – 66
127 ASGE Technology Committee. Conway JD, Adler DG et al. Endoscopic 152 Yau AH, Ou G, Galorport C et al. Safety and efficacy of Hemospray in
hemostatic devices. Gastrointest Endosc 2009; 69: 987 – 996 upper gastrointestinal bleeding. Can J Gastroenterol Hepatol 2014;
128 Laine L. Therapeutic endoscopy and bleeding ulcers. Bipolar/multipo- 28: 72 – 76
lar electrocoagulation. Gastrointest Endosc 1990; 36: S38 – S41 153 Sung JJ, Luo D, Wu JC et al. Early clinical experience of the safety and
129 Ginsberg GG, Barkun AN, Bosco JJ et al. The argon plasma coagulator. effectiveness of Hemospray in achieving hemostasis in patients with
Gastrointest Endosc 2002; 55: 807 – 810 acute peptic ulcer bleeding. Endoscopy 2011; 43: 291 – 295
130 Watson JP, Bennett MK, Griffin SM et al. The tissue effect of argon plas- 154 Smith LA, Stanley AJ, Bergman JJ et al. Hemospray application in non-
ma coagulation on esophageal and gastric mucosa. Gastrointest En- variceal upper gastrointestinal bleeding: results of the survey to eval-
dosc 2000; 52: 342 – 345 uate the application of hemospray in the luminal tract. J Clin Gastro-
131 Raju GS, Gajula L. Endoclips for GI endoscopy. Gastrointest Endosc enterol 2014; 48: 89 – 92
2004; 59: 267 – 279
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a20 Guideline
155 Barkun A, Sabbah S, Enns R et al. The Canadian Registry on Nonvari- 177 Iacopini F, Petruzziello L, Marchese M et al. Hemostasis of Dieulafoy’s
ceal Upper Gastrointestinal Bleeding and Endoscopy (RUGBE): Endo- lesions by argon plasma coagulation (with video). Gastrointest En-
scopic hemostasis and proton pump inhibition are associated with dosc 2007; 66: 20 – 26
improved outcomes in a real-life setting. Am J Gastroenterol 2004; 178 Alis H, Oner OZ, Kalayci MU et al. Is endoscopic band ligation superior
99: 1238 – 1246 to injection therapy for Dieulafoy lesion? Surg Endosc 2009; 23:
156 Nahon S, Nouel O, Hagège H et al. Favorable prognosis of upper-gas- 1465 – 1469
trointestinal bleeding in 1041 older patients: results of a prospective 179 Sone Y, Kumada T, Toyoda H et al. Endoscopic management and follow
multicenter study. Clin Gastroenterol Hepatol 2008; 6: 886 – 892 up of Dieulafoy’s lesion in the upper gastrointestinal tract. Endoscopy
157 Loperfido S, Baldo V, Piovesana E et al. Changing trends in acute up- 2005; 37: 449 – 453
per-GI bleeding: a population-based study. Gastrointest Endosc 180 Lim W, Kim TO, Park SB et al. Endoscopic treatment of Dieulafoy le-
2009; 70: 212 – 224 sions and risk factors for rebleeding. Korean J Intern Med 2009; 24:
158 Guntipalli P, Chason R, Elliott A et al. Upper gastrointestinal bleeding 318 – 322
caused by severe esophagitis: a unique clinical syndrome. Dig Dis Sci 181 Durham JD, Kumpe DA, Rothbarth LJ et al. Dieulafoy disease: arterio-
2014; 59: 2997 – 3003 graphic findings and treatment. Radiology 1990; 174: 937 – 941
159 Wang WH, Huang JQ, Zheng GF et al. Head-to-head comparison of H2- 182 Alshumrani G, Almuaikeel M. Angiographic findings and endovascular
receptor antagonists and proton pump inhibitors in the treatment of embolization in Dieulafoy disease: a case report and literature re-
erosive esophagitis: A meta-analysis. World J Gastroenterol 2005; 11: view. Diagn Interv Radiol 2006; 12: 151 – 154
4067 – 4077 183 Jackson CS, Gerson LB. Management of gastrointestinal angiodysplas-
160 Gralnek IM, Dulai GS, Fennerty MB et al. Esomeprazole versus other tic lesions (GIADs): a systematic review and meta-analysis. Am J Gas-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
proton pump inhibitors in erosive esophagitis: a meta-analysis of troenterol 2014; 109: 474 – 483
randomized clinical trials. Clin Gastroenterol Hepatol 2006; 4: 184 Swanson E, Mahgoub A, MacDonald R et al. Medical and endoscopic
1452 – 1458 therapies for angiodysplasia and gastric antral vascular ectasia: a sys-
161 Ljubičić N, Budimir I, Pavić T et al. Mortality in high-risk patients with tematic review. Clin Gastroenterol Hepatol 2014; 12: 571 – 582
bleeding Mallory–Weiss syndrome is similar to that of peptic ulcer 185 Heller SJ, Tokar JL, Nguyen MT et al. Management of bleeding GI tu-
bleeding. Results of a prospective database study. Scand J Gastroen- mors. Gastrointest Endosc 2010; 72: 817 – 824
terol 2014; 49: 458 – 464 186 Sheibani S, Kim JJ, Chen B et al. Natural history of acute upper GI
162 Bharucha AE, Gostout CJ, Balm RK. Clinical and endoscopic risk factors bleeding due to tumours: short-term success and long-term recur-
in the Mallory-Weiss syndrome. Am J Gastroenterol 1997; 92: 805 – rence with or without endoscopic therapy. Aliment Pharmacol Ther
808 2013; 38: 144 – 150
163 Kortas DY, Haas LS, Simpson WG et al. Mallory–Weiss tear: predispos- 187 Kim YI, Choi IJ, Cho SJ et al. Outcome of endoscopic therapy for cancer
ing factors and predictors of a complicated course. Am J Gastroenterol bleeding in patients with unresectable gastric cancer. J Gastroenterol
2001; 96: 2863 – 2865 Hepatol 2013; 28: 1489 – 1495
164 Chung IK, Kim EJ, Hwang KY et al. Evaluation of endoscopic hemostasis 188 Koh KH, Kim K, Kwon DH et al. The successful endoscopic hemostasis
in upper gastrointestinal bleeding related to Mallory–Weiss syn- factors in bleeding from advanced gastric cancer. Gastric Cancer
drome. Endoscopy 2002; 34: 474 – 479 2013; 16: 397 – 403
165 Kim JW, Kim HS, Byun JW et al. Predictive factors of recurrent bleeding 189 Leontiadis GI, Sharma VK, Howden CW. Proton pump inhibitor treat-
in Mallory–Weiss syndrome. Korean J Gastroenterol 2005; 46: 447 – ment for acute peptic ulcer bleeding. Cochrane Database Syst Rev
454 2006; 1: CD002094
166 Fujisawa N, Inamori M, Sekino Y et al. Risk factors for mortality in pa- 190 Leontiadis G, Martin J, Sharma V et al. T1942 Proton pump inhibitor
tients with Mallory–Weiss syndrome. Hepatogastroenterology 2011; (PPI) treatment for peptic ulcer (PU) bleeding: an updated Cochrane
58: 417 – 420 meta-analysis of randomized controlled trials (RCTs) [abstract]. Gas-
167 Huang SP, Wang HP, Lee YC et al. Endoscopic hemoclip placement and troenterology 2009: DOI http://dx.doi.org/10.1016/S0016-5085(09)
epinephrine injection for Mallory-Weiss syndrome with active bleed- 62789-X
ing. Gastrointest Endosc 2002; 55: 842 – 846 191 Sachar H, Vaidya K, Laine L. Intermittent vs continuous proton pump
168 Park CH, Min SW, Sohn YH et al. A prospective, randomized trial of inhibitor therapy for high-risk bleeding ulcers: a systematic review
endoscopic band ligation vs. epinephrine injection for actively bleed- and meta-analysis. JAMA Intern Med 2014; 174: 1755 – 1762
ing Mallory–Weiss syndrome. Gastrointest Endosc 2004; 60: 22 – 27 192 Javid G, Zargar SA, U-Saif R et al. Comparison of p. o. or i.v. proton
169 Cho YS, Chae HS, Kim HK et al. Endoscopic band ligation and endo- pump inhibitors on 72-h intragastric pH in bleeding peptic ulcer. J
scopic hemoclip placement for patients with Mallory–Weiss syn- Gastroenterol Hepatol 2009; 24: 1236 – 1243
drome and active bleeding. World J Gastroenterol 2008; 14: 2080 – 193 Sung JJ, Suen BY, Wu JC et al. Effects of intravenous and oral esome-
2084 prazole in the prevention of recurrent bleeding from peptic ulcers
170 Lecleire S, Antonietti M, Iwanicki-Caron I et al. Endoscopic band liga- after endoscopic therapy. Am J Gastroenterol 2014; 109: 1005 – 1010
tion could decrease recurrent bleeding in Mallory–Weiss syndrome 194 Lau JYW, Sung JJY, Lam YH et al. Endoscopic retreatment compared
as compared to haemostasis by hemoclips plus epinephrine. Aliment with surgery in patients with recurrent bleeding after initial endo-
Pharmacol Ther 2009; 30: 399 – 405 scopic control of bleeding ulcer. N Engl J Med 1999; 340: 751 – 756
171 Lara LF, Sreenarasimhaiah J, Tang SJ et al. Dieulafoy lesions of the GI 195 Wong TCF, Wong TT, Chiu PWY et al. A comparison of angiographic
tract: localization and therapeutic outcomes. Dig Dis Sci 2010; 55: embolization with surgery after failed endoscopic hemostasis to
3436 – 3441 bleeding peptic ulcers. Gastrointest Endosc 2011; 73: 900 – 908
172 Chung IK, Kim EJ, Lee MS et al. Bleeding Dieulafoy’s lesions and the 196 Kyaw M, Tse Y, Ang D et al. Embolization versus surgery for peptic ul-
choice of endoscopic method: comparing the hemostatic efficacy of cer bleeding after failed endoscopic hemostasis: a meta-analysis. En-
mechanical and injection methods. Gastrointest Endosc 2000; 52: dos Int Open 2014; 2: E6 – E14
721 – 724 197 Beggs AD, Dilworth MP, Powell SL et al. A systematic review of transar-
173 Kasapidis P, Georgopoulos P, Delis V et al. Endoscopic management and terial embolization versus emergency surgery in treatment of major
long-term follow-up of Dieulafoy’s lesions in the upper GI tract. Gas- nonvariceal upper gastrointestinal bleeding. Clin Exp Gastroenterol
trointest Endosc 2002; 55: 527 – 531 2014; 7: 93 – 104
174 Cheng CL, Liu NJ, Lee CS et al. Endoscopic management of Dieulafoy le- 198 Sulz M, Frei R, Meyenberger C et al. Routine use of Hemospray for gas-
sions in acute nonvariceal upper gastrointestinal bleeding. Dig Dis Sci trointestinal bleeding: prospective two-center experience in Swit-
2004; 49: 1139 – 1144 zerland. Endoscopy 2014; 46: 619 – 624
175 Park CH, Sohn YH, Lee WS et al. The usefulness of endoscopic hemo- 199 El OualiS, Barkun AN, Wyse J et al. Is routine second-look endoscopy
clipping for bleeding Dieulafoy lesions. Endoscopy 2003; 35: 388 – effective after endoscopic hemostasis in acute peptic ulcer bleeding?
392 A meta-analysis Gastrointest Endosc 2012; 76: 283 – 292
176 Katsinelos P, Paroutoglou G, Mimidis K et al. Endoscopic treatment and 200 Imperiale TF, Kong N. Second look endoscopy for bleeding peptic ulcer
follow-up of gastrointestinal Dieulafoy’s lesions. World J Gastroenter- disease: a decision and cost-effectiveness analysis. J Clin Gastroenter-
ol 2005; 11: 6022 – 6026 ol 2012; 46: e71 – e75
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a21
201 Holster IL, Kuipers EJ. Management of acute nonvariceal upper gastro- 216 Anderson JL, Adams CD, Antman EM et al. 2012 ACCF/AHA focused up-
intestinal bleeding: current policies and future perspectives. World J date incorporated into the ACCF/AHA 2007 guidelines for the man-
Gastroenterol 2012; 18: 1202 – 1207 agement of patients with unstable angina/non-ST-elevation myocar-
202 Sbrozzi-Vanni A, Zullo A, Di Giulio E et al. Low prevalence of idiopathic dial infarction: a report of the American College of Cardiology Foun-
peptic ulcer disease: an Italian endoscopic survey. Dig Liver Dis 2010; dation/American Heart Association Task Force on Practice Guidelines.
42: 773 – 776 J Am Coll Cardiol 2013; 61: 179 – 347
203 Gisbert JP, Khorrami S, Carballo F et al. Meta-analysis: Helicobacter py- 217 Garcia-Rodriguez LA, Lin KJ, Hernandez-Diaz S et al. Risk of upper gas-
lori eradication therapy vs. anti-secretory non-eradication therapy trointestinal bleeding with low dose acetylsalicylic acid alone and in
for the prevention of recurrent bleeding from peptic ulcer. Aliment combination with clopidogrel and other medications. Circulation
Pharmacol Ther 2004; 19: 617 – 629 2011; 123: 1108 – 1115
204 Gisbert JP, Calvet X, Cosme A et al. Long-term follow-up of 1,000 pa- 218 Lanas A, Garcia-Rodriguez LA, Arroyo MT et al. Effect of anti-secretory
tients cured of Helicobacter pylori infection following an episode of drugs and nitrates on the risk of ulcer bleeding associated with non-
peptic ulcer bleeding. Am J Gastroenterol 2012; 107: 1197 – 1204 steroidal anti-inflammatory drugs, antiplatelet agents, and anticoa-
205 Gisbert JP, Abraira V. Accuracy of Helicobacter pylori diagnostic tests gulants. Am J Gastroenterol 2007; 102: 507 – 515
in patients with bleeding peptic ulcer: a systematic review and meta- 219 Bhatt DL, Scheiman J, Abraham NS et al. ACCF/ACG/AHA 2008 expert
analysis. Am J Gastroenterol 2006; 101: 848 – 863 consensus document on reducing the gastrointestinal risks of anti-
206 Sánchez-Delgado J, Gené E, Suárez D et al. Has H. pylori prevalence in platelet therapy and NSAID use: a report of the American College of
bleeding peptic ulcer been underestimated? A meta-regression Am J Cardiology Foundation Task Force on Clinical Expert Consensus
Gastroenterol 2011; 106: 398 – 405 Documents. Circulation 2008; 118: 1894 – 1909
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
207 Witt DM, Delate T, Garcia DA et al. Risk of thromboembolism, recur- 220 Kwok CS, Nijar RS, Loke YK. Effects of proton pump inhibitors on ad-
rent hemorrhage, and death after warfarin therapy interruption for verse gastrointestinal events in patients receiving clopidogrel: a sys-
gastrointestinal tract bleeding. Arch Intern Med 2012; 172: 1484 – tematic review and meta-analysis. Drug Saf 2011; 34: 47 – 57
1491 221 Gilard M, Arnaud B, Cornily JC et al. Influence of omeprazole on the
208 Lee JK, Kang HW, Kim SG et al. Risks related with withholding and re- antiplatelet action of clopidogrel associated with aspirin: the ran-
suming anticoagulation in patients with non-variceal upper gastroin- domized, double-blind OCLA (Omeprazole CLopidogrel Aspirin)
testinal bleeding while on warfarin therapy. Int J Clin Pract 2012; 66: study. J Am Coll Cardiol 2008; 51: 256 – 260
64 – 68 222 Cuisset T, Frere C, Quilici J et al. Comparison of omeprazole and panto-
209 Qureshi W, Mittal C, Patsias I et al. Restarting anticoagulation and out- prazole influence on a high 150-mg clopidogrel maintenance dose:
comes after major gastrointestinal bleeding in atrial fibrillation. Am J the PACA (Proton Pump Inhibitors And Clopidogrel Association) pro-
Cardiol 2014; 113: 662 – 668 spective randomized study. J Am Coll Cardiol 2009; 54: 1149 – 1153
210 Douketis JD, Spyropoulos AC, Spencer FA et al. Perioperative manage- 223 Siller-Matula JM, Spiel AO, Lang IM et al. Effects of pantoprazole and
ment of antithrombotic therapy and prevention of thrombosis, 9th esomeprazole on platelet inhibition by clopidogrel. Am Heart J
ed: American College of Chest Physicians evidence-based clinical 2009; 157: 148 – 145
practice guidelines. Chest 2012; 141: Suppl e326S – e350S 224 O’Donoghue ML, Braunwald E, Antman EM et al. Pharmacodynamic ef-
211 Sibon I, Orgogozo JM. Antiplatelet drug discontinuation is a risk factor fect and clinical efficacy of clopidogrel and prasugrel with or without
for ischemic stroke. Neurology 2004; 62: 1187 – 1189 a proton-pump inhibitor: an analysis of two randomised trials. Lan-
212 Biondi-Zoccai GG, Lotrionte M, Agostoni P et al. A systematic review cet 2009; 374: 989 – 997
and meta-analysis on the hazards of discontinuing or not adhering 225 Chen J, Chen SY, Lian JJ et al. Pharmacodynamic impacts of proton
to aspirin among 50,279 patients at risk for coronary artery disease. pump inhibitors on the efficacy of clopidogrel in vivo – a systematic
Eur Heart J 2006; 27: 2667 – 2674 review. Clin Cardiol 2013; 36: 184 – 189
213 Garcia-Rodriguez LA, Cea-Soriano L, Martin-Merino E et al. Disconti- 226 Kwok CS, Loke YK. Meta-analysis: the effects of proton pump inhibi-
nuation of low dose aspirin and risk of myocardial infarction: case– tors on cardiovascular events and mortality in patients receiving clo-
control study in UK primary care. BMJ 2011; 343: d4094 DOI pidogrel. Aliment Pharmacol Ther 2010; 31: 810 – 823
10.1136/bmj.d4094 227 Siller-Matula JM, Jilma B, Schror K et al. Effect of proton pump inhibi-
214 Cea Soriano L, Bueno H, Lanas A et al. Cardiovascular and upper gas- tors on clinical outcome in patients treated with clopidogrel: a sys-
trointestinal bleeding consequences of low dose acetylsalicylic acid tematic review and meta-analysis. J Thromb Haemost 2010; 8:
discontinuation. Thromb Haemost 2013; 110: 1298 – 1304 2624 – 2641
215 King SBIII, Smith SCJr, Hirshfeld JWJr et al. 2007 focused update of the 228 Cardoso RN, Benjo AM, DiNicolantonio JJ et al. Incidence of cardiovas-
ACC/AHA/SCAI 2005 guideline update for percutaneous coronary in- cular events and gastrointestinal bleeding in patients receiving clopi-
tervention: a report of the American College of Cardiology/American dogrel with and without proton pump inhibitors: an updated meta-
Heart Association Task Force on Practice guidelines. J Am Coll Cardiol analysis. Open Heart 2015; 2: e000248 DOI 10.1136/openhrt-2015-
2008; 51: 172 – 209 000248
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a22 Guideline
Appendix e1 Nonvariceal upper gastrointestinal hemorrhage (NVUGIH): task forces and key questions.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
What is the role of pre-endoscopy somatostatin therapy? Franco Radaelli
What is the role of naso-/orogastric tube aspiration/lavage? Gianluca Rotondano
What is the role of endotracheal intubation before upper endoscopy?
Is there a role for antifibrinolytic medications?
What is the role of prokinetic agents prior to upper endoscopy?
Is there a role for capsule endoscopy in the emergency department in evaluating acute UGI bleeding?
What is appropriate timing for upper endoscopy?
Task force 3: Endoscopic management Ernst J. Kuipers
Which endoscopic classification should be used for describing high and low risk endoscopic stigmata of recent hemorrhage Ricardo Cardoso
in peptic ulcer bleeding? What are high risk vs. low risk endoscopic stigmata and their importance in risk stratification? Livio Cipolletta
Is there a role for doppler ultrasonography, magnification endoscopy, chromoendoscopy in helping to better evaluate Mário Dinis-Ribeiro
endoscopic stigmata of recent hemorrhage for peptic ulcer bleeding? Luís Maia
Which ulcer stigmata require endoscopic hemostasis? Which do not? Gianluca Rotondano
Which endoscopic hemostasis modality should be used (with focus on peptic ulcer bleeding)? Paulo Salgueiro
Injection therapy?
Thermal contact therapy?
Thermal noncontact therapy?
Mechanical therapy?
Combination therapy?
Topical spray/powder therapy
What to do in situations of nonvariceal, nonulcer bleeding lesions?
Task force 4: Post-endoscopic management Angel Lanas
What is the medical management post endoscopic hemostasis? Lars Aabakken
What to do when rebleeding occurs? What is the role of repeat upper endoscopy? Alberto Arezzo
Is there a role for scheduled second-look endoscopy? Roberto de Franchis
Rebleeding/failed endoscopic hemostasis: When should the interventional radiologist be involved/when should the Cesare Hassan
surgeon be involved? Ralf-Thorsten Hoffmann
Diagnosis and treatment of Helicobacter pylori? When? In whom? What if testing for H. pylori in the acute setting of Tomas Hucl
bleeding is negative? Documentation of eradication? Gilles Lesur
How to manage the NVUGIH patient using antiplatelet and anticoagulant drugs (collectively known as antithrombotics Franco Radaelli
agents) post endoscopy? How and when to reinstitute these medications? Andrew Veitch
Angelo Zullo
When to discharge patients home?
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Appendix e2 Criteria for outpatient management of patients with nonvariceal upper gastrointestinal hemorrhage (NVUGIH).
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a23
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a24
Appendix e2 (Continuation)
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Endoscopic criteria.
No clot obscuring complete visualization;
Clean ulcer base;
Mallory-Weiss tear;
Esophagitis, gastritis, duodenitis, or other benign findings.
Hemodynamically stable, alert, wanted to go home
Gralnek, Retrospective Complete Rockall scores (after endoscopy) of ≤ 2 175 53 (30 %) 2 (3.7 %) 0 0
2004 [235]
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e2 (Continuation)
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a25
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a26
Appendix e3 Role of proton pump inhibitors (PPIs) prior to upper endoscopy in acute upper gastrointestinal hemorrhage.
First author, Study type Patient group Key outcomes Key results Limitations Conclusions
Guideline
year [ref.]
Sreedharan, Cochrane meta- Total of 2223 patients were – 30-day mortality No significant difference between PPI Oral and intravenous PPI studies PPI treatment initiated before
2012 [237] analysis of 6 RCTs included with unselected – Rebleeding and control treatment for: are mixed endoscopy for upper GI bleeding
Searches were re-run upper GI bleeding. – Surgery – Mortality: 6.1 % vs. 5.5 %; OR 1.12, – No data regarding the length might reduce the proportion of
in Feb 2006 and Oct Prior to endoscopy patients – Stigmata of recent hemor- 95 %CI 0.72 – 1.73 of pre-emptive PPI and place- patients with SRH at index endos-
2008 were given PPI (oral or rhage (SRH) at index endos- – Rebleeding rates: 13.9 % vs. bo treatment copy and significantly reduces
intravenous) or H2RA or copy 16.6 %; OR 0.81, 95 %CI 0.61 – – Variceal and nonvariceal requirement for endoscopic ther-
placebo – Need for endoscopic therapy 1.09 bleeding sources are not apy during index endoscopy.
– Need for transfusion – Surgery: 9.9 % – 10.2 %; OR 0.96, analyzed selectively However, there is no evidence
95 %CI 0.68 – 1.35 that PPI treatment affects clini-
Significant reduction: cally important outcomes, name-
– Proportion with SRH: ly mortality or need for surgery.
37.2 % – 46.5 %; OR 0.67, 95 %CI
0.54 – 0.84
– Need for endoscopic therapy:
8.6 % – 11.7 %; OR 0.68, 95 %CI
0.50 – 0.93
Lau, 2007 Double-blind, place- Omeprazole group (n = Primary: Among patients with peptic ulcer Long-term aspirin users were Pre-emptive omeprazole appears
[238] bo-controlled, 319): 80 mg omeprazole – Need for endoscopic therapy bleeding, endoscopic treatment re- excluded to accelerate the resolution of
randomized trial intravenous bolus + 8 mg/h at the first endoscopic quired in: signs of bleeding.
control infusion until examination – Omeprazole group 22.5 % Fewer cases of actively bleeding
endoscopy Secondary: – Placebo group 36.8 % ulcers were seen among patients
Placebo group (n = 319): – Signs of bleeding – OR 0.61, 95 %CI 0.44 – 0.84; receiving omeprazole than
placebo bolus + placebo – Need for urgent endoscopy P = 0.002 among those who received
infusion until endoscopy – Emergency surgery Active ulcer bleeding seen less placebo.
Endoscopies were per- – Recurrent bleeding frequently in omeprazole group In patients awaiting endoscopy
formed next morning. than in placebo group: pre-emptive use of high dose in-
Mean duration of infusion 6.4 % vs. 14.7 %; P = 0.001. travenous omeprazole is recom-
before endoscopy: Rebleedings: 3.5 % in omeprazole mended.
– Omeprazole group group vs. 2.5 % in placebo group;
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
14.7 ± 6.3 h P = 0.49
– Placebo group
15.2 ± 6.2 h
Liu, 2012 Randomized, single- PPI parenteral (either ome- Rebleeding rate Rebleeding: Both omeprazole and esomepra- High dose intravenous PPI initi-
[239] center, prospective, prazole or esomeprazole) Need for endoscopic therapy at 11 % in standard regimen group vs. zole were used ated before endoscopy reduces
double-blind started prior to endoscopy first endoscopy 6.4 % in intensive regimen group Endoscopic therapy was not rebleeding rates, blood transfu-
(“early” endoscopy ≤ 24 h, (P = 0.03) standardized sion volume and hospital stay,
“late” > 24 – 72 h) Early endoscopy therapy need: No detailed data about the timing especially when endoscopy is
Intensive regimen, 80 mg + 16.7 % vs. 10.0 % (P = 0.05) of endoscopy (how many hours delayed beyond 24 h of presen-
8 mg/h infusion: Rebleeding rates at late endoscopies: after the initiation of pre-emptive tation.
410 patients 7.5 % in standard vs. 4.0 % in intensive PPI therapy) No alteration on need for endo-
Standard regimen, PPI group (P = 0.03) scopic treatment.
2 × 40 mg PPI/day: 456 pa-
tients
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e3 (Continuation)
First author, Study type Patient group Key outcomes Key results Limitations Conclusions
year [ref.]
Barkun, 2010 International Consen- NA Rebleeding Pre-endoscopic PPI treatment signifi- NA Pre-endoscopic PPI therapy may
[240] sus Guideline Patients receiving pre- Need for surgery cantly reduced the proportion of be considered in order to down-
emptive PPI before endos- Mortality patients with high risk stigmata (OR stage the endoscopic lesion and
copy. Need for intervention 0.67, CI 0.54 – 0.84), and the need for decrease the need for endoscopic
Supportive cost-effective analysis pre-endoscopic therapy (OR 0.68, intervention but should not delay
CI 0.50 – 0.93) compared with the endoscopy.
control group (placebo or H2RA). Comment: ESGE Board endorsed
Economic dominance of pre-endo- the consensus recommendation.
scopic high dose intravenous PPI
therapy.
Tsoi, 2008 Decision analysis Omeprazole group (80 mg Number of patients who avoided 248 patients in the PPI group and 227 The estimated values were based The use of high dose PPI before
[241] model bolus intravenous + 8 mg/h endoscopic therapy within the patients in the placebo group avoided on an assumed linear relationship an endoscopy is shown to be an
Cost-effectiveness infusion): follow-up period endoscopic therapy. effective and cost-saving way to
comparison n = 314 patients Direct costs of medical treat- Overall direct costs: 2813 US dollars treat patients with upper GI
187 peptic ulcer cases ment, diagnostic endoscopy, (USD) in the PPI group vs. 2948 USD in bleeding.
– initiated prior to hemostasis, emergency surgery, the placebo group
endoscopy hospitalization Costs were reduced by 7.4 % with
Placebo group: pre-emptive PPI therapy
n = 317 patients
190 peptic ulcer cases
– initiated prior to
endoscopy
Barkun, 2008 Editorial NA Key questions: NA NA The earlier the endoscopy is per-
[242] To contrast the pre-endoscopic formed, the less the cost-effec-
with post-endoscopic hemostatic tiveness of high dose intravenous
uses of PPI (in high dose intrave- PPI.
nous administration) in patients Pre-emptive PPI use should not
with peptic ulcer bleeding replace the performance of early
Almost no data about the time endoscopy.
elapsed until endoscopy (i. e., the
duration of intravenous PPI ad-
ministration before gastroscopy)
What is the optimal duration of
pre-emptive high dose intrave-
nous PPI therapy regarding
cost-effectiveness
Which subgroups of ulcer
patients benefits most from
pre-endoscopy PPI therapy
Guideline
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a27
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a28
Appendix e3 (Continuation)
First author, Study type Patient group Key outcomes Key results Limitations Conclusions
year [ref.]
Guideline
Rácz, 2012 Single-center, retro- 240 PUB patients, panto- Active ulcer bleeding at the first Mean duration of infusion before Retrospective study Pre-emptive infusion of high dose
[243] spective, comparative prazole bolus 80 mg intra- endoscopy endoscopy: Aspirin, NSAIDs, and anticoagu- pantoprazole longer than 4 h
cohort study venous + 8 mg/h pantopra- Need for endoscopic hemostasis 5.4 ± 12.9 h in the saline group vs. 6.9 lants were discontinued at time of before endoscopy decreased the
zole infusion before endos- at the first endoscopy ± 13.2 h in the pantoprazole group hospital admission. ratio of active bleeding only in
copy Need for urgent endoscopy (P = 0.29). Only data from patients admitted gastric and not in duodenal ulcer
Control group: 93 PUB All outcomes were analyzed in Active bleeding at first endoscopy: during evening hours (6pm to patients.
patients received saline subgroups of patients with differ- 19.2 % in the pantoprazole group vs. 8am) were analyzed
infusion until endoscopy ent durations of intravenous pan- 24.7 % in the saline infusion group
toprazole before endoscopy. (P = 0.26).
Rebleeding rates Endoscopic treatment requirement at
Separate analyses in duodenal first endoscopy:
ulcer and gastric ulcer patients 61.3 % in the pantoprazole group vs.
56.9 % in the saline group (P = 0.82)
Lanas, 2013 Summary of main NA – – – It has been reported that the
[244] studies administration of a PPI prior to
endoscopy or the early perform-
ance of endoscopy within 6 h of
admission in patients with UGIB
does not improve the prognosis of
the event.
Sung, 2011 Consensus report NA NA – – A pre-endoscopy proton pump
[245] inhibitor (PPI) is recommended as
a stopgap treatment when
endoscopy within 24 h is not
available.
Al-Sabah, Decision model com- Cost-effectiveness – – Intravenous PPIs given before
2008 [246] pared high dose PPI in- endoscopy are slightly more ef-
itiated while awaiting fective than no administration.
endoscopy with ad-
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
ministration on the
basis of endoscopic
findings
Ghassemi. Overview NA NA NA – Intravenous PPI therapy before
2009 [247] endoscopy seems reasonable in
patients presenting with severe
upper gastrointestinal bleeding
suspected from a peptic ulcer, if
a delay is anticipated in urgent
endoscopy.
Laursen, 2012 Guideline approved by NA NA NA – Treatment with PPI prior to
[248] the Society of Danish endoscopy cannot be recommen-
Society for Gastro- ded and must not delay the timing
enterology of upper endoscopy.
Lin, 2010 Overview NA NA NA – There was no evidence that PPI
[249] before endoscopy improves
clinical outcomes.
ESGE, European Society of Gastrointestinal Endoscopy; H2RA, histamine-2 receptor antagonist; trial; NA, not available; NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor; PUB, peptic ulcer bleeding; RCT, randomized controlled trial;
UGIB, upper gastrointestinal bleeding.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e4 Role of tranexamic acid (TXA) in upper gastrointestinal hemorrhage.
First author, Study type Study population, Interven- Key outcomes Key results Limitations Conclusion
year [ref.] tions
Gluud, 2012 Cochrane meta-analysis of Patients admitted with sus- – Mortality (1654 patients, TXA versus placebo: Methodological quality of stud- TXA cannot be recommended
[250] 7 RCTs vs. placebo pected upper gastrointestinal 7 studies) Significant difference in ies included (unclear or high risk for routine use.
1 RCT included also a treatment bleeding confirmed by gastric – Rebleeding or continued mortality: 5 % vs. 8.2 %; RR 0.61 of bias of selection bias in many Additional trials in which TXA is
arm with cimetidine (Ref. 5) lavage, hematemesis or melena bleeding (1604 patients, (95 %CI 0.42 – 0.89) of them) used in combination with the
1 RCT included also a treatment Interventions: 6 studies) No significant difference in a Scarce applicability of evidence currently recommended inter-
arm with lansoprazole and – Oral TXA administration – Surgery (1504 patients, subgroup analysis stratified for (most trials do not include the ventions are required.
lansoprazole + TXA (3 studies; 6 studies) quality of bias control (trials currently recommended inter-
– Intravenous TXA adminis- – Need for transfusion (1504 with adequate allocation vention for management of
tration (4 studies. patients, 6 studies) sequence generation or alloca- NVUGIB)
Treatment duration (range): – Any thromboembolic event tion concealment included):
2 – 7 days (1048 patients, 3 studies RR 0.78, 95 %CI 0.58 – 1.05)
Total daily TXA dose, range: No significant difference in:
4 – 8 g (divided in four to six – Rebleeding: RR 0.73,
daily doses) 95 %CI 0.5 – 1.07
1 RCT offered endoscopic – Surgery: RR 0.62,
treatment to all randomized 95 %CI 0.35 – 1.09
patients – Need for transfusion: 1.02,
95 %CI 0.93 – 1.11
– Any thromboembolic
event: RR 1.86, 95 %CI
0.66 – 5.24
TXA versus cimetidine or
lansoprazole:
No significant difference in all
outcome variables
CI, confidence interval; NVUGIB, nonvariceal upper gastrointestinal bleeding; RCT, randomized controlled trial; RR, relative risk.
Guideline
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a29
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a30
First author, year [ref.] Study type Patient group Key outcomes Key results Limitation Conclusion
Guideline
Magnusson, 1985 [251] Randomized, 95 patients with acute non- Need for surgical treatment Patients who needed surgery Only included patients Somatostatin infusion was
double‐blind variceal GI bleeds with “mas- (indication: > 6 units PRBC (P < 0.04): with signs of shock superior to placebo regard-
sive” bleeding (clinical signs needed to keep hemodyna- – 14 placebo Endoscopic treatment ing surgical need in patients
of shock or preshock) mically stable – 5 somatostatin was not standardized with nonvariceal upper gas-
– 46 patients: 72 h soma- Evidence of active bleeding P < 0.04 trointestinal bleeding.
tostatin infusion 1 day after treatment Rebleeding:
– 49 patients: placebo – 5 placebo
– 6 somatostatin
Choi, 2011 [252] Retrospective analysis of 101 patients with acute ulcer Clinically significant early Rebleeding rates: Not randomized study Adjunctive somatostatin for
a prospective database bleeding, high risk stigmata rebleeding – PPI: 12.2 % management of nonvariceal
(FIa,Ib, IIa) Loss of endoscopic high risk – 14.3 % PPI + somatostatin upper gastrointestinal
Endoscopic therapy in all stigmata at second-look P = 0.766 bleeding did not show an ad-
patients endoscopy Loss of stigmata: ditive effect in reducing early
– 52 patients: pantopra- – 94.2 % PPI rebleeding.
zole 80-mg bolus + con- – 95.9 % PPI + somatostatin
tinuous infusion 8 mg/h P = 0.696
for 72 h
– 49 patients: pantopra-
zole as above, and also
somatostatin as 250-µg
bolus + continuous infu-
sion 250 µg/h for 72 h
Avgerinos, 2005 [253] Randomized, Adult patients admitted Gastric pH compared with Successful maintenance of pH > 4.0 186 patients screened; During the first 12 h of the
double‐blind within 24 h of bleeding patient baseline gastric pH in patients having somatostatin 143 excluded infusion somatostatin was
Endoscopic stage IIc and III during drug infusion (P < 0.0001) and PPI (P < 0.0001) Patients with stigmata more effective than panto-
peptic ulcer bleeding Intragastric pH > 4 in the During first half of treatment, time of recent hemorrhage prazole maintaining high
Not on PPI or H2 blockers in fundus during the 24-h with above pH 4.0 and 5.4, respec- were excluded intragastric pH.
the previous week infusion period tively, was higher with somatostatin Evaluation of rebleeding
No hypovolemia than PPI (P < 0.005) and (P < 0.02) was only a secondary
Normal platelets, no end point
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
coagulopathy
14 somatostatin
14 PPI
15 placebo
Archimandritis, 2000 [254] Prospective, 84 patients with acute non- Need for surgical inter- No difference between groups for: Small numbers
randomized, open variceal upper GI bleed vention – Hospital stay length (P = 0.25) Not blinded
– 44 ranitidine Hospital stay length – Amount of blood units
– 40 ranitidine + Blood units transfused (P = 0.16)
octreotide Need for emergency surgery did
not differ (P = 1.0)
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e5 (Continuation)
First author, year [ref.] Study type Patient group Key outcomes Key results Limitation Conclusion
Lin, 1995 [255] Randomized, open Patients with active peptic Hemostasis Hemostasis achieved in 35/42 Not blinded
ulcer bleeding, or Nonbleed- Blood units transfused (83.3 %) octreotide group vs. 23/42
ing visible vessel at ulcer (54.8 %) in ranitidine group;
bases (P < 0.01)
– 42 octreotide Octreotide group had:
– 42 ranitidine – Lower volume of blood
(P < 0.05)
– Fewer patients needing endo-
scopic hemostasis/surgery
(P < 0.05)
– Fewer days in hospital
(P < 0.001)
Kim, 2008 [256] Retrospective analysis Patients with peptic ulcer Rebleeding in: 72 h rebleeding: Not randomized Combined therapy with PPI
bleeding, Forrest Ia, b, II a,b – 72 h, and – PPI alone: 11 % Endoscopic treatment and somatostatin did not re-
Group A: 45 patients; 48 h – 30 days – PPI + somatostatin: 13 % was not standardized sult in better outcomes than
pantoprazole infusion Not significant PPI alone.
Group B: 45 patients; 48 h 30-day rebleeding:
pantoprazole + 250 µg/h – 13 % vs. 16 %
infusion of somatostatin Not significant
Antonioli, 1986 [257] Multicenter, randomized, 56 PUB patients Secondary hemostasis Hemostasis: Small patient numbers Somatostatin infusion for
prospective, controlled trial Group A: 250 µg/h infusion Transfusion requirement – Somatostatin group: 93.3 % Rebleeding rate was not 48 h is superior to cimetidine
somatostatin for 48 h – Cimetidine group: 61.5 % defined infusion regarding
Group B: 1600 µg/24 h cime- P < 0.01 definitive hemostasis.
tidine for 48 h after endos- Blood requirement:
copy – Somatostatin: 1.14 PRBC
– Cimetidine: 2.46 PRBC
P < 0.05
Tisbouris, 2007 [258] Randomized, double-blind, Patients with peptic ulcer Rebleeding 3 days after initial Rebleeding and need for endo- Endoscopic treatment Infusion of pantoprazole
controlled trial bleeding, Forrest Ia, Ib, Iia hemostasis scopic re-treatment: was not standardized after initial hemostasis was
Group P: 82 patients, 8 mg/h Need for endoscopic – Group P: 5 % vs. Not placebo-controlled superior to somatostatin
pantoprazole infusion for re-treatment – Group S: 17 % infusion to prevent bleeding
48 h (P = 0.046) recurrence
Group S: 82 patients, 250 µg Achieved pH > 6 for more than No difference in the need for
somatostatin infusion for 85 % of the time: surgery and in mortality.
48 h – Group P: 56.7 % vs.
– Group S: 46.7 %
(P = 0.44)
Guideline
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a31
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a32
Appendix e5 (Continuation)
First author, year [ref.] Study type Patient group Key outcomes Key results Limitation Conclusion
Guideline
Okan, 2000 [259] Prospective, randomized, Total of 48 patients with up- Blood transfusion require- Transfusion requirement: no sign. Low patient numbers Somatostatin is more effec-
double-blind controlled per gastrointestinal bleeding ment difference tive than ranitidine in con-
study Group I: 15 Forrest Ib Time to bleeding cessation In group I (Forrest Ib patients), time trolling acute nonvariceal
patients Rebleeding rate to bleeding cessation was signifi- upper gastrointestinal
Group II: 30 Forrest II cantly shorter in those receiving bleeding in patients with
patients somatostatin vs. those receiving Forrest Ib bleeding activity.
Randomly administered: ranitidine (3.24 vs. 11.25 h, No additional benefit with
– Somatostatin 250 µg bo- P = 0.038) Forrest II bleeding activity.
lus + continuous infusion No differences in rebleeding rate or
6 mg/day for 72 h, or mortality
– Ranitidine 300 mg/day
continuous infusion for
72 h
Rutgeerts, 2006 [260] Double-blind randomized 369 PUB patients Failure rate (Forrest Ia or Ib at Failure rates in S and P groups: Unusual presentation of The results of this multicen-
controlled trial Before endoscopy, random- diagnostic endoscopy, clini- – No statistically significant results ter large trial were published
ized for intravenous treat- cal signs of active bleeding difference, 34 % vs. 36 % “Failure rate” is not an only in abstract form. How-
ment: ≥ 1 h after the start of infu- Stigmata of bleeding: established measure of ever, the results appear to
– S group: 250 µg bolus + sion – Statistically significant effect in results in studies support the use of intrave-
12 mg/day somatostatin Use of rescue therapy (endo- favor of somatostatin nous somatostatin adminis-
for 72 h scopic treatment, surgery, – P = 0.034 tered early and prior to diag-
– P group: placebo infu- vasoactive drugs) to stop Rescue treatment: nostic endoscopy and endo-
sion for 72 h bleeding – Significantly less in S vs. P group therapy to control bleeding
Diagnostic endoscopy within P = 0.012 in patients with suspected
8 h of randomization acute and severe PUB.
PPI, proton pump inhibitor; PRBC, packed red blood cells; PUB, peptic ulcer bleeding.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e6 Role of prokinetic agents in acute overt upper gastrointestinal hemorrhage.
First author, year Study type Patient group Key outcomes Key results Limitation Conclusion
[ref.]
Carbonell, 2006, Prospective, random- Upper gastrointestinal bleeding Improvement of gastric cleans- – Better visualization of entire Low patient numbers Intravenous erythromycin
[261] ized, double-blind patients, all with nasogastric ing prior to endoscopy gastric mucosa in erythro- Subjective criteria based on the before endoscopy improves
tube Identification of bleeding point mycin patients: 65 % vs. 44 % endoscopist’s judgment stomach cleansing and the
– 49 patients, 250 mg ery- Effectiveness of hemostasis P < 0.05 quality of endoscopic examina-
thromycin prior to endos- treatment – Clots in the stomach: 30 % tion.
copy erythromycin vs. 52 % place- Limited clinical benefit.
– 50 patients, placebo prior to bo
endoscopy P < 0.05
– Hemostatic treatment: 67 %
erythromycin vs. 62 % place-
bo
P = 0.67
Coffin, 2002 [262] Prospective, random- Acute upper gastrointestinal Quality of EGD examination Significantly better visualization Low patient numbers Erythromycin infusion before
ized, endoscopist- bleeding within 12 h requiring evaluated using a score during EGD in the erythromycin Qualitative evaluation of results endoscopy significantly im-
blinded, controlled esophagogastro-duodenoscopy Need for second EGD group (P = 0.02) (by scoring system) proved the quality of EGD.
trial (EGD) Effectiveness of endoscopy No interference with the hemo- Tendency to reduce the need for
Before EGD patients were given: therapy static procedure second-look endoscopy.
– 3 mg/kg erythromycin in Need for second endoscopy:
125 mL saline as a 30-min- erythromycin:15.8 % vs. con-
ute infusion, or trol:45.4 % (P = 0.089)
– Observation without motili-
ty agent
EGD within 60 – 120 min follow-
ing erythromycin infusion
Erythromycin group:
19 patients
Control group: 22 patients
Frossard, 2002 Prospective, random- – Erythromycin intravenous Effect of intravenous erythro- Clear stomach found more often Low patient number with non- Erythromycin infusion before
[263] ized, double-blind, 250 mg: 51 patients mycin on endoscopic yield in erythromycin group: 82 % vs. variceal upper gastrointestinal endoscopy in patients with
monocentric – Placebo: 54 patients Need for second-look endos- 33 %, P < 0.001 bleeding recent hematemesis makes
20 minutes before endoscopy copy Erythromycin shortened the endoscopy shorter and easier,
Within 12 h after hematemesis Transfusion requirement endoscopy duration and the and reduces the need for
need for second-look endoscopy second-look endoscopy.
Theivanayagam, Meta-analysis of 558 patients with upper gastro- Visualization of gastric mucosa Erythromycin infusion before Inclusion of variceal bleedings Erythromycin infusion before
2013 [264] 6 studies intestinal bleeding (patients Need for second endoscopy endoscopy significantly im- endoscopy in patients with up-
with variceal bleeds were also proved visualization vs. no ery- per gastrointestinal bleeding
included) thromycin: significantly improves visuali-
OR 3.43, 95 %CI 1.81 – 6.50 zation of gastric mucosa and
Significant decrease in the need decreases the need for a second
for second endoscopy (P = 0.01) endoscopy.
Guideline
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a33
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a34
Appendix e6 (Continuation)
First author, year Study type Patient group Key outcomes Key results Limitation Conclusion
[ref.]
Guideline
Sussman, 2008 Single-center, pro- Patients with overt hema- Mucosal visualization during Better visualization after meto- Low patient numbers Intravenous metoclopramide
[265] spective, random- temesis or melena EGD (Avgerinos score) clopramide, P = not significant Subjective quality score may increase visualization of the
ized, controlled study A total of 26 patients Duration of EGD No significant difference re- proximal stomach during EGD,
Intravenous metoclo- Rebleeding rate garding blood transfusion, re- but not significantly.
pramide bolus (10 mg peat EGD No translation into better
intravenous) clinical outcome.
Barkun, 2010 [266] Meta-analysis 3 erythromycin studies Primary: A prokinetic agent significantly Small numbers of patients Intravenous erythromycin or
3 fully published 2 metoclopramide studies – Need for repeat EGD reduced the need for repeat metoclopramide immediately
articles, 2 abstracts 316 patients Secondary: EGD: before EGD in patients with
– Blood transfusion OR 0.55, 95 %CI 0.32 – 0.94 acute upper gastrointestinal
– Need for surgery No significant alterations: bleeding decreases the need for
– Blood transfusion, and repeat EGD, but does not im-
– Need for surgery: prove clinical outcomes.
Habashi, 2007 [267] Prospective, random- 54 patients with hematemesis Visualization of the mucosa Gastric mucosa entirely Small sample sizes No significant difference be-
ized, controlled trial (42 % PUB patients) Quality of the EGD visualized: Subjective scores tween placebo and either medi-
– Intravenous erythromycin: – erythromycin group: 86 %, Dosage of prokinetics not cation regarding key outcomes.
15 patients vs. reported
– Intravenous metoclopra- – metoclopramide group:
mide: 15 patients 66 %
– Placebo: 15 patients Quality of EGD significantly
better in erythromycin group vs.
metoclopramide and placebo
groups
Winstead, 2007 Cost-effectiveness a- 3 RCTs were analyzed with a total Quality-adjusted life-years Intravenous erythromycin be- The estimates need for repeat Intravenous erythromycin
[268] nalysis of 126 patients (QALYs) fore endoscopy was cost-effec- endoscopy. before endoscopy in acute
(Frossard, Coffin and Carbonell Need for second-look endos- tive when the rate of second- The complex model may fail to upper gastrointestinal hemor-
studies) copy look endoscopy was less than capture all the slight variations rhage (UGIH) is cost-saving and
0.29 in patterns of care among dif- increases QALYs.
Intravenous erythromycin was ferent hospitals. Erythromycin is recommended
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
cost-effective when the charges prior to EGD in UGIH.
for uncomplicated peptic ulcer
disease PUD were less than
8000 US dollars
CI, confidence interval; EGD, esophagogastroduodenoscopy; PUB, peptic ulcer bleeding; OR, odds ratio; RCT, randomized controlled trial.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Guideline a35
Appendix e7 Summary of the evidence regarding impact of early endoscopy (≤ 24 h) on the outcome of patients with nonvariceal upper gastrointestinal
hemorrhage (NVUGIH).
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
utilization
Tsoi, 2009 – Systematic review 5 677 Early endoscopy aids risk stratification and reduces the need for hospi-
[270] 1996 – 2007 talization; however it may increase the use of unnecessary therapeutic
8 studies procedures.
– 3 RCT Endoscopy performed ≤ 8 h of presentation has no advantage over
– 5 retrospective endoscopy performed within 12 – 24 h of presentation in reducing
recurrent bleeding or improving survival.
Sarin, 2009 Canada Retrospective 502 No advantage for early endoscopy (< 6 h) compared with endoscopy
[271] 2004 – 2006 within 24 h in terms of mortality, need for surgery, or transfusion
requirements.
Lim, 2011 Singapore Retrospective 837 low risk Endoscopy within 13 h of presentation is associated with lower mortality
[272] 97 high risk in high risk but not low-risk patients with NVUGIH.
Marmo, Italy Multicenter, prospective 3 207 Significant increase of mortality in high risk patients when endoscopy is
2011 [273] cohort studies (3 databases) performed ≤ 12 h compared with endoscopy performed 13 – 24 h after
2004 – 2009 presentation (14.3 % – 16.6 % vs. 5.2 %, P = 0.001).
Wysocki, US Retrospective 435 765 Increased mortality risk in patients who do not receive endoscopy within
2012 [274] Administrative data 1 day of admission: OR 1.32, 95 %CI 1.26 – 1.38.
NIS 2002 – 2007
Jairath, UK Multicenter, prospective 4 478 Compared with later endoscopy (> 24 to 48 h), endoscopy performed
2012 [275] cohort study ≤ 12 h did not affect mortality (OR 0.98, 95 %CI 0.88 – 1.09), but led to
2007 a decreased risk-adjusted length of hospital stay (1.7 days,
95 %CI 1.39 – 1.99).
CI, confidence interval; OR, odds ratio; RCT, randomized controlled trial.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a36
First author, year Study design Intervention Participants Outcomes Results Level of evidence, conclusions
Guideline
[ref.]
Barkun, 2010 [240] International consensus After endoscopic therapy: Patients with endoscopic Mortality Rebleeding: OR 0.45, CI 0.36 – 0.57 Intravenous PPI 80 mg bolus
guideline – Proton pump inhibitor (PPI) vs. signs of high risk of Rebleeding 3 – 7 days Surgery: OR 0.56, CI 0.45 – 0.70 + 8 mg/h for 72 h recommended.
– Placebo or H2RA rebleeding Mortality: OR 0.90 CI 0.67 – 1.19 Evidence: High
Sachar, 2014 [276] Systematic review After endoscopic therapy: Patients with endoscopic Rebleeding 7 days Rebleeding: risk ratio (RR) 0.72 No differences between intermit-
– PPI intravenous continuous signs of high risk of Mortality (upper boundary of 1-sided 95 %CI tent high dose PPI (all routes) vs.
infusion (80 mg + 8 mg/h for rebleeding 0.97). intravenous PPI 80 mg bolus
72 h) vs. RRs for rebleeding 3 – 30 days, + 8 mg/h for 72 h.
– Intermittent PPI mortality, urgent interventions Evidence: High
were less than 1.
Sung, 2014 [277] Randomized, controlled After endoscopic therapy: Patients with endoscopic Rebleeding at 30 days Intravenous esomeprazole group: High dose oral esomeprazole may
trial, double blind – Intravenous esomeprazole signs of high risk of 118 patients be considered as a useful alterna-
bolus 80 mg + 8 mg/h for 72 h rebleeding Oral esomeprazole group: tive to intravenous therapy.
vs. 126 patients Evidence: Moderate
– Oral esomeprazole 40 mg/12 h Recurrent bleeding:
– Intravenous group: 7.7 %, vs.
– Oral group: 6.4 %
Mostaghni, 2011 Randomized, open-label – Oral omeprazole (80 mg twice- Patients with endoscopic Rebleeding (30 days?; – Oral omeprazole: 44 patients Oral omeprazole and intravenous
[278] daily for 3 days), or signs of high risk of timing not well-defined) – Intravenous pantoprazole: 41 pantoprazole had equal effects on
– Intravenous pantoprazole rebleeding patients prevention of rebleeding after
(80 mg bolus and 8 mg/hour Rebleeding: endoscopic therapy in patients
infusion for 3 days) – Oral omeprazole: 11.4 %, vs. with high risk bleeding peptic
Followed in all patients by – Intravenous pantoprazole: ulcers.
omeprazole (20 mg each day 9.8 % Evidence: Low
for 30 days). Mean hospital stay and blood
transfusion were not different
between the two groups.
Yen, 2012 [279] Randomized, open-label Esomeprazole group: 40 mg con- Patients with endoscopic Rebleeding rate within 100 patients enrolled. There is no evidence of a difference
tinuous infusion of esomeprazole signs of high risk of 14 days. Rebleeding rates: in clinical outcomes between oral
every 6 h for 3 days. Then oral rebleeding Secondary outcomes – esomeprazole intravenous and intravenous PPI treatment.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
esomeprazole 40 mg once-daily for included: group: 4 % (2 /50), and (not powered to prove equivalence
2 months. – Hospital stay – Lansoprazole oral group: 4 % or noninferiority). Patients receiv-
Lansoprazole group: oral 30 mg – Volume of blood (2/50) ing oral PPI have a shorter hospital
four times daily for 3 days followed transfusion No difference between the two stay.
by once daily for 2 months. – Surgical intervention, groups with regard to: Evidence: Low
and – Hospital stay
– Mortality within – Volume of blood transfusion
1 month – Surgery, or
– Mortality rate
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e8 (Continuation)
First author, year Study design Intervention Participants Outcomes Results Level of evidence, conclusions
[ref.]
Tsai, 2008 [280] Randomized, open-label – Oral rabeprazole (20 mg twice- Patients with endoscopic Bleeding up to 14 days. 156 patients with 78 in each group Oral rabeprazole and intravenous
daily for 3 days), or signs of high risk of Also compared: Rebleeding: regular-dose omeprazole are
– Intravenous omeprazole rebleeding – Hospital stay – Omeprazole group: 12 patients equally effective in preventing re-
(40 mg intravenous infusion – Volume of blood (15.4 %) bleeding in patients with high risk
every 12 h for 3 days). transfusion – Rabeprazole group: bleeding peptic ulcers after suc-
– Surgery, and 13 patients (16.7 %) cessful endoscopic injection with
– Mortality within – 95 %CI of difference: epinephrine.
14 days 12.82 – 10.22 Evidence: Low
No differences in the other
evaluated outcomes
Wang, 2010 [281] Systematic review High dose PPIs (80 mg bolus + Patients with bleeding – Rebleeding 7 studies and 1157 patients Compared with non – high dose
8 mg/h for 3 days vs. peptic ulcer – Surgical intervention, – Rebleeding: OR 1.30, 95 %CI PPIs, high dose PPIs do not further
Non-high dose PPIs (40 mg160 mg/ and 0.88 – 1.91 reduce the rates of rebleeding,
day for 3 days) – Mortality – Surgical intervention: OR 1.49, surgical intervention, or mortality
0.66 – 3.37 after endoscopic treatment.
– Mortality: OR 0.89, 0.37 – 2.13 Evidence: High
Masjedizadeh, Randomized, open-label – High dose pantoprazole (80 mg Patients with endoscopic Rebleeding at 30 days 166 patients (83 patients per For controlling peptic ulcer bleed-
2014 [282] bolus, 8 mg per hour) infused signs of high risk of re- group) ing, there is no difference between
for 3 days vs. bleeding Rebleeding observed in: high dose and low dose pantopra-
– Low dose pantoprazole (40 mg – High dose group: 27 patients zole infusion.
bolus, 4 mg per hour) infused (32.53 %) Evidence: Low
for 3 days – Low dose group: 21 patients
(25.30 %)
P = 0.30
CI, confidence interval; H2RA, histamine-2 receptor antagonist; OR, odds ratio; PPI, proton pump inhibitor; RR, risk ratio.
Guideline
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a37
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a38
First author, Study type Patient group Key outcomes Key results Limitations Level of evidence, Conclusion
Guideline
year [ref.]
Lau, 1999 Prospective, random- Endoscopic re-treat- Definitive hemostasis Definitive hemostasis: 35 /48 Very high expertise center High quality
[283] ized trial ment with the same Complications Salvage surgery: 13 Not blinded In patients with peptic ulcers and
Comparison between device (epinephrine Surgery – Persistent bleeding: 11 2 perforations related to rebleeding after initial hemostatic suc-
immediate endo- injection + heater Mortality – Perforation: 2 re-treatment with heater cess, endoscopic re-treatment reduces
scopic re-treatment probe): Complications: probe the need for surgery without increasing
vs. surgery – n = 48 patients 7 patients in the endoscopy group (includ- risk of death and is associated with
Surgery: ing 6 who underwent salvage surgery) had fewer complications than surgery.
– n = 44 complications, as compared with 16 in the
surgery group (P = 0.03)
7
Mortality: n.s.
Hypotension at randomization and ulcer
> 2 cm were independent predictors of
re-treatment failure
Wong, 2011 Retrospective Transcatheterangio- Definitive hemostasis Bleeding recurrence: Results of initial failure and Moderate quality
[284] graphic embolization Complications – TAE group 34.4 % rebleeding are mixed In patients with ulcer bleeding after
(TAE): n = 23 Mortality – Surgery group 12.5 % failed endoscopic hemostasis, TAE
Surgery: n = 56 P = 0.1 reduces the need for surgery without
More complications after surgery: increasing the overall mortality and is
67.9 % vs. 40.6 %, P = 0.1 associated with fewer complications.
No difference in mortality: 25 %
vs. 30.4 %
Kyaw, 2014 Meta-analysis TAE patients were Definitive hemostasis Risk of rebleeding was significantly higher in Retrospective studies High quality
[285] 6 retrospective older, mean age: Complications TAE vs. surgery: A higher rebleeding rate was observed
studies that involved – TAE: 75 years vs. Mortality RR 1.82, 95 %CI 1.23 – 2.67 after TAE, suggesting surgery more
423 patients – Surgery: 68 years After age exclusion, the high risk of rebleed- definitively secured hemostasis, with
ing remained in the TAE group: no significant differences in mortality
RR 2.64, 95 %CI 1.48 – 4.71 rate or requirement for additional
No difference in: interventions.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
– Mortality, or
– Requirement for additional
interventions
Beggs, 2014 Systematic review TAE: n = 347 Mortality Deaths: Moderate quality
[286] (with homogeneity) Surgery: n = 364 Rebleeding – TAE: 61 When compared with surgery, TAE had
of cohort studies Length of hospital stay – Surgery: 101 a significantly increased risk of rebleed-
Adverse events Rebleeding: ing. However, there were no differences
– TAE: 78 in mortality.
– Surgery: 45
Length of hospital stay:
– Mean difference between TAE and sur-
gery: 0.75 days
Complications:
– TAE: 92
– Surgery: 81
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Appendix e9 (Continuation)
First author, Study type Patient group Key outcomes Key results Limitations Level of evidence, Conclusion
year [ref.]
Smith, 2014 European register Patients who other- Definitive hemostasis Definitive hemostasis: Small number of patients Low quality
[287] wise might have – n = 8 (100 %) Hemospray may be useful as a rescue
required either therapy.
surgery or TAE
n=8
Sulz, 2014 Prospective, case n = 16 Definitive hemostasis Initial hemostasis: 15/16 Small number of patients Low quality
[288] series Salvage therapy: 13/14 Hemospray may be useful as a salvage
Monotherapy: 2/2 therapy.
Rebleeding: 12.5 %
Skinner, 2014 Retrospective Over-the-scope clip n = 12 Primary hemostasis: 11/12 Retrospective Low quality
[289] after failure of con- Cause of bleed: Rebleeding n = 2 (day1, day7) Small number of patients Over-the-scope clip effective and safe
ventional endoscopic – Duodenal ulcer: 6 No complication Expertise needed for severe acute gastrointestinal bleed-
treatment – Gastric ulcer: 2 ing when conventional endoscopic
– Dieulafoy lesion: 2 therapy fails.
– Anastomotic ulceration: 1
– Mallory-Weiss: 1
– Prior endoscopies
– Mean of 2
– Shock present:
– n = 9 (75 %)
RBC:
– Mean 5.1 units (2 – 12)
Manta, 2013 Retrospective Over-the-scope clip- n = 23 Primary hemostasis: 22 (96 %) Retrospective Low quality
[290] ping after failure of – Duodenal ulcer: 12 – Duodenal ulcer: 11 (1 rebleeding) Small number of patients Over-the-scope clipping is effective and
conventional endo- Forrest Ia n = 5 Expertise needed safe for severe acute gastrointestinal
scopic treatment Forrest Ib n = 4 bleeding when conventional endo-
– Gastric ulcer 6 – Gastric ulcer: 6 (1 rebleeding) scopic therapy fails.
Forrest Ia: n = 2
Forrest Ib: n = 2
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a39
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a40 Guideline
Appendix e10 Helicobacter pylori and nonvariceal upper gastrointestinal hemorrhage (NVUGIH).
First author, year Study design, Participants Outcomes Results Level of evidence,
[ref.] study objective conclusions
Sánchez-Delgado, Meta-regression NVUGIB patients H. pylori infection Delayed testing for H. pylori increases Moderate
2011 [291] analysis rate after the event the detection rate (OR 2.08, 95 %CI
1.10 – 3.93)
Gisbert, 2006 [292] Meta-analysis NVUGIB patients H. pylori infection Low sensitivity of histology, RUT, Moderate
rate after the event culture, UBT, and serology
Barkun, 2010 [240] International NVUGIB patients Re-bleeding H. pylori therapy and eradication High
consensus re- confirmation needed
commendations
Gisbert, 2012 [293] Prospective study NVUGIB patients Re-bleeding after Rebleeding was 0.15 % per patient- Moderate
on 1000 patients H. pylori eradication year of follow-up and it was associated
with either re-infection or NSAIDs use.
Gisbert, 2004 [294] Meta-analysis NVUGIB patients Re-bleeding after Rebleeding was significantly lower in High
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
eradication the H. pylori eradication group than in
antisecretory therapy group (1.6 % vs.
5.6 %)
Dixon, 1996 [295] International Gastritis NA Presence of neutrophil histology Moderate
workshop strongly suggests H. pylori infection
CI, confidence interval; NA, not available; NVUGIB, nonvariceal upper gastrointestinal bleeding; OR, odds ratio; RUT, rapid urease test; UBT, urea breath test.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Appendix e11 Risk of thromboembolism, recurrent gastrointestinal (GI) bleeding and death after warfarin therapy interruption for GI bleeding.
First author, Study type Study population Key outcomes Key results Limitations Conclusion
year [ref.] Intervention
Witt, (2012) Retrospective, 442 patients with warfar- Thromboembolic events 90-day thromboembolic event rate: Retrospective study The decision to not resume war-
[296] cohort in-associated GI bleeding Recurrent GI bleeding – 0.4 % (1/260) in warfarin group Data from administrative farin therapy in the 90 days fol-
Intervention: Death – 5.5 % (10/182) in no-warfarin group databases lowing a GI bleeding event is asso-
Withhold warfarin HR (95 %CI): 0.05 (0.001 – 0.58) Selection bias (greater ciated with increased risk for
(no-warfarin group): 182 No thromboembolic events in patients who co-morbidity burdens in thrombosis and death.
(41.2 %) resumed therapy within 14 days no-warfarin group, which Resuming warfarin within 7 days is
Resumed warfarin (war- 90-day recurrent GI bleeding rate: may have contributed to associated with a twofold higher
farin group): 260 (58.8 %) – Warfarin group: 10 % (26/260) their worse outcomes) risk of rebleeding.
– No-warfarin group: 5.5 % (10/182) Detection and survivor-
HR (95 %CI): 1.32 (0.50 – 3.57) ship biases
Higher risk of recurrent GI bleeding in patients who
resumed warfarin within 7 days from index bleeding
as compared with those who resumed warfarin
later:
– 12.4 % vs. 6.2 %
P = 0.03
90-day mortality rate:
– Warfarin group: 5.8 % (15/260) in
– No-warfarin group: 20.3 % (37/182)
HR (95 %CI): 0.31 (0.15 – 0.62)
Qureshi, 2014 Retrospective, 1329 atrial fibrillation Thromboembolic events Warfarin group vs. no-warfarin group, adjusted HR Retrospective study. Data The decision to not resume war-
[297] cohort study patients with warfarin- Recurrent GI bleeding (95 %CI): from administrative farin therapy after a GI bleeding
associated major GI Death – Thromboembolism: 0.71 (054 – 0.93) database event is associated with increased
bleeding – Recurrent GI bleeding; 1.20 (0.78 – 1.86) Selection bias (greater risk for thrombosis and death.
Intervention: – Mortality: 0.72 (0.60 – 0.86) co-morbidity burdens in There is a trend toward reduced
676 (50.9 %) withhold Incidence of adverse outcomes per 100 person- no-warfarin group) incidence of thromboembolic
warfarin (no-warfarin years in the warfarin group, stratified by the Detection and survivor- events the earlier the warfarin is
group) duration of warfarin interruption: ship biases introduced; this trend is more evi-
653 (49.1 %) resumed – < 7 days (n = 62): dent within the first 15 days.
warfarin (warfarin group) Thromboembolism: 11.6 (8.3 – 16.2) Resuming warfarin within 7 days is
Time duration of inter- Recurrent GI bleeding: 19.3 (14.6 – 25.5) associated with a twofold higher
ruption: – 7 – 15 days (n = 51): risk of rebleeding.
– < 7 days: 62 patients Thromboembolism: 12.0 (8.2 – 17.5) Decision to restart warfarin after
– 7 – 30 days: 162 pa- Recurrent GI bleeding: 10.8 (7.2 – 16.3) 7 days of interruption is associated
tients – 15 – 21 days (n = 58): with improved survival and de-
– 30 days: 429 patients Thromboembolism: 18.1 (13.4 – 24.5) creased thromboembolism with-
Recurrent GI bleeding: 10.9 (7.2 – 16.4) out increased risk of recurrent GI
– 21 – 30 days (n = 53): bleeding.
Thromboembolism 20.7 (15.5 – 27.7)
Recurrent GI bleeding 9.9 (6.3 – 15.5)
– > 30 days (n = 429):
Thromboembolism 20.4 (17.8 – 23.5)
Recurrent GI bleeding 9.9 (8.0 – 12.3)
Guideline
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a41
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
a42
First author, Study type Study population Key outcomes Key results Limitations Conclusion
year [ref.] Intervention
Guideline
Lee, 2011 Retrospective, 58 patients with NVUGIB Recurrent GI bleeding Warfarin vs. no-warfarin group: Retrospective design Anticoagulation is recommended
[298] case-control study on warfarin for native Thromboembolic events – Rebleeding rate: 7 % vs.0 % Time of resuming antico- to be resumed within 20 days from
valvular heart disease Mean follow-up: 255 + – Thromboembolism rate: 0 % vs. 17 % (6 throm- agulation not specified the cessation to prevent throm-
Controls (aspirin group): 14 days (range 182 – 330) boembolic events occurred, at days 21, 27, 28, Individual thromboem- boembolic events.
41 age- and gender-mat- 31, 58, 75 from admission)Aspirin group: bolic risk not specified
ched patients with NVU- – Rebleeding rate: 0 %
GIB on aspirin for is- – Thromboembolic event rate: 2 % (1 myocardial
chemic heart disease infarction at day 95 from admission)
presenting
Intervention
Cases:
– Discontinued
warfarin (no-warfarin
group): 36/58
– Restarted warfarin
(warfarin group): 22/58
Controls:
– Discontinued aspirin:
41/41
CI, confidence interval; HR, hazard ratio; NVUGIB, nonvariceal upper gastrointestinal bleeding.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Guideline a43
Appendix e12 Observational studies assessing the effect of proton pump inhibitors (PPIs) on clinical cardiovascular outcomes in patients prescribed clopidogrel.
First author, year [ref.] Design Population Patients, n End point Results
PPI No PPI
Studies with evidence of a clinically significant interaction
Goodman, 2012 [299] Retrospective cohort Acute coronary syn- 3 255 6 021 Cardiovascular death, myocar- Clopidogrel cohort:
within RCT (PLATO) drome (ACS) dial infarction, cerebrovascular HR 1.20 (95 %CI
accident 1.04 – 1.38)
Ticagrelor cohort:
HR 1.24 (95 %CI,
1.07 – 1.45)
Stockl, 2010 [300] Retrospective Post myocardial infarc- 1 033 1 033 Myocardial infarction HR 1.93 (95 %CI,
Propensity matching tion or post percuta- Myocardial infarction or 1.05 – 3.54)
neous coronary inter- revascularization HR 1.91 (95 %CI
vention 1.19 – 3.06)
Kreutz, 2010[301] Retrospective cohort Post percutaneous cor- 6 828 9 862 Cardiovascular deaths, acute HR 1.51 (95 %CI,
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
onary intervention coronary syndrome, cerebro- 1.39 – 1.64)
vascular accident revasculari-
zation
Ho, 2009 [302] Retrospective cohort Post myocardial infarc- 5 244 2 961 Deaths, acute coronary syn- HR 1.25(95 %CI,
tion drome 1.11 – 1.41)
Huang, 2010 [303] Registry Post percutaneous cor- 572 2 706 Death HR 1.65(95 % CI,
onary intervention 1.35 – 2.01)
Zou, 2014 [304] Retrospective cohort Post percutaneous cor- 61 288 1 465 Myocardial infarction, stent HR 1.33 (95 % CI
onary intervention thrombosis, cardiovascular 1.12 – 1.57)
deaths
Van Boxel, 2010 [305] Retrospective cohort Clopidogrel users 5 734 12 405 Deaths, acute coronary HR 1.75 (95 % CI,
syndrome, cerebrovascular 1.58 – 1.94)
accident
Munoz-Torrero, 2011 Registry Vascular disease 519 703 Deaths, myocardial infarction, HR 1.8 (95 % CI,
[306] acute coronary syndrome, cer- 1.1 – 2.7)
ebrovascular accident, chronic
limb ischemia
Studies without evidence of a clinically significant interaction
O’Donoghue, 2009 Retrospective cohort Acute coronary 2 257 4 538 Cardiovascular death, myocar- No effect
[307] within RCT syndrome and post dial infarction, cerebrovascular
(TRITON-TIMI 38) percutaneous coronary accident
intervention
Hsiao, 2011 [308] Retrospective Post myocardial 622 9 131 Acute coronary syndrome No effect
Propensity matching infarction
Banerjee, 2011 [309] Retrospective Post percutaneous 867 3 678 Death, myocardial infarction, No effect
Propensity matching coronary intervention revascularization
Harjai, 2011 [310] Registry Post percutaneous 751 1 900 Death, myocardial infarction, No effect
Propensity matching coronary intervention revascularization
Aihara, 2012 [311] Registry Post percutaneous 819 1 068 Death, myocardial infarction No effect
Propensity matching coronary intervention
Tentzeris, 2010 [312] Registry Post percutaneous 691 519 Death, acute coronary No effect
Propensity matching coronary intervention syndrome
Schmidt, 2012 [313] Retrospective cohort Post percutaneous 2 742 10 259 Cardiovascular deaths, acute No effect
coronary intervention coronary syndrome, cerebro-
vascular accident, revasculari-
zation
Rassen, 2009 [314] Retrospective cohort Post percutaneous 3 996 14 569 Death, myocardial infarction, No effect
coronary intervention revascularization
or post acute coronary
syndrome
Ray, Retrospective cohort Post percutaneous 7 593 13 003 Death, myocardial infarction, No effect
2010 [315] coronary intervention revascularization
or post acute coronary
syndrome
CI, confidence interval; HR, hazard ratio; PLATO, Platelet Inhibition and Patient Outcomes; RCT, randomized controlled trial.
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a44 Guideline
Appendix e13 Meta-analyses evaluating the effect of proton pump inhibitors (PPIs) on clinical outcomes in patients treated with clopidogrel.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Siller-Matula, 2 nested case-control 159 138 Myocardial infarction
2010 [317] 20 retrospective cohort (13 studies)
(3 studies used a propensity Death RR 1.04 (95 %CI 0.93 – 1.24)
scoring method for the analysis) (13 studies)
2 post.hoc analyses of RCT MACE RR 1.29 (95 %CI 1.15 – 1.44)
1 prospective RCT (19 studies)
CI, confidence interval; RCT, randomized controlled trial; RR, risk ratio
References for Appendices 242 Barkun AN. Should every patient with suspected upper GI bleeding
receive a proton pump inhibitor while awaiting endoscopy? Gastro-
intest Endosc 2008; 67: 1064 – 1066
229 Longstreth GF, Feitelberg SP. Outpatient care of selected patients with
243 Rácz I, Szalai M, Dancs N et al. Pantoprazole before endoscopy in pa-
acute non-variceal upper gastrointestinal haemorrhage. Lancet 1995;
tients with gastroduodenal ulcer bleeding: Does the duration of infu-
345: 108 – 111
sion and ulcer location influence the effects? Gastroenterol Res Pract
230 Longstreth GF, Feitelberg SP. Successful outpatient management of
2012: Article ID561207
acute upper gastrointestinal hemorrhage: use of practice guidelines
244 Lanas A. Update on non variceal gastrointestinal bleeding. Gastroen-
in a large patient series. Gastrointest Endosc 1998; 47: 219 – 222
terol Hepatol 2013; 36: 57 – 65
231 Cebollero-Santamaria F, Smith J, Gioe S et al. Selective outpatient man-
245 Sung JJ, Chan FK, Chen M et al. Asia-Pacific Working consensus on
agement of upper gastrointestinal bleeding in the elderly. Am J Gas-
non-variceal upper gastrointestinal bleeding. Gut 2011; 60: 1170 –
troenterol 1999; 94: 1242 – 1247
1177
232 Brullet E, Campo R, Calvet X et al. A randomized study of the safety of
246 Al-Sabah S, Barkun AN, Herba K et al. Cost-effectiveness of proton
outpatient care for patients with bleeding peptic ulcer treated by
pump inhibition before endoscopy in upper gastrointestinal bleed-
endoscopic injection. Gastrointest Endosc 2004; 60: 15 – 21
ing. Clin Gastroenterol Hepatol 2008; 6: 418 – 425
233 Lai KC, Hui WM, Wong BC et al. A retrospective and prospective study
247 Ghassemi KA, Kovacs TOG, Jensen DM. Gastric acid inhibition in the
on the safety of discharging selected patients with duodenal ulcer
treatment of peptic ulcer haemorrhage. Curr Gastroenterol Rep
bleeding on the same day as endoscopy. Gastrointest Endosc 1997;
2009; 11: 462 – 469
45: 26 – 30
248 Laursen SB, Jorgensen HS, Schaffalitzky de Muckadell OB. Management
234 Lee JG, Turnipseed S, Romano PS et al. Endoscopy-based triage signifi-
of bleeding gastroduodenal ulcers. Dan Med J 2012; 59: C4473
cantly reduces hospitalization rates and costs of treating upper GI
249 Lin HJ. Role of proton pump inhibitors in the management of peptic
bleeding: a randomized controlled trial. Gastrointest Endosc 1999;
ulcer bleeding. World J Gastrointest Pharmacol Ther 2010; 1: 51 – 53
50: 755 – 761
250 Gluud LL, Klingenberg SL, Langholz E. Tranexamic acid for upper gas-
235 Gralnek IM, Dulai GS. Incremental value of upper endoscopy for triage
trointestinal bleeding. Cochrane Database Syst Rev 2012; 1:
of patients with acute non-variceal upper GI hemorrhage. Gastroin-
CD006640
test Endosc 2004; 60: 9 – 14
251 Magnusson I, Ihre T, Johansson C et al. Randomized double blind trial
236 Cipolletta L, Bianco MA, Rotondano G et al. Outpatient management
of somatostatin in the treatment of massive upper gastrointestinal
for low-risk nonvariceal upper GI bleeding: a randomized controlled
hemorrhage. Gut 1985; 26: 221 – 226
trial. Gastrointest Endosc 2002; 55: 1 – 5
252 Choi CW, Kang DH, Kim HW et al. Somatostatin adjunctive therapy for
237 Sreedharan A, Martin J, Leontiadis GI et al. Proton pump inhibitor
non-variceal upper gastrointestinal rebleeding after endoscopic ther-
treatment initiated prior to endoscopic diagnosis in upper gastroin-
apy. World J Gastroenterol 2011; 17: 3441 – 3447
testinal bleeding. Cochrane Database Syst Rev 2010; 7: CD005415
253 Avgerinos A, Sgouros S, Viazis N et al. Somatostatin inhibits gastric acid
238 Lau JY, Leung WK, Wu JC et al. Omeprazole before endoscopy in pa-
secretion more effectively than pantoprazole in patients with peptic
tients with gastrointestinal bleeding. N Engl J Med 2007; 356:
ulcer bleeding: A prospective, randomized, placebo controlled trial.
1631 – 40
Scan J Gastroenterol 2005; 40: 515 – 522
239 Liu N, Liu L, Zhang HH et al. Effect of intravenous proton pump inhib-
254 Archimandritis A, Tsirantonaki M, Tryphonos M et al. Ranitidine versus
itor requirements and timing of endoscopy of peptic ulcer bleeding. J
ranitidine plus octreotide in the treatment of acute nonvariceal up-
Gastroenterol Hepatol 2012; 27: 1473 – 79
per gastrointestinal bleeding: a prospective randomized study. Curr
240 Barkun AN, Bardou M, Kuipers EJ et al. International consensus recom-
Med Res Opin 2000; 16: 178 – 183
mendations on the management of patients with non-variceal upper
255 Lin H, Perng C, Wang K et al. Octreotide for arrest of peptic ulcer he-
gastrointestinal bleeding. Ann Intern Med 2010; 152: 101 – 113
morrhage – a prospective randomized controlled trial. Hepatogas-
241 Tsoi KKF, Lau JYW, Sung JJY. Cost-effectiveness analysis of high-dose
troenterology 1995; 42: 856 – 860
omeprazole infusion before endoscopy for patients with upper GI
bleeding. Gastrointest Endosc 2008; 67: 1056 – 1063
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Guideline a45
256 Kim I, Lee YS, Koh BS et al. Does adding somatostatin to proton pump peptic ulcers: a prospective, randomized, controlled clinical trial.
inhibitor improve the outcome of peptic ulcer bleeding? Korean J Crit Iran Red Crescent Med J 2011; 13: 458 – 463
Care Med 2008; 23: 75 – 78 279 Yen HH, Yang CW, Su WW et al. Oral versus intravenous proton pump
257 Antonioli A, Gandolfo M, Rigo GD et al. Somatostatin and cimetidine in inhibitors in preventing re-bleeding for patients with peptic ulcer
the control of acute upper gastrointestinal bleeding. A controlled bleeding after successful endoscopic therapy. BMC Gastroenterology
multicentre study. Hepatogastroenterology 1986; 33: 71 – 94 2012; 12: 66
258 Tisbouris P, Zintazas E, Lappas C et al. High-dose pantoprazole infu- 280 Tsai JJ, Hsu YC, Perng CL et al. Oral or intravenous proton pump inhib-
sion is superior to somatostatin after endoscopic hemostasis in pa- itor in patients with peptic ulcer bleeding after successful endoscopic
tients with peptic ulcer bleeding. Am J Gastroenterol 2007; 102: epinephrine injection. Br J Clin Pharmacol 2009; 67: 326 – 332
1192 – 1199 281 Wang CH, Ma MH, Chou HC et al. High-dose vs non-high-dose proton
259 Okan A, Simsek I, Akpinar H et al. Somatostatin and ranitidine in the pump inhibitors after endoscopic treatment in patients with bleed-
treatment of non-variceal upper gastrointestinal bleeding: a pro- ing peptic ulcer: a systematic review and meta-analysis of random-
spective randomized, double-blind, controlled study. Hepatogas- ized controlled trials. Arch Intern Med 2010; 170: 751 – 758
troenterology 2000; 47: 1325 – 1327 282 Mazjedizadeh AR, Hajiani E, Alavinejad P et al. High dose versus low
260 Rutgeerts P, Avgerinos A et al. Early administration of somatostatin be- dose intravenous pantoprazole in bleeding peptic ulcer: a random-
fore endoscopy to non-cirrhotic patients with suspected peptic ulcer ized clinical trial. Middle East J Dig Dis 2014; 6: 137 – 143
bleeding: The PUB double-blind, randomized, placebo-controlled 283 Lau JYW, Sung JJY, Lam YH et al. Endoscopic retreatment compared
trial. Gut 2006; 55: A47 with surgery in patients with recurrent bleeding after initial endo-
261 Carbonell N, Pauwels A, Serfaty L et al. Erythromycin infusion prior to scopic control of bleeding ulcer. N Engl J Med 1999; 340: 751 – 756
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
endoscopy for acute upper gastrointestinal bleeding: a randomized, 284 Wong TCF, Wong TT, Chiu PWY et al. A comparison of angiographic
controlled, double-blind trial. Am J Gastroenterol 2006; 101: 1211 – embolization with surgery after failed endoscopic hemostasis to
1215 bleeding peptic ulcers. Gastrointest Endosc 2011; 73: 900 – 908
262 Coffin B, Pocard M, Panis Y et al. Erythromycin improves the quality of 285 Kyaw M, Tse Y, Ang D et al. Embolization versus surgery for peptic
EGD in patients with acute upper GI bleeding: a randomized con- ulcer bleeding after failed endoscopic hemostasis: a meta-analysis.
trolled study. Gastrointest Endosc 2002; 56: 174 – 179 Endosc Int Open 2014; 2: E6 – E14
263 Frossard JL, Spahr L, Queneau PE et al. Erythromycin intravenous bolus 286 Beggs AD, Dilworth MP, Powell SL et al. A systematic review of transar-
infusion in acute upper gastrointestinal bleeding: a randomized, con- terial embolization versus emergency surgery in treatment of major
trolled, double-blind trial. Gastroenterology 2002; 123: 17 – 23 nonvariceal upper gastrointestinal bleeding. Clin Exp Gastroenterol
264 Theivanayagam S, Lim RG, Cobell WJ et al. Administration of erythro- 2014; 7: 93 – 104
mycin before endoscopy in upper gastrointestinal bleeding: a meta- 287 Smith LA, Stanley AJ, Bergman JJ et al. Hemospray application in non-
analysis of randomized controlled trials. Saudi J Gastroenterol 2013; variceal upper gastrointestinal bleeding: results of the survey to eval-
19: 205 – 210 uate the application of hemospray in the luminal tract. J Clin Gastro-
265 Sussman DA, Deshpande AR, Parra JL et al. Intravenous metoclopra- enterol 2014; 48: 89 – 92
mide to increase mucosal visualization during endoscopy in patients 288 Sulz M, Frei R, Meyenberger C et al. Routine use of Hemospray for gas-
with acute upper gastrointestinal bleeding: a randomized, controlled trointestinal bleeding: prospective two-center experience in Swit-
study. Gastrontest Endosc 2008; 67: AB247 zerland. Endoscopy 2014; 46: 619 – 624
266 Barkun AN, Bardou M, Martel M et al. Prokinetics in acute upper GI 289 Skinner M, Gutteriez JP, Neumann H et al. Over-the-scope clip place-
bleeding: a meta-analysis. Gastrointest Endosc 2010; 72: 1138 – 1145 ment is effective rescue therapy for severe acute upper gastrointesti-
267 Habashi SL, Lambiase LR, Kottoor R. Prokinetics infusion prior endos- nal bleeding. Endosc Int Open 2014; 02: E37 – E40
copy for acute upper gastrointestinal bleeding: A randomized, con- 290 Manta R, Galloro G, Mangiavillano B et al. Over-the-scope clip (OTSC)
trolled, double-blind and placebo-controlled trial [abstract]. Am J represents an effective endoscopic treatment for acute GI bleeding
Gastroenterol 2007; 102: S526 after failure of conventional techniques. Surg Endosc 2013; 27:
268 Winstead NS, Wilcox CM. Erythromycin prior to endoscopy for acute 3162 – 3164
upper gastrointestinal haemorrhage: a cost-effectiveness analysis. 291 Sanchez-Delgado J, Gene E, Suarez D et al. Has H. pylori prevalence in
Aliment Pharmacol Ther 2007; 26: 1371 – 1377 bleeding peptic ulcer been underestimated? A meta-regression Am J
269 Spiegel BM, Vakil NB, Ofman JJ. Endoscopy for acute nonvariceal upper Gastroenterol 2011; 106: 398 – 405
gastrointestinal tract hemorrhage: is sooner better? A systematic re- 292 Gisbert JP, Abraira V. Accuracy of Helicobacter pylori diagnostic tests
view Arch Intern Med 2001; 161: 1393 – 1404 in patients with bleeding peptic ulcer: a systematic review and
270 Tsoi KKF, Ma TKW, Sung JJY. Endoscopy for upper gastrointestinal meta-analysis. Am J Gastroenterol 2006; 101: 848 – 863
bleeding: how urgent is it? Nat Rev Gastroenterol Hepatol 2009; 6: 293 Gisbert JP, Calvet X, Cosme A et al. Long-term follow-up of 1,000 pa-
463 – 469 tients cured of Helicobacter pylori infection following an episode of
271 Sarin N, Monga N, Adams PC. Time to endoscopy and outcomes in up- peptic ulcer bleeding. Am J Gastroenterol 2012; 107: 1197 – 1204
per gastrointestinal bleeding. Can J Gastroenterol 2009; 23: 489 – 493 294 Gisbert JP, Khorrami S, Carballo F et al. Meta-analysis: Helicobacter py-
272 Lim L, Ho K, Chan Y et al. Urgent endoscopy is associated with lower lori eradication therapy vs. anti-secretory non-eradication therapy
mortality in high-risk but not low-risk nonvariceal upper gastroin- for the prevention of recurrent bleeding from peptic ulcer. Aliment
testinal bleeding. Endoscopy 2011; 43: 300 – 306 Pharmacol Ther 2004; 19: 617 – 629
273 Marmo R, Del Piano M, Rotondano G et al. Mortality from nonvariceal 295 Dixon MF, Genta RM, Yardley JH et al. Classification and grading of gas-
upper gastrointestinal bleeding: is it time to differentiate the timing tritis. The updated Sydney System. Am J Surg Pathol 1996; 20: 1161 –
of endoscopy? Gastrointest Endosc 2011; 73: AB224 1181
274 Wysocki JD, Srivastav S, Winstead NS. A nationwide analysis of risk 296 Witt DM, Delate T, Garcia DA et al. Risk of thromboembolism, recur-
factors for mortality and time to endoscopy in upper gastrointestinal rent hemorrhage, and death after warfarin therapy interruption for
haemorrhage. Aliment Pharmacol Ther 2012; 36: 30 – 36 gastrointestinal tract bleeding. Arch Intern Med 2012; 172: 1484 –
275 Jairath V, Kakan BC, Logan RF et al. Outcomes following acute nonvar- 1491
iceal upper gastrointestinal bleeding in relation to time to endos- 297 Qureshi W, Mittal C, Patsias I et al. Restarting anticoagulation and out-
copy: results from a nationwide study. Endoscopy 2012; 44: 723 – comes after major gastrointestinal bleeding in atrial fibrillation. Am J
730 Cardiol 2014; 113: 662 – 668
276 Sachar H, Vaidya K, Laine L. Intermittent vs. continuous proton pump 298 Lee JK, Kang HW, Kim SG et al. Risks related with withholding and re-
inhibitor therapy for high-risk bleeding ulcers: a systematic review suming anticoagulation in patients with non-variceal upper gastroin-
and meta-analysis. JAMA Intern Med 2014; 174: 1755 – 1762 testinal bleeding while on warfarin therapy. Int J Clin Pract 2012; 66:
277 Sung JJ, Suen BY, Wu JC et al. Effects of intravenous and oral esome- 64 – 68
prazole in the prevention of recurrent bleeding from peptic ulcers 299 Goodman SG, Clare R, Pieper KS et al. Association of proton pump in-
after endoscopic therapy. Am J Gastroenterol 2014; 109: 1005 – 1010 hibitor use on cardiovascular outcomes with clopidogrel and ticagre-
278 Mostaghni AA, Hashemi SA, Heydari ST. Comparison of oral and intra- lor: insights from the platelet inhibition and patient outcomes trial.
venous proton pump inhibitor on patients with high risk bleeding Circulation 2012; 125: 978 – 986
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
a46 Guideline
300 Stockl KM, Le L, Zakharyan A et al. Risk of rehospitalization for pa- 309 Banerjee S, Weideman RA, Weideman MW et al. Effect of concomitant
tients using clopidogrel with a proton pump inhibitor. Arch Intern use of clopidogrel and proton pump inhibitors after percutaneous
Med 2010; 170: 704 – 710 coronary intervention. Am J Cardiol 2011; 107: 871 – 878
301 Kreutz RP, Stanek EJ, Aubert R et al. Impact of proton pump inhibitors 310 Harjai KJ, Shenoy C, Orshaw P et al. Clinical outcomes in patients with
on the effectiveness of clopidogrel after coronary stent placement: the concomitant use of clopidogrel and proton pump inhibitors after
the clopidogrel Medco outcomes study. Pharmacotherapy 2010; 30: percutaneous coronary intervention: an analysis from the Guthrie
787 – 796 Health Off-Label Stent (GHOST) investigators. Circ Cardiovasc Interv
302 Ho PM, Maddox TM, Wang L et al. Risk of adverse outcomes associated 2011; 4: 162 – 170
with concomitant use of clopidogrel and proton pump inhibitors fol- 311 Aihara H, Sato A, Takeyasu N et al. Effect of individual proton pump
lowing acute coronary syndrome. JAMA 2009; 301: 937 – 944 inhibitors on cardiovascular events in patients treated with clopido-
303 Huang CC, Chen YC, Leu HB et al. Risk of adverse outcomes in Taiwan grel following coronary stenting: results from the Ibaraki Cardiac As-
associated with concomitant use of clopidogrel and proton pump in- sessment Study Registry. Catheter Cardiovasc Interv 2012; 80: 556 –
hibitors in patients who received percutaneous coronary interven- 563
tion. Am J Cardiol 2010; 105: 1705 – 1709 312 Tentzeris I, Jarai R, Farhan S et al. Impact of concomitant treatment
304 Zou JJ, Chen SL, Tan J et al. Increased risk for developing major adverse with proton pump inhibitors and clopidogrel on clinical outcome in
cardiovascular events in stented Chinese patients treated with dual patients after coronary stent implantation. Thromb Haemost 2010;
antiplatelet therapy after concomitant use of the proton pump inhib- 104: 1211 – 1218
itor. PLoS One 2014; 9: e84985 DOI 10.1371/journal.pone. 0084985 313 Schmidt M, Johansen MB, Robertson DJ et al. Concomitant use of clopi-
305 van Boxel OS, van Oijen MG, Hagenaars MP et al. Cardiovascular and dogrel and proton pump inhibitors is not associated with major ad-
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
gastrointestinal outcomes in clopidogrel users on proton pump inhi- verse cardiovascular events following coronary stent implantation.
bitors: results of a large Dutch cohort study. Am J Gastroenterol 2010; Aliment Pharmacol Ther 2012; 35: 165 – 174
105: 2430 – 2436 314 Rassen JA, Choudhry NK, Avorn J et al. Cardiovascular outcomes and
306 Munoz-Torrero JF, Escudero D, Suarez C et al. Concomitant use of pro- mortality in patients using clopidogrel with proton pump inhibitors
ton pump inhibitors and clopidogrel in patients with coronary, cere- after percutaneous coronary intervention or acute coronary syn-
brovascular, or peripheral artery disease in the factores de Riesgo y drome. Circulation 2009; 120: 2322 – 2329
ENfermedad Arterial (FRENA) registry. J Cardiovasc Pharmacol 2011; 315 Ray WA, Murray KT, Griffin MR et al. Outcomes with concurrent use of
57: 13 – 19 clopidogrel and proton-pump inhibitors: a cohort study. Ann Intern
307 O’Donoghue ML, Braunwald E, Antman EM et al. Pharmacodynamic ef- Med 2010; 152: 337 – 345
fect and clinical efficacy of clopidogrel and prasugrel with or without 316 Kwok CS, Loke YK. Meta-analysis: the effects of proton pump inhibi-
a proton-pump inhibitor: an analysis of two randomised trials. Lan- tors on cardiovascular events and mortality in patients receiving clo-
cet 2009; 374: 989 – 997 pidogrel. Aliment Pharmacol Ther 2010; 31: 810 – 823
308 Hsiao FY, Mullins CD, Wen YW et al. Relationship between cardiovas- 317 Siller-Matula JM, Jilma B, Schror K et al. Effect of proton pump inhibi-
cular outcomes and proton pump inhibitor use in patients receiving tors on clinical outcome in patients treated with clopidogrel: a sys-
dual antiplatelet therapy after acute coronary syndrome. Pharmacoe- tematic review and meta-analysis. J Thromb Haemost 2010; 8:
pidemiol Drug Saf 2011; 20: 1043 – 1049 2624 – 2641
Gralnek Ian M et al. Nonvariceal upper gastrointestinal hemorrhage: ESGE Guideline … Endoscopy 2015; 47: a1–a46
Annals of Internal Medicine Clinical Guidelines
International Consensus Recommendations on the Management of
Patients With Nonvariceal Upper Gastrointestinal Bleeding
Alan N. Barkun, MD, MSc (Clinical Epidemiology); Marc Bardou, MD, PhD; Ernst J. Kuipers, MD; Joseph Sung, MD; Richard H. Hunt, MD;
Myriam Martel, BSc; and Paul Sinclair, MSc, for the International Consensus Upper Gastrointestinal Bleeding Conference Group*
Description: A multidisciplinary group of 34 experts from 15 coun- alone is not recommended. Second-look endoscopy may be useful
tries developed this update and expansion of the recommendations in selected high-risk patients but is not routinely recommended.
on the management of acute nonvariceal upper gastrointestinal Preendoscopy proton-pump inhibitor (PPI) therapy may downstage
bleeding (UGIB) from 2003. the lesion; intravenous high-dose PPI therapy after successful en-
doscopic hemostasis decreases both rebleeding and mortality in
Methods: The Appraisal of Guidelines for Research and Evaluation patients with high-risk stigmata. Although selected patients can be
(AGREE) process and independent ethics protocols were used. discharged promptly after endoscopy, high-risk patients should be
Sources of data included original and published systematic reviews; hospitalized for at least 72 hours after endoscopic hemostasis. For
randomized, controlled trials; and abstracts up to October 2008. patients with UGIB who require a nonsteroidal anti-inflammatory
Quality of evidence and strength of recommendations have been drug, a PPI with a cyclooxygenase-2 inhibitor is preferred to reduce
rated by using the Grading of Recommendations Assessment, De- rebleeding. Patients with UGIB who require secondary cardiovascu-
velopment, and Evaluation (GRADE) criteria. lar prophylaxis should start receiving acetylsalicylic acid (ASA) again
as soon as cardiovascular risks outweigh gastrointestinal risks (usu-
Recommendations: Recommendations emphasize early risk strati- ally within 7 days); ASA plus PPI therapy is preferred over clopi-
fication, by using validated prognostic scales, and early endoscopy dogrel alone to reduce rebleeding.
(within 24 hours). Endoscopic hemostasis remains indicated for
high-risk lesions, whereas data support attempts to dislodge clots
with hemostatic, pharmacologic, or combination treatment of the Ann Intern Med. 2010;152:101-113. www.annals.org
underlying stigmata. Clips or thermocoagulation, alone or with epi- For author affiliations, see end of text.
nephrine injection, are effective methods; epinephrine injection * For a list of voting participants, see Appendix 1, available at www.annals.org.
untarily recuse themselves from the discussion if they Section A: Resuscitation, Risk Assessment, and
had a conflict of interest. The participants voted anon- Preendoscopy Management
ymously by using touchpad technology after completing Statement A2
scientific discussions for each statement (Appendix 2, Prognostic scales are recommended for early stratification
available at www.annals.org). Substantial numbers of of patients into low- and high-risk categories for rebleeding
participants with declared conflicts of interests (ⱖ24%) and mortality.
recused themselves from the discussions for 6 statements (Agree, 97% [Vote: a, 56%; b, 35%; c, 6%; d, 3%].
(A8, C3, C4, E1, E2, E3, and E4) (Appendix 2). How- Grade: Low, 1c, “do it”)
ever, vote tallies with and without those of participants As stated in the 2003 guidelines (15), patients should
with declared conflicts of interest revealed no differences be stratified into low and high risk by using prognostic
scales, on the basis of clinical, laboratory, and endoscopic
in the final outcomes of the “agree” decisions.
criteria. Early identification of high-risk patients allows ap-
Additional Domains Addressed by the Consensus propriate intervention, which minimizes morbidity and
Meeting mortality.
Knowledge gaps requiring further research were iden- Clinical predictors of increased risk for rebleeding or
tified, and dissemination of the guidelines was discussed. mortality include age greater than 65 years; shock; poor
A large, Canadian, randomized, controlled trial (RCT) overall health status; comorbid illnesses; low initial hemo-
globin levels; melena; transfusion requirement; fresh red
(ClinicalTrials.gov registration number: NCT00840008)
blood on rectal examination, in the emesis, or in the naso-
is assessing whether adherence to existing consensus recom-
gastric aspirate; sepsis; and elevated urea, creatinine, or se-
mendations on UGIB is improved with a multifaceted,
rum aminotransferase levels (15). Other factors predictive
tailored implementation strategy. In addition, subcommit-
of outcomes include chronic alcoholism, active cancer, or
tees are developing manuscripts on methodology of RCTs unsuitable sociofamily conditions (31), and an Acute Phys-
in UGIB, quality indicators, the effect of the adopted eth- iology and Chronic Health Evaluation (APACHE) II score
ical process, and endoscopic classification of ulcer bleeding of 11 or greater (32).
stigmata. The Blatchford and preendoscopic Rockall scores (33)
use only clinical and laboratory data (before endoscopy) to
Role of the Funding Source
identify patients who require intervention, whereas the
The conference was underwritten by unrestricted, complete Rockall score (34) also use endoscopic variables
pooled funds contributed to the supporting societies. to predict rebleeding or mortality.
Funding or in-kind support was provided by the Canadian The Blatchford score includes hemoglobin level, blood
Association of Gastroenterology (CAG); European Associ- urea level, pulse, systolic blood pressure, the presence of
ation for Gastroenterology and Endoscopy; Asian Pacific syncope or melena, and evidence of hepatic disease or car-
Society of Digestive Endoscopy; and Institute of Diabetes, diac failure and accurately identifies patients at low risk for
Metabolism, and Nutrition (of the Canadian Institutes of clinical intervention (35–37), even without inclusion of
Health Research) and from at-arms-length contributions urea level or syncope (38). Selected patients can be safely
from AstraZeneca Mölndal (Sweden), Abbott Canada, and managed as outpatients without early endoscopy by using
Olympus Canada provided to the CAG. The CAG ad- the Blatchford score (36, 37). This scale also compares
ministered all aspects of the meeting. The funding favorably with the preendoscopic and complete Rockall
sources had no role in identifying statements, abstract- scores (37–39).
ing data, synthesizing results, grading evidence, or pre- Endoscopic predictors of increased risk for rebleeding
paring the manuscript or in the decision to submit the and mortality include active bleeding (especially arterial
manuscript for publication. bleeding rather than oozing), nonbleeding visible vessel
(NBVV) or adherent clot, ulcer size (generally ⬎2 cm)
(40 – 43), ulcer location (posterior lesser gastric curvature
or posterior duodenal wall), and lesion type (for example,
RECOMMENDATION STATEMENTS ulcer, varices, or cancer) (15). A comparison of the Baylor
Each statement is followed by the grade of supporting College, Rockall, and Cedars–Sinai Medical Center predic-
evidence, the result of the vote, and a discussion of the tive indexes found that the Rockall score best identified
evidence. patients at low risk (44). This score has been validated in
The Table summarizes the recommendations that are multiple countries (44 – 47) but has better discriminative
new from this consensus and those that are revised from ability for mortality than for rebleeding (38, 44, 46, 47).
the 2003 guidelines (15), as well as those that are un- Use of the Rockall score has been shown to yield a more
changed because the majority of the group felt that they accurate diagnosis (significantly fewer undefined causes and
did not require revision at this time. (These are not dis- increased identification of peptic ulcer) and shorter dura-
cussed within the text [15].) tion of hospitalization (48).
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 103
Table. Summary of Consensus Recommendations for the Management of Patients With Nonvariceal Upper Gastrointestinal
Bleeding
B. Endoscopic management
B1. Develop institution-specific protocols for multidisciplinary management.* Include access to an endoscopist trained in endoscopic hemostasis.*
B2. Have available on an urgent basis support staff trained to assist in endoscopy.*
B3. Early endoscopy (within 24 hours of presentation) is recommended for most patients with acute upper gastrointestinal bleeding.†
B4. Endoscopic hemostatic therapy is not indicated for patients with low-risk stigmata (a clean-based ulcer or a nonprotuberant pigmented dot in an ulcer bed).*
B5. A finding of a clot in an ulcer bed warrants targeted irrigation in an attempt at dislodgement, with appropriate treatment of the underlying lesion.†
B6. The role of endoscopic therapy for ulcers with adherent clots is controversial. Endoscopic therapy may be considered, although intensive PPI therapy alone
may be sufficient.†
B7. Endoscopic hemostatic therapy is indicated for patients with high-risk stigmata (active bleeding or a visible vessel in an ulcer bed).*
B8. Epinephrine injection alone provides suboptimal efficacy and should be used in combination with another method.†
B9. No single method of endoscopic thermal coaptive therapy is superior to another.*
B10. Clips, thermocoagulation, or sclerosant injection should be used in patients with high-risk lesions, alone or in combination with epinephrine injection.†
B11. Routine second-look endoscopy is not recommended.†
B12. A second attempt at endoscopic therapy is generally recommended in cases of rebleeding.*
C. Pharmacologic management
C1. Histamine-2 receptor antagonists are not recommended for patients with acute ulcer bleeding.*
C2. Somatostatin and octreotide are not routinely recommended for patients with acute ulcer bleeding.*
C3. An intravenous bolus followed by continuous-infusion PPI therapy should be used to decrease rebleeding and mortality in patients with high-risk stigmata
who have undergone successful endoscopic therapy.†
C4. Patients should be discharged with a prescription for a single daily-dose oral PPI for a duration as dictated by the underlying etiology.
ASA ⫽ acetylsalicylic acid; COX-2 ⫽ cyclooxygenase-2; H. pylori ⫽ Helicobacter pylori; NSAID ⫽ nonsteroidal anti-inflammatory drug; PPI ⫽ proton-pump inhibitor.
* Recommendation unchanged from 2003 guidelines. See reference 15 for supporting evidence and discussions.
† Recommendation revised from 2003 guidelines.
confounding by need for transfusion could not be (55), which may reflect its greater importance as a comor-
excluded. bid index.
International guidelines (51) recommend initiating Another study in patients with any UGIB (57) found
red blood cell transfusions for most critically ill patients that intensive measures to correct INR can reduce mortal-
when hemoglobin levels decrease to less than 70 g/L, with ity. Baradarian and colleagues (57), who used a historical
a target level of 70 to 90 g/L, in the absence of tissue cohort comparison, suggested that correcting an INR to
hypoperfusion, coronary artery disease, or acute hemor- less than 1.8 as part of intensive resuscitation led to lower
rhage. The Transfusion Requirements in Critical Care trial mortality and fewer myocardial infarctions in the interven-
(52) in 838 critically ill patients suggested lower mortality tion group. The groups did not differ in length of stay,
with hemoglobin levels of 70 to 90 g/L than with levels of units of blood transfused, or time to endoscopy. Other
100 to 120 g/L (52). Unfortunately, these data excluded data suggest that endoscopic treatment with injection or
patients with UGIB. The actual transfusion requirement heater probe may be safely performed in patients with an
and threshold hemoglobin for transfusion in patients with INR less than 2.5 (58). A cohort study in patients who
acute UGIB (assuming a value after equilibration) may be underwent endoscopic treatment (58) found no differences
higher because of hemodynamic instability, inaccurate he- in rebleeding, surgery, mortality, or complication rates be-
moglobin measures, or the presence of continued or recur- tween patients receiving warfarin (baseline INR, 1.5 to
rent bleeding that leads to a rapid decrease to dangerously 6.0), whose INRs were corrected to 1.5 to 2.5 by using
low hemoglobin levels. In a prospective cohort study (53), fresh frozen plasma, and a control group who did not re-
hemoglobin levels less than 82 g/L in patients with UGIB ceive anticoagulants.
predicted elevated cardiac troponin I levels. Considering the paucity of data on INR correction
Because patients with UGIB are often elderly or have and the recognized benefits of early endoscopy (see state-
comorbid cardiovascular conditions, they may have poor ment B3), the participants felt that treating coagulopathy
tolerance for anemia. Threshold hemoglobin levels of 60 to was necessary in patients who received anticoagulants but
100 g/L may warrant transfusion in patients with underly- that endoscopy should not be delayed while doing so un-
ing cardiac disease (ischemic heart disease, peripheral vas- less the INR (or prothrombin time where INR is unavail-
cular surgery, or heart failure) (49). However, a prospective able) is supratherapeutic, because correction in these pa-
study (54) found no difference in postoperative morbidity tients may facilitate endoscopic treatment. This approach
and mortality between transfusion threshold levels less than should not be generalized to patients with cirrhosis because
80 g/L and less than 90 g/L in patients undergoing coro- the prothrombin time does not seem to predict bleeding
nary artery bypass graft surgery. risk in this setting (59). Correction of coagulopathy from
other causes may be necessary on a case-by-case basis.
Statement A5
In patients receiving anticoagulants, correction of coagu- Statement A6
lopathy is recommended but should not delay endoscopy. Promotility agents should not be used routinely before
(Agree, 97% [Vote: a, 38%; b, 44%; c, 15%; d, 3%]. endoscopy to increase the diagnostic yield.
Grade: Low, 2c, “probably do it”) (Agree, 82% [Vote: a, 35%; b, 35%; c, 12%; d, 6%;
Data on correction of coagulopathy are few and con- e, 3%; f, 9%]. Grade: Moderate, 2b, “probably don’t
flicting, as identified by a systematic review done for the do it”)
conference (55). A retrospective cohort study (56) that in- Although the use of preendoscopy promotility agents
cluded 233 patients with nonvariceal UGIB found that may improve diagnostic yield in selected patients with sus-
95% of the patients who received anticoagulants had an pected blood in the stomach, they are not warranted for
international normalized ratio (INR) between 1.3 and 2.7 routine use in all patients who present with UGIB.
(56). A baseline INR less than 1.3 versus 1.3 or greater did A meta-analysis (21) of 3 trials that evaluated erythro-
not predict rebleeding, transfusion requirement, surgery, mycin (60 – 62), comprising 316 patients, and 2 abstracts
length of stay, or mortality. Therefore, data suggest that it that evaluated metoclopramide (63, 64) found that use of a
may not be necessary to delay endoscopic therapy in pa- prokinetic agent significantly reduced the need for repeated
tients with mild to moderate coagulation defects. Further- endoscopy (odds ratio [OR], 0.51 [95% CI, 0.30 to 0.88])
more, an exploratory analysis of 1869 patients in the in patients suspected of having blood in their stomach,
RUGBE (Registry on Non-variceal Upper Gastrointestinal compared with placebo or no treatment (Appendix Table
Bleeding and Endoscopy) Canadian cohort study (9) 2). The groups did not differ in length of stay, units of
found that neither INR nor platelet count predicted re- blood transfused, or need for surgery. An analysis of data
bleeding. Although platelet count did not significantly pre- from the 3 erythromycin trials (65) found that preendo-
dict mortality, a presentation INR of 1.5 or greater was a scopic erythromycin resulted in a cost-effective outcome in
significant predictor of mortality in patients with UGIB most of the trials.
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 105
Because adequate visualization allows proper treatment (Agree, 94% [Vote: a, 32%; b, 38%; c, 24%; d, 3%; e,
in most patients and on the basis of the characteristics of 3%]. Grade: Moderate, 1b, “do it”)
patients selected for inclusion in the aforementioned trials, Although preendoscopic PPI therapy has not been
participants felt that promotility agents were not warranted shown to affect rebleeding, surgery, or mortality, the ben-
for routine use but may be useful in patients who are sus- eficial effects on the need for intervention, supportive cost-
pected to have substantial amounts of blood or clot in their effectiveness analyses, and excellent safety profile suggest
upper gastrointestinal tract or those who have recently that these agents may be useful, particularly in those sus-
eaten. pected of having high-risk stigmata.
A Cochrane meta-analysis through February 2006 that
included 5 RCTs (77) was updated with an additional trial
Statement A7 (78) and the full publication (79) of an abstract (22) (Ap-
Selected patients with acute ulcer bleeding who are at low pendix Table 2). The updated meta-analysis in 2223 pa-
risk for rebleeding on the basis of clinical and endoscopic cri- tients included 1 study that assessed oral PPI strategies and
teria may be discharged promptly after endoscopy. 5 studies that assessed intravenous strategies, only 1 of
(Agree, 97% [Vote: a, 53%; b, 35%; c, 9%; d, 3%]. which used a high-dose regimen (79). The investigators
Grade: High, 1a, “do it”) found no statistically significant differences in rates of mor-
Although some highly selected patients may be dis- tality, rebleeding, or surgery between the PPI therapy and
charged before undergoing endoscopy (see statement A2), control groups. However, preendoscopic PPI treatment
most patients will require risk stratification by endoscopic significantly reduced the proportion of patients with high-
as well as clinical criteria. Those classified as being at low risk stigmata (OR, 0.67 [CI, 0.54 to 0.84]) and the need
risk for rebleeding can be discharged promptly after endos- for endoscopic therapy (OR, 0.68 [CI, 0.50 to 0.93]) com-
copy (15). pared with patients in the control group who received pla-
An RCT in 95 low-risk patients (66) prospectively cebo or a histamine-2 receptor antagonist (22).
assessed the role of early discharge and found no differ- A North American analysis found that preendoscopic
ences in rates of recurrent bleeding but that early discharge PPI therapy was more effective and more costly in the
significantly reduced costs compared with admission (me- United States, whereas in Canada it became more effective
dian costs, $340 and $3940, respectively; P ⫽ 0.001). Cri- and less costly as the duration of hospitalization for high-
teria for early discharge included a clean ulcer base or flat risk patients increased or that of low-risk patients decreased
pigmented spot, hemodynamic stability, no serious concur- (80). Identifying patients with a greater likelihood of hav-
rent medical illness, easy accessibility to hospital, and ade- ing a high-risk lesion, such as those who present with red
quate sociofamily support at home. No patient who was blood in the emesis or nasogastric aspirate, may optimize
discharged early had any serious adverse event, underwent the cost-effectiveness of this approach (81). Other cost-
surgery, or died during the 30-day follow-up. effectiveness analyses have suggested either the economic
Substantial observational data (67–71) also support dominance of preendoscopic high-dose intravenous PPI
early discharge of low-risk patients after endoscopy. Pa- therapy (82) or the cost-effectiveness of oral PPI in this
tients stratified as low risk who were discharged early did setting (83), but certain model assumptions limit these
not differ in complications (for example, rebleeding, sur- conclusions.
gery, mortality), health status, or satisfaction from those The observed lesion downstaging attributable to PPI
who were admitted (72–74). Unfortunately, recommenda- therapy before endoscopy may be even more beneficial in
tions for early discharge based on endoscopic findings are situations in which early endoscopy may be delayed or
often not followed (75). when available endoscopic expertise may be suboptimal.
Patients are not suitable for early discharge if they have
serious comorbid conditions (heart failure, recent cardio- Section B: Endoscopic Management
vascular or cerebrovascular event, chronic alcoholism, or Statement B3
active cancer), are hemodynamically unstable, have an en- Early endoscopy (within 24 hours of presentation) is rec-
doscopic finding of high-risk stigmata (active bleeding, ommended for most patients with acute upper gastrointestinal
NBVV, or adherent clot), or have unsuitable sociofamily bleeding.
conditions (31, 76). Patient location (distance to nearest (Agree, 100% [Vote: a, 85%; b, 12%; c, 3%]. Grade:
emergency care center), local legal regulations, and social Moderate, 1b, “do it”)
support should also be considered (76). The definition of early endoscopy ranges from 2 to 24
hours after initial presentation (71, 75, 84, 85). Among the
1869 patients of the RUGBE cohort, 76% received their
Statement A8 first endoscopy within 24 hours of presentation (mean, 23
Preendoscopic PPI therapy may be considered to down- hours [SD, 38]) (9). In contrast, in a United Kingdom
stage the endoscopic lesion and decrease the need for endoscopic survey of 6750 patients (13), only 50% received endoscopy
intervention but should not delay endoscopy. within 24 hours. Early endoscopy (within the first 24
106 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
hours), with risk classification by clinical and endoscopic indirect findings from recent administrative data (87) sug-
criteria, allows for safe and prompt discharge of patients gest that early endoscopy may be associated with lower
classified as low risk, improves patient outcomes for pa- mortality.
tients classified as high risk, and reduces use of resources On the basis of available data, the participants recom-
for patients classified as either low or high risk (15). mended a target time to endoscopy of within 24 hours of
Although early endoscopy is encouraged for most pa- presentation.
tients, endoscopy may need to be delayed or deferred in
selected high-risk patients, such as those with active acute
Statement B5
coronary syndrome or suspected perforation. Also, a very
A finding of a clot in an ulcer bed warrants targeted
low Blatchford score can identify very low-risk patients
irrigation in an attempt at dislodgement, with appropriate
who are unlikely to have high-risk stigmata or benefit from
treatment of the underlying lesion.
endoscopic therapy (38) or who can be safely managed as
(Agree, 97% [Vote: a, 59%; b, 29%; c, 9%; d, 0%; f,
outpatients (36, 37) without the need for early endoscopy;
3%]. Grade: Moderate, 2b, “probably do it”)
however, this remains controversial (see statement A2).
Data suggest that early endoscopy is safe and effective
for all risk groups. A systematic review (71) found no ma- Statement B6
jor complications in patients triaged to outpatient care af- The role of endoscopic therapy for ulcers with adherent
ter early endoscopy. Early endoscopy is associated with sig- clots is controversial. Endoscopic therapy may be considered,
nificant reductions in length of hospital stay in patients at although intensive PPI therapy alone may be sufficient.
low risk (66, 72, 85), high risk (84), and combined patient (Agree, 86% [Vote: a, 24%; b, 50%; c, 12%; d, 9%; e,
groups (71, 86, 87), compared with delayed endoscopy. 3%; f, 3%]. Grade: Moderate, 2b, “probably do it”)
Recent administrative data found that the performance of Vigorous irrigation (for example, water pump) of a
early endoscopy was associated with a decreased need for clot in an ulcer bed has successfully exposed the underlying
surgery in elderly patients (88) and that patients with non- stigmata in 26% to 43% of cases (93, 94), and the revealed
variceal UGIB who were admitted on weekends had higher stigmata were high risk in 70% of those cases (94). The
adjusted in-hospital mortality and were less likely to un- endoscopic findings present after clot removal should be
dergo early endoscopy within 1 day of hospitalization (87). appropriately managed.
A large United Kingdom cohort analysis (13) has also The risk for rebleeding with clots that remain adherent
shown a strong trend in risk-adjusted mortality ratio that after washing without endoscopic therapy (with or without
just failed to show a statistically significant link between PPI therapy) has been reported to be as low as 0% to 8%
decreased mortality and the practice of after-hours (94, 95) but also as high as 25% to 35% (93, 96 –98) in
endoscopy. clinically high-risk patients. The disparity of these data has
Further analysis found no additional benefit from very led to controversy as to the optimal management of adher-
early or urgent (⬍12 hours) endoscopy over early (⬎12 ent clots (99).
hours) endoscopy. A meta-analysis of 3 trials (75, 84, 85), For the most part, endoscopic therapy for adherent
comprising 528 patients, found no significant reduction in clots refers to preinjecting them with epinephrine before
rebleeding (OR, 0.71 [CI, 0.28 to 1.81]), surgery (OR, shaving them down with a cold guillotining snare tech-
1.16 [CI, 0.39 to 3.51]), or mortality (OR, 0.70 [CI, 0.14 nique—without disrupting the pedicle of the clot—and
to 3.57]) with urgent (1 to 12 hours) endoscopy compared then applying combination treatment to the residual stig-
with later (⬎12 hours) endoscopy (Appendix Table 2). mata of hemorrhage (97, 98).
One study (85) reported significantly shorter hospital stays One meta-analysis of 5 RCTs (99), comprising 189
and lower costs with very early (1 to 2 hours) versus elec- patients with adherent clots, found no significant benefits
tive (1 to 2 days) endoscopy. In a subgroup of patients for endoscopic versus no endoscopic therapy (relative risk
with a bloody gastric aspirate, blood transfusions and hos- [RR], 0.31 [CI, 0.06 to 1.77]). Similarly, another meta-
pital stay were significantly reduced with urgent (⬍12 analysis that included 6 RCTs (100), comprising 240 pa-
hours vs. ⬎12 hours) endoscopy (84). Retrospective anal- tients, also found that endoscopic therapy did not signifi-
yses that assessed urgent (0 to 8 hours) versus early (6 or 8 cantly reduce rebleeding (RR, 0.48 [CI, 0.18 to 1.30])
to 24 hours) endoscopy (89 –91) reported no between- compared with medical therapy. A patient-level analysis of
group differences in clinical outcomes; however, these data from 4 fully published trials (100) did find a signifi-
studies did not control for other endoscopist-related fac- cant benefit for rebleeding (RR, 0.30 [CI, 0.10 to 0.77]).
tors, type of therapeutic interventions, or co-interventions. These discrepant results fuel the controversy and may be
A study identified 4 independent predictors (P ⬍ 0.050) of attributed to both varying patient populations and statisti-
active bleeding and the need for very early endoscopy (⬍12 cal heterogeneity not fully accounted for in the meta-
hours): fresh blood in the nasogastric tube, hemodynamic analytic methods (101). No data suggest excess risk; a
instability, a hemoglobin level less than 80 g/L, and a leu- systematic review (102) found a low incidence of compli-
kocyte count greater than 12 ⫻ 109 cells/L (92). Of note, cations resulting from endoscopic therapy.
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 107
The 1 RCT that compared endoscopic therapy plus group) and perforation (0.6% vs. 0%; P ⫽ 0.003) (102);
high-dose intravenous PPI therapy with high-dose intra- however, perforation has also been reported with mono-
venous PPI therapy alone found no rebleeding among 24 therapy in some RCTs (108, 109).
Asian patients with clots resistant to vigorous irrigation
(95). Therefore, among patients with adherent clots resis-
tant to vigorous irrigation, endoscopic therapy may be ben- Statement B11
eficial in patients at high risk for rebleeding (such as those Routine second-look endoscopy is not recommended.
with serious concurrent illness), whereas intensive PPI (Agree, 91% [Vote: a, 50%; b, 21%; c, 21%; d, 3%; e,
therapy without endoscopic treatment may be sufficient in 3%; f, 3%]. Grade: Moderate, 2b, “probably don’t do it”)
patients at low risk (particularly those who are Asian or A routine second-look endoscopy is generally defined
Helicobacter pylori–positive) (99). as a preplanned systematic second endoscopy performed 16
to 24 hours after the initial endoscopy, with appropriate
therapy in patients with evidence of active bleeding or
Statement B8
NBVV.
Epinephrine injection alone provides suboptimal efficacy
Although data support some benefits associated with
and should be used in combination with another method.
second-look endoscopy, they are generally older data and
(Agree, 100% [Vote: a, 71%; b, 24%; c, 6%]. Grade:
do not include the use of PPI therapy or optimal hemo-
Moderate, 1b, “don’t do it”)
static strategies. The findings are therefore not generaliz-
able to current clinical practice. In addition, cost-
Statement B10 effectiveness data do not seem to support the routine use of
Clips, thermocoagulation, or sclerosant injection should second-look endoscopy.
be used in patients with high-risk lesions, alone or in combi- Five published studies (110 –114) and 4 abstracts of
nation with epinephrine injection. randomized trials (115–118) have assessed a second-look
(Agree, 97% [Vote: a, 50%; b, 35%; c, 12%; d, 3%]. approach, with only 1 in each group demonstrating statis-
Grade: High, 1a, “do it”) tically significant benefits. The conclusions of 2 previous
Several recent meta-analyses have better quantified the meta-analyses of these trials (119, 120), which suggest ben-
efficacy of endoscopic therapies (99, 102–107). Although efits in rebleeding, are hampered by methodological limi-
monotherapy with epinephrine injection is more effective tations. In 1 case (119), the investigators noted analytical
than medical therapy in patients with high-risk stigmata, it shortcomings in data abstraction from 2 trials (111, 113),
is inferior to other monotherapies or to combination ther- whereas selection bias may have resulted in the other case
apy that uses 2 or more methods (99, 102–107). Numer- (120) from the inclusion of a decade-old abstract of a pos-
ous meta-analyses indicate that adding a second procedure, itive study (117), not yet fully published, but not of an-
such as a second injectate (for example, alcohol, thrombin, other abstract from the same era (118) that failed to
or fibrin glue), thermal contact, or clips, is superior to show efficacy.
epinephrine injection alone (99, 102, 103, 105, 107). Epi- A more recent meta-analysis of 6 trials (121) found
nephrine plus a second method for treating high-risk stig- that routine second-look endoscopy, with heater probe
mata significantly reduced rebleeding (OR, 0.51 [CI, 0.39 therapy when appropriate, significantly reduced the risk for
to 0.66]), surgery (OR, 0.63 [CI, 0.45 to 0.89]), and mor- rebleeding (RR, 0.29 [CI, 0.11 to 0.73]) compared with
tality compared with epinephrine monotherapy (OR, 0.50 single endoscopy; however, performing second-look endos-
[CI, 0.30 to 0.82]) (105). copy with injection monotherapy conferred no advantage.
Monotherapy with thermal devices, sclerosants, clips, A meta-analysis performed for the meeting (23) included 6
thrombin, or fibrin glue provides more effective endo- trials comprising 750 patients. It excluded 2 older abstracts
scopic hemostasis than epinephrine alone (99) or pharma- (117, 118), which have not been fully published, and Rut-
cotherapy alone (106). Clips were superior to injection geerts and colleagues’ study (114), which included second-
monotherapy in 4 (99, 102, 103, 106) of 5 meta-analyses look endoscopy in both study groups. In the meta-analysis,
(99, 102–104, 106). Clips with injection were superior to routine second-look endoscopy significantly decreased re-
injection alone but not to clips alone (103, 106). Combi- bleeding (OR, 0.59 [CI, 0.38 to 0.91]) and surgery (OR,
nation therapy (injection plus second injectate, thermal, or 0.43 [CI, 0.19 to 0.96]) but not mortality (OR, 0.65 [CI,
clips) was superior to injection therapy alone, but not to 0.26 to 1.62]) (Appendix Table 2). These findings must be
clips or thermal therapy alone (102, 106). The participants interpreted in light of differences across trials with regard
felt that the data were insufficient to show superiority or to patient selection, adopted methodologies, and both in-
equivalence of the recommended treatments but that the tervention and control treatments, as well as sensitivity
data were strongest for the use of thermal devices, clips, or analyses that show poor robustness of the results. The most
combination treatments. favorable results were from studies with the greatest pro-
Complications with dual versus single endoscopic portions of high-risk patients (110, 113). Indeed, Chiu
therapy included induction of bleeding (1.7% in each and colleagues (110) included 47% of patients who pre-
108 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
sented with shock and more than 40% with active bleed- A 2006 Cochrane meta-analysis of data as of Novem-
ing. Similarly, Saeed and colleagues (113) assessed patients ber 2004 (124) included 24 RCTs and was updated (24)
with a very high risk for rebleeding (on the basis of Forrest with 7 additional trials (125–131) (Appendix Table 2).
high-risk stigmata, as well as additional clinical and endo- The updated meta-analysis included 5792 patients. Over-
scopic criteria), of whom 70% had active bleeding. Al- all, PPI therapy with or without endoscopic therapy re-
though the investigators reported the noted decrease in duced rebleeding (OR, 0.45 [CI, 0.36 to 0.57]) and sur-
rebleeding (OR, 0.08 [CI, 0.00 to 1.50]) as statistically gery (OR, 0.56 [CI, 0.45 to 0.70]) but not mortality (OR,
significant in the final results of the latter trial, it did not 0.90 [CI, 0.67 to 1.19]) compared with placebo or
remain so when conventional 2-sided inferential testing, histamine-2 receptor agonist. Proton-pump inhibitor ther-
adapted to the small sample size, was applied. apy reduced mortality among patients with active bleeding
Of note, the most recently published trial—and the or NBVV and in trials conducted in Asia. Further analysis
only one with a control group that received high-dose in- showed that in patients with active bleeding or NBVV who
travenous PPI therapy (115)—found no benefit with received endoscopic hemostatic therapy, high-dose intra-
second-look endoscopy. High-dose intravenous PPI ther- venous PPI therapy (80 mg bolus plus 8 mg/h continuous
apy is the current standard in many health care systems, infusion) reduced rebleeding (OR, 0.43 [CI, 0.27 to
which suggests that second-look endoscopy may not pro- 0.67]), surgery (OR, 0.60 [CI, 0.37 to 0.96]), and mortal-
vide additional benefits when PPI therapy is available. ity (OR, 0.57 [CI, 0.34 to 0.96]). Lower doses of PPI
A U.S. cost-effectiveness analysis (122) found that a (either intravenous or oral) reduced rebleeding but no ev-
strategy of selective (not routine) second-look endoscopy at idence was found of an effect on mortality.
24 hours only in patients at high risk for rebleeding was Similarly, the meta-analysis by Laine and McQuaid
(99) found significant benefit in rebleeding (RR, 0.40
more effective and less expensive than repeated endoscopy
[CI, 0.28 to 0.59]), surgery (RR, 0.43 [CI, 0.24 to
in patients with rebleeding (with or without intravenous
0.58]), and mortality (RR, 0.41 [CI, 0.20 to 0.84]) with
PPIs) or routine repeated endoscopy in all patients, al-
high-dose intravenous PPI therapy after endoscopic
though extrapolations were made from trials that did not
therapy, whereas lower doses were associated with sig-
actually use high-dose intravenous PPI therapy. Intra-
nificant benefits in rebleeding (RR, 0.53 [CI, 0.35 to
venous PPI therapy became the dominant strategy if the 0.78]) but not surgery or mortality compared with pla-
rebleeding rate was 9% (closer to real-life outcomes) or if cebo or no treatment.
the cost of PPI dropped below $10 per day. When consid- Strong evidence demonstrates the efficacy of high-dose
ering baseline assumptions in the model and how they intravenous PPI therapy after successful endoscopy, but it
relate to current practice, PPI therapy seemed to be the is not possible to make conclusions regarding the efficacy
most cost-effective alternative. of either lower intravenous doses or high-dose oral therapy.
In the only study that fully reported risks (110), no Indeed, head-to-head comparisons and subgroup analyses
complications directly attributable to the second-look en- of high versus lower intravenous doses are underpowered,
doscopy were reported. and no direct comparisons of high-dose intravenous ther-
In conclusion, although older data supported a apy and high-dose oral therapy have been made. However,
second-look approach, these trials did not use contempo- lower intravenous doses or high-dose oral PPI therapy (at
rary management strategies associated with decreased re- doses equivalent to at least 4 times the standard daily oral
bleeding, such as initial endoscopic hemostasis with hemo- dose) are also effective (especially in Asian populations)
clips or combination therapy (123), or post– endoscopic and can be considered when high-dose intravenous therapy
hemostasis high-dose PPI therapy (24). Furthermore, the is not available or feasible.
few existing contemporary data do not favor the use of In patients with UGIB who have undergone successful
routine second-look endoscopy at this time. A subgroup of endoscopic hemostasis, administering high-dose intra-
patients with particularly high-risk presentations may ben- venous PPI therapy for 3 days is both more effective and
efit, but this requires further study. In light of these con- less costly than not doing so, as demonstrated by cost anal-
siderations, the participants did not recommend the rou- yses (132–134)— or can become so, as demonstrated by
tine use of second-look endoscopy. sensitivity analyses (83, 122). The intervention group usu-
ally receives high-dose intravenous PPI therapy, and the
Section C: Pharmacologic Management comparator group usually receives placebo; few trials have
Statement C3 included low-dose intravenous or oral PPI as comparators.
An intravenous bolus followed by continuous-infusion High-dose intravenous PPI therapy is a dominant strategy
PPI therapy should be used to decrease rebleeding and mortal- mainly because the cost of the medications is relatively
ity in patients with high-risk stigmata who have undergone lower than the incremental expenses of 1 additional re-
successful endoscopic therapy. bleeding episode. Comparisons of intravenous and oral PPI
(Agree, 94% [Vote: a, 65%; b, 24%; c, 6%; d, 3%; e, use remain theoretical, because only a few underpowered
0%; f, 3%]. Grade: High, 1a, “do it”) RCTs (135–137) have assessed this comparison.
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 109
Although recent data have linked PPI use to in- stigmata, should be admitted to a hospital for at least 72
hospital Clostridium difficile infection (138, 139), the par- hours.
ticipants felt that the benefits outweighed the risks in pa- However, 1 RCT showed that selected patients at
tients who have acute UGIB that requires PPI therapy. higher risk might be safely managed as outpatients after
endoscopic therapy without an increased risk for compli-
cations (149). The study in 82 patients with NBVV, ulcer
Statement C4
size less than 1.5 cm, no hypovolemia, no associated severe
Patients should be discharged with a prescription for a disease, and appropriate family support found no differ-
single daily-dose oral PPI for a duration as dictated by the ences in rebleeding, morbidity, or mortality at the 30-day
underlying etiology. follow-up among patients randomly assigned to outpatient
(Agree, 94% [Vote: a, 56%; b, 32%; c, 6%; d, 3%; e, or hospital care (149). Mean cost of care per patient was
3%]. Grade: Low, 1c, “do it”) significantly lower for the outpatient group than for the
Because rebleeding episodes may occur more than 3 hospital group ($970 vs. $1595; P ⬍ 0.001). However, the
days after endoscopy (125, 140, 141), most RCTs that study was underpowered to be able to confidently recom-
assess the role of postendoscopic PPI therapy have also mend early discharge of patients at higher risk at this time.
included a prescription for once-daily PPI therapy that The consensus participants felt that further research was
starts 72 hours after endoscopic hemostasis (125, 129, needed, especially in light of recent data that suggest some
140 –142). In the nonacute setting, once-daily PPI therapy patients may be discharged too early (150).
has demonstrated effective ulcer healing for patients with Patients should be admitted to a monitored setting for
peptic ulcer disease (143), and inadequate evidence sup- at least the first 24 hours on the basis of risk (hemody-
ports the need for twice-daily therapy in the maintenance namic instability, increasing age, severe comorbidity, active
setting. In the absence of direct comparative trial data of bleeding at endoscopy, or large ulcer size [for example, ⬎2
single versus daily PPI administration, this recommenda- cm]) (5, 32, 40, 151, 152) or clinical condition.
tion is supported by observational results in studies that One study (153) showed that use of a checklist with
used a standard once-daily dosage (125, 129, 140 –142). specific recommendations on diet, PPI therapy, H. pylori
However, the duration and dose of the PPI will be deter- eradication therapy, NSAID use, discharge, and follow-up
mined by the underlying etiology; for example, healing can significantly reduce the length of hospital stay for pa-
rates for complicated or severe esophagitis are relatively low tients with UGIB.
in some studies, and twice-daily doses may be warranted in
the context of UGIB (144). In addition, patients who re-
quire ASA or NSAID therapy may require long-term sec- Statement D4
ondary prophylaxis, as discussed in statements E1 to E4. Where available, percutaneous embolization can be con-
In community-based population studies, the use of sidered as an alternative to surgery for patients for whom
PPIs has been associated with potential side effects, includ- endoscopic therapy has failed.
ing C. difficile infection (145), pneumonia (146), and (Agree, 100% [Vote: a, 62%; b, 32%; c, 6%]. Grade:
osteoporosis-related fractures (147). These findings remain Low, 2c, “probably do it”)
controversial, and the benefits of PPI administration for Percutaneous or transcatheter arterial embolization has
acute ulcer healing outweigh these potential risks in the been investigated as an alternative to surgery in patients for
acute treatment setting; however, caution may be war- whom endoscopic therapy has failed, especially those who
ranted during long-term use. are high-risk candidates for surgery. Gelatin sponges, poly-
Section D: Nonendoscopic and Nonpharmacologic vinyl alcohol, cyanoacrylic glues, and coils have been used
In-Hospital Management to embolize the vessels feeding bleeding lesions (154).
Statement D2 In uncontrolled trials, primary rates of technical suc-
Most patients who have undergone endoscopic hemostasis cess range from 52% to 98%, with recurrent bleeding oc-
for high-risk stigmata should be hospitalized for at least 72 curring in about 10% to 20% of patients (155–159). A
hours thereafter. retrospective, single-center study (160) showed no signifi-
(Agree, 100% [Vote: a, 68%; b, 24%; c, 9%]. Grade: cant differences between embolization therapy and surgery
Low, 1c, “do it”) for rates of rebleeding, surgery, or mortality, despite pa-
Studies of the natural history of ulcer lesions show that tients in the embolization group being older and having a
it takes 72 hours for most high-risk lesions to become higher prevalence of heart disease.
low-risk lesions after endoscopic therapy (96, 148). In Although uncommon with modern, highly selective
many contemporary trials (125, 140, 141), 60% to 76% of techniques, complications include bowel ischemia; second-
patients who had rebleeding within 30 days after endo- ary duodenal stenosis; and gastric, hepatic, and splenic in-
scopic hemostasis plus high-dose PPI therapy did so within farction (154, 155, 159, 161). The high periprocedural
the first 72 hours. Thus, patients identified as being at high mortality of 25% to 30% is largely attributed to patients
risk for rebleeding, such as those with high-risk endoscopic being selected for this procedure because they are at high
110 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
surgical risk because of advanced age and underlying con- from U.S. and international groups on the reduction of
ditions (156, 158 –160). gastrointestinal risks associated with NSAID and antiplate-
For patients who had rebleeding after initial successful, let therapy (189, 190).
endoscopic hemostasis, a second attempt at endoscopic
therapy remains the preferred strategy (15, 162). When
Statement E1
persistent or recurrent bleeding cannot be controlled by
In patients with previous ulcer bleeding who require an
endoscopic therapy, percutaneous embolization can be
NSAID, it should be recognized that treatment with a tradi-
considered as an alternative to surgery, if such expertise is
tional NSAID plus PPI or a cyclooxygenase-2 (COX-2) inhib-
available.
itor alone is still associated with a clinically important risk for
recurrent ulcer bleeding.
Statement D5 (Agree, 97% [Vote: a, 73%; b, 21%; c, 3%; d, 3%].
Patients with bleeding peptic ulcers should be tested for Grade: Moderate, 1b)
H. pylori and receive eradication therapy if it is present, with
confirmation of eradication.
Statement E2
(Agree, 94% [Vote: a, 82%; b, 12%; c, 0%; d, 3%; e,
3%]. Grade: High, 1a, “do it”) In patients with previous ulcer bleeding who require an
NSAID, the combination of a PPI and a COX-2 inhibitor is
recommended to reduce the risk for recurrent bleeding from
Statement D6
that of COX-2 inhibitors alone.
Negative H. pylori diagnostic tests obtained in the acute (Agree, 94% [Vote: a, 52%; b, 33%; c, 9%; d, 0%; e,
setting should be repeated. 6%]. Grade: Moderate, 1b, “do it”)
(Agree, 94% [Vote: a, 68%; b, 21%; c, 6%; d, 3%; e, Two small RCTs conducted in Asia (191–193) found
3%]. Grade: Moderate, 1b, “do it”) no significant difference in the rate of recurrent bleeding or
As recommended in the previous consensus (15), pa- ulcer complications (about 4% to 6%) at 6 months with
tients with UGIB should be tested for H. pylori and receive COX-2 inhibitor therapy alone versus therapy with a tra-
eradication therapy if infection is present. A meta-analysis ditional NSAID plus PPI. The relatively small numbers of
(163) demonstrated that eradication of H. pylori was sig- patients in these studies do not exclude a benefit of one
nificantly more effective than PPI therapy alone in prevent- strategy over the other. These studies did not assess tradi-
ing rebleeding from peptic ulcer disease. The rebleeding tional NSAID therapy alone or complete withdrawal of
rate was even lower among the subgroup of patients with NSAID therapy, and although a strategy of COX-2 inhib-
successful eradication, which emphasizes the importance of itor therapy alone or therapy with a traditional NSAID plus
confirming eradication. PPI may have lowered the rates of recurrent bleeding com-
Tests for H. pylori may show increased false-negative pared with historical rates from therapy with traditional
rates in the context of acute bleeding, although the data NSAIDs alone, the risk was not eliminated (191–193).
vary. Although the biological mechanisms involved are Population-based studies (194, 195) also support add-
poorly understood—and may indeed vary depending on ing a PPI to traditional NSAID therapy or administering a
the test— one suggestion is the pH buffering effect of the COX-2 inhibitor to reduce the risk for upper gastrointes-
blood, because more alkaline settings are known to be as- tinal complications; however, the combination of a COX-2
sociated with more false-negative results (164). A system- inhibitor with a PPI was associated with the greatest risk
atic review of 23 studies (165–187), done for the consensus reduction.
meeting, found that diagnostic tests for H. pylori infection One RCT (196) demonstrated a significantly lower
(including serology, histology, urea breath test, rapid ure- rate of recurrent UGIB with a COX-2 inhibitor plus a PPI
ase test, stool antigen, and culture) demonstrate high pos- (0%) compared with a COX-2 inhibitor alone (8.9%) over
itive predictive value (0.85 to 0.99) but low negative pre- 1 year (difference, 8.9 percentage points [CI, 4.1 to 13.7
dictive value (0.45 to 0.75) in the setting of acute UGIB, percentage points). A subgroup analysis (197) of pooled
with 25% to 55% of H. pylori–infected patients yielding data from 3 RCTs with similar study designs, comprising
false-negative results (Appendix Table 3) (25). This sug- 34 701 patients, suggested a lower incidence of clinical gas-
gests caution in the interpretation of initially negative re- trointestinal events with a COX-2 inhibitor plus a PPI
sults and the need for repeated testing at follow-up. compared with a COX-2 inhibitor alone; however, no sta-
Section E: Postdischarge, ASA, and NSAIDs tistical analysis was performed. Several studies (198, 199)
The following statements are similar to those included have also shown lower risks for endoscopic ulcers in pa-
in the NSAID guidelines developed by a Canadian consen- tients who receive a COX-2 inhibitor plus a PPI compared
sus group, which included 8 of the participants present at with those who receive a COX-2 inhibitor alone.
this consensus (188). Statements E1 and E2 were included Two meta-analyses (200, 201) have demonstrated an
in the NSAID guidelines and are here in condensed for- excess risk for serious cardiovascular events associated with
mat. We also refer the reader to consensus publications COX-2 inhibitors compared with placebo. Optimal man-
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 111
agement of patients who require long-term NSAID ther- were treated with a PPI and randomly assigned to restart
apy should consider both gastrointestinal and cardiovascu- antiplatelet therapy with ASA or clopidogrel within 1
lar risks (188). No evidence has been found of a further day of endoscopy. Therefore, the participants agreed
increased risk for renal, cardiovascular, or dermatologic ad- that no ASA-free period should be mandated; instead,
verse events with COX-2 inhibitor plus PPI therapy com- patients who require ASA for cardiovascular protection
pared with COX-2 inhibitor therapy alone (196). should restart ASA therapy as soon as the risks for car-
Several cost-effectiveness analyses (202–204) in pa- diovascular complication are thought to outweigh the risks
tients at high risk for gastrointestinal events found that a for bleeding.
COX-2 inhibitor was more cost-effective than a traditional
NSAID plus PPI, and a traditional NSAID plus PPI was
more cost-effective than a traditional NSAID alone; how-
Statement E4
ever, these analyses did not include a strategy of therapy
with a COX-2 inhibitor plus PPI. In patients with previous ulcer bleeding who require car-
In summary, patients with previous ulcer bleeding re- diovascular prophylaxis, it should be recognized that clopi-
quire more careful follow-up and alternative strategies, in- dogrel alone has a higher risk for rebleeding than ASA com-
cluding discontinuation of NSAID therapy when possible bined with a PPI.
or therapy with a COX-2 inhibitor plus PPI. Statement C4 (Agree, 100% [Vote: a, 70%; b, 24%; c, 6%]. Grade:
discusses the potential side effects associated with long- Moderate, 1b, “do it” [adding PPI to ASA])
term PPI use. Clopidogrel is often perceived as relatively safe in
terms of gastrointestinal adverse events, but data show that
even as monotherapy, clopidogrel is associated with a high
Statement E3 risk for rebleeding (9% to 14%) (211, 212).
In patients who receive low-dose ASA and develop acute Pooled results of 2 RCTs (211, 212) showed a signif-
ulcer bleeding, ASA therapy should be restarted as soon as the icant reduction in rebleeding with ASA plus a PPI com-
risk for cardiovascular complication is thought to outweigh the pared with clopidogrel therapy alone (OR, 0.06 [CI, 0.01
risk for bleeding. to 0.32]) but no significant effect on mortality (OR, 0.63
(Agree, 100% [Vote: a, 70%; b, 30%]. Grade: Mod- [CI, 0.24 to 1.64]) (Appendix Table 2). The 2 groups did
erate, 1b, “do it”) not differ in the development or relapse of cardiovascular
Patients receiving low-dose ASA who develop UGIB or cerebrovascular events.
are often advised to discontinue ASA therapy until the Physicians should be aware that PPIs may decrease the
ulcers have healed. However, prolonged discontinuation of platelet inhibitory effect of clopidogrel (213–215). The
ASA therapy increases thrombotic risk in patients who re- PPI and clopidogrel may compete for the cytochrome
quire cardioprotective ASA therapy (205, 206). In a meta- P450 isoenzyme CYP2C19, which is required to convert
analysis (206), ASA nonadherence or withdrawal was asso- the prodrug clopidogrel to its active metabolite (216 –218).
ciated with a 3-fold higher risk for major adverse cardiac Some observational studies in clopidogrel recipients (218 –
events. The delay to a thrombotic event is generally re- 220) show a small but significant association between PPI
ported as between 7 and 30 days, and usually between 7 use and cardiovascular events, whereas others (215, 221,
and 10 days (205, 207, 208). This temporal pattern has
222) do not. No randomized trials addressing this issue are
biological plausibility because the inhibited platelets circu-
available.
late in the blood for about 10 days (205). The American
The American College of Cardiology, American Heart
Heart Association recommends (189) that the decision to
Association, and American College of Gastroenterology
discontinue ASA therapy in the setting of acute ulcer
bleeding be made on an individual basis, on the basis of currently recommend that patients receiving these medica-
cardiac and gastrointestinal risks. tions not change their treatment regimen unless advised by
Data from RCTs (209, 210) suggest that the cardio- their health care provider (223). The U.S. Food and Drug
vascular benefits of early reintroduction of ASA or clopi- Administration found sufficient evidence of an interaction
dogrel may outweigh the gastrointestinal risks. An RCT in to require the inclusion of a statement on the clopidogrel
156 patients with ASA-induced ulcer bleeding who under- product label that concomitant administration of drugs
went endoscopic therapy (209) found that immediate re- that inhibit CYP2C19 (such as omeprazole) should be dis-
introduction of ASA in the presence of intravenous and couraged (224). On the basis of pharmacologic profiles,
oral PPI therapy was associated with a 2-fold (but statisti- some experts suggest a staggered schedule of intake for
cally insignificant) increase in the risk for recurrent bleed- clopidogrel and the PPI (216, 225), but further research in
ing from peptic ulcers, but discontinuation of ASA therapy this area is needed.
was associated with a significantly increased 8-week mor- With regard to the potential side effects associated
tality rate. Another RCT (210) found no cases of rebleed- with long-term PPI use, we refer the reader to statement
ing in patients with ASA-associated endoscopic ulcers who C4.
112 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 113
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 W-21
W-22 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 W-23
W-24 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
www.annals.org 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 W-25
W-26 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
W-28 19 January 2010 Annals of Internal Medicine Volume 152 • Number 2 www.annals.org
Clinical guideline
Published: 13 June 2012
www.nice.org.uk/guidance/cg141
Your responsibility
The recommendations in this guideline represent the view of NICE, arrived at after careful
consideration of the evidence available. When exercising their judgement, professionals and
practitioners are expected to take this guideline fully into account, alongside the individual needs,
preferences and values of their patients or the people using their service. It is not mandatory to
apply the recommendations, and the guideline does not override the responsibility to make
decisions appropriate to the circumstances of the individual, in consultation with them and their
families and carers or guardian.
Local commissioners and providers of healthcare have a responsibility to enable the guideline to be
applied when individual professionals and people using services wish to use it. They should do so in
the context of local and national priorities for funding and developing services, and in light of their
duties to have due regard to the need to eliminate unlawful discrimination, to advance equality of
opportunity and to reduce health inequalities. Nothing in this guideline should be interpreted in a
way that would be inconsistent with complying with those duties.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 2
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
Contents
Introduction ......................................................................................................................................................................... 4
1 Guidance ............................................................................................................................................................................ 8
1.6 Control of bleeding and prevention of re-bleeding in patients on NSAIDs, aspirin or clopidogrel ......... 11
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 3
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
Introduction
Acute upper gastrointestinal bleeding is a common medical emergency that has a 10% hospital
mortality rate. Despite changes in management, mortality has not significantly improved over the
past 50 years.
Elderly patients and people with chronic medical diseases withstand acute upper gastrointestinal
bleeding less well than younger, fitter patients, and have a higher risk of death. Almost all people
who develop acute upper gastrointestinal bleeding are treated in hospital and the guideline
therefore focuses on hospital care. The most common causes are peptic ulcer and oesophago-
gastric varices.
Endoscopy is the primary diagnostic investigation in patients with acute upper gastrointestinal
bleeding but it has not always been clear whether urgent endoscopy is cost effective as well as
clinically valuable. Endoscopy aids diagnosis, yields information that helps predict outcome and
most importantly allows treatments to be delivered that can stop bleeding and reduce the risk of
re-bleeding.
Drugs may have a complementary role in reducing gastric acid secretion and portal vein pressure.
Not every patient responds to endoscopic and drug treatments; emergency surgery and a range of
radiological procedures may be needed to control bleeding.
This guideline aims to identify which diagnostic and therapeutic steps are useful in managing acute
upper gastrointestinal bleeding. This should enable hospitals to develop a structure in which
clinical teams can deliver an optimum service for people who develop this condition.
The guideline will assume that prescribers will use a drug's summary of product characteristics to
inform decisions made with individual patients.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 4
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
Patient-centred care
This guideline offers best practice advice on the care of adults and young people aged 16 years and
older with acute variceal and non-variceal upper gastrointestinal bleeding.
Treatment and care should take into account patients' needs, preferences and religious beliefs.
People with acute upper gastrointestinal bleeding should have the opportunity to make informed
decisions about their care and treatment, in partnership with their healthcare professionals. If
patients do not have the capacity to make decisions, healthcare professionals should follow the
Department of Health's advice on consent and the code of practice that accompanies the Mental
Capacity Act. In Wales, healthcare professionals should follow advice on consent from the Welsh
Government. In taking account of patients' religious beliefs in the context of blood transfusion,
healthcare professionals should follow the advice from UK Blood Transfusion and Tissue
Transplantation Services.
If the patient agrees, families and carers should have the opportunity to be involved in decisions
about treatment and care.
Families and carers should also be given the information and support they need.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 5
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
Risk assessment
• Use the following formal risk assessment scores for all patients with acute upper
gastrointestinal bleeding:
Timing of endoscopy
• Offer endoscopy to unstable patients with severe acute upper gastrointestinal bleeding
immediately after resuscitation.
• Offer endoscopy within 24 hours of admission to all other patients with upper gastrointestinal
bleeding.
• Units seeing more than 330 cases a year should offer daily endoscopy lists. Units seeing fewer
than 330 cases a year should arrange their service according to local circumstances.
• Do not use adrenaline as monotherapy for the endoscopic treatment of non-variceal upper
gastrointestinal bleeding.
• For the endoscopic treatment of non-variceal upper gastrointestinal bleeding, use one of the
following:
• Offer interventional radiology to unstable patients who re-bleed after endoscopic treatment.
Refer urgently for surgery if interventional radiology is not promptly available.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 6
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
• Continue low-dose aspirin for secondary prevention of vascular events in patients with upper
gastrointestinal bleeding in whom haemostasis has been achieved.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 7
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
1 Guidance
The following guidance is based on the best available evidence. The full guideline gives details of
the methods and the evidence used to develop the guidance.
1.2.2 Base decisions on blood transfusion on the full clinical picture, recognising that
over-transfusion may be as damaging as under-transfusion.
1.2.3 Do not offer platelet transfusion to patients who are not actively bleeding and
are haemodynamically stable.
1.2.4 Offer platelet transfusion to patients who are actively bleeding and have a
platelet count of less than 50 x 109/litre.
• Offer fresh frozen plasma to patients who are actively bleeding and have a
prothrombin time (or international normalised ratio) or activated partial
thromboplastin time greater than 1.5 times normal. If a patient's fibrinogen level
remains less than 1.5 g/litre despite fresh frozen plasma use, offer cryoprecipitate as
well.
1.2.5 Offer prothrombin complex concentrate to patients who are taking warfarin
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 8
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
1.2.6 Treat patients who are taking warfarin and whose upper gastrointestinal
bleeding has stopped in line with local warfarin protocols.
1.2.7 Do not use recombinant factor Vlla except when all other methods have failed.
1.3.2 Offer endoscopy within 24 hours of admission to all other patients with upper
gastrointestinal bleeding.
1.3.3 Units seeing more than 330 cases a year should offer daily endoscopy lists. Units
seeing fewer than 330 cases a year should arrange their service according to
local circumstances.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 9
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
1.4.6 Offer a repeat endoscopy to patients who re-bleed with a view to further
endoscopic treatment or emergency surgery.
Oesophageal varices
1.5.3 Use band ligation in patients with upper gastrointestinal bleeding from
oesophageal varices.
Gastric varices
1.5.5 Offer endoscopic injection of N-butyl-2-cyanoacrylate to patients with upper
gastrointestinal bleeding from gastric varices.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 10
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
1.5.6 Offer TIPS if bleeding from gastric varices is not controlled by endoscopic
injection of N-butyl-2-cyanoacrylate.
1.6.3 Discuss the risks and benefits of continuing clopidogrel (or any other
thienopyridine antiplatelet agents) in patients with upper gastrointestinal
bleeding with the appropriate specialist (for example, a cardiologist or a stroke
specialist) and with the patient.
1.7.2 Review the ongoing need for acid-suppression drugs for primary prevention of
upper gastrointestinal bleeding in acutely ill patients when they recover or are
discharged from critical care.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 11
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
• ensuring patients and their families and carers receive consistent information.
More information
You can also see this guideline in the NICE pathway on Acute upper gastrointestinal bleeding.
To find out what NICE has said on topics related to this guideline, see our web page on Digestive
tract conditions.
See also the guideline committee's discussion and the evidence reviews (in the full guideline),
and information about how the guideline was developed, including details of the committee.
[1]
At the time of publication (June 2012), terlipressin was indicated for the treatment of bleeding
from oesophageal varices, with a maximum duration of treatment of 72 hours (3 days). Prescribers
should consult the relevant summary of product characteristics. Informed consent for off-label use
of terlipressin should be obtained and documented.
[2]
As of August 2016, only the H2-receptor antagonists ranitidine and cimetidine are licensed for
prophylaxis of gastrointestinal bleeding in acutely ill patients. The proton pump inhibitors
omeprazole, esomeprazole, lansoprazole, pantoprazole and rabeprazole are not licensed for
prophylaxis of gastrointestinal bleeding in acutely ill patients. The use of proton pump inhibitors or
H2-receptor antagonists other than ranitidine and cimetidine for this indication would be off label.
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 12
conditions#notice-of-rights). Last updated 25 August 2016 of 13
Acute upper gastrointestinal bleeding in over 16s: management (CG141)
Update information
August 2016
2016: A footnote was added to recommendation 1.7.1 about which proton pump inhibitors
and H2-receptor antagonists are licensed for use and which are classed as off-label.
April 2015
2015: Recommendation 1.2.5 has been amended to add the use of cryoprecipitate as further
treatment.
ISBN: 978-1-4731-2041-9
Accreditation
© NICE 2020. All rights reserved. Subject to Notice of rights (https://www.nice.org.uk/terms-and- Page 13
conditions#notice-of-rights). Last updated 25 August 2016 of 13
GUIDELINE
This is one of a series of statements discussing the use of patients with non-variceal UGIB has been demonstrated to
GI endoscopy in common clinical situations. The Stan- improve patient outcomes.4 This updated ASGE guideline
dards of Practice Committee of the American Society for focuses on the role of GI endoscopy in patients with acute
Gastrointestinal Endoscopy (ASGE) prepared this text. In non-variceal UGIB. This guideline will not address ob-
preparing this guideline, a search of the medical literature scure GI bleeding or the role of endoscopy in the man-
was performed by using PubMed. Additional references agement of variceal bleeding, both of which are addressed
were obtained from the bibliographies of the identified in existing ASGE practice guidelines.5-6 UGIB refers to GI
articles and from recommendations of expert consultants. blood loss having an origin proximal to the ligament of
When few or no data exist from well-designed prospective Treitz. Acute UGIB can manifest as hematemesis, “coffee
trials, emphasis is given to results from large series and ground’’ emesis, the return of red blood via a nasogastric
reports from recognized experts. Guidelines for appropri- tube, and/or melena with or without hemodynamic com-
ate use of endoscopy are based on a critical review of the promise. Hematochezia may occur in patients with ex-
available data and expert consensus at the time that the tremely brisk UGIB.7-8
guidelines are drafted. Further controlled clinical studies
may be needed to clarify aspects of this guideline. This guide-
line may be revised as necessary to account for changes in INITIAL ASSESSMENT AND MANAGEMENT
technology, new data, or other aspects of clinical practice.
The recommendations are based on reviewed studies and are Initial assessment
graded on the strength of the supporting evidence1 (Table 1). A primary goal of the initial assessment is to determine
The strength of individual recommendations is based on whether the patient requires urgent intervention (eg, en-
both the aggregate evidence quality and an assessment of the doscopic, surgical, transfusion) or can undergo delayed
anticipated benefits and harms. Weaker recommendations endoscopy or even be discharged to outpatient manage-
are indicated by phrases such as “We suggest . . . ,” whereas ment. Although numerous factors from the patient history,
stronger recommendations are typically stated as “We rec- physical examination, and initial tests have been exam-
ommend . . . .” ined for an association with a need for intervention, no
This guideline is intended to be an educational device single factor is sufficiently predictive of UGIB severity to
to provide information that may assist endoscopists in be used for triage. The most predictive individual factors
providing care to patients. This guideline is not a rule and are a history of malignancy or cirrhosis,9 presentation with
should not be construed as establishing a legal standard of hematemesis,9-10 and signs of hypovolemia including hy-
care or as encouraging, advocating, requiring, or discour- potension,9,11 tachycardia and shock, and a hemoglobin
aging any particular treatment. Clinical decisions in any ⬍8 g/dL.9-10 Some factors, such as a history of aspirin or
particular case involve a complex analysis of the patient’s nonsteroidal anti-inflammatory drug (NSAID) use, may not
condition and available courses of action. Therefore, clin- be useful for immediate disposition but are still important
ical considerations may lead an endoscopist to take a to assess for future management (eg, if peptic ulcer disease
course of action that varies from these guidelines. [PUD] were the etiology of UGIB, then NSAID use should
be discontinued).11 Patients who have significant comor-
INTRODUCTION bidities may require admission regardless of the severity of
the UGIB.
Upper GI bleeding (UGIB) results in over 300,000 hos- Because individual clinical factors are generally not
pital admissions annually in the United States, with a diagnostic of UGIB severity, there have been attempts to
mortality of 3.5% to 10%.2-3 Appropriate management of create prediction rules. In 3 studies comparing clinical
prediction rule scores in the same study population, the
Blatchford score performed better than the Clinical Rockall
score for predicting patients at high risk for clinical
Copyright © 2012 by the American Society for Gastrointestinal Endoscopy
0016-5107/$36.00 intervention.12-14 The Blatchford score15 and the Clinical
doi:10.1016/j.gie.2012.02.033 Rockall score16 have been examined in several studies and
may determine the need for urgent endoscopy. The
TABLE 1. GRADE system for rating the quality of TABLE 2. Blatchford scoring: Admission risk markers
evidence for guidelines and associated score component values
90-99 2
Blatchford score uses data on blood urea and hemoglobin ⬍90 3
levels, systolic blood pressure, pulse, presentation with
Other markers
melena, presentation with syncope, history of hepatic dis-
ease, and history of heart failure (Table 2).15 A Blatchford Pulse ⱖ100/min 1
score ⬎0 was 99% to 100% sensitive for identifying a Presentation with melena 1
severe bleed in 5 studies.12-15,17 The specificity of the
Blatchford scoring system is low (4%-44%), but clinically it Presentation with syncope 2
is more important to be comfortable identifying all severe Hepatic disease 2
UGIB at the expense of admitting some patients with
Cardiac failure 2
minor bleeding episodes. Patients found to have minor
Adapted with permission from Blatchford O, Murray WR, Blatchford
bleeding episodes typically may be discharged soon after
M. A risk score to predict need for treatment for upper-
endoscopy. Use of the Blatchford score may allow early gastrointestinal haemorrhage. Lancet 2000;356:1318-21.15
discharge of 16% to 25% of all patients presenting with
UGIB.12-15
then ruptures. These lesions are usually in the stomach but therapy reduces the rate of recurrent bleeding to approx-
may occur throughout the GI tract.66 Endoscopic methods imately 10%.54,56 Patients with recurrent bleeding generally
to treat Dieulafoy lesions include banding, clipping, elec- respond favorably to repeat endoscopic therapy.77-78 Rou-
trocautery, cyanoacrylate glue, sclerosant injection, epi- tine second-look endoscopy, defined as a planned endos-
nephrine injection, heater probe, and laser therapy. Large, copy performed within 24 hours of the initial endoscopy,
single-center experiences have not identified one modality is not recommended.79 In cases where the initial endos-
as being superior to others; however, epinephrine injec- copy failed to identify the source (eg, because of a large
tion monotherapy is associated with a higher rate of re- clot in the stomach) or if there are concerns that inade-
current bleeding.67-69 Given the difficulty in identifying quate therapy was delivered, repeat endoscopy may be
these lesions absent active hemorrhage, tattooing of the appropriate.
lesion should be considered to facilitate identification and
treatment in the event of recurrent bleeding. If endoscopic RECOMMENDATIONS
therapy fails, interventional radiology or surgical ap- ● We recommend that patients with UGIB be adequately
proaches may be required. Placement of a clip can help resuscitated before endoscopy. QŒŒŒ
identify the lesion should recurrent bleeding cease before ● We recommend antisecretory therapy with PPIs for
these non-endoscopic interventions. patients with bleeding caused by peptic ulcers or in
those with suspected peptic ulcer bleeding awaiting
Aortoenteric fistulas endoscopy. QQQQ
Aortoenteric fistulas may be primary (caused by arte- ● We suggest prokinetic agents in patients with a high
riosclerosis, aortic aneurysms, aortic infections),68 or sec- probability of having fresh blood or a clot in the stom-
ondary (aortic repair with a synthetic graft).69 This condi- ach when undergoing endoscopy. QQŒŒ
tion is a medical emergency that may present with what ● We recommend endoscopy to diagnose the etiology of
appears to be self-limited bleeding (“herald” bleed). The acute UGIB. QQQŒ
clinical suspicion of an aortoenteric fistula should prompt The timing of endoscopy should depend on clinical
emergency CT imaging. CT scans and angiography can factors. Urgent endoscopy (within 24 hours of presen-
demonstrate the fistula if contrast extravasation into the tation) is recommended for patients with a history of
bowel is visualized.70-72 There is no endoscopic therapy malignancy or cirrhosis, presentation with hematem-
for a bleeding aortoenteric fistula, although endoscopy esis, and signs of hypovolemia including hypotension,
may be required to confirm the diagnosis or exclude other tachycardia and shock, and a hemoglobin ⬍8 g/dL.
causes of UGIB. Most aortoenteric fistulas occur at the ● We recommend endoscopic therapy for peptic ulcers
level of the distal duodenum or the jejunum and may be with high-risk stigmata (active spurting, visible vessel).
beyond the reach of a standard upper endoscope. In QQQQ
some cases, aortic graft material may be seen protruding The management of PUD with an adherent clot is con-
into the bowel lumen. Emergency surgical consultation troversial. Recommended endoscopic treatment modal-
should be obtained. ities include injection (sclerosants, thrombin, fibrin, or
cyanoacrylate glue), cautery, and mechanical therapies.
GI tumors ● We recommend against epinephrine injection alone for
Benign or malignant GI tumors, whether primary or
peptic ulcer bleeding. If epinephrine injection is per-
metastatic, cause approximately 5% of cases of UGIB.73
formed, it should be combined with a second endo-
Case series of endoscopic therapy have reported initial
scopic treatment modality (eg, cautery or clips). QQQQ
hemostasis rates similar to, or lower than, those seen in
● We recommend that patients with low-risk lesions be
PUD and high recurrent bleeding rates, between 16% and
considered for outpatient management. QQQŒ
80%.74-76 Procedure-related complications also appear to
● We recommend against routine second-look endos-
be more common.75-76 The optimal treatment modality has
copy in patients who have received adequate endo-
not been defined and depends on the goals of therapy.
scopic therapy. QŒŒŒ
Surgery or angiography may be better approaches to en-
● We recommend repeat endoscopy for patients with
suring long-term hemostasis. Any lesion appearing malig-
evidence of recurrent bleeding. QQQŒ
nant when seen in the context of an episode of UGIB
should be biopsied.
DISCLOSURES
RECURRENT BLEEDING AFTER ENDOSCOPIC
THERAPY T. Ben-Menachem is a consultant for Boston Scientific.
D. Fisher is a consultant for Epigenomics, Inc. K. Chathadi
Despite adequate initial endoscopic therapy, recurrent is a speaker for Boston Scientific. Rajeev Jain is a consul-
UGIB can occur in up to 24% of high-risk patients. The use tant for Boston Scientific and does research for Barxx. J.
of PPI therapy in addition to combination endoscopic Saltzman is a consultant for Hemoclip Development and
has a relationship with Cook Endoscopy. No other finan- 19. Cuellar RE, Gavaler JS, Alexander JA, et al. Gastrointestinal tract hemor-
cial relationships relevant to this publication were rhage. The value of a nasogastric aspirate. Arch Intern Med 1990;150:
1381-4.
disclosed. 20. Palamidessi N, Sinert R, Falzon L, et al. Nasogastric aspiration and lavage
in emergency department patients with hematochezia or melena with-
Abbreviations: NSAID, nonsteroidal anti-inflammatory drug; PPI, proton out hematemesis. Acad Emerg Med 2010;17:126-132.
pump inhibitor; PUD, peptic ulcer disease; UGIB, upper GI bleeding. 21. Singer AJ, Richman PB, Kowalska A, et al. Comparison of patient and
practitioner assessments of pain from commonly performed emer-
gency department procedures. Ann Emerg Med 1999;33:652-8.
REFERENCES 22. Sreedharan A, Martin J, Leontiadis GI, et al. Proton pump inhibitor treat-
ment initiated prior to endoscopic diagnosis in upper gastrointestinal
1. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on bleeding. Cochrane Database Syst Rev 2010;CD005415.
rating quality of evidence and strength of recommendations. BMJ 2008; 23. Barkun AN, Bardou M, Martel M, et al. Prokinetics in acute upper GI
336:924-6. bleeding: a meta-analysis. Gastrointest Endosc 2010;72:1138-45.
2. Targownik LE, Nabalamba A. Trends in management and outcomes of 24. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recom-
acute nonvariceal upper gastrointestinal bleeding: 1993-2003. Clin Gas- mendations on the management of patients with nonvariceal upper
troenterol Hepatol 2006;4:1459-66. gastrointestinal bleeding. Ann Intern Med 2010;152:101-13.
3. Yavorski RT, Wong RK, Maydonovitch C, et al. Analysis of 3,294 cases of 25. Carbonell N, Pauwels A, Serfaty L, et al. Erythromycin infusion prior to
upper gastrointestinal bleeding in military medical facilities. Am J Gas- endoscopy for acute upper gastrointestinal bleeding: a randomized,
troenterol 1995;90:568-73. controlled, double-blind trial. Am J Gastroenterol 2006;101:1211-5.
4. Barkun A, Sabbah S, Enns R, et al. The Canadian Registry on Nonvariceal 26. Coffin B, Pocard M, Panis Y, et al. Erythromycin improves the quality of
Upper Gastrointestinal Bleeding and Endoscopy (RUGBE): Endoscopic EGD in patients with acute upper GI bleeding: a randomized controlled
hemostasis and proton pump inhibition are associated with improved study. Gastrointest Endosc 2002;56:174-9.
outcomes in a real-life setting. Am J Gastroenterol 2004;99:1238-46. 27. Frossard JL, Spahr L, Queneau PE, et al. Erythromycin intravenous bolus
5. Fisher L, Lee KM, Anderson MA, et al. The role of endoscopy in the man- infusion in acute upper gastrointestinal bleeding: a randomized, con-
trolled, double-blind trial. Gastroenterology 2002;123:17-23.
agement of obscure GI bleeding. Gastrointest Endosc 2010;72:471-9.
28. Chak A, Cooper GS, Lloyd LE, et al. Effectiveness of endoscopy in pa-
6. Qureshi W, Adler DG, Davila R, et al. ASGE Guideline: the role of endos-
tients admitted to the intensive care unit with upper GI hemorrhage.
copy in the management of variceal hemorrhage, updated July 2005.
Gastrointest Endosc 2001;53:6-13.
Gastrointest Endosc 2005;62:651-5.
29. Lin HJ, Wang K, Perng CL, et al. Early or delayed endoscopy for patients
7. Wilcox CM, Alexander LN, Cotsonis G. A prospective characterization of
with peptic ulcer bleeding. A prospective randomized study. J Clin Gas-
upper gastrointestinal hemorrhage presenting with hematochezia.
troenterol 1996;22:267-71.
Am J Gastroenterol 1997;92:231-5.
30. Cipolletta L, Bianco MA, Rotondano G, et al. Outpatient management for
8. Laine L, Shah A. Randomized trial of urgent vs. elective colonoscopy in
low-risk nonvariceal upper GI bleeding: a randomized controlled trial.
patients hospitalized with lower GI bleeding. Am J Gastroenterol 2010;
Gastrointest Endosc 2002;55:1-5.
105:2636-41.
31. Gisbert JP, Legido J, Castel I, et al. Risk assessment and outpatient man-
9. Adamopoulos AB, Baibas NM, Efstathiou SP, et al. Differentiation be-
agement in bleeding peptic ulcer. J Clin Gastroenterol 2006;40:129-34.
tween patients with acute upper gastrointestinal bleeding who need
32. Lee JG, Turnipseed S, Romano PS, et al. Endoscopy-based triage signif-
early urgent upper gastrointestinal endoscopy and those who do not: a
icantly reduces hospitalization rates and costs of treating upper GI
prospective study. Eur J Gastroenterol Hepatol 2003;15:381-7. bleeding: a randomized controlled trial. Gastrointest Endosc 1999;50:
10. Aljebreen AM, Fallone CA, Barkun AN. Nasogastric aspirate predicts 755-61.
high-risk endoscopic lesions in patients with acute upper-GI bleeding. 33. Longstreth GF, Feitelberg SP. Successful outpatient management of
Gastrointest Endosc 2004;59:172-8. acute upper gastrointestinal hemorrhage: use of practice guidelines in a
11. Stoltzing H, Ohmann C, Krick M, et al. Diagnostic emergency endoscopy large patient series. Gastrointest Endosc 1998;47:219-22.
in upper gastrointestinal bleeding—Do we have any decision aids for 34. Conway JD, Adler DG, Diehl DL, et al. Endoscopic hemostatic devices.
patient selection? Hepatogastroenterology 1991;38:224-7. Gastrointest Endosc 2009;69:987-96.
12. Chen IC, Hung MS, Chiu TF, et al. Risk scoring systems to predict need for 35. Eisen GM, Baron TH, Dominitz JA, et al. Complications of upper GI en-
clinical intervention for patients with nonvariceal upper gastrointesti- doscopy. Gastrointest Endosc 2002;55:784-93.
nal tract bleeding. Am J Emerg Med 2007;25:774-9. 36. Lin HJ, Hsieh YH, Tseng GY, et al. A prospective, randomized trial of
13. Srirajaskanthan R, Conn R, Bulwer C, et al. The Glasgow Blatchford scor- large-versus small-volume endoscopic injection of epinephrine for pep-
ing system enables accurate risk stratification of patients with upper tic ulcer bleeding. Gastrointest Endosc 2002;55:615-9.
gastrointestinal haemorrhage. Int J Clin Pract 2010;64:868-874. 37. Church NI, Dallal HJ, Masson J, et al. A randomized trial comparing
14. Stanley AJ, Ashley D, Dalton HR, et al. Outpatient management of pa- heater probe plus thrombin with heater probe plus placebo for bleed-
tients with low-risk upper-gastrointestinal haemorrhage: multicentre ing peptic ulcer. Gastroenterology 2003;125:396-403.
validation and prospective evaluation. Lancet 2009;373:42-7. 38. Ginsberg GG, Barkun AN, Bosco JJ, et al. The argon plasma coagulator:
15. Blatchford O, Murray WR, Blatchford M. A risk score to predict need for February 2002. Gastrointest Endosc 2002;55:807-10.
treatment for upper-gastrointestinal haemorrhage. Lancet 2000;356: 39. Raju GS, Gajula L. Endoclips for GI endoscopy. Gastrointest Endosc 2004;
1318-21. 59:267-9.
16. Rockall TA, Logan RFA, Devlin HB, et al. Variation in outcome after acute 40. Alis H, Oner OZ, Kalayci MU, et al. Is endoscopic band ligation superior to
upper gastrointestinal haemorrhage. Lancet 1995;346:346-50. injection therapy for Dieulafoy lesion? Surg Endosc 2009;23:1465-9.
17. Masaoka T, Suzuki H, Hori S, et al. Blatchford scoring system is a useful 41. Rockall TA, Logan RF, Devlin HB, et al. Incidence of and mortality from
scoring system for detecting patients with upper gastrointestinal acute upper gastrointestinal haemorrhage in the United Kingdom.
bleeding who do not need endoscopic intervention. J Gastroenterol Steering Committee and members of the National Audit of Acute Upper
Hepatol 2007;22:1404-8. Gastrointestinal Haemorrhage. BMJ 1995;311:222-6.
18. Anderson MA, Ben-Menachem T, Gan SI, et al. Management of anti- 42. Boonpongmanee S, Fleischer DE, Pezzullo JC, et al. The frequency of
thrombotic agents for endoscopic procedures. Gastrointest Endosc peptic ulcer as a cause of upper-GI bleeding is exaggerated. Gastroin-
2009;70:1060-70. test Endosc 2004;59:788-94.
43. Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an 66. Gadenstatter M, Wetscher G, Crookes PF, et al. Dieulafoy’s disease of the
evidence-based approach based on meta-analyses of randomized con- large and small bowel. J Clin Gastroenterol 1998;27:169-72.
trolled trials. Clin Gastroenterol Hepatol 2009;7:33-47. 67. Norton ID, Petersen BT, Sorbi D, et al. Management and long-term prog-
44. Vergara M, Calvet X, Gisbert JP. Epinephrine injection versus epineph- nosis of Dieulafoy lesion. Gastrointest Endosc 1999;50:762-7.
rine injection and a second endoscopic method in high risk bleeding 68. Lemos DW, Raffetto JD, Moore TC, et al. Primary aortoduodenal fistula: a
ulcers. Cochrane Database Syst Rev 2007;CD005584. case report and review of the literature. J Vasc Surg 2003;37:686-9.
45. Barkun AN, Martel M, Toubouti Y, et al. Endoscopic hemostasis in peptic 69. Nagy SW, Marshall JB. Aortoenteric fistulas: recognizing a potentially
ulcer bleeding for patients with high-risk lesions: a series of meta- catastrophic cause of gastrointestinal bleeding. Postgrad Med 1993;93:
analyses. Gastrointest Endosc 2009;69:786-99. 211-6, 219.
46. Marmo R, Rotondano G, Piscopo R, et al. Dual therapy versus mono- 70. Iino M, Kuribayashi S, Imakita S, et al. Sensitivity and specificity of CT in
therapy in the endoscopic treatment of high-risk bleeding ulcers: a the diagnosis of inflammatory abdominal aortic aneurysms. J Comput
meta-analysis of controlled trials. Am J Gastroenterol 2007;102:279-89. Assist Tomogr 2002;26:1006-12.
47. Jensen DM. Spots and clots—leave them or treat them? Why and how to 71. Voorhoeve R, Moll FL, de Letter JA, et al. Primary aortoenteric fistula:
treat. Can J Gastroenterol 1999;13:413-5. report of eight new cases and review of the literature. Ann Vasc Surg
48. Johnston JH. Endoscopic risk factors for bleeding peptic ulcer. Gastro- 1996;10:40-8.
intest Endosc 1990;36:S16-S20. 72. Saers SJ, Scheltinga MR. Primary aortoenteric fistula. Br J Surg 2005;92:
49. Savides TJ, Jensen DM. Therapeutic endoscopy for nonvariceal gastro- 143-52.
intestinal bleeding. Gastroenterol Clin N Am 2000;29:465-87. 73. Savides TJ, Jensen DM, Cohen J, et al. Severe upper gastrointestinal
50. Kovacs TO, Jensen DM. Endoscopic treatment of ulcer bleeding. Curr tumor bleeding: endoscopic findings, treatment, and outcome. Endos-
Treat Options Gastroenterol 2007;10:143-8. copy 1996;28:244-8.
51. Sung JJ, Chan FK, Lau JY, et al. The effect of endoscopic therapy in pa- 74. Loftus EV, Alexander GL, Ahlquist DA, et al. Endoscopic treatment of
tients receiving omeprazole for bleeding ulcers with nonbleeding visi- major bleeding from advanced gastroduodenal malignant lesions.
Mayo Clin Proc 1994;69:736-40.
ble vessels or adherent clots: a randomized comparison. Ann Intern Med
75. Mathus-Vliegen EM, Tytgat GN. Analysis of failures and complications of
2003;139:237-43.
neodymium: YAG laser photocoagulation in gastrointestinal tract tu-
52. Banerjee S, Cash BD, Dominitz JA, et al. The role of endoscopy in the
mors: a retrospective survey of 18 years’ experience. Endoscopy 1990;
management of patients with peptic ulcer disease. Gastrointest Endosc
22:17-23.
2010;71:663-8.
76. Suzuki H, Miho O, Watanabe Y, et al. Endoscopic laser therapy in the
53. Laine L, McQuaid KR. Endoscopic therapy for bleeding ulcers: an
curative and palliative treatment of upper gastrointestinal cancer.
evidence-based approach based on meta-analyses of randomized con-
World J Surg 1989;13:158-64.
trolled trials. Clin Gastroenterol Hepatol 2009;7:33-47.
77. Saeed ZA, Cole RA, Ramirez FC, et al. Endoscopic retreatment after suc-
54. Lau JY, Sung JJ, Lee KK, et al. Effect of intravenous omeprazole on recur-
cessful initial hemostasis prevents ulcer rebleeding: a prospective ran-
rent bleeding after endoscopic treatment of bleeding peptic ulcers.
domized trial. Endoscopy 1996;28:288-94.
N Engl J Med 2000;343:310-6.
78. Lau JY, Sung JJ, Lam YH, et al. Endoscopic retreatment compared with
55. Leontiadis GI, Sharma VK, Howden CW. Proton pump inhibitor treat- surgery in patients with recurrent bleeding after initial endoscopic con-
ment for acute peptic ulcer bleeding. Cochrane Database Syst Rev 2006; trol of bleeding ulcers. N Engl J Med 1999;340:751-6.
CD002094. 79. Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recom-
56. Sung JJ, Barkun A, Kuipers EJ, et al. Intravenous esomeprazole for pre- mendations on the management of patients with nonvariceal upper
vention of recurrent peptic ulcer bleeding: a randomized trial. Ann In- gastrointestinal bleeding. Ann Intern Med 2010;152:101-13.
tern Med 2009;150:455-64.
57. Rockall TA, Logan RF, Devlin HB, et al. Incidence of and mortality from
acute upper gastrointestinal haemorrhage in the United Kingdom. Joo Ha Hwang, MD, PhD
Steering Committee and members of the National Audit of Acute Upper Deborah A. Fisher, MD, MHS
Gastrointestinal Haemorrhage. BMJ 1995;311:222-6. Tamir Ben-Menachem, MD
58. Bharucha AE, Gostout CJ, Balm RK. Clinical and endoscopic risk factors in Vinay Chandrasekhara, MD
the Mallory-Weiss syndrome. Am J Gastroenterol 1997;92:805-8. Krishnavel Chathadi, MD
59. Gunay K, Cabioglu N, Barbaros U, et al. Endoscopic ligation for patients G. Anton Decker, MD
with active bleeding Mallory-Weiss tears. Surg Endosc 2001;15:1305-7. Dayna S. Early, MD
60. Huang SP, Wang HP, Lee YC, et al. Endoscopic hemoclip placement and John A. Evans, MD
epinephrine injection for Mallory-Weiss syndrome with active bleeding. Robert D. Fanelli, MD, SAGES Representative
Kimberly Foley, RN, BSN, CGRN, SGNA Representative
Gastrointest Endosc 2002;55:842-6.
Norio Fukami, MD
61. Morales P, Baum AE. Therapeutic alternatives for the Mallory-Weiss Tear.
Rajeev Jain, MD
Curr Treat Options Gastroenterol 2003;6:75-83.
Terry L. Jue, MD
62. Myung SJ, Kim HR, Moon YS. Severe Mallory-Weiss tear after endoscopy
Kahlid M. Khan, MD
treated by endoscopic band ligation. Gastrointest Endosc 2000;52:99-
Jenifer Lightdale, MD, MPH
101.
Phyllis M. Malpas, RN, SGNA Representative
63. Yamaguchi Y, Yamato T, Katsumi N, et al. Endoscopic hemoclipping for
John T. Maple, DO
upper GI bleeding due to Mallory-Weiss syndrome. Gastrointest Endosc
Shabana Pasha, MD
2001;53:427-30.
John Saltzman, MD
64. Wong RM, Ota S, Katoh A, et al. Endoscopic ligation for non-esophageal
Ravi Sharaf, MD
variceal upper gastrointestinal hemorrhage. Endoscopy 1998;30:774-7. Amandeep K. Shergill, MD
65. Sebastian S, O’Morain CA, Buckley MJ. Review article: current therapeu- Jason A. Dominitz, MD, MHS, Prior Chair
tic options for gastric antral vascular ectasia. Aliment Pharmacol Ther Brooks D. Cash, MD, Chair
2003;18:157-65.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
Dhiraj Tripathi,1 Adrian J Stanley,2 Peter C Hayes,3 David Patch,4 Charles Millson,5
Homoyon Mehrzad,6 Andrew Austin,7 James W Ferguson,1 Simon P Olliff,6
Mark Hudson,8 John M Christie9
1
Liver Unit, University Hospitals ABSTRACT maximum dose of 240 mg (level 1a,
Birmingham NHS Foundation These updated guidelines on the management of grade A).
Trust, Birmingham, UK
2
GI Unit, Glasgow Royal
variceal haemorrhage have been commissioned by the 1.3.3. Carvedilol: 6.25 mg once daily to
Infirmary, Glasgow, UK Clinical Services and Standards Committee (CSSC) of the increase to maintenance of 12.5 mg
3
Liver Unit, Royal Infirmary of British Society of Gastroenterology (BSG) under the after a week if tolerated or once HR
Edinburgh, Edinburgh, UK auspices of the liver section of the BSG. The original of <50–55 bpm is reached (level 1a,
4
The Royal Free Sheila Sherlock
guidelines which this document supersedes were written grade A).
Liver Centre, Royal Free
Hospital and University College in 2000 and have undergone extensive revision by 13 1.3.4. It is suggested that NSBB are discon-
London, London, UK members of the Guidelines Development Group (GDG). tinued at the time of spontaneous
5
Gastrointestinal and Liver The GDG comprises elected members of the BSG liver bacterial peritonitis, renal impair-
Services, York Teaching section, representation from British Association for the ment and hypotension (level 2b,
Hospitals NHS Foundation
Trust, York, UK Study of the Liver (BASL) and Liver QuEST, a nursing grade B).
6
Department of Interventional representative and a patient representative. The quality 1.4. In cases of contraindications or intolerance
Radiology, University Hospitals of evidence and grading of recommendations was to NSBB, we recommend variceal band
Birmingham NHS Foundation appraised using the AGREE II tool. ligation (level 1a, grade A).
Trust, Birmingham, UK
7 The nature of variceal haemorrhage in cirrhotic 2. Who should have surveillance for variceal
Department of
Gastroenterology, Derby patients with its complex range of complications makes bleeding?
Hospitals NHS Foundation rigid guidelines inappropriate. These guidelines deal 2.1. We recommend all patients with cirrhosis
Trust, Derby, UK
8
specifically with the management of varices in patients should be endoscoped at the time of diag-
Liver Unit, Freeman Hospital, with cirrhosis under the following subheadings: nosis (level 1a, grade A). There is no indica-
Newcastle upon Tyne, UK
9
Department of
(1) primary prophylaxis; (2) acute variceal haemorrhage; tion to repeat endoscopy in patients
Gastroenterology, Royal Devon (3) secondary prophylaxis of variceal haemorrhage; and receiving NSBB.
and Exeter Hospital, Devon, (4) gastric varices. They are not designed to deal with 3. How often should cirrhotic patients be
UK (1) the management of the underlying liver disease; endoscoped?
(2) the management of variceal haemorrhage in children; 3.1. If at the time of first endoscopy no varices
Correspondence to
Dr Dhiraj Tripathi, Liver Unit, or (3) variceal haemorrhage from other aetiological are seen, we suggest that patients
Queen Elizabeth Hospital, conditions. with cirrhosis should be endoscoped at
Edgbaston, Birmingham B15 2–3-year intervals (level 2a, grade B).
2TH, UK; Dhiraj.Tripathi@uhb. 3.2. If grade I varices are diagnosed, we suggest
nhs.uk, d.tripathi@bham.ac.uk
Summary of all recommendations that patients should be endoscoped at yearly
Received 28 January 2015 Recommendations: primary prophylaxis of variceal intervals (level 2a, grade B).
Revised 11 March 2015 haemorrhage in cirrhosis (Figure 2) 3.3. If there is clear evidence of disease progres-
Accepted 17 March 2015 1. What is the best method for primary prophylaxis? sion we suggest that the intervals can be
Published Online First
17 April 2015 1.1. We recommend non-cardioselective β modified by the clinician. Endoscopy should
blockers (NSBB) or variceal band ligation also be offered at time of decompensation
(VBL). We suggest pharmacological treat- (level 2a, grade B).
ment with propranolol as first line. VBL is 4. Which patients with cirrhosis should have
offered if there are contraindications to primary prophylaxis?
NSBB. The choice of VBL or NSBB should 4.1. If grade I varices and red signs or grade 2–
also take into account patient choice (level 3 varices are diagnosed, we recommend
1a, grade A). that patients have primary prophylaxis irre-
Open Access
Scan to access more 1.2. We suggest carvedilol or nadolol as alterna- spective of the severity of the liver disease
free content
tives to propranolol (level 1b, grade A). (level 1a, grade A).
1.3. Dose: 5. Treatments not recommended:
1.3.1. Propranolol: 40 mg twice daily. Dose 5.1. Proton pump inhibitors are not recom-
titrated to maximum tolerated or once mended unless otherwise required for
heart rate (HR) of 50–55 bpm is peptic ulcer disease (level 1b, grade B).
reached to a maximum dose of 5.2. Isosorbide mononitrate monotherapy is not
320 mg (level 1a, grade A). recommended as primary prophylaxis (level
To cite: Tripathi D, 1.3.2. Nadolol: 40 mg daily dose. Dose 1b, grade A). There is insufficient evidence
Stanley AJ, Hayes PC, et al. titrated to maximum tolerated or to recommend isosorbide mononitrate in
Gut 2015;64:1680–1704. once HR of 50–55 bpm is reached a combination with NSBB (level 1b, grade A).
1680 Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262
Guidelines
5.3. Shunt surgery or transjugular intrahepatic portosystemic 2. Suggestions for timing of upper gastrointestinal endoscopy:
stent shunt (TIPSS) is not recommended as primary 2.1. Offer endoscopy to unstable patients with severe acute
prophylaxis (level 1a, grade A). upper gastrointestinal bleeding immediately after resusci-
5.4. Sclerotherapy is not recommended as primary prophy- tation (level 5, grade A).
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
laxis (level 1a, grade A). 2.2. Offer endoscopy within 24 h of admission to all other
6. Areas requiring further study: patients with upper gastrointestinal bleeding (level 2b,
6.1. Role of NSBB in patients without varices, with focus on grade A).
carvedilol. 2.3. Units seeing more than 330 cases a year should offer
6.2. Role of NSBB in patients with small varices, with focus daily endoscopy lists. Units seeing fewer than 330 cases
on carvedilol. a year should arrange their service according to local cir-
6.3. Comparison of carvedilol versus propranolol in primary cumstances (level 5, grade D).
prophylaxis. 3. Control of bleeding:
6.4. Identification of, and trials assessing, new drugs for 3.1. Antibiotics are recommended for all patients with sus-
primary prophylaxis such as statins. pected or confirmed variceal bleeding (level 1a,
7. Quality indicator: grade A).
7.1. Percentage of patients at diagnosis of cirrhosis who have 3.2. In all patients, vasoconstrictors such as terlipressin or
had an endoscopy to screen for varices (level 1a, somatostatin are recommended and should be started as
grade A). soon variceal bleeding is suspected and continued until
Numerator; patients diagnosed with cirrhosis who have haemostasis is achieved or for up to 5 days. Octreotide
had an endoscopy either before or after diagnosis within (unlicensed) is suggested if terlipressin or somatostatin
6 months. are unavailable (level 1a, grade A).
Denominator; patients newly diagnosed with cirrhosis. 3.3. Variceal band ligation is recommended as the preferred
7.2. Percentage of patients receiving primary prophylaxis endoscopic method (level 1a, grade A).
among those newly diagnosed with grade I varices and 3.4. After satisfactory haemostasis with the methods above,
red signs or grade 2–3 varices. and depending on local resources, early covered TIPSS
Numerator; patients who have grade 1 varices with red (<72 h after index variceal bleed) can be considered in
signs or grade 2–3 varices receiving primary prophylaxis. selected patients with Child’s B cirrhosis and active
Denominator; patients diagnosed with cirrhosis who bleeding or Child’s C cirrhosis with Child’s score <14
have grade I varices with red signs or grade 2–3 varices. (level 1b, grade B).
3.5. Proton pump inhibitors are not recommended unless
Recommendations: control of active variceal haemorrhage in otherwise required for peptic ulcer disease (level 1b,
cirrhosis (Figure 3) grade B).
1. Suggestions for resuscitation and initial management 4. Failure to control active bleeding:
1.1. Units offering an emergency acute upper gastrointestinal 4.1. If bleeding is difficult to control, a Sengstaken–
bleeding service should have expertise in VBL, balloon Blakemore tube should be inserted until further
tamponade and management of gastric variceal bleeding endoscopic treatment, TIPSS or surgery is performed
(level 5, grade D). depending on local resources and expertise (level 1b,
1.2. Transfuse patients with massive bleeding with blood, pla- grade B).
telets and clotting factors in line with local protocols for 4.2. Specialist help should be sought at this time and transfer
managing massive bleeding (level 5, grade D). to a specialist centre should be considered. Units that do
1.3. Base decisions on blood transfusion on the full clinical not offer a TIPSS service should identify a specialist
picture, recognising that overtransfusion may be as dam- centre which offers a 24 h emergency TIPSS service and
aging as undertransfusion. A restrictive transfusion policy have appropriate arrangements for safe transfer of
aiming for a haemoglobin of 70–80 g/L is suggested in patients in place (level 2a, grade B).
haemodynamically stable patients (level 1b, grade B). 5. Areas requiring further study:
1.4. Do not offer platelet transfusion to patients who are not 5.1. The efficacy of restrictive blood transfusion in variceal
actively bleeding and are haemodynamically stable (level haemorrhage.
5, grade D). 5.2. The role of blood products in variceal haemorrhage.
1.5. Offer platelet transfusion to patients who are actively 5.3. The utility of early TIPSS (<72 h) in acute variceal
bleeding and have a platelet count of <50×109/L (level haemorrhage.
5, grade D). 5.4. The role of removable oesophageal stents in acute vari-
1.6. Offer fresh frozen plasma to patients who have either: ceal haemorrhage.
▸ a fibrinogen level of <1 g/L (level 5, grade D), or 5.5. The role of haemostatic powders in acute variceal
▸ a prothrombin time (international normalised ratio) haemorrhage.
or activated partial thromboplastin time >1.5 times 5.6. The role of proton pump inhibitors in variceal
normal (level 5, grade D). haemorrhage.
1.7. Offer prothrombin complex concentrate to patients who 6. Quality indicators
are taking warfarin and actively bleeding (level 5, grade D). 6.1. Antibiotic administration in acute variceal bleeding
1.8. Treat patients who are taking warfarin and whose upper within 1 day either before or after the procedure
gastrointestinal bleeding has stopped in line with local (level 1a, grade A).
warfarin protocols (level 5, grade D). Numerator; patients with an acute variceal bleed who
1.9. There is insufficient evidence for the use of recombinant have received antibiotics within 1 day either before or
factor VIIa in acute variceal haemorrhage (level 1b, after the procedure.
grade B). Denominator; patients with an acute variceal bleed.
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1681
Guidelines
6.2. Endoscopy performed within 24 h of presentation of an Recommendations: management of active haemorrhage from
acute variceal bleed (level 2b, grade A). gastric varices (Figure 3)
Numerator; patients with an acute variceal bleed who 1. What is the optimal management of bleeding gastro-
have received endoscopy within 24 h of presentation. oesophageal varices?
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
Denominator; patients with an acute variceal bleed. 1.1. Gastro-oesophageal varices (GOV)-1: treat as for
oesophageal varices (level 2b, grade B).
Recommendations: secondary prophylaxis of variceal haemor- 1.2. GOV-2 and isolated gastric varices (IGV):
rhage in cirrhosis (figure 3) 1.2.1. We recommend initial endoscopic therapy with
1. Should VBL be used in combination with NSBB? cyanoacrylate injection (level 1a, grade A).
1.1. NSBB ( propranolol or nadolol)+VBL combination 1.2.2. Thrombin may also be considered (level 4, grade C).
therapy are recommended as secondary prophylaxis 1.3. TIPSS can be considered, depending on local resources
(level 1a, grade A). and clinical judgement (level 3a, grade B).
1.2. NSBB or VBL monotherapy are suggested as alternative 2. If control of bleeding fails:
options taking into account patient preference and clin- 2.1. Balloon tamponade is suggested for GOV and IGV-1
ical judgement (level 1a, grade B). until definitive treatment is undertaken (level 2b, grade B).
1.3. Carvedilol is suggested as an alternative to propranolol 2.2. Salvage TIPSS is suggested as the first-line definite treat-
and nadolol (level 1b, grade B). ment, where feasible (level 3a, grade B).
1.4. If NSBB alone are used, there is no need to undertake 2.3. Balloon-occluded retrograde transvenous obliteration
further endoscopy unless clinically indicated (level 1a, (B-RTO) or surgical shunting can be considered if TIPSS
grade A). is not possible (eg, portal vein thrombosis present) and
1.5. We recommend that VBL alone is used to eradicate depending on local resources (level 3a, grade B).
varices if there are contraindications or intolerance to 3. What are the therapeutic options for prevention of rebleed-
combined use with NSBB (level 1a, grade A). ing from gastric varices?
2. What is the optimal protocol for VBL? 3.1. We recommend that patients with GOV-1 are entered into
2.1. It is suggested that varices are banded at 2–4-weekly a VBL surveillance programme (level 2b, grade B).
intervals until eradication (level 1b, grade B). 3.2. We recommend endoscopic surveillance with cyano-
2.2. After successful eradication of the varices, patients acrylate injection as needed for GOV-2 and IGV (level
should be endoscoped at 3 months, then 6 monthly 2b, grade B). Note the optimum endoscopic follow-up
thereafter. Any recurrent varices should be treated with strategy remains unclear. Thrombin can also be consid-
further VBL until eradication (level 1b, grade B). ered (level 4, grade C).
2.3. Proton pump inhibitors are not recommended unless 3.3. NSBB can be considered in certain circumstances after
otherwise required for peptic disease (level 1b, grade B). taking into account the patient’s preferences and clinical
3. When is TIPSS indicated? judgement (level 1b, grade B).
3.1. We suggest that TIPSS is used for patients who rebleed 3.4. We suggest TIPSS if patients rebleed despite cyanoacryl-
despite combined VBL and NSBB therapy (or when ate injection. TIPSS can also be considered in other
monotherapy with VBL or NSBB is used owing to selected patients (eg, those with large or multiple gastric
intolerance or contraindications to combination therapy), varices) (level 1b, grade B).
and in selected cases owing to patient choice. PTFE- 3.5. Shunt surgery may be used in selected patients with
covered stents are recommended (level 1a, grade A). well-compensated cirrhosis and depending on local
3.2. Where TIPSS is not feasible in Child’s A and B patients, resources (level 3c, grade B).
we suggest shunt surgery can be used where local 3.6. Splenectomy or splenic artery embolisation should be
expertise and resources allow (level 1b, grade B). considered in all patients where there is splenic vein
4. Areas requiring further study: thrombosis or left-sided portal hypertension (level 4,
4.1. Combination of VBL and carvedilol (or other NSBB) grade C).
versus carvedilol as monotherapy. 4. Is there a role for primary prophylaxis of gastric variceal
4.2. Comparison of carvedilol with propranolol in secondary bleeding?
prophylaxis. 4.1. NSBB (level 2a, grade B) can be considered in selected
4.3. Optimum time interval between VBL sessions. high-risk patients with large GOV-2 after taking into
4.4. Strategy of VBL or NSBB discontinuation after variceal account the patient’s preferences and clinical judgement.
eradication during combination therapy with VBL 4.2. Cyanoacrylate injection is not recommended outside
+NSBB. clinical trials (level 2a, grade A).
4.5. Strategy of VBL add-on therapy to failure of NSBB 5. Areas requiring further study:
monotherapy. 5.1. Role of thrombin in gastric varices, comparing this with
4.6. Strategy of NSBB add-on therapy to failure of VBL tissue adhesives in both acute gastric variceal bleeding
monotherapy. and secondary prophylaxis.
4.7. Role of early TIPSS in secondary prophylaxis. 5.2. Role of TIPSS in acute gastric variceal bleeding and sec-
4.8. Role of statins in secondary prophylaxis. ondary prophylaxis.
5. Quality indicator: 5.3. Role of haemostatic powders in controlling refractory
5.1. Institution of secondary prophylaxis after acute variceal active gastric variceal bleeding.
bleeding (level 1a, grade A) 5.4. Role of NSBB in the prevention of rebleeding from
Numerator; patients with an acute variceal bleed who gastric varices.
have received either NSBB or banding or both within 5.5. Role of B-RTO as monotherapy or in combination with
4 weeks of the index bleed. endoscopic injection of tissue adhesives in prevention of
Denominator; patients with an acute variceal bleed. bleeding from gastric varices.
1682 Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
econferences during that time. The guidelines rely considerably
INTRODUCTION on consensus statements published by the Baveno V Consensus
The guidelines refer closely to the Baveno V consensus state- and NICE.1 2 The style of graded recommendations is deter-
ment published in 20101 and the 2012 NICE Guidelines on mined by the level of supporting evidence (graded level 1 to 5)
Acute Upper GI bleeding (CG141).2 These documents are as described by the Oxford Centre For Evidence Based
widely used and offer useful evidence-based guidance. However, Medicine4 (table 1) and is as follows:
we feel that owing to significant recent advances, further addi- A: consistent level 1 studies;
tions and refinements to the published guidance, with particular B: consistent level 2 or 3 studies or extrapolations from level 1
focus on resource implications, service development and the studies;
patient pathway, are necessary. The previously mentioned docu- C: level 4 studies or extrapolations from level 2 or 3 studies;
ments1 2 do not cover all the recent advances—in particular, in D: level 5 evidence or troublingly inconsistent or inconclusive
the field of acute variceal bleeding and the role of transjugular studies of any level.
intrahepatic portosystemic stent shunt (TIPSS). There have also Areas of disagreement about the recommendation grade were
been developments and better insights into drug treatment for subjected to discussion and, if necessary, voting by members of
prevention of varices and variceal bleeding—in particular, the the guidelines group. Where possible, the health benefits, side
role of non-cardioselective β blockers (NSBB). effects and risks of recommendations have been discussed. The
guidelines were subject to peer review after submission for con-
Guideline development sideration of publication in Gut.
These guidelines were drafted after discussions within the liver
section of the British Society of Gastroenterology (BSG) and Clarity and presentation
acceptance of the proposal by the Clinical Services and Recommendations are intended to be specific to particular situa-
Standards Committee (CSSC). There followed division of sec- tions and patient groups; where necessary, different options are
tions to be researched by designated authors and an exhaustive listed. Key recommendations are linked to discussion threads on
literature review. The Baveno V consensus and NICE guidelines a discussion forum hosted on the BSG website.
were closely followed and guideline quality was assessed using
the AGREE tool3 (section ‘Assessing the quality of guidelines: Applicability
the AGREE II instrument’). Where necessary, we have discussed organisational changes that
A preliminary guideline document was drafted by the authors may be needed in order to apply recommendations. We have
following discussion and, where necessary, voting by members attempted to identify key criteria for monitoring and audit
of the Guidelines Development Group (GDG). The draft guide- purposes.
lines were submitted for review by CSSG, then BSG council
members. Finally, full peer review was undertaken by reviewers Editorial independence and conflict of interest
selected by the editor of Gut. Guideline group members have declared any conflicts of
Attempts were made to preserve the format of the original interest.
guidelines, with additional sections relating to service develop-
ment, the patient pathway and pre-primary prophylaxis. The
Scheduled review of guidelines
section on the management of acute variceal bleeding has been
The proposed time for review of the guidelines is 5 years to
extensively rewritten to take into account recent important
take into account new developments. To ensure that there is a
developments in interventional radiology, drug treatment and
facility for feedback after publication, links to the BSG discus-
resuscitation.
sion forums corresponding to the particular section of these
guidelines are included with this document. This facility to
Assessing the quality of guidelines: the AGREE II instrument provide new evidence is provided to all BSG members. In
The AGREE II instrument is an accepted method for appraising accordance with the AGREE II tool the BSG forum will provide
clinical guidelines.3 Six domains are listed: feedback.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
1a SR (with homogeneity*) of SR (with homogeneity*) of inception SR (with homogeneity*) of level 1 SR (with homogeneity*) of
randomised controlled trial (RCT) cohort studies; CDR† validated in diagnostic studies; CDR† with 1b studies prospective cohort studies
different populations from different clinical centres
1b Individual RCT (with narrow CI) Individual inception cohort study with Validating‡ cohort study with good§ Prospective cohort study with good
≥80% follow-up; CDR† validated in a reference standards; or CDR† tested follow-up¶
single population within one clinical centre
1c All or none** All or none case series Absolute SpPins and SnNouts†† All or none case series
2a SR (with homogeneity*) of cohort SR (with homogeneity*) of either SR (with homogeneity*) of level >2 SR (with homogeneity*) of 2b and
studies retrospective cohort studies or untreated diagnostic studies better studies
control groups in RCTs
2b Individual cohort study (including Retrospective cohort study or follow-up Exploratory‡ cohort study with good§ Retrospective cohort study, or poor
low-quality RCT; eg, <80% of untreated control patients in an RCT; reference standards; CDR† after follow-up
follow-up) derivation of CDR† or validated on split derivation, or validated only on split
sample‡‡ only sample‡‡ or databases
2c ‘Outcomes’ research; ecological ‘Outcomes’ research Ecological studies
studies
3a SR (with homogeneity*) of case– SR (with homogeneity*) of 3b and better SR (with homogeneity*) of 3b and
control studies studies better studies
3b Individual case–control study Non-consecutive study; or without Non-consecutive cohort study or
consistently applied reference standards very limited population
4 Case series (and poor quality Case series (and poor quality prognostic Case–control study, poor or Case series or superseded reference
cohort and case-control studies§§) cohort studies¶¶) non-independent reference standard standards
5 Expert opinion without explicit Expert opinion without explicit critical Expert opinion without explicit critical Expert opinion without explicit
critical appraisal or based on appraisal or based on physiology, bench appraisal or based on physiology, bench critical appraisal or based on
physiology, bench research or ‘first research or ‘first principles’ research or ‘first principles’ physiology, bench research or ‘first
principles’ principles’
*Homogeneity means a systematic review (SR) that is free of worrisome variations (heterogeneity) in the directions and degrees of results between individual studies. Not all SRs with
statistically significant heterogeneity need be worrisome, and not all worrisome heterogeneity need be statistically significant.
†CDR, Clinical Decision Rule (algorithms or scoring systems which lead to a prognostic estimation or a diagnostic category).
‡Validating studies test the quality of a specific diagnostic test based on prior evidence. An exploratory study collects information and trawls the data (eg, using a regression analysis) to
find which factors are ‘significant’.
§Good reference standards are independent of the test, and applied blindly or objectively to all patients. Poor reference standards are haphazardly applied, but still independent of the
test. Use of a non-independent reference standard (where the ‘test’ is included in the ‘reference’, or where the ‘testing’ affects the ‘reference’) implies a level 4 study.
¶Good follow-up in a differential diagnosis study is >80%, with adequate time for alternative diagnoses to emerge (eg, 1–6 months acute, 1–5 years chronic).
**Met when all patients died before the treatment became available but some now survive while receiving it; or when some patients died before the treatment became available but
none now die while receiving it.
††An ‘absolute SpPin’: a diagnostic finding whose Specificity is so high that a Positive result rules in the diagnosis. An ‘Absolute SnNout’: a diagnostic finding whose Sensitivity is so
high that a Negative result rules out the diagnosis.
‡‡Split-sample validation is achieved by collecting all the information in a single tranche, then artificially dividing this into ‘derivation’ and ‘validation’ samples.
§§Poor quality cohort study: one that failed to clearly define comparison groups and/or failed to measure exposures and outcomes in the same (preferably blinded) objective way in
both exposed and non-exposed individuals and/or failed to identify or appropriately control known confounders and/or failed to carry out a sufficiently long and complete follow-up of
patients. Poor quality case–control study: one that failed to clearly define comparison groups and/or failed to measure exposures and outcomes in the same (preferably blinded)
objective way in both cases and controls and/or failed to identify or appropriately control known confounders.
¶¶Poor quality prognostic cohort study: one in which sampling was biased in favour of patients who already had the target outcome, or the measurement of outcomes was
accomplished in <80% of study patients, or outcomes were determined in an unblinded non-objective way, or there was no correction for confounding factors.
administered in only 27% of patients before endoscopy, and emergency endoscopy and use of prophylactic antibiotics and
administration of vasoactive drugs before endoscopy was only vasoactive drugs was better in tertiary centres, although this
slightly higher at 44%. Furthermore, only four patients (<1%) did not appear to affect survival.8 9
were referred for TIPSS, which may reflect the lack of access Acute variceal haemorrhage refractory to endoscopic and
to interventional radiology, and that the audit was conducted pharmacological treatments, where TIPSS is usually indicated,
before the trial of early TIPSS.6 The National Confidential must be managed with appropriate resources. TIPSS is an estab-
Enquiry into Patient Outcome and Death (NCEPOD) report lished interventional treatment for refractory or recurrent vari-
‘Measuring the units’ assessed clinical management before ceal haemorrhage. It remains a highly specialised procedure,
death of 594 patients with alcoholic liver disease over a requiring adequate training and experience. Knowledge of the
6-month period in the UK.7 Gastrointestinal bleeding was relevant equipment, anatomy and how to deal with any compli-
noted in 35% of cases, with approximately 50% having vari- cations is essential. It should therefore be performed in centres
ceal bleeding. Delays in endoscopy were noted in 10% of with adequate personnel, multidisciplinary support and equip-
cases, and several aspects of clinical and/or organisational care ment required to optimise management and minimise risks.10
were judged to be poor or unacceptable in 18% of patients Regional centres with easily accessible interventional radiology
presenting with GI bleeding. There were deficiencies noted in services are generally best equipped to perform this procedure.
the out-of-hours rotas for GI bleeding, with 27% of hospitals Setting up regional agreements and pathways to allow transfer of
not having a dedicated-out-of hours GI bleeding service. appropriate patients to hospitals that undertake TIPSS proce-
Studies from other countries have also reported deficiencies, dures is an important step. These pathways could also be used to
with delays in admission to hospital and administration of anti- provide emergency endoscopic management if necessary due to
biotics. Two observational studies showed that access to problems with out-of-hours endoscopic cover in smaller
hospitals. This model referred to as “spoke and wheel” or zero (time zero is the time of admission to the first hospital to
network model, is well established for other complex procedures which the patient is taken).
and helps to expedite and streamline the process. In the
NCEPOD report ‘Measuring the units’ just 15% of hospitals had Time frame of acute bleeding
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
on-site access to TIPSS, while 72% had access to TIPSS in other The acute bleeding episode is represented by an interval of
centres.7 120 h (5 days) from time zero. Evidence of any bleeding after
There have been significant efforts to address the need to 120 h is the first rebleeding episode.
improve the upper GI bleeding (UGIB) service. A toolkit was
produced in collaboration with BSG; Association of Upper Failure to control active bleeding
Gastrointestinal Surgeons (AUGIS); Royal Colleges of Failure to control active bleeding is defined as death or need to
Physicians, Radiology and Nursing; and Academy of Medical change treatment defined by one of the following criteria:16 17
Royal Colleges.11 The key nine service standards recommended 1. Fresh haematemesis or nasogastric aspiration of ≥100 mL of
by the document are detailed below: fresh blood ≥2 h after the start of a specific drug treatment
1. There will be a nominated individual with the authority to or therapeutic endoscopy.
ensure implementation by the contracted provider. 2. Development of hypovolaemic shock.
2. Contracted providers will ensure the minimum service is 3. 30 g/L drop in haemoglobin (9% drop of haematocrit)
adequately resourced. within any 24 h period if no transfusion is given. This time
3. All patients with suspected UGIB should be properly frame needs to be further validated.
assessed and their risk scored on presentation.
4. All patients should be resuscitated before therapeutic Variceal rebleeding
intervention. Variceal rebleeding is defined as the occurrence of a single
5. All high-risk patients with UGIB should be endoscoped episode of clinically significant rebleeding from portal hyperten-
within 24 h, preferably on a planned list in the first instance. sive sources from day 5. Clinically significant rebleeding is
6. For patients who require more urgent intervention either for defined as recurrent melaena or haematemesis in any of the fol-
endoscopy, interventional radiology or surgery formal 24/7 lowing settings:
arrangements must be available. 1. hospital admission;
7. The necessary team, meeting an agreed competency level, 2. blood transfusion;
should be available throughout the complete patient pathway. 3. 30 g/L drop in haemoglobin;
8. Each stage of the patient pathway should be carried out in 4. death within 6 weeks.
an area with ‘appropriate’ facilities, equipment and support
including staff experienced in the management of UGIB. Early mortality
9. All hospitals must collect a minimum dataset in order to Death within 6 weeks of the initial episode of bleeding.
measure service provision against auditable outcomes
(case-mix adjusted as appropriate). NATURAL HISTORY OF VARICES IN CIRRHOSIS
NICE recommendations for endoscopy provision are detailed Development of varices
in the section ‘Management of active variceal haemorrhage’ The rise in portal pressure is associated with the development
recommendations.2 The BSG has also produced a care bundle of collateral circulation, which allows the portal blood to be
for patients admitted with decompensated cirrhosis in light of diverted into the systemic circulation. These spontaneous shunts
the NCEPOD report with a check list method which includes occur (a) at the cardia through the intrinsic and extrinsic gastro-
gastrointestinal bleeding.12 oesophageal veins; (b) in the anal canal where the superior
Since the 2008 Darzi report, quality has become a priority haemorrhoidal vein belonging to the portal system anastomoses
for the NHS.13 With these guidelines there is real opportunity with the middle and inferior haemorrhoidal veins which belong
to introduce quality outcomes based on good clinical evidence. to the caval system; (c) in the falciform ligament of the liver
Furthermore by incorporating them into the liver accreditation through the para-umbilical veins, which are the remains of the
scheme, Liver Quest, one can improve and assure quality in umbilical circulation of the fetus; (d) in the abdominal wall and
liver services across the UK.14 Therefore a small number of the retroperitoneal tissues, from the liver to the diaphragm,
quality outcomes measures have been chosen and form part of veins in the lienorenal ligament, in the omentum and lumbar
the key recommendations.15 veins; and (e) blood diversion from the diaphragm, gastric, pan-
creatic, splenic, and adrenal veins, which may drain into the left
renal vein.
DEFINITIONS
Numerous lines of evidence suggest that varices develop and
It is important to define the terms that should be used in the
enlarge with time. Christensen et al18 followed up a cohort of
context of a variceal bleed. These are the Baveno V consensus
532 patients with cirrhosis and showed that the cumulative inci-
definitions.1
dence of patients with varices increased from 12% to 90% over
12 years. In a study involving 80 patients followed up for
Variceal haemorrhage 16 months, Cales and Pascal19 showed that 20% of patients
Variceal haemorrhage is defined as bleeding from an oesopha- who did not have varices developed new varices and 42% of
geal or gastric varix at the time of endoscopy or the presence of patients with small varices showed definite enlargement. Czaja
large oesophageal varices with blood in the stomach and no et al20 also showed that the prevalence of varices increased from
other recognisable cause of bleeding. An episode of bleeding is 8% to 13% over 5 years in a cohort of patients with chronic
clinically significant when there is a transfusion requirement for active hepatitis even though they were treated with prednisol-
2 units of blood or more within 24 h of the time zero, together one. Merli et al,21 in a study of 213 patients with cirrhosis with
with a systolic blood pressure of <100 mm Hg or a postural no or small varices, demonstrated that the annual progression of
change of >20 mm Hg and/or pulse rate >100 bpm at time varices was 12%. A recent database analysis by D’Amico et al22
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1685
Guidelines
using a competing risk model showed that the cumulative inci- Capsule endoscopy
dence of varices at 10 and 20 years was 44% and 53%, respect- Capsule endoscopy uses a 26 mm pill-shaped device which
ively, suggesting an overestimation in previous studies not using transmits video footage which is stored and later analysed.
a competing risk model. Patients are not sedated, but patient cooperation is essential.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
The main factors that appear to determine the development In a large study by de Franchis et al,30 capsule endoscopy was
of varices are continued hepatic injury, the degree of portosyste- compared with standard gastroscopy. The primary end point of
mic shunting, endoscopic appearances and portal pressure. 90% or greater concordance was not achieved. Lapalus et al,31
Evidence for the role of hepatic injury is derived from studies in in a prospective study of 120 patients, demonstrated similar
which varices were shown to regress with time. Baker et al23 fol- results with capsule endoscopy. Therefore, capsule endoscopy
lowed up a cohort of 115 patients with oesophageal varices and cannot be considered an alternative to standard endoscopy,
showed that varices had disappeared in nine patients, regressed although may have a role in patient who refuse gastroscopy.
in seven and remained unchanged in six. They concluded that
the disappearance and regression of varices might be related to Transient elastography
abstinence from alcohol. This observation was confirmed in a Transient elastography ((FibroScan, Echosens, Paris, France) uses
study by Dagradi24 who followed up a cohort of patients with the principles of ultrasound to derive tissue stiffness by measur-
alcoholic cirrhosis over 3 years and showed a reduction in vari- ing the speed of propagation of a low-frequency wave, which
ceal size in 12 of the 15 patients with alcoholic cirrhosis who then correlates with liver fibrosis. Vizzutti et al32 in a study of
stopped drinking and an enlargement in variceal size in 17 61 patients with hepatitis C showed a sensitivity for prediction
patients who continued to drink. On the other hand, Cales and of oesophageal varices of 90% using a threshold 17.6 kPa.
Pascal19 showed that regression of varices occurred in 16% of However, specificity was poor at 43%. A study of 298 patients
patients with alcoholic cirrhosis who continued to imbibe found the optimal cut-off point for the prediction of oesopha-
alcohol. This might be related to the development of large por- geal varices was 21.5 kPa (sensitivity 76% and specificity
tosystemic collaterals, which decompress the portal system and 78%).33 In one uncontrolled study the use of transient elasto-
reduce the risk of the development of large oesophageal varices. graphy was found to be as effective as HVPG at predicting
The degree of portosystemic shunting can be quantified by portal hypertension-related complications.34 Therefore, the role
measuring the diameter of portal veins and collaterals, and can of transient elastography in predicting varices is controversial
be significant in those with gastrorenal or splenorenal shunt- due to the lack of consistent results and controlled studies. This
ing.25 26 Others have shown that the presence of alcoholic cir- modality may be more useful for predicting decompensation in
rhosis, Child’s B or C cirrhosis and red whale signs on index patients with cirrhosis.
endoscopy predicted progression of varices.21 Groszmann
et al27 in a placebo-controlled randomised trial of timolol in Radiological and serum parameters.
213 cirrhotic patients without varices showed that a baseline A prospective study of 311 patients with chronic hepatitis C
hepatic venous pressure gradient (HVPG) of >10 mm Hg or a showed that a platelet-to-spleen size ratio with a threshold of
≥10% increase in HVPG during follow-up were both predictive 909 had positive and negative predictive values of 100% and
of the development of varices. 94%, respectively.35 These good results have not been repro-
duced by others as demonstrated in a meta-analysis.36
Diagnosis of gastro-oesophageal varices
Until recently, endoscopy has been used exclusively to diagnose Risk factors for first variceal bleeding
varices. Non-invasive methods of screening for varices include The factors that predispose to, and precipitate, variceal haemor-
capsule endoscopy, transient elastography and use of laboratory rhage are still not clear. The suggestion that oesophagitis may
and radiological findings. precipitate variceal haemorrhage has been discarded.37
Presently, the most important factors that have been held
Endoscopy responsible include (i) pressure within the varix, (ii) variceal
There is universal acceptance that endoscopy is the ‘gold stand- size, (iii) tension on the variceal wall and (iv) severity of the
ard’ for diagnosing gastro-oesophageal varices. The main limita- liver disease.
tions are intraobserver variability in the diagnosis of small or
grade I oesophageal varices (figure 1A–C). Recently, unsedated Portal pressure
nasal gastroscopy has been found to have similar accuracy to In most cases, portal pressure reflects intravariceal pressure38
conventional endoscopy and has the advantage of tolerability and a HVPG >10 mm Hg is necessary for the development of
and potential cost saving since it can be performed in the clinic oesophageal varices.27 There is no linear relationship between
setting in some institutions.28 29 However, there are no con- the severity of portal hypertension and the risk of variceal
trolled studies and banding of varices is not possible. haemorrhage, although HVPG >12 mm Hg is an accepted
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
The Child–Pugh score is the sum of severity scores for Child class,
showed that bleeding from varices did not occur if the portal
variceal size and red wale markings the variables shown below.
pressure gradient (PPG) could be reduced to <12 mm Hg.
Others have shown that a 20% reduction in portal pressure pro- Category 1 2 3
tects against further bleeding.42 These haemodynamic goals
Encephalopathy 0 I/II III/IV
have been accepted as the aim of pharmacological treatment of
Ascites Absent Mild-moderate Severe
portal hypertension. It is important to appreciate that gastric
Bilirubin (μmol/L) <34 34–51 >51
varices can bleed at pressures <12 mm Hg, and the influence of
Albumin (g/L) >35 28–35 <28
wall tension of the varix plays a greater role in the risk of bleed-
INR <1.3 1.3–1.5 >1.5
ing.43 A greater pressure reduction may be necessary to protect
against bleeding. This is further discussed in the section ‘Gastric Child–Pugh class A represents a score of ≤6, class B a score of 7–9, and class C, ≥10.
The MELD score is a formula that includes three laboratory-based variables reflecting
varices’. At present, measurement of portal pressure in guiding the severity of liver disease. It was originally used to predict the short-term mortality
pharmacological treatments is limited to clinical trials in the UK. after placement of a transjugular intrahepatic portosystemic stent-shunt for variceal
bleeding. Subsequently, it has been used in selecting candidates for liver
transplantation.
Variceal size MELD score: please use the online calculator https://www.esot.org/Elita/meldCalculator.
Variceal size is best assessed endoscopically (figure 1A–C). aspx.
INR, international normalised ratio.
Published results are variable owing to the lack of a definition dis-
tinguishing between large and small varices. Small (grade I)
varices tend to be narrow and flatten easily with air, whereas treatment, although the predictive value does not appear to
larger (grade 2 and 3) varices are usually broader and flatten with improve on the NIEC index and presence of red whale marking.54
difficulty, if at all. Numerous studies40 44 have shown that the
risk of variceal haemorrhage increases with the size of varices.45 Risk and mortality of first variceal bleed
Data describing the overall risk of bleeding from varices must be
Variceal wall and tension viewed with caution and have some pitfalls in interpretation.
Polio and Groszmann46 using an in vitro model showed that The natural history of patients who have varices that are diag-
rupture of varices was related to the tension on the variceal nosed as part of their baseline investiations is different from
wall. The tension depends on the radius of the varix. In this that of patients who have complications of liver disease such as
model, increasing the size of the varix and decreasing the thick- ascites and encephalopathy. Most studies do not comment on
ness of the variceal wall caused variceal rupture. Recently, endo- either the severity of liver disease or whether patients with alco-
scopic ultrasound and manometry have been used to estimate holic cirrhosis are continuing to drink. Both these factors have a
wall tension of varices.47 significant effect on the risk of variceal haemorrhage.
Endoscopic features such as ‘red spots’ and ‘whale’ markings Most studies report bleeding from varices in about 20–50%
were first described by Dagradi.24 They have been described as of patients with cirrhosis during the period of follow-up. Baker
being important in the prediction of variceal haemorrhage. et al23 reported variceal bleeding in 33 of 115 patients that they
These features represent changes in variceal wall structure and followed up for a mean of 3.3 years, with a mortality of 48%
tension associated with the development of microtelangiectasias from the first variceal haemorrhage. These data were confirmed
and reduced wall thickness. In a retrospective study by the by Christensen et al.18 About 70% of the episodes of bleeding
Japanese Research Society for Portal Hypertension, Beppu occur within 2 years of diagnosis. Recent studies demonstrate a
et al48 showed that 80% of patients who had blue varices or dramatic reduction in mortality following variceal bleeding of
cherry red spots bled from varices, suggesting that this was an 20% 6-week mortality55 and 15% in-hospital mortality,5 with
important predictor of variceal haemorrhage in cirrhosis. contributions from improved endoscopic, pharmacological and
radiological therapies, notably TIPSS. Intensive care treatment
Severity of liver disease and bleeding indices has also improved, with outcomes being particularly good for
Two independent groups prospectively assessed factors predict- those requiring minimal organ support.
ing first variceal haemorrhage in cirrhosis (table 2). The North Analysis of the non-active treatment arms in the primary
Italian Endoscopic Club (NIEC)49 reported their findings in prophylaxis trials comparing propranolol with placebo show
1988, followed in 1990 by data from the Japanese.50 Both these results similar to those of the primary prophylaxis shunt trials,
studies showed that the risk of bleeding was based on three with most episodes of bleeding occurring within the first 2 years
factors: severity of liver disease as measured by Child class, vari- of follow-up. In these studies the rate of first variceal haemor-
ceal size and red wale markings. The NIEC study showed a rhage ranged from 22% to 61%.56–60 This large difference in
wide range for the risk of bleeding of 6–76%, depending on the the rate of first bleed relates almost certainly to the number of
presence or absence of the different factors. Using the same vari- patients with severe liver disease included in the study (Pascal,
ables the NIEC index was simplified by de Franchis et al51 and Child C—46%, bleeding—61%; Italian Multicenter Project for
shown to correlate with the original index. Further studies Propranolol in Prevention of Bleeding (IMPP), Child C—6%,
showed that the HVPG and intravariceal pressure were also bleeding—32%; Conn, Child C—6%, bleeding—22%).
independent predictors of first variceal haemorrhage when ana- Mortality varied from 24% to 49% over 2 years (Pascal, mortal-
lysed in conjunction with the NIEC index.52 53 ity—49%; IMPP, mortality—24%; Conn, mortality—24%).
In summary, the most important factors that determine the risk of
variceal haemorrhage are the severity of liver disease, size of varices, Primary prophylaxis
and presence of red signs. Measurement of HVPG is a useful guide Since 30–50% of patients with portal hypertension will bleed
for selection of patients for treatment and their response to from varices and about 20% will die from the effects of the first
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1687
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
expected bleeding and death rates in the control group, the problems with the interpretation of this study because of the
minimum number of patients needed to detect a 50% reduction use of different procedures in each of the 22 centres. These
in bleeding would be 270, and 850 patients in each arm to results require confirmation.
detect the same reduction in mortality. A proposed algorithm
for surveillance and prophylaxis of varices is shown in figure 2.
Pharmacological treatment
At this time there is insufficient evidence to support treating
Non-cardioselective β blockers
patients without varices or ‘pre-primary prophylaxis’. A large
The mainstay of the pharmacological approach to the primary
randomised placebo-controlled trial of timolol in patients
prophylaxis of variceal haemorrhage has been NSBB.
without varices and portal hypertension defined as HVPG
Propranolol which has been shown to reduce the PPG, reduce
>6 mm Hg did not show any effect on the development of
azygos blood flow, and also variceal pressure. It achieves this by
varices or variceal bleeding.27 The role of drug treatment in pre-
causing splanchnic vasoconstriction and reducing cardiac
venting bleeding in patients with small varices is unclear. Three
output. There is no clear dose-related reduction in HVPG or
randomised placebo-controlled trials have studied this. Cales
correlation of HPVG reduction with reduction in heart rate.69
et al61 showed that propranolol in patients with small, or no,
Observational studies have shown that a 10–12% reduction in
varices resulted in greater development of varices. However,
HVPG after acute administration of propranolol was associated
patients without varices were included and there was significant
with reduced bleeding and hepatic decompensation.54 70
loss of patients to follow-up. The second trial showed that
However, HVPG monitoring is not routinely available in most
nadolol reduced variceal bleeding without survival benefit and
centres outside of larger institutions. A meta-analysis of nine
increased adverse events.62 Sarin et al63 did not show any effect
placebo-controlled randomised trials (964 patients) showed that
with propranolol, despite a significant effect on portal pressure.
the pooled risk difference for bleeding was −11% (95% CI
−21% to −1%), and for death was −9% (95% CI −18% to
Surgery −1%) in favour of propranolol.71
Portacaval shunts Nadolol exerts similar effects on portal haemodynamics,
Four trials of portacaval shunts have been published, which ran- although the effect on blood pressure may not be as pro-
domised a total of 302 patients64–67 either to prophylactic shunt nounced. Two placebo-controlled trials58 59 have shown
surgery or to non-active treatment. A meta-analysis of these reduced bleeding, although in one study this was only seen on
studies showed a significant benefit in the reduction of variceal per protocol analysis.59 There was no effect on overall survival.
bleeding (OR=0.31, 95% CI 0.17 to 0.56) but also a signifi- Carvedilol is a NSBB like propranolol, and a vasodilator due
cantly greater risk of hepatic encephalopathy (OR=2, 95% CI to α1 receptor blockade. The latter reduces portocollateral
1.2 to 3.1) and mortality (OR=1.6, 95% CI 1.02 to 2.57) in resistance, and by actions on hepatic stellate cells leads to a
patients treated with shunt surgery.68 At this time, there is no reduction in intrahepatic resistance. Haemodynamic studies
evidence for the use of TIPSS for primary prophylaxis.1 demonstrate a greater reduction in portal pressure with
carvedilol than with propranolol, although blood pressure is plus placebo failed to show any differences between the two
reduced.72 73 The optimum dose is 6.25–12.5 mg/day.74 Higher arms. Combination therapy is associated with more side effects.
doses are not more effective and are associated with more
adverse events—in particular, hypotension. Carvedilol at a dose Proton pump inhibitors
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
of 12.5 mg/day at current UK prices is considerably cheaper A placebo-controlled randomised trial reported reduced bleed-
than propranolol 40 mg twice a day and nadolol 80 mg/day ing and mortality with rabeprazole after eradication of varices.86
(monthly cost, £1.20, £5.62 and £5, respectively). Two RCTs of However, the study had a heterogeneous population with VBL
carvedilol versus variceal band ligation (VBL) in primary performed for both primary and secondary prophylaxis and
prophylaxis have been published.75 76 The first study75 showed small numbers (n=43), limiting the validity of the conclusions.
significantly reduced bleeding in the carvedilol arm (10% vs Furthermore, there was no arm comparing proton pump inhibi-
23%, relative hazard 0.41; 95% CI 0.19 to 0.96), with no effect tors with NSBB. The use of proton pump inhibitors in patients
on survival. The second trial by Shah et al76 did not show any with cirrhosis and ascites was associated with increased risk of
differences in bleeding or mortality. Compliance with VBL was spontaneous bacterial peritonitis in a large retrospective study.87
better in the latter trial, and unlike the first trial, there were sig- This was not confirmed in a larger prospective non-randomised
nificantly more patients with viral hepatitis than alcoholic cir- study.88 However, a recent prospective observational study has
rhosis. A further study74 assessed the effect of carvedilol in shown proton pump use to be associated with increased mortal-
patients who were haemodynamic non-responders to propran- ity in cirrhosis.89 Proton pump inhibitors are also associated
olol, where haemodynamic response was defined as HVPG with increased risk of Clostridium difficile infection.90 There
reduction to ≤12 mm Hg or by >20% of baseline after 4 weeks remains continuing concern about proton pump inhibitors in
of treatment. Patients who were haemodynamic non-responders patients with cirrhosis, therefore caution should be used.
or intolerant to carvedilol were treated with VBL. Carvedilol
resulted in significantly lower variceal bleeding compared with Endoscopic therapy
VBL, and haemodynamic responders to carvedilol or propran- Variceal band ligation
olol had significantly lower mortality than those treated with VBL has been compared with NSBB in 19 trials in a recent
VBL. It is worth noting that the study was not randomised. Cochrane meta-analysis of 1504 patients.91 Despite reduced
There have been recent suggestions based on low-level evi- bleeding (RR=0.67, 95% CI 0.46 to 0.98) with VBL, there was
dence that NSBB may result in a poorer outcome in patients no difference in overall mortality and bleeding-related mortality.
with cirrhosis and refractory ascites.77 The ‘window hypothesis’ The difference in bleeding was not seen when only trials with
for β blockers in cirrhosis has also recently been described, sug- low selection or attrition bias were included. Banding can have
gesting that NSBB are helpful in the compensated and early serious complications. The risk of fatal banding-induced bleed-
decompensated cirrhotic period, but may not be helpful in very ing was highlighted in a meta-analysis showing reduced fatal
early cirrhosis, such as in a patient with no varices, and may be adverse events with NSBB (OR=0.14, 95% 0.02 to 0.99).92
harmful in patients with end-stage cirrhosis with refractory The optimal timing of banding intervals is discussed in the
ascites.78 However, recent large observational studies question section ‘Secondary prophylaxis of variceal haemorrhage’. A ran-
the last hypothesis, with improved survival seen in patients with domised trial of 96 patients who underwent endoscopic surveil-
refractory ascites treated with NSBB,79 unless patients have an lance at 6 or 3 months after eradication of varices with VBL did
episode of spontaneous bacterial peritonitis.80 Therefore, until not demonstrate a difference in bleeding on mortality.93
there are further prospective controlled studies, NSBB should be However, the trial had a heterogeneous study group of patients
continued in patients with refractory ascites. The clinician must who underwent VBL both for primary (65%) and secondary
carefully monitor haemodynamic parameters such as blood pres- prevention (35%).
sure, and discontinue NSBB in patients with hypotension and
renal impairment as can occur after an episode of spontaneous
Sclerotherapy
bacterial peritonitis. Other potentially severe adverse events
Nineteen trials have compared endoscopic variceal sclerotherapy
with NSBB include symptomatic bradycardia, asthma and
with no treatment.68 Owing to the marked heterogeneity
cardiac failure. Less severe side effects such as fatigue, insomnia
between these studies a meta-analysis is clinically inappropri-
and sexual dysfunction may also result.
ate.68 Sclerotherapy does not offer any benefit in combination
with NSBB or VBL compared with VBL or NSBB alone, and
Isosorbide mononitrate increases iatrogenic complications such as strictures.94–96 At this
Interest in the use of vasodilators such as isosorbide mononitrate time sclerotherapy cannot be recommended for prophylaxis of
(ISMN) developed after the demonstration that it reduces portal variceal haemorrhage in patients with cirrhosis.
pressure as effectively as propranolol,81 but has subsequently
Recommendations: primary prophylaxis of variceal haemor-
waned. A trial comparing ISMN with propranolol showed no
rhage in cirrhosis (figure 2)
significant difference between these agents.82 Another rando-
1. What is the best method for primary prophylaxis?
mised trial of ISMN versus placebo did show any difference in
1.1. We recommend NSBB or variceal band ligation (VBL).
the two arms.83 Therefore, ISMN is not recommended as
We suggest pharmacological treatment with propranolol
monotherapy in primary prophylaxis.
as first line. VBL is offered if there are contraindications
to NSBB. The choice of VBL or NSBB should also take
β Blocker and ISMN into account patient choice (level 1a, grade A).
The combination of nadolol and ISMN has been compared with 1.2. We suggest carvedilol or nadolol as alternatives to pro-
nadolol in a RCT. The combination therapy reduced the fre- pranolol (level 1b, grade A).
quency of bleeding significantly but no significant differences 1.3. Dose:
were detected in mortality.84 However, Garcia-Pagan et al85 in a 1.3.1. Propranolol: 40 mg twice daily. Dose titrated to
double-blind RCT of propranolol plus ISMN versus propranolol maximum tolerated or once heart rate (HR) of
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1689
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
reached a maximum dose of 240 mg (level 1a, red signs or grade 2–3 varices.
grade A). Numerator; patients who have grade 1 varices with red
1.3.3. Carvedilol: 6.25 mg once daily to increase to signs or grade 2–3 varices receiving primary prophy-
maintenance of 12.5 mg after a week if tolerated laxis.
or once HR of <50–55 bpm is reached (level 1a, Denominator; patients diagnosed with cirrhosis who
grade A). have grade I varices with red signs or grade 2–3 varices.
1.3.4. It is suggested that NSBB are discontinued at the
time of spontaneous bacterial peritonitis, renal
impairment and hypotension (level 2b, grade B). MANAGEMENT OF ACTIVE VARICEAL HAEMORRHAGE
1.4. In cases of contraindications or intolerance to NSBB, we The average 6-week mortality of the first episode of variceal
recommend variceal band ligation (level 1a, grade A). bleeding in most studies is reported to be up to 20%. There has
2. Who should have surveillance for variceal bleeding? been considerable improvement in survival since the early 1980s
2.1. We recommend all patients with cirrhosis should be when the in-hospital mortality was 40–50%,97 compared with
endoscoped at the time of diagnosis (level 1a, grade A). 15% from a recent UK audit.5 Such is the improvement in out-
There is no indication to repeat endoscopy in patients comes, that a patient with Child’s A cirrhosis is very unlikely to
receiving NSBB. succumb to an index variceal bleed. Studies have shown the
3. How often should cirrhotic patients be endoscoped? Child–Pugh score, MELD score, and HVPG to be strong predic-
3.1. If at the time of first endoscopy no varices are seen, we tors of outcomes.98–103 The MELD score has been shown to
suggest that patients with cirrhosis should be endos- outperform Child’s score in a recent study, with a score >19
coped at 2–3-year intervals (level 2a, grade B). associated with 20% 6 week mortality.103 Furthermore, the
3.2. If grade I varices are diagnosed, we suggest that patients MELD score has been shown to perform as well as the trad-
should be endoscoped at yearly intervals (level 2a, itional intensive care unit scores in predicting mortality in
grade B). patients admitted to intensive care in the UK.104 MELD >18,
3.3. If there is clear evidence of disease progression we active bleeding, transfusing >4 units of packed red blood cells
suggest that the intervals can be modified by a clinician. have been shown to be predictors of mortality and early
Endoscopy should also be offered at time of decompen- rebleeding.99 101 102 HVPG has also been shown to predict
sation (level 2a, grade B). outcome when measured at 2 weeks after a bleed,44 and a value
4. Which patients with cirrhosis should have primary of ≥20 mm Hg when measured acutely within 48 h has been
prophylaxis? shown to provide significant prognostic information.100
4.1. If grade I varices and red signs or grade 2–3 varices are However, this technique is not used routinely in the manage-
diagnosed, we recommend that patients have primary ment of patients around the world, and substitution of clinical
prophylaxis irrespective of the severity of the liver data in the latter study was shown to provide the same clinical
disease (level 1a, grade A). predictive value.100 These scoring systems are not purely aca-
5. Treatments not recommended: demic; they allow the referring clinician to predict those
5.1. Proton pump inhibitors are not recommended unless patients with a high chance of rebleeding to be transferred to a
otherwise required for peptic ulcer disease (level 1b, specialist centre offering, for instance, TIPSS before the patient
grade B). rebleeds.
5.2. Isosorbide mononitrate monotherapy is not recommended Nonetheless, probably the most important step in the man-
as primary prophylaxis (level 1b, grade A). There is insuffi- agement of acute variceal haemorrhage is the initial resuscitation
cient evidence to recommend isosorbide mononitrate in assessed according to standard ‘ABC’ practice, together with
combination with NSBB (level 1b, grade A). protection of the airway to prevent aspiration. Although early
5.3. Shunt surgery or TIPSS is not recommended as primary endoscopy allows for accurate diagnosis of the bleeding site and
prophylaxis (level 1a, grade A). decisions about management (figure 3), therapeutic intervention
5.4. Sclerotherapy is not recommended as primary prophy- in acute variceal bleeding can be initiated, safely in most cases,
laxis (level 1a, grade A). before diagnostic endoscopy. As similar efficacy is demonstrated
6. Areas requiring further study: with pharmacological treatment as with sclerotherapy, the
6.1. Role of NSBB in patients without varices, with focus on former should be first-line therapy.99 β Blockade should not be
carvedilol. started in the acute setting, and those already taking β blockers
6.2. Role of NSBB in patients with small varices, with focus as prophylaxis should probably stop taking them for 48–72 h in
on carvedilol. order that the patient’s physiological response to blood loss can
6.3. Comparison of carvedilol versus propranolol in primary be allowed to manifest.
prophylaxis.
6.4. Identification of, and trials assessing, new drugs for General considerations
primary prophylaxis such as statins. Patient evaluation
7. Quality indicator: The majority of patients with a variceal bleed will be sufficiently
7.1. Percentage of patients at diagnosis of cirrhosis who have stable to enable a full history and examination to take place.
had an endoscopy to screen for varices (level 1a, History of alcohol excess and or intravenous drug use should be
grade A). sought and may become particularly relevant if the patient has
Numerator; patients diagnosed with cirrhosis who have withdrawal symptoms after admission. Comorbidity is important
had an endoscopy either before or after diagnosis within when estimating risk and deciding on use of vasopressors. The
1690 Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
following risk factors doubled mortality after an acute variceal excluded. Evidence of ascites requires a diagnostic tap to search
bleed in one US study: older age, comorbidities, male gender and for infection.
not undergoing a gastroscopy within 24 h.105 Investigations including full blood count, coagulation
A full examination is helpful for the important negatives as profile, liver and renal function and blood group and save and
much as the positives. Baseline observations should include the cross-match. Blood and urine should also be cultured. An ultra-
temperature, as infection is a serious complication with signifi- sound scan later in the admission is helpful to identify subclin-
cant mortality. Confusion may be present because of encephal- ical ascites, flow in portal vein and any obvious emergence of an
opathy, intoxication with alcohol or drugs or withdrawal from hepatocellular carcinoma (HCC).
alcohol or drugs. The patient should be on continuous BP and
pulse monitor and their haemodynamic status recorded. Location of patient
An oxygen saturation monitor is helpful. Stigmata of chronic A decision must be made as to where the patient is best managed.
liver disease and concurrent jaundice provide insight into the Variceal bleeding is unpredictable, generally occurs in patients
current status of a patient’s liver, and also give warning of with significant liver disease and is associated with significant mor-
potential further decompensation if significant bleeding persists tality. Hence, a high-dependency unit is usually the most appropri-
(see scoring systems above). Pneumonia must be actively ate initial location, although a properly staffed ‘gastrointestinal
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1691
Guidelines
bleeding bed’ may be appropriate. If a patient is vomiting blood, blood flow. In comparison with no active treatment, the pooled
or there is a perceived risk of a haemodynamically unstable patient results of four randomised trials showed that it reduced failure
having blood in the stomach, then the patient must be intubated to control variceal bleeding (OR=0.22, 95% CI 0.12 to 0.43),
before endoscopy, and return to an intensive care or high- although survival was unaffected.68 Meta-analysis of five trials
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
dependency unit will be necessary until extubation. comparing sclerotherapy with vasopressin has shown a signifi-
cant effect on reduction in failure to control bleeding
Volume resuscitation and blood products (OR=0.51, 95% CI 0.27 to 0.97), with no effect on survival.68
Intravenous access (two 16–18G cannulae) should have been
secured on admission with a reported GI bleed. Further intra- Vasopressin with nitroglycerine
venous access may be necessary. In patients with poor venous The addition of nitroglycerine enhances the effect of vasopres-
access, advanced liver disease, or renal failure associated with sin on portal pressure and reduces cardiovascular side effects.112
their liver disease, central venous access may be helpful with Meta-analysis of three randomised trials comparing vasopressin
guiding fluid infusions. However, the drawbacks include the risk alone with vasopressin and nitroglycerine showed that the com-
of the procedure and a potential source of infection. Therefore, bination was associated with a significant reduction in failure to
there is no absolute requirement for a central line, and no evi- control bleeding (OR=0.39, 95% CI 0.22 to 0.72), although no
dence of unequivocal benefit. Intravenous fluid resuscitation survival benefit was shown.68
should be initiated with plasma expanders aiming to maintain a
systolic blood pressure of 100 mm Hg. Care with monitoring is Terlipressin
paramount in this group of patients. Terlipressin is a synthetic analogue of vasopressin, which has an
Overtransfusion has been shown to have a deleterious effect immediate systemic vasoconstrictor action followed by portal
on outcome. In a recent single-centre RCT, a restrictive transfu- haemodynamic effects due to slow conversion to vasopressin. In
sion policy of maintaining haemoglobin between 70 and 80 g/L a Cochrane meta-analysis of seven placebo-controlled trials, ter-
improved the control of variceal bleeding (11% vs 22%, lipressin was shown to reduce failure to control bleeding
p=0.05), and lowered HVPG compared with a liberal transfu- (RR=0.66, 95% CI 0.55 to 0.93) and also to improve survival
sion policy without effect on 45-day survival.106 However, it (RR=0.66, 95% CI 0.49 to 0.88).113 In the same meta-analysis,
should be noted that these results were from a single Spanish there was no difference between terlipressin versus vasopressin,
centre, which was a tertiary unit for variceal bleeding, where all balloon tamponade or endoscopic therapy in failure to control
patients underwent endoscopy within 6 h. Nonetheless, a bleeding or survival.113 The role of terlipressin in combination
restrictive transfusion policy has been recommended for some with VBL is explored in the section ‘Endoscopic therapy in
time1 and there is now good evidence to support not transfusing combination with pharmacological therapy’.
a stable patient with a haemoglobin of ≥80 g/L. However, under- The recommended dose of terlipressin is 2 mg IV every 4 h,
resuscitation should also be avoided and while goal-oriented although many units reduce the dose to 6 hourly as it may cause
fluid replacement has generally not been useful in an intensive peripheral vasoconstriction which manifests as painful hands
therapy unit setting, a venous saturation >70% remains an easily and feet. While 5 days of IV treatment has been advocated in
measurable target with some evidence to support it.107 the Baveno V guidelines,1 this prolonged treatment has not
Interpretation and management of clotting profile is challen- been shown to have a survival benefit, and for pragmatic
ging in liver disease, where there is usually a balanced deficiency reasons many units will stop treatment shortly after satisfactory
of both procoagulant and anticoagulant factors.108 The NICE haemostasis. In a randomised trial terlipressin given for 24 h
guidelines recommend activation of a hospital’s massive transfu- after satisfactory haemostasis with VBL after oesophageal vari-
sion policy when there is major haemorrhage, and platelet ceal bleeding was as effective as 72 h of treatment.114
support when the value is <50, and clotting factor support In patients intolerant of terlipressin or in countries where ter-
when the international normalised ratio (INR) is >1.5 times lipressin is not available, alternatives should be considered.
normal.2 There is no evidence for the use of ‘prophylactic’ clot-
ting or platelet support to reduce the risk of rebleeding. There Somatostatin and octreotide
is insufficient evidence to support the routine use of transexamic Somatostatin causes selective splanchnic vasoconstriction and
acid, or recombinant factor VIIa.109 reduces portal pressure and portal blood flow.115 Octreotide is a
somatostatin analogue. The mechanism of action of these two
Pharmacological treatment agents is not clear. Inhibition of glucagon increases vasodilatation
The two major classes of drugs that have been used in the rather than a direct vasoconstrictive effect and post-prandial gut
control of acute variceal bleeding are vasopressin or its analo- hyperaemia is also reduced. The actions of octreotide on hepatic
gues (either alone or in combination with nitroglycerine) and and systemic hemodynamics are transient, making continuous
somatostatin or its analogues. Terlipressin is the only agent that infusion necessary. Octreotide is given as a 50 μg bolus followed
has been shown to reduce mortality in placebo-controlled trials. by an infusion of 25–50 μg/h. Somatostatin is given as a 250 mg
However, in trials comparing terlipressin, somatostatin and intravenous bolus followed by an infusion of 250 mg/h.
octreotide, no difference in efficacy was identified in a system- Somatostatin and octreotide have been shown to be as
atic review110 and in a recent large RCT.111 Prophylactic anti- effective as terlipressin in acute variceal bleeding in a
biotics can result in a similar survival benefit following acute meta-analysis.110 Seo et al111 in a large RCT of 780 patients com-
variceal bleeding. paring these three agents failed to show a difference in treatment
success (range 83.8–86.2%), rebleeding (range 3.4–4.4%) and
Vasopressin mortality (range 8–8.8%). A low systolic blood pressure at pres-
Vasopressin reduces portal blood flow, portal systemic collateral entation, high serum creatinine level, active bleeding in the emer-
blood flow and variceal pressure. It does, however, have signifi- gency endoscopy, gastric variceal bleeding and Child–Pugh grade
cant systemic side effects such as an increase in peripheral resist- C were independent factors predicting 5-day treatment
ance, and reduction in cardiac output, heart rate and coronary failure.111
1692 Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
placebo-controlled trials (RR=0.79, 95% CI 0.63 to 0.98).116 at 15 days.123
Antibiotics were also shown to reduce bacterial infections
(RR=0.43, 95% CI 0.19 to 0.97) and early rebleeding (RR=0.53, Endoscopic therapy in combination with pharmacological therapy
95% CI 0.38 to 0.74).116 Therefore, short-term antibiotics should The role of combining vasoactive drugs with endoscopic
be considered standard practice in all cirrhotic patients who have a therapy (VBL or sclerotherapy) was reported in a meta-analysis
variceal bleed, irrespective of the presence of confirmed infection. of eight trials.124 Combination therapy resulted in better initial
Third-generation cephalosporins, such as ceftriaxone (1 g IV, control of bleeding (RR=1.12, 95% CI 1.02 to 1.23), and
daily), have been shown to be more effective at reducing 5-day haemostasis (RR=1.28, 95% CI 1.18 to 1.39), without
Gram-negative sepsis than oral norfloxacin,117 but choice of anti- any difference in survival. Adverse events were similar in both
biotics must be dictated by local resistance patterns and availability. groups. Two RCTs have compared VBL with sclerotherapy in
combination with vasoactive agents in acute variceal bleed-
Proton pump inhibitors ing.125 126 Lo et al125 used vasopressin and found that VBL
One RCT compared a short course of proton pump inhibitors resulted in better 72 h haemostasis (97% vs 76%, p=0.009),
with vasoconstrictor therapies after haemostasis in acute variceal with fewer complications (5% vs 29%, p=0.007). Villanueva
bleeding.118 Despite larger ulcers noted in the vasoconstrictor et al used somatostatin, and reported lower failure to control
arm, there were no differences in bleeding or survival. Nearly acute bleeding with VBL (4% vs 15%, p=0.02), with fewer
50% of patients had ascites, which might have implications in serious complications (4% vs 13%, p=0.04). Overall survival
light of the reports of increased incidence of spontaneous bac- was similar in both trials.125 126
terial peritonitis as mentioned earlier.
Balloon tamponade
Balloon tamponade is highly effective and controls acute bleed-
Endoscopic therapy
ing in up to 90% of patients although about 50% rebleed when
Endoscopy should take place within 24 h of admission and
the balloon is deflated.127 128 It is, however, associated with
earlier if there is excessive bleeding, based on low-level evi-
serious complications such as oesophageal ulceration and aspir-
dence.105 While many guidelines and reviews suggest that
ation pneumonia in up to 15–20% of patients. Despite this, it
endoscopy should be carried out within 12 h the only study that
may be a life-saving treatment in cases of massive uncontrolled
examined the influence of timing on outcome failed to demon-
variceal haemorrhage pending other forms of treatment. An
strate any advantage of endoscopy before 12 h.119 The optimal
appropriately placed Sengstaken–Blakemore tube allows for
time is after sufficient resuscitation, and pharmacological treat-
resuscitation, safe transportation and either repeat endoscopy or
ment, with the endoscopy performed by a skilled endoscopy
radiological shunting in a patient with a stable cardiovascular
team, in a suitably equipped theatre environment and with
system. The oesophageal balloon is rarely required, must never
airway protection. Airway protection is essential where risk of
be used on its own and should be used only if there is continu-
aspiration is high, and affords the endoscopist time for thorough
ing bleeding despite an adequately inflated gastric balloon cor-
evaluation, including complete clot aspiration and controlled
rectly placed and with appropriate tension. Placement of the
application of treatment, including tamponade if required. The
tube endoscopically or over a guide wire might reduce the risk
endoscopy team must comprise an experienced endoscopy
of complications, especially oesophageal rupture.
nurse acquainted with the equipment necessary for endoscopy
therapy of varices, and a skilled endoscopist, competent in using
banding devices and deployment of balloon tamponade.
Removable oesophageal stents
The SX-Ella Danis stent (ELLA-CS, Hradec Kralove, Czech
Republic) is a removable covered metal mesh stent placed endo-
Variceal band ligation scopically in the lower oesophagus without radiological screen-
This technique is a modification of that used for the elastic band ing. It has no role in the management of gastric variceal
ligation of internal haemorrhoids. Its use in humans was first bleeding. These stents can be left in situ for up to 2 weeks
described in 1988.120 A meta-analysis of seven trials comparing unlike the Sengstaken–Blakemore tube which should be
VBL with sclerotherapy in acute bleeding showed that VBL removed after a maximum of 24–48 h.129 130 No published con-
reduced rebleeding from varices (OR=0.47, 95% CI 0.29 to trolled trials have compared this modality with balloon
0.78), reduced mortality (OR=0.67, 95% CI 0.46 to 0.98) and tamponade.
resulted in fewer oesophageal strictures (OR=0.10, 95% CI
0.03 to 0.29).121 The number of sessions required to obliterate Transjugular intrahepatic portosystemic stent-shunt
varices was lower with VBL (2.2 fewer sessions (95% CI 0.9 Several uncontrolled studies have examined the role of salvage
to 3.5)). bare TIPSS in acute variceal bleeding. In a review of 15 studies,
control of bleeding was achieved in 90–100%, with rebleeding
Sclerotherapy in 6–16%.131 Mortality varied between 75% (in hospital) and
Sclerotherapy has been replaced by VBL and should no longer 15% (30 day). It is important to appreciate that sclerotherapy
be offered as standard of care in acute variceal haemorrhage. was used as first-line endoscopic therapy in most of these
studies. Long-term follow-up of a study that compared TIPSS
Other endoscopic measures with H-graft portacaval shunts in patients for whom non-
In an RCT, cyanoacrylate offered no benefit over VBL, with the operative management had failed suggested that H-grafts were a
additional risk of embolisation and trend towards increased useful method of reducing portal pressure and had a signifi-
rebleeding with cyanoacrylate.122 cantly lower failure rate ( p=0.04), but had no significant
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1693
Guidelines
improvement in overall survival despite a benefit seen in Child’s 1.1. Units offering an emergency acute upper gastrointestinal
A and B disease.132 A recent RCT compared emergency porto- bleeding service should have expertise in VBL, balloon
caval surgery with bare TIPSS within 24 h of presenting with tamponade and management of gastric variceal bleeding
acute oesophageal variceal bleeding in unselected cirrhotic (level 5, grade D).
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
patients. Emergency portocaval surgery resulted in better out- 1.2. Transfuse patients with massive bleeding with blood,
comes for long-term bleeding control, encephalopathy and sur- platelets and clotting factors in line with local protocols
vival ( p<0.001).133 Before wider application of surgery for for managing massive bleeding (level 5, grade D).
acute variceal bleeding, more data are needed in light of the 1.3. Base decisions on blood transfusion on the full clinical
recent adoption of covered stents. picture, recognising that overtransfusion may be as dam-
There has also been a generalised established change in prac- aging as undertransfusion. A restrictive transfusion
tice in using covered TIPSS stents ( polytetrafluoroethylene policy aiming for a haemoglobin of 70–80 g/L is sug-
(PTFE)) rather than a bare metal stent, with evidence to support gested in haemodynamically stable patients (level 1b,
this change. In randomised controlled studies, these stents were grade B).
shown to have higher primary patency rates than bare stents, 1.4. Do not offer platelet transfusion to patients who are not
without significant differences in survival, and the potential for actively bleeding and are haemodynamically stable
reduced incidence of hepatic encephalopathy.134 135 (level 5, grade D).
There is, however, growing evidence from two RCTs for the 1.5. Offer platelet transfusion to patients who are actively
earlier use of TIPSS in selected patients stratified by HVPG, bleeding and have a platelet count of <50×109/L
Child–Pugh class and active bleeding, and not just use as a (level 5, grade D).
salvage option.6 136 Monescillo et al136 randomised patients 1.6. Offer fresh frozen plasma to patients who have either:
presenting with acute oesophageal variceal haemorrhage to bare ▸ a fibrinogen level of <1 g/L (level 5, grade D), or
TIPSS or standard of care if the HVPG was ≥20 mm Hg within ▸ a prothrombin time (international normalised ratio)
24 h of admission. Significantly reduced treatment failure, as or activated partial thromboplastin time >1.5 times
defined by failure to control acute bleeding and/or early rebleed- normal (level 5, grade D).
ing (12% vs 50%), was seen and improved survival (62% vs 1.7. Offer prothrombin complex concentrate to patients who
35%) in the patients randomised to undergo a TIPSS procedure. are taking warfarin and actively bleeding (level 5, grade D).
However, the standard of care was sclerotherapy and not com- 1.8. Treat patients who are taking warfarin and whose upper
bination endoscopic and pharmacological treatment. This limita- gastrointestinal bleeding has stopped in line with local
tion and the lack of availability of HVPG measurement in most warfarin protocols (level 5, grade D).
centres meant this trial did not have a significant impact on clin- 1.9. There is insufficient evidence for the use of recombinant
ical practice. factor VIIa in acute variceal haemorrhage (level 1b,
Garcia-Pagan et al6 selected patients with active bleeding and grade B).
Child’s B cirrhosis or patients with Child’s C cirrhosis (Child’s 2. Suggestions for timing of upper gastrointestinal endoscopy:
score <14) for randomisation to early PTFE-covered TIPSS 2.1. Offer endoscopy to unstable patients with severe acute
within 72 h or standard of care with VBL and pharmacological upper gastrointestinal bleeding immediately after resusci-
treatment. This has shown encouraging results with reduced risk tation (level 5, grade A).
of treatment failure (3% vs 50%), improved survival (86% vs 2.2. Offer endoscopy within 24 h of admission to all other
61% at 1 year), yet without increased risk of hepatic encephal- patients with upper gastrointestinal bleeding (level 2b,
opathy. The results were supported by an observational study grade A).
from the same group, although a survival benefit was not 2.3. Units seeing more than 330 cases a year should offer
seen.137 Furthermore, a recent well-conducted observational daily endoscopy lists. Units seeing fewer than 330 cases
study did not demonstrate such high survival rates with early a year should arrange their service according to local
TIPSS, with 11-year survival of 67%, which was similar to that circumstances (level 5, grade D).
of patients given endoscopic and pharmacological treatments 3. Control of bleeding:
only.138 Therefore, larger multicentred RCTs need to be under- 3.1. Antibiotics are recommended for all patients with suspected
taken to further evaluate the role of early TIPSS. It is important or confirmed variceal bleeding (level 1a, grade A).
to make the distinction between salvage TIPSS and early TIPSS 3.2. In all patients, vasoconstrictors such as terlipressin or
to prevent rebleeding. somatostatin are recommended and should be started as
soon variceal bleeding is suspected and continued until
Liver transplantation haemostasis is achieved or for up to 5 days. Octreotide
This is probably appropriate only for patients who bleed while (unlicensed) is suggested if terlipressin or somatostatin
awaiting liver transplantation, although studies comparing VBL are unavailable (level 1a, grade A).
or TIPSS placement with urgent liver transplantation in this 3.3. Variceal band ligation is recommended as the preferred
situation need to be done. Liver transplantation is an exceed- endoscopic method (level 1a, grade A).
ingly rare option for the vast majority of patients, both because 3.4. After satisfactory haemostasis with the methods above,
it is not commonly available and because of shortages and and depending on local resources, early covered TIPSS
delays in organ procurement. No controlled trials of liver trans- (<72 h after index variceal bleed) can be considered in
plantation in uncontrolled/active bleeding are available. selected patients with Child’s B cirrhosis and active
Recommendations for the control of variceal bleeding in cir- bleeding or Child’s C cirrhosis with Child’s score <14
rhosis are given below and in figure 3. (level 1b, grade B).
3.5. Proton pump inhibitors are not recommended unless
Recommendations: control of active variceal haemorrhage in otherwise required for peptic ulcer disease (level 1b,
cirrhosis (figure 3) grade B).
1. Suggestions for resuscitation and initial management 4. Failure to control active bleeding:
1694 Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
4.2. Specialist help should be sought at this time and transfer as treatment for the prevention of variceal rebleeding.144 There
to a specialist centre should be considered. Units that do was no difference in rebleeding and the survival benefit was
not offer a TIPSS service should identify a specialist restricted to Child A and B patients. Serious adverse events were
centre which offers a 24 h emergency TIPSS service and similar in both groups. More data are required to investigate
have appropriate arrangements for safe transfer of this interesting observation of a survival benefit from simvastatin
patients in place (level 2a, grade B). in this situation, which may relate to its effects on hepatocellular
5. Areas requiring further study: function, fibrosis and portal pressure.
5.1. The efficacy of restrictive blood transfusion in variceal
haemorrhage. Proton pump inhibitors
5.2. The role of blood products in variceal haemorrhage. A double-blind randomised placebo-controlled trial showed that
5.3. The utility of early TIPSS (<72 h) in acute variceal pantoprazole reduced the size of ulcers in patients who under-
haemorrhage. went VBL. However, the total number of ulcers and other out-
5.4. The role of removable oesophageal stents in acute vari- comes were similar in the two groups.145
ceal haemorrhage.
5.5. The role of haemostatic powders in acute variceal
Endoscopic therapy
haemorrhage.
VBL has been accepted as the preferred endoscopic treatment
5.6. The role of proton pump inhibitors in variceal
for the prevention of variceal rebleeding, with a lower rate of
haemorrhage.
rebleeding, mortality and complications than sclerother-
6. Quality indicators:
apy.146 147 The time interval between VBL sessions to achieve
6.1. Antibiotic administration in acute variceal bleeding
eradication of varices is debateable. However, a recent RCT
within 1 day either before or after the procedure (level
comparing monthly with biweekly VBL after initial haemostasis
1a, grade A).
with VBL in 70 patients suggested that there were fewer
Numerator; patients with an acute variceal bleed who
post-VBL ulcers in the monthly group (11% vs 57%;
have received antibiotics within 1 day either before or
p<0.001).148 Variceal recurrence, rebleeding and mortality were
after the procedure.
similar in both groups.
Denominator; patients with an acute variceal bleed.
Two meta-analyses showed there is no evidence that the add-
6.2. Endoscopy performed within 24 h of presentation of an
ition of sclerotherapy to VBL improves clinically relevant out-
acute variceal bleed (level 2b, grade A).
comes, including variceal rebleeding and death, and the
Numerator; patients with an acute variceal bleed who
combination led to higher stricture rates.149 150
have received endoscopy within 24 h of presentation.
Denominator; patients with an acute variceal bleed.
Endoscopic therapy versus drug therapy
VBL has been reported to be more effective than combined
NSBB and ISMN drug therapy.151 However, an 8-year
SECONDARY PROPHYLAXIS OF VARICEAL HAEMORRHAGE follow-up study of this RCT found that although VBL was
β Blockers superior in reducing variceal rebleeding, survival rates were sig-
A meta-analysis of 12 trials comparing propranolol or nadolol139 nificantly higher in the group treated with combined drug treat-
with no active treatment showed a significant reduction in ment.152 Other studies have found no superiority of VBL over
rebleeding but no significant reduction in mortality.140 The combined drug therapy for prevention of variceal rebleeding or
greater reduction in portal pressure with carvedilol compared mortality.153 154 A recent small multicentre RCT reported carve-
with propranolol has been described in the section ‘Primary dilol to be similar to VBL in the prevention of variceal rebleed-
prophylaxis’ of this guideline. ing, with a trend in favour of survival with carvedilol (73% vs
48%, p=0.110).155
Several meta-analyses have compared drug therapy with VBL
Nitrates in the prevention of variceal rebleeding. One meta-analysis of
The addition of ISMN to NSBB has been shown to reduce vari- six RCTs showed no significant difference in variceal rebleeding
ceal rebleeding compared with NSBB alone, although no sur- rates when comparing VBL alone with combined NSBB and
vival benefit was seen.141 In addition, adverse events leading to ISMN therapy. However, all-cause mortality was significantly
drug withdrawal were more common in the group receiving higher in patients treated with the VBL (RR=1.25, 95% CI
combined drug treatment. A meta-analysis of ISMN alone or 1.01 to 1.55).156 Three meta-analyses comparing drug therapy
with either NSBB or endoscopic therapy reported that there was (NSBB alone or with ISMN) with endoscopic therapy alone
no mortality benefit from combining nitrates and NSBB com- reported no difference in variceal rebleeding or mortality.157–159
pared with NSBB alone.142
Side effects of ISMN include dizziness and headache. Owing Endoscopic+drug therapy versus either alone
to the side effects and relative lack of data, ISMN is not com- Numerous studies and several meta-analyses have compared
monly used in clinical practice. combined endoscopic and drug therapy with monotherapy
A recent RCT of 121 patients reported carvedilol to be (endoscopic or drugs alone) in the prevention of variceal
similar to combined ISMN and NSBB therapy in the prevention rebleeding. A meta-analysis of 23 trials assessing sclerotherapy
of variceal rebleeding and mortality, although severe adverse or VBL combined with NSBB reported that combination
events were less common with carvedilol.143 therapy reduced rebleeding more than either endoscopic
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1695
Guidelines
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
and either alone.157 A further meta-analysis reported reduced benefit.68 Non-selective shunts resulted in similar outcomes
variceal rebleeding (RR=0.601, 95% CI 0.440 to 0.820) but compared with distal splenorenal shunts.68 Extended follow-up
similar mortality with combined drug and endoscopic therapy of a randomised study comparing portocaval shunt surgery with
versus endoscopic therapy alone.159 Another meta-analysis of sclerotherapy following acute variceal bleeding, reported better
17 trials (14 using sclerotherapy and three using VBL) reported long-term bleeding control (100% vs 20%, p<0.001) and
that combined endoscopic and NSBB therapy reduced rebleed- improved survival (5-year survival 71% vs 21%, p<0.001) in
ing (OR=2.20, 95% CI 1.69 to 2.85) and overall mortality the portocaval shunt arm.169 Distal splenorenal shunt surgery
(OR=1.43, 95% CI 1.03 to 1.98) compared with endoscopic was compared with TIPSS in a multicentre RCT including 140
therapy alone.161 patients with Child’s A and B cirrhosis.170 Results showed
A further meta-analysis of 10 RCTs suggested that combin- similar rebleeding and survival, but higher rates of shunt dys-
ation therapy reduces the risk of rebleeding from oesophageal function and re-intervention in the TIPSS group, although
varices compared with VBL (RR=0.68, 95% CI 0.45 to 0.93) covered stents were not used. A follow-up study suggested that
or medical treatment (RR=0.60, 95% CI 0.43 to 0.84).162 This TIPSS was more cost-effective.171
meta-analysis included seven trials comparing combination Portosystemic shunts (total surgical, distal splenorenal or bare
therapy with VBL and three trials comparing combination TIPSS) were compared with endoscopic therapy for variceal
therapy with drug treatment. Combined VBL and drug therapy rebleeding in a Cochrane database systematic review.172
gave a survival benefit when compared with VBL alone Twenty-two trials incorporating 1409 patients were included. All
(RR=0.52, 95% CI 0.27 to 0.99), but not when compared with shunt therapies reduced rebleeding (OR=0.24, 95% CI 0.18 to
medical treatment alone. 0.30) at the expense of higher rates of encephalopathy
Another recent meta-analysis assessed five studies comparing (OR=2.09, 95% CI 1.20 to 3.62), with no survival advantage.
VBL alone with combination VBL and drug therapy, and four TIPSS was complicated by a high incidence of shunt dysfunction.
studies comparing drugs alone or combined with VBL.163 This Laparoscopic splenectomy plus VBL was also compared with
found that adding drugs to VBL reduced rebleeding (RR=0.44, TIPSS for variceal rebleeding in a recent non-randomised trial
95% CI 0.28 to 0.69) with a trend towards reduced mortality, of 83 patients.173 This reported surgery plus VBL to be better
but adding VBL to drug treatment did not significantly affect than TIPSS in preventing variceal rebleeding, with low rates of
either rebleeding or mortality. encephalopathy.
The meta-analyses are not entirely consistent, although it Liver transplantation should be considered in eligible patients
would appear that combined VBL and drug treatment might following a variceal bleed determined by the selection criteria of
improve survival, but is likely to increase adverse effects com- the country.174 There is no clear evidence that prior shunt
pared with VBL alone. There appears to be less clear benefit surgery has a significant impact on transplant outcome.169
from combined VBL and drug treatment compared with drug Recommendations for the secondary prophylaxis of variceal
treatment alone. bleeding in cirrhosis are given below and in figure 3.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
contraindications to combination therapy), and in selected logical procedures, surgery and long-term NSBB. Splenic vein
cases owing to patient choice. PTFE-covered stents are thrombosis should be considered and appropriate investigations
recommended (level 1a, grade A). undertaken in patients presenting with gastric variceal bleeding.
3.2. Where TIPSS is not feasible in Child’s A and B patients,
we suggest shunt surgery can be used where local Endoscopic therapy
expertise and resources allow (level 1b, grade B). Endoscopic sclerotherapy
4. Areas requiring further study: Sclerotherapy has been largely replaced by VBL and tissue adhe-
4.1. Combination of VBL and carvedilol (or other NSBB) sives or thrombin when appropriate for gastric varices, owing to
versus carvedilol as monotherapy. the lower complication and rebleeding rates.
4.2. Comparison of carvedilol with propranolol in secondary
prophylaxis.
4.3. Optimum time interval between VBL sessions. Endoscopic VBL
4.4. Strategy of VBL or NSBB discontinuation after variceal Standard VBL or the use of detachable snares has been shown
eradication during combination therapy with VBL to control active bleeding from gastric varices, but rebleeding
+NSBB. and recurrence rates are high.178 179 As GOV-1 are generally
4.5. Strategy of VBL add-on therapy to failure of NSBB considered extensions of oesophageal varices, VBL is often used
monotherapy. to treat bleeding from here. However, given the larger diameter
4.6. Strategy of NSBB add-on therapy to failure of VBL and the anatomy of other types of gastric varices, and the
monotherapy. limited data on use of VBL in this situation, this technique is
4.7. Role of early TIPSS in secondary prophylaxis. generally not recommended for these.
4.8. Role of statins in secondary prophylaxis.
5. Quality indicator: Endoscopic injection therapy with tissue adhesives
5.1. Institution of secondary prophylaxis after acute variceal Numerous studies have reported the use of tissue adhesives,
bleeding (level 1a, grade A) most commonly histoacryl (N-butyl-cyanoacrylate), in the treat-
Numerator; patients with an acute variceal bleed who ment of gastric varices.180–194 Variations in technique, dilution
have received either NSBB or banding or both within with lipiodol and follow-up strategy have been described. These
4 weeks of the index bleed. studies have reported an initial haemostasis success rate with
Denominator; patients with an acute variceal bleed. tissue glue of 86–100%, with rebleeding rates of 7–28%.
Uncommon, but severe complications, including emboli to the
GASTRIC VARICES pulmonary and cerebral circulations, have been described.181
Natural history A randomised study compared cyanoacrylate injection with
At first endoscopy in patients with portal hypertension, 20% are VBL in 60 patients with gastric variceal bleeding.186 Patients
shown to have gastric varices.175 They are commonly seen in treated with cyanoacrylate had a higher haemostasis rate (87%
patients with portal hypertension due to portal or splenic vein vs 45%), lower rebleeding (31% vs 54%) and lower mortality
obstruction.175 Only 10–20% of all variceal bleeding occurs (29% vs 48%) than those treated with VBL. Another rando-
from gastric varices, but outcome is worse than with bleeding mised study comparing cyanoacrylate with VBL in 97 patients
from oesophageal varices.175 176 with gastric variceal bleeding, reported equal haemostasis rates
Gastric varices can be classified on the basis of their location at 93%, but significantly higher rebleeding with VBL (72% vs
in the stomach and relationship with oesophageal varices. This 27%).193 This study reported no difference in survival or com-
classification has implications for management. The commonly plications between groups.
used Sarin classification divides them into (a) gastro-oesophageal A non-randomised study comparing cyanoacrylate with VBL
varices (GOV), which are associated with oesophageal varices; for gastric variceal bleeding reported similar haemostasis rates,
and (b) isolated gastric varices (IGV), which occur independ- but lower rebleeding with cyanoacrylate (32% vs 72%).194
ently of oesophageal varices.175 Both GOV and IGV are subdi- Survival and complication rates were similar in both groups. In
vided into two groups. Type 1 GOV are continuous with a controlled but non-randomised study comparing cyanoacrylate
oesophageal varices and extend for 2–5 cm below the gastro- with sclerotherapy for gastric variceal bleeding, Oho et al188
oesophageal junction along the lesser curvature of the stomach. showed that the haemostasis rate was significantly higher in the
Type 2 GOV extend beyond the gastro-oesophageal junction cyanoacrylate group. Survival was also significantly greater in
into the fundus of the stomach. Type 1 IGV refers to varices patients treated with cyanoacrylate.
that occur in the fundus of the stomach and type 2 IGV Mishra et al187 reported a randomised study comparing
describes varices anywhere else in the stomach, including the cyanoacrylate injection with β blockers in the prevention of
body, antrum and pylorus. The most common type of varices rebleeding in 67 patients with bleeding GOV-2 or IGV-1.
seen in cirrhosis is GOV type 1. Patients who bleed from IGV During a median 26-month follow-up, patients in the cyano-
are at a significantly higher risk of dying from an episode of acrylate group had significantly lower rates of both variceal
variceal bleeding than patients bleeding from GOV.177 rebleeding (15% vs 55%) and mortality (3% vs 25%).
Treatment modality, presence of portal hypertensive gastropathy
Management of acute gastric variceal bleeding and gastric variceal size >20 mm correlated with mortality.
Although no studies have reported the use of vasopressors and Another recent RCT compared repeated gastric variceal obtura-
antibiotics specifically for the initial management of gastric vari- tion with or without NSBB in patients with bleeding GOV-2 or
ceal haemorrhage, any patient with suspected variceal bleeding IGV-1.182 Mortality and rebleeding rates were similar in the two
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1697
Guidelines
groups, although adverse effects were more common in the Transjugular intrahepatic portosystemic stent-shunt
combination group. An initial TIPSS series using bare stents reported control of
In a non-randomised study, Lee et al185 suggested that endo- active bleeding from gastric varices in almost all patients in
scopic ultrasound (EUS)-guided biweekly cyanoacrylate injection whom the shunt was performed successfully.202–206 Tripathi
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
versus ‘on demand’ injection after recurrent bleeding led to sig- et al43 described 272 patients who had a TIPSS procedure for
nificantly lower rebleeding (19% vs 45%) from gastric varices, either gastric or oesophageal variceal bleeding. They reported
although survival was similar. However, others have not con- similar rebleeding rates after TIPSS for either gastric or
firmed this approach.189 EUS-guided coil therapy has recently oesophageal varices. Initial PPG was lower in patients with
been described as having similar efficacy, but fewer adverse bleeding from gastric varices. In addition, mortality was lower
events, compared with cyanoacrylate injection in a small non- in those patients with initial PPG >12 mm Hg, who had TIPSS
randomised study.191 for gastric compared with oesophageal variceal bleeding. Shunt
Binmoeller et al180 described a new method for the manage- insufficiency and encephalopathy rates were similar in both
ment of fundal gastric varices in 30 patients, using EUS and a groups. The authors suggested aiming to reduce HVPG to
combination of 2-octyl-cyanoacrylate and coils. Haemostasis <7 mm Hg in gastric variceal bleeding.
was achieved in 100% of patients with no procedure-related Lo et al207 undertook a randomised trial in 72 patients com-
complications. Use of coils appeared to reduce the volume of paring TIPSS with cyanoacrylate injection in the prevention of
cyanoacrylate required to obliterate varices. gastric variceal rebleeding. Control of active bleeding had
been achieved with cyanoacrylate in all patients before random-
isation. They reported a significantly lower rate of gastric vari-
Endoscopic injection of thrombin ceal rebleeding with TIPSS (11% vs 38%), although overall
Injection of bovine thrombin to successfully control gastric vari- upper gastrointestinal rebleeding was similar in both groups.
ceal bleeding was initially described in a small cohort in Encephalopathy was more common in those patients treated
1994.195 Varices were eradicated in all patients after a mean of with TIPSS (26% vs 3%), but overall complications and survival
two injections. Przemioslo et al196 reported 94% haemostasis were similar in both groups.
and 18% rebleeding in 52 patients with gastric variceal bleeding A non-randomised study compared TIPSS with cyanoacrylate
treated with bovine thrombin. Ramesh et al197 also studied injection for gastric variceal bleeding.208 No differences were found
bovine thrombin for bleeding gastric varices. They reported in haemostasis, rebleeding or survival, but the group treated with
92% haemostasis, with no rebleeding during follow-up. No TIPSS had increased encephalopathy. Another comparative study
adverse events or technical problems were noted. More recent described lower rebleeding with TIPSS, but reduced in-patient
studies have used human rather than bovine thrombin because length of stay with cyanoacrylate, and similar mortality.209 This
of safety concerns with the latter. McAvoy et al198 reported on study also reported cyanoacrylate to be more cost-effective.
the largest series of patients treated with human thrombin injec-
tion for gastric or ectopic variceal bleeding. They reported 11%
rebleeding in the 33 patients who had gastric variceal haemor- Other radiological procedures
rhage, with no significant adverse events. A recent series by The use of balloon-occluded retrograde transvenous obliteration
Smith et al199 reported a high rate of initial haemostasis in acute (B-RTO) for the treatment of bleeding gastric varices was pio-
bleeding. However, failure to control bleeding or rebleeding was neered by the Japanese.184 210 This procedure involves insertion
reported in >50%, suggesting that thrombin has a role in bridg- of a balloon catheter into an outflow shunt (gastrorenal or
ing to definitive treatment in acute bleeding. Where thrombin gastric-inferior vena caval) via the femoral or internal jugular
was used as prophylaxis, rebleeding occurred in 20%. To date, vein. Blood flow is blocked by balloon inflation, then the veins
no randomised studies assessing thrombin injection for gastric draining gastric varices are embolised with microcoils and a
variceal bleeding have been reported. sclerosant injected to obliterate the varices.
In a small randomised study, B-RTO was compared with
New endoscopic therapies TIPSS in the management of 14 patients with active gastric vari-
Two recent reports have described the successful use of ceal bleeding and gastrorenal shunts.211 Immediate haemostasis,
Hemospray (Cook Medical, USA) in the management of active rebleeding and encephalopathy were similar in both groups. In a
gastric variceal bleeding refractory to cyanoacrylate injection non-randomised study of 27 high-risk patients, Hong et al212
therapy.200 201 In the latter case this was used as a bridge to a compared B-RTO with cyanoacrylate injection in acute gastric
TIPSS procedure,201 but in the former case TIPSS was not variceal bleeding. Active bleeding at baseline was more common
undertaken owing to pre-existing cardiomyopathy.200 No in the cyanoacrylate group. Haemostasis rates after B-RTO and
rebleeding was reported in either case at a 30-day follow-up. cyanoacrylate were similar at 77% and 100%. Rebleeding was
Further data on the use of haemostatic powders in gastric vari- higher in the cyanoacrylate group (71% vs 15%), with compli-
ceal bleeding are required. cations and mortality similar in both groups. This rebleeding
rate after cyanoacrylate is much higher than figures reported
from other studies.
Balloon tamponade A large Korean retrospective study evaluated B-RTO for the
Insertion of a Sengstaken–Blakemore or Linton–Nachlas tube management of gastric variceal haemorrhage.213 Technical
may sometimes help to temporarily stabilise the patient with success of B-RTO was 97% with procedure-related complica-
severe gastric variceal bleeding, which is uncontrolled by stand- tions seen in 4% and rebleeding in 22%. Another retrospective
ard endoscopic methods as described above.127 The Linton– study of B-RTO for bleeding gastric varices described 95% tech-
Nachlas tube has been reported to have greater efficacy in nical success and 50% 5-year survival.214 Cho et al215 assessed
gastric varices haemorrhage in a controlled trial.128 However, B-RTO in 49 patients who had gastric varices with spontaneous
rebleeding is almost universal if another treatment modality is gastro-systemic shunts. Procedural success rate was 84% but two
not instituted. procedure-related deaths occurred. No variceal recurrence or
1698 Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262
Guidelines
rebleeding was noted. It has been reported that B-RTO can 1.1. GOV-1: treat as for oesophageal varices (level 2b,
increase PPG and may aggravate pre-existing oesophageal grade B).
varices and ascites.215 216 Although B-RTO appears to be an 1.2. GOV-2 and IGV:
effective alternative to TIPSS in patients with gastric variceal 1.2.1. We recommend initial endoscopic therapy with
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
bleeding who have appropriate shunts,217 it is rarely performed cyanoacrylate injection (level 1a, grade A).
outside Asian centres.218 1.2.2. Thrombin may also be considered (level 4, grade C).
Percutaneous transhepatic variceal embolisation with cyano- 1.3. TIPSS can be considered, depending on local resources
acrylate and standard endoscopic cyanoacrylate injection have also and clinical judgement (level 3a, grade B).
been compared in a non-randomised study of 77 patients.219 The 2. In control of bleeding fails:
authors reported lower rebleeding with the percutaneous 2.1. Balloon tamponade is suggested for GOV, IGV-1 until
approach, although mortality was similar in both groups. definitive treatment is undertaken (level 2b, grade B).
2.2. Salvage TIPSS is suggested as the first-line definite treat-
Surgery ment, where feasible (level 3a, grade B).
Surgery for portal hypertension should be performed by experi- 2.3. B-RTO or surgical shunting can be considered if TIPSS
enced surgeons in lower-risk patients, ideally in specialist is not possible (eg, portal vein thrombosis present) and
units.220 Because of the increasing use of simpler endoscopic depending on local resources (level 3a, grade B).
and radiological procedures as described above, the need for 3. What are the therapeutic options for prevention of rebleed-
such an intervention has reduced dramatically, and is mainly ing from gastric varices?
confined to splenectomy or splenic artery embolisation in 3.1. We recommend that patients with GOV-1 are entered
patients with splenic vein thrombosis.221 222 into a VBL surveillance programme (level 2b, grade B).
Under-running of gastric varices has been shown to control 3.2. We recommend endoscopic surveillance with cyano-
active bleeding but is followed by recurrence of bleeding in acrylate injection as needed for GOV-2 and IGV (note
50% of patients and is associated with a perioperative mortality the optimum endoscopic follow-up strategy remains
of >40%.223 Complete devascularisation of the cardia, stomach unclear)(level 2b, grade B). Thrombin can also be con-
and distal oesophagus for bleeding from gastric varices is asso- sidered (level 4, grade C).
ciated with good control of bleeding but is followed by rebleed- 3.3. NSBB can be considered in certain circumstances after
ing in >40% of patients and early mortality in about 50%.224 taking into account the patient’s preferences and clinical
The use of distal splenorenal shunting for bleeding from gastric judgement (level 1b, grade B).
varices in patients with cirrhosis was reported in six patients 3.4. We suggest TIPSS if patients rebleed despite cyanoacryl-
with Child class A or B cirrhosis.225 Although good control of ate injection. TIPSS can also be considered in other
bleeding was attained, two patients died in the postoperative selected patients (eg, those with large or multiple gastric
period. Orloff et al169 reported that a portal-systemic shunt can varices) (level 1b, grade B).
be an effective treatment for bleeding varices in patients with 3.5. Shunt surgery may be used in selected patients with
portal vein thrombosis and preserved liver function. well-compensated cirrhosis and depending on local
resources (level 3c, grade B).
Primary prophylaxis of gastric variceal bleeding 3.6. Splenectomy or splenic artery embolisation should be
A randomised study of 89 patients compared β blockers, cyano- considered in all patients where there is splenic vein
acrylate injection and no active treatment in the primary preven- thrombosis or left-sided portal hypertension (level 4,
tion of bleeding from larger (>10 mm) GOV-2 and IGV-1.226 grade C).
Over a 26-month follow-up period, bleeding occurred in 38%, 4. Is there a role for primary prophylaxis of gastric variceal
10% and 53% of patients in the β blocker, cyanoacrylate and bleeding?
no-treatment groups, respectively. The cyanoacrylate group had 4.1. NSBB (level 2a, grade B) can be considered in selected
significantly lower bleeding rates than the other groups for high-risk patients with large GOV-2 after taking into
GOV-2, but not for IGV-1 patients. Mortality was lower in the account the patient’s preferences and clinical judgement.
group treated with cyanoacrylate (7%) than in those given no 4.2. Cyanoacrylate injection is not recommended outside
treatment (26%) but was similar to that in the β blocker group clinical trials (level 2a, grade A).
(17%). However, this was a small, single-centre study with an 5. Areas requiring further study:
unusually high failure rate for NSBB. Many clinicians have sig- 5.1. Role of thrombin in gastric varices, comparing this with
nificant concerns about the safety of cyanoacrylate injection in tissue adhesives in both acute gastric variceal bleeding
the context of primary prophylaxis. and secondary prophylaxis.
In a retrospective study, Kang et al suggested that cyanoacryl- 5.2. Role of TIPSS in acute gastric variceal bleeding and sec-
ate injection may be an effective prophylactic treatment for ondary prophylaxis.
higher-risk gastric varices.227 5.3. Role of haemostatic powders in controlling refractory
A retrospective study evaluated the clinical outcomes of active gastric variceal bleeding.
B-RTO for gastric varices, in which the procedure was per- 5.4. Role of NSBB in the prevention of rebleeding from
formed as a primary prophylactic treatment in 40 patients.228 gastric varices.
The procedure was successful in 79% of patients, although pro- 5.5. Role of B-RTO as monotherapy or in combination with
cedural complications were reported in 9%. Survival at 1 and 5 endoscopic injection of tissue adhesives in prevention of
years was 92% and 73%, respectively. bleeding from gastric varices.
5.6. Role of EUS-guided injection of tissue adhesives or
Recommendations: management of active haemorrhage from thrombin.
gastric varices (figure 3) 5.7. Primary prevention of gastric variceal bleeding with
1. What is the optimal management of bleeding gastro- tissues adhesives and NSBB.
oesophageal varices?
Tripathi D, et al. Gut 2015;64:1680–1704. doi:10.1136/gutjnl-2015-309262 1699
Guidelines
Correction notice This article has been corrected since it published Online First. 10 Krajina A, Hulek P, Fejfar T, et al. Quality improvement guidelines for transjugular
The article now has an Open Access licence. intrahepatic portosystemic shunt (TIPS). Cardiovasc Intervent Radiol
2012;35:1295–300.
Acknowledgements BSG Liver Section 2013–2015—in particular, Dr Grace 11 Scope for improvement: a toolkit for a safer upper gastrointestinal bleeding (UGIB)
Dolman, trainee member of the BSG liver section for review of the guideline on service. 2011. http://www aomrc org uk/doc_download/9338-upper-
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
behalf of BSG liver section and Dr Stuart McPherson, member of the BSG liver gastrointestinal-bleeding-toolkit
section for review of the guideline on behalf of the BSG liver section. 12 McPherson S, Dyson J, Austin A, et al. Response to the NCEPOD report: development
of a care bundle for patients admitted with decompensated cirrhosis—the first 24 h.
Collaborators Mr Benedict Lowsley-Williams, UK. Patient representative. Member
Frontline Gastroenterology Published Online First: 2 Dec 2014 doi:10.1136/flgastro-
of guidelines development group. Sister Kathy Guo, research nurse, University
2014-100491
Hospitals Birmingham, Birmingham, UK. Member of guidelines development group.
13 Darzi A. High quality care for all: NHS Next Stage Review final report. London:
Contributors DT: consultant hepatologist, University Hospitals Birmingham. DoH, 2008.
Member of BSG liver section. Chair of GDG and lead author. AJS: consultant 14 Liver Quality Enhancement Service Tool (QuEST). 2015. http://www liverquest org uk/
gastroenterologist, Glasgow Royal Infirmary. Coauthor of sections ‘Secondary 15 [No authors listed]. Tackling liver disease in the UK: a Lancet Commission. Lancet
prophylaxis of variceal haemorrhage’ and ‘Gastric varices’. Review of entire 2014;384:1902.
guideline. PCH: consultant hepatologist, Royal Infirmary of Edinburgh. Coauthor of 16 Thabut D, Rudler M, Dib N, et al. Multicenter prospective validation of the Baveno
sections ‘Natural history of varices in cirrhosis’ and ‘Management of active variceal IV and Baveno II/III criteria in cirrhosis patients with variceal bleeding. Hepatology
haemorrhage’. Review of entire guideline. DP: consultant hepatologist, Royal Free 2015;61:1024–32.
Hospital, London. Coauthor of section ‘Management of active variceal 17 Ahn SY, Park SY, Tak WY, et al. Prospective validation of Baveno V definitions and
haemorrhage’. Review of entire guideline. CM: consultant hepatologist, The York criteria for failure to control bleeding in portal hypertension. Hepatology
Hospital. Member of BSG liver section. Coauthor of section ‘Management of active 2015;61:1033–40.
variceal haemorrhage’. Review of entire guideline. HM: consultant interventional 18 Christensen E, Fauerholdt L, Schlichting P, et al. Aspects of the natural history of
radiologist, University Hospitals Birmingham. Coauthor of sections ‘Service delivery gastrointestinal bleeding in cirrhosis and the effect of prednisone. Gastroenterology
and development’ and ‘Management of active variceal haemorrhage’ in relation to 1981;81:944–52.
TIPSS. AA: consultant hepatologist, Royal Derby Hospital. Derby. Chair of BSG liver 19 Cales P, Pascal JP. [Natural history of esophageal varices in cirrhosis (from origin
section. Contribution to sections ‘Secondary prophylaxis of variceal haemorrhage’ to rupture)]. Gastroenterol Clin Biol 1988;12:245–54.
and ‘Gastric varices’. Review of entire guideline. JWF: consultant hepatologist, 20 Czaja AJ, Wolf AM, Summerskill WH. Development and early prognosis of
University Hospitals Birmingham. Clinical lead of Liver QuEST. Development of esophageal varices in severe chronic active liver disease (CALD) treated with
quality indicators and implementation through Liver QuEST. SPO: consultant prednisone. Gastroenterology 1979;77(4 Pt 1):629–33.
interventional radiologist, University Hospitals Birmingham. Review of entire 21 Merli M, Nicolini G, Angeloni S, et al. Incidence and natural history of small
guideline with focus on radiological aspects. MH: consultant hepatologist, The esophageal varices in cirrhotic patients. J Hepatol 2003;38:266–72.
Freeman Hospital, Newcastle. Member of BSG liver section and BASL. Review of 22 D’Amico G, Pasta L, Morabito A, et al. Competing risks and prognostic stages of
entire guideline. JMC: consultant gastroenterologist, Royal Devon and Exeter cirrhosis: a 25-year inception cohort study of 494 patients. Aliment Pharmacol
Hospital. Member of BSG liver section. Review of entire guideline. Benedict Ther 2014;39:1180–93.
Lowsley-Williams: patient representative. Advice given relating to patient aspects. 23 Baker LA, Smith C, Lieberman G. The natural history of esophageal varices; a
Sister Kathy Guo, RGN: research nurse, University Hospitals Birmingham. Nursing study of 115 cirrhotic patients in whom varices were diagnosed prior to bleeding.
representative. Advice given relating to nursing aspects. Am J Med 1959;26:228–37.
Funding This guideline was commissioned by Clinical Services and Standards 24 Dagradi AE. The natural history of esophageal varices in patients with alcoholic
Committee (CSSC) of the British Society of Gastroenterology (BSG) under the liver cirrhosis. An endoscopic and clinical study. Am J Gastroenterol
auspices of the liver section of the BSG. 1972;57:520–40.
25 Ohnishi K, Sato S, Saito M, et al. Clinical and portal hemodynamic features in
Competing interests DT: speaker fees for Gore Medical. AA: educational grant cirrhotic patients having a large spontaneous splenorenal and/or gastrorenal shunt.
from Ferring Pharmaceuticals. PCH: speaker fees for Gore Medical. SPO: speaker Am J Gastroenterol 1986;81:450–5.
fees for Gore Medical. DP: involvement in NICE guidelines on Acute GI Bleeding. 26 Takashi M, Igarashi M, Hino S, et al. Portal hemodynamics in chronic portal-
Provenance and peer review Not commissioned; externally peer reviewed. systemic encephalopathy. Angiographic study in seven cases. J Hepatol
1985;1:467–76.
Open Access This is an Open Access article distributed in accordance with the 27 Groszmann RJ, Garcia-Tsao G, Bosch J, et al. Beta-blockers to prevent
Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which gastroesophageal varices in patients with cirrhosis. N Engl J Med
permits others to distribute, remix, adapt, build upon this work non-commercially, 2005;353:2254–61.
and license their derivative works on different terms, provided the original work is 28 Chung JW, Park S, Chung MJ, et al. A novel disposable, transnasal
properly cited and the use is non-commercial. See: http://creativecommons.org/ esophagoscope: a pilot trial of feasibility, safety, and tolerance. Endoscopy
licenses/by-nc/4.0/ 2012;44:206–9.
29 Choe WH, Kim JH, Ko SY, et al. Comparison of transnasal small-caliber vs. peroral
conventional esophagogastroduodenoscopy for evaluating varices in unsedated
REFERENCES cirrhotic patients. Endoscopy 2011;43:649–56.
1 de Franchis R. Revising consensus in portal hypertension: report of the Baveno V 30 de Franchis R, Eisen GM, Laine L, et al. Esophageal capsule endoscopy for
consensus workshop on methodology of diagnosis and therapy in portal screening and surveillance of esophageal varices in patients with portal
hypertension. J Hepatol 2010;53:762–8. hypertension. Hepatology 2008;47:1595–603.
2 Acute upper gastrointestinal bleeding: management. NICE, 2012. http://www.nice. 31 Lapalus MG, Ben Soussan E, Gaudric M, et al. Esophageal capsule endoscopy vs.
org.uk/Guidance/CG141 EGD for the evaluation of portal hypertension: a French prospective multicenter
3 Brouwers MC, Kho ME, Browman GP, et al. AGREE II: advancing guideline comparative study. Am J Gastroenterol 2009;104:1112–18.
development, reporting and evaluation in health care. CMAJ 2010;182:E839–42. 32 Vizzutti F, Arena U, Romanelli RG, et al. Liver stiffness measurement predicts
4 Centre for evidence-based medicine. Oxford: Oxford University, 2009. severe portal hypertension in patients with HCV-related cirrhosis. Hepatology
5 Jairath V, Rehal S, Logan R, et al. Acute variceal haemorrhage in the United 2007;45:1290–7.
Kingdom: patient characteristics, management and outcomes in a nationwide 33 Castera L, Le Bail B, Roudot-Thoraval F, et al. Early detection in routine clinical
audit. Dig Liver Dis 2014;46:419–26. practice of cirrhosis and oesophageal varices in chronic hepatitis C: comparison of
6 Garcia-Pagan JC, Caca K, Bureau C, et al. Early use of TIPS in patients with transient elastography (FibroScan) with standard laboratory tests and non-invasive
cirrhosis and variceal bleeding. N Engl J Med 2010;362:2370–9. scores. J Hepatol 2009;50:59–68.
7 Measuring the Units. A review of patients who died with alcohol-related liver 34 Robic MA, Procopet B, Metivier S, et al. Liver stiffness accurately predicts portal
disease. NCEPOD, 2013. http://www.ncepod.org.uk/2013report1/downloads/ hypertension related complications in patients with chronic liver disease:
Measuring%20the%20Units_full%20report.pdf a prospective study. J Hepatol 2011;55:1017–24.
8 Hobolth L, Krag A, Malchow-Moller A, et al. Adherence to guidelines in bleeding 35 Giannini EG, Zaman A, Kreil A, et al. Platelet count/spleen diameter ratio for the
oesophageal varices and effects on outcome: comparison between a specialized noninvasive diagnosis of esophageal varices: results of a multicenter, prospective,
unit and a community hospital. Eur J Gastroenterol Hepatol 2010;22:1221–7. validation study. Am J Gastroenterol 2006;101:2511–19.
9 Schlansky B, Lee B, Hartwell L, et al. Guideline adherence and outcomes in 36 Chawla S, Katz A, Attar BM, et al. Platelet count/spleen diameter ratio to predict
esophageal variceal hemorrhage: comparison of tertiary care and non-tertiary care the presence of esophageal varices in patients with cirrhosis: a systematic review.
settings. J Clin Gastroenterol 2012;46:235–42. Eur J Gastroenterol Hepatol 2012;24:431–6.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
39 Garcia-Tsao G, Groszmann RJ, Fisher RL, et al. Portal pressure, presence Med 1965;272:1255–63.
of gastroesophageal varices and variceal bleeding. Hepatology 1985;5: 65 Conn HO, Lindenmuth WW, May CJ, et al. Prophylactic portacaval anastomosis.
419–24. Medicine (Baltimore) 1972;51:27–40.
40 Lebrec D, De Fleury P, Rueff B, et al. Portal hypertension, size of esophageal 66 Jackson FC, Perrin EB, Smith AG, et al. A clinical investigation of the portacaval
varices, and risk of gastrointestinal bleeding in alcoholic cirrhosis. Gastroenterology shunt. II. Survival analysis of the prophylactic operation. Am J Surg
1980;79:1139–44. 1968;115:22–42.
41 Groszmann RJ, Bosch J, Grace ND, et al. Hemodynamic events in a prospective 67 Resnick RH, Chalmers TC, Ishihara AM, et al. A controlled study of the
randomized trial of propranolol versus placebo in the prevention of a first variceal prophylactic portacaval shunt. A final report. Ann Intern Med 1969;70:675–88.
hemorrhage. Gastroenterology 1990;99:1401–7. 68 D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-
42 Feu F, Garcia-Pagan JC, Bosch J, et al. Relation between portal pressure response analytic review. Hepatology 1995;22:332–54.
to pharmacotherapy and risk of recurrent variceal haemorrhage in patients with 69 Garcia-Tsao G, Grace ND, Groszmann RJ, et al. Short-term effects of propranolol
cirrhosis. Lancet 1995;346:1056–9. on portal venous pressure. Hepatology 1986;6:101–6.
43 Tripathi D, Therapondos G, Jackson E, et al. The role of the transjugular 70 La Mura V, Abraldes JG, Raffa S, et al. Prognostic value of acute hemodynamic
intrahepatic portosystemic stent shunt (TIPSS) in the management of bleeding response to i.v. propranolol in patients with cirrhosis and portal hypertension.
gastric varices: clinical and haemodynamic correlations. Gut 2002;51:270–4. J Hepatol 2009;51:279–87.
44 Vinel JP, Cassigneul J, Levade M, et al. Assessment of short-term prognosis after 71 Cheng JW, Zhu L, Gu MJ, et al. Meta analysis of propranolol effects on
variceal bleeding in patients with alcoholic cirrhosis by early measurement of gastrointestinal hemorrhage in cirrhotic patients. World J Gastroenterol
portohepatic gradient. Hepatology 1986;6:116–17. 2003;9:1836–9.
45 Palmer ED, Brick IB. Correlation between the severity of esophageal varices in 72 Banares R, Moitinho E, Matilla A, et al. Randomized comparison of long-term
portal cirrhosis and their propensity toward hemorrhage. Gastroenterology carvedilol and propranolol administration in the treatment of portal hypertension in
1956;30:85–90. cirrhosis. Hepatology 2002;36:1367–73.
46 Polio J, Groszmann RJ. Hemodynamic factors involved in the development and 73 Hobolth L, Moller S, Gronbaek H, et al. Carvedilol or propranolol in portal
rupture of esophageal varices: a pathophysiologic approach to treatment. Semin hypertension? A randomized comparison. Scand J Gastroenterol 2012;47:467–74.
Liver Dis 1986;6:318–31. 74 Reiberger T, Ulbrich G, Ferlitsch A, et al. Carvedilol for primary prophylaxis of
47 Vegesna AK, Chung CY, Bajaj A, et al. Minimally invasive measurement of variceal bleeding in cirrhotic patients with haemodynamic non-response to
esophageal variceal pressure and wall tension (with video). Gastrointest Endosc propranolol. Gut 2013;62:1634–41.
2009;70:407–13. 75 Tripathi D, Ferguson JW, Kochar N, et al. Randomized controlled trial of carvedilol
48 Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal hemorrhage by versus variceal band ligation for the prevention of the first variceal bleed.
esophageal endoscopy. Gastrointest Endosc 1981;27:213–18. Hepatology 2009;50:825–33.
49 North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. 76 Shah HA, Azam Z, Rauf J, et al. Carvedilol vs. esophageal variceal band ligation in
Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver the primary prophylaxis of variceal hemorrhage: a multicentre randomized
and esophageal varices. A prospective multicenter study. N Engl J Med controlled trial. J Hepatol 2014;60:757–64.
1988;319:983–9. 77 Serste T, Melot C, Francoz C, et al. Deleterious effects of beta-blockers on
50 Inokuchi K. Improved survival after prophylactic portal nondecompression surgery survival in patients with cirrhosis and refractory ascites. Hepatology
for esophageal varices: a randomized clinical trial. Cooperative Study Group of 2010;52:1017–22.
Portal Hypertension of Japan. Hepatology 1990;12:1–6. 78 Krag A, Wiest R, Albillos A, et al. The window hypothesis: haemodynamic and
51 de Franchis R, Primignani M, Arcidiacono PG, et al. Prophylactic sclerotherapy in non-haemodynamic effects of beta-blockers improve survival of patients with
high-risk cirrhotics selected by endoscopic criteria. A multicenter randomized cirrhosis during a window in the disease. Gut 2012;61:967–9.
controlled trial. Gastroenterology 1991;101:1087–93. 79 Leithead JA, Rajoriya N, Tehami N, et al. Non-selective beta-blockers are
52 Feu F, Ruiz del Arbol L, Banares R, et al. Double-blind randomized controlled trial associated with improved survival in patients with ascites listed for liver
comparing terlipressin and somatostatin for acute variceal hemorrhage. Variceal transplantation. Gut 2015;64:1111–9.
Bleeding Study Group. Gastroenterology 1996;111:1291–9. 80 Mandorfer M, Bota S, Schwabl P, et al. Nonselective beta blockers increase risk
53 Merkel C, Bolognesi M, Bellon S, et al. Prognostic usefulness of hepatic vein for hepatorenal syndrome and death in patients with cirrhosis and spontaneous
catheterization in patients with cirrhosis and esophageal varices. Gastroenterology bacterial peritonitis. Gastroenterology 2014;146:1680–90.
1992;102:973–9. 81 Navasa M, Chesta J, Bosch J, et al. Reduction of portal pressure by isosorbide-5-
54 Villanueva C, Aracil C, Colomo A, et al. Acute hemodynamic response to beta- mononitrate in patients with cirrhosis. Effects on splanchnic and systemic
blockers and prediction of long-term outcome in primary prophylaxis of variceal hemodynamics and liver function. Gastroenterology 1989;96:1110–18.
bleeding. Gastroenterology 2009;137:119–28. 82 Angelico M, Carli L, Piat C, et al. Isosorbide-5-mononitrate versus propranolol
55 Stokkeland K, Brandt L, Ekbom A, et al. Improved prognosis for patients in the prevention of first bleeding in cirrhosis. Gastroenterology 1993;104:
hospitalized with esophageal varices in Sweden 1969–2002. Hepatology 1460–5.
2006;43:500–5. 83 Garcia-Pagan JC. Isosorbide-5-monitrate (ISMN) vs placebo (PLA) in the prevention
56 [No authors listed]. Propranolol prevents first gastrointestinal bleeding in non- of the first variceal bleeding (FVB) in patients with contraindications or intolerance
ascitic cirrhotic patients. Final report of a multicenter randomized trial. The Italian to beta-blockers. J Hepatol 2000;32:28.
Multicenter Project for Propranolol in Prevention of Bleeding (IMPP). J Hepatol 84 Merkel C, Marin R, Enzo E, et al. Randomised trial of nadolol alone or with isosorbide
1989;9:75–83. mononitrate for primary prophylaxis of variceal bleeding in cirrhosis. Gruppo-Triveneto
57 Conn HO, Grace ND, Bosch J, et al. Propranolol in the prevention of the first per L’ipertensione portale (GTIP). Lancet 1996;348:1677–81.
hemorrhage from esophagogastric varices: a multicenter, randomized clinical trial. 85 Garcia-Pagan JC, Morillas R, Banares R, et al. Propranolol plus placebo versus
The Boston-New Haven-Barcelona Portal Hypertension Study Group. Hepatology propranolol plus isosorbide-5-mononitrate in the prevention of a first variceal
1991;13:902–12. bleed: a double-blind RCT. Hepatology 2003;37:1260–6.
58 Ideo G, Bellati G, Fesce E, et al. Nadolol can prevent the first gastrointestinal 86 Hidaka H, Nakazawa T, Wang G, et al. Long-term administration of PPI reduces
bleeding in cirrhotics: a prospective, randomized study. Hepatology 1988;8:6–9. treatment failures after esophageal variceal band ligation: a randomized, controlled
59 Lebrec D, Poynard T, Capron JP, et al. Nadolol for prophylaxis of gastrointestinal trial. J Gastroenterol 2012;47:118–26.
bleeding in patients with cirrhosis. A randomized trial. J Hepatol 1988;7:118–25. 87 Min YW, Lim KS, Min BH, et al. Proton pump inhibitor use significantly increases
60 Pascal JP, Cales P. Propranolol in the prevention of first upper gastrointestinal tract the risk of spontaneous bacterial peritonitis in 1965 patients with cirrhosis and
hemorrhage in patients with cirrhosis of the liver and esophageal varices. N Engl J ascites: a propensity score matched cohort study. Aliment Pharmacol Ther
Med 1987;317:856–61. 2014;40:695–704.
61 Cales P, Oberti F, Payen JL, et al. Lack of effect of propranolol in the prevention of 88 Terg R, Casciato P, Garbe C, et al. Proton pump inhibitor therapy does
large oesophageal varices in patients with cirrhosis: a randomized trial. French- not increase the incidence of spontaneous bacterial peritonitis in cirrhosis:
Speaking Club for the Study of Portal Hypertension. Eur J Gastroenterol Hepatol a multicenter prospective study. J Hepatol 2014. Doi:10.1016/j.jhep.
1999;11:741–5. 2014.11.036
62 Merkel C, Marin R, Angeli P, et al. A placebo-controlled clinical trial of nadolol in 89 Dultz G, Piiper A, Zeuzem S, et al. Proton pump inhibitor treatment is associated
the prophylaxis of growth of small esophageal varices in cirrhosis. with the severity of liver disease and increased mortality in patients with cirrhosis.
Gastroenterology 2004;127:476–84. Aliment Pharmacol Ther 2015;41:459–66.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
92 Funakoshi N, Duny Y, Valats JC, et al. Meta-analysis: beta-blockers versus banding of esophageal variceal bleeding. A meta-analysis. Ann Intern Med
ligation for primary prophylaxis of esophageal variceal bleeding. Ann Hepatol 1995;123:280–7.
2012;11:369–83. 122 Ljubicic N, Biscanin A, Nikolic M, et al. A randomized-controlled trial of endoscopic
93 Debernardi VW, Elia C, Stradella D, et al. Prospective randomized trial: endoscopic treatment of acute esophageal variceal hemorrhage: N-butyl-2-cyanoacrylate
follow up 3 vs 6 months after esophageal variceal eradication by band ligation in injection vs. variceal ligation. Hepatogastroenterology 2011;58:438–43.
cirrhosis. Eur J Intern Med 2014;25:674–9. 123 Ibrahim M, El Mikkawy A, Mostafa I, et al. Endoscopic treatment of acute variceal
94 Avgerinos A, Armonis A, Manolakopoulos S, et al. Endoscopic sclerotherapy plus hemorrhage by using hemostatic powder TC-325: a prospective pilot study.
propranolol versus propranolol alone in the primary prevention of bleeding in high Gastrointest Endosc 2013;78:769–73.
risk cirrhotic patients with esophageal varices: a prospective multicenter 124 Banares R, Albillos A, Rincon D, et al. Endoscopic treatment versus endoscopic
randomized trial. Gastrointest Endosc 2000;51:652–8. plus pharmacologic treatment for acute variceal bleeding: a meta-analysis.
95 Gotoh Y, Iwakiri R, Sakata Y, et al. Evaluation of endoscopic variceal ligation in Hepatology 2002;35:609–15.
prophylactic therapy for bleeding of oesophageal varices: a prospective, controlled 125 Lo GH, Lai KH, Cheng JS, et al. Emergency banding ligation versus sclerotherapy
trial compared with endoscopic injection sclerotherapy. J Gastroenterol Hepatol for the control of active bleeding from esophageal varices. Hepatology
1999;14:241–4. 1997;25:1101–4.
96 Masumoto H, Toyonaga A, Oho K, et al. Ligation plus low-volume sclerotherapy 126 Villanueva C, Piqueras M, Aracil C, et al. A randomized controlled trial comparing
for high-risk esophageal varices: comparisons with ligation therapy or ligation and sclerotherapy as emergency endoscopic treatment added to
sclerotherapy alone. J Gastroenterol 1998;33:1–5. somatostatin in acute variceal bleeding. J Hepatol 2006;45:560–7.
97 Graham DY, Smith JL. The course of patients after variceal hemorrhage. 127 Panes J, Teres J, Bosch J, et al. Efficacy of balloon tamponade in treatment of
Gastroenterology 1981;80:800–9. bleeding gastric and esophageal varices. Results in 151 consecutive episodes.
98 Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict survival in Dig Dis Sci 1988;33:454–9.
patients with end-stage liver disease. Hepatology 2001;33:464–70. 128 Teres J, Cecilia A, Bordas JM, et al. Esophageal tamponade for bleeding varices.
99 D’Amico G, de Franchis R. Upper digestive bleeding in cirrhosis. Post-therapeutic Controlled trial between the Sengstaken-Blakemore tube and the Linton-Nachlas
outcome and prognostic indicators. Hepatology 2003;38:599–612. tube. Gastroenterology 1978;75:566–9.
100 Abraldes JG, Villanueva C, Banares R, et al. Hepatic venous pressure gradient and 129 Dechene A, El Fouly AH, Bechmann LP, et al. Acute management of refractory
prognosis in patients with acute variceal bleeding treated with pharmacologic and variceal bleeding in liver cirrhosis by self-expanding metal stents. Digestion
endoscopic therapy. J Hepatol 2008;48:229–36. 2012;85:185–91.
101 Bambha K, Kim WR, Pedersen R, et al. Predictors of early re-bleeding and mortality 130 Wright G, Lewis H, Hogan B, et al. A self-expanding metal stent for complicated
after acute variceal haemorrhage in patients with cirrhosis. Gut 2008;57:814–20. variceal hemorrhage: experience at a single center. Gastrointest Endosc 2010;71:
102 Hunter SS, Hamdy S. Predictors of early re-bleeding and mortality after acute 71–8.
variceal haemorrhage. Arab J Gastroenterol 2013;14:63–7. 131 Vangeli M, Patch D, Burroughs AK. Salvage tips for uncontrolled variceal bleeding.
103 Reverter E, Tandon P, Augustin S, et al. A MELD-based model to determine risk of J Hepatol 2002;37:703–4.
mortality among patients with acute variceal bleeding. Gastroenterology 132 Rosemurgy AS, Frohman HA, Teta AF, et al. Prosthetic H-graft portacaval shunts vs
2014;146:412–19. transjugular intrahepatic portasystemic stent shunts: 18-year follow-up of a
104 Al Freah MA, Gera A, Martini S, et al. Comparison of scoring systems and randomized trial. J Am Coll Surg 2012;214:445–53.
outcome of patients admitted to a liver intensive care unit of a tertiary referral 133 Orloff MJ, Vaida F, Haynes KS, et al. Randomized controlled trial of emergency
centre with severe variceal bleeding. Aliment Pharmacol Ther 2014;39:1286–300. transjugular intrahepatic portosystemic shunt versus emergency portacaval shunt
105 Chen PH, Chen WC, Hou MC, et al. Delayed endoscopy increases re-bleeding and treatment of acute bleeding esophageal varices in cirrhosis. J Gastrointest Surg
mortality in patients with hematemesis and active esophageal variceal bleeding: 2012;16:2094–111.
a cohort study. J Hepatol 2012;57:1207–13. 134 Bureau C, Pagan JC, Layrargues GP, et al. Patency of stents covered with
106 Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper polytetrafluoroethylene in patients treated by transjugular intrahepatic
gastrointestinal bleeding. N Engl J Med 2013;368:11–21. portosystemic shunts: long-term results of a randomized multicentre study. Liver Int
107 Scalea TM, Holman M, Fuortes M, et al. Central venous blood oxygen saturation: 2007;27:742–7.
an early, accurate measurement of volume during hemorrhage. J Trauma 135 Perarnau JM, Le Gouge A, Nicolas C, et al. Covered vs. uncovered stents for
1988;28:725–32. transjugular intrahepatic portosystemic shunt: a randomized controlled trial.
108 Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl J Med J Hepatol 2014;60:962–8.
2011;365:147–56. 136 Monescillo A, Martinez-Lagares F, Ruiz-del-Arbol L, et al. Influence of portal
109 Bendtsen F, D’Amico G, Rusch E, et al. Effect of recombinant Factor VIIa on hypertension and its early decompression by TIPS placement on the outcome of
outcome of acute variceal bleeding: an individual patient based meta-analysis of variceal bleeding. Hepatology 2004;40:793–801.
two controlled trials. J Hepatol 2014;61:252–9. 137 Garcia-Pagan JC, Di Pascoli M, Caca K, et al. Use of early-TIPS for high-risk
110 Wells M, Chande N, Adams P, et al. Meta-analysis: vasoactive medications for variceal bleeding: results of a post-RCT surveillance study. J Hepatol
the management of acute variceal bleeds. Aliment Pharmacol Ther 2012;35:1267–78. 2013;58:45–50.
111 Seo YS, Park SY, Kim MY, et al. Lack of difference among terlipressin, 138 Rudler M, Cluzel P, Corvec TL, et al. Early-TIPSS placement prevents rebleeding in
somatostatin, and octreotide in the control of acute gastroesophageal variceal high-risk patients with variceal bleeding, without improving survival. Aliment
hemorrhage. Hepatology 2014;60:954–63. Pharmacol Ther 2014;40:1074–80.
112 Groszmann RJ, Kravetz D, Bosch J, et al. Nitroglycerin improves the hemodynamic 139 Gatta A, Merkel C, Sacerdoti D, et al. Nadolol for prevention of variceal rebleeding
response to vasopressin in portal hypertension. Hepatology 1982;2:757–62. in cirrhosis: a controlled clinical trial. Digestion 1987;37:22–8.
113 Ioannou G, Doust J, Rockey DC. Terlipressin for acute esophageal variceal 140 D’Amico G, Pagliaro L, Bosch J. Pharmacological treatment of portal hypertension:
hemorrhage. Cochrane Database Syst Rev 2003;(1):CD002147. an evidence-based approach. Semin Liver Dis 1999;19:475–505.
114 Azam Z, Hamid S, Jafri W, et al. Short course adjuvant terlipressin in acute 141 Gournay J, Masliah C, Martin T, et al. Isosorbide mononitrate and propranolol
variceal bleeding: a randomized double blind dummy controlled trial. J Hepatol compared with propranolol alone for the prevention of variceal rebleeding.
2012;56:819–24. Hepatology 2000;31:1239–45.
115 Bosch J, Kravetz D, Rodes J. Effects of somatostatin on hepatic and systemic 142 Gluud LL, Langholz E, Krag A. Meta-analysis: isosorbide-mononitrate alone or with
hemodynamics in patients with cirrhosis of the liver: comparison with vasopressin. either beta-blockers or endoscopic therapy for the management of oesophageal
Gastroenterology 1981;80:518–25. varices. Aliment Pharmacol Ther 2010;32:859–71.
116 Chavez-Tapia NC, Barrientos-Gutierrez T, Tellez-Avila F, et al. Meta-analysis: 143 Lo GH, Chen WC, Wang HM, et al. Randomized, controlled trial of carvedilol
antibiotic prophylaxis for cirrhotic patients with upper gastrointestinal bleeding— versus nadolol plus isosorbide mononitrate for the prevention of variceal
an updated Cochrane review. Aliment Pharmacol Ther 2011;34:509–18. rebleeding. J Gastroenterol Hepatol 2012;27:1681–7.
117 Fernandez J, Ruiz del Arbol L, Gomez C, et al. Norfloxacin vs ceftriaxone in the 144 Abraldes JG, Villanueva C, Aracil C, et al. Addition of simvastatin to standard
prophylaxis of infections in patients with advanced cirrhosis and hemorrhage. treatment improves survival after variceal bleeding in patients with cirrhosis.
Gastroenterology 2006;131:1049–56. A double-blind randomized trial. J Hepatol 2014;60:S525.
118 Lo GH, Perng DS, Chang CY, et al. Controlled trial of ligation plus vasoconstrictor 145 Shaheen NJ, Stuart E, Schmitz SM, et al. Pantoprazole reduces the size of
versus proton pump inhibitor in the control of acute esophageal variceal bleeding. postbanding ulcers after variceal band ligation: a randomized, controlled trial.
J Gastroenterol Hepatol 2013;28:684–9. Hepatology 2005;41:588–94.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
biweekly endoscopic variceal ligation for the prevention of esophageal variceal (EVL) in the treatment of recurrent variceal bleeding. Surg Endosc
rebleeding. J Gastroenterol Hepatol 2014;29:1229–36. 2013;27:2712–20.
149 Karsan HA, Morton SC, Shekelle PG, et al. Combination endoscopic band ligation 174 Neuberger J, Gimson A, Davies M, et al. Selection of patients for liver
and sclerotherapy compared with endoscopic band ligation alone for the transplantation and allocation of donated livers in the UK. Gut 2008;57:252–7.
secondary prophylaxis of esophageal variceal hemorrhage: a meta-analysis. Dig Dis 175 Sarin SK, Lahoti D, Saxena SP, et al. Prevalence, classification and natural history
Sci 2005;50:399–406. of gastric varices: a long-term follow-up study in 568 portal hypertension patients.
150 Singh P, Pooran N, Indaram A, et al. Combined ligation and sclerotherapy versus Hepatology 1992;16:1343–9.
ligation alone for secondary prophylaxis of esophageal variceal bleeding: a meta- 176 Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices.
analysis. Am J Gastroenterol 2002;97:623–9. Gastrointest Endosc 1986;32:264–8.
151 Lo GH, Chen WC, Chen MH, et al. Banding ligation versus nadolol and isosorbide 177 Gimson AE, Westaby D, Williams R. Endoscopic sclerotherapy in the management
mononitrate for the prevention of esophageal variceal rebleeding. Gastroenterology of gastric variceal haemorrhage. J Hepatol 1991;13:274–8.
2002;123:728–34. 178 Harada T, Yoshida T, Shigemitsu T, et al. Therapeutic results of endoscopic variceal
152 Lo GH, Chen WC, Lin CK, et al. Improved survival in patients receiving medical ligation for acute bleeding of oesophageal and gastric varices. J Gastroenterol
therapy as compared with banding ligation for the prevention of esophageal Hepatol 1997;12:331–5.
variceal rebleeding. Hepatology 2008;48:580–7. 179 Takeuchi M, Nakai Y, Syu A, et al. Endoscopic ligation of gastric varices. Lancet
153 Patch D, Sabin CA, Goulis J, et al. A randomized, controlled trial of medical 1996;348:1038.
therapy versus endoscopic ligation for the prevention of variceal rebleeding in 180 Binmoeller KF, Weilert F, Shah JN, et al. EUS-guided transesophageal treatment of
patients with cirrhosis. Gastroenterology 2002;123:1013–19. gastric fundal varices with combined coiling and cyanoacrylate glue injection (with
154 Romero G, Kravetz D, Argonz J, et al. Comparative study between nadolol and videos). Gastrointest Endosc 2011;74:1019–25.
5-isosorbide mononitrate vs. endoscopic band ligation plus sclerotherapy in the 181 Cheng LF, Wang ZQ, Li CZ, et al. Low incidence of complications from endoscopic
prevention of variceal rebleeding in cirrhotic patients: a randomized controlled gastric variceal obturation with butyl cyanoacrylate. Clin Gastroenterol Hepatol
trial. Aliment Pharmacol Ther 2006;24:601–11. 2010;8:760–6.
155 Stanley AJ, Dickson S, Hayes PC, et al. Multicentre randomised controlled study 182 Hung HH, Chang CJ, Hou MC, et al. Efficacy of non-selective beta-blockers as
comparing carvedilol with variceal band ligation in the prevention of variceal adjunct to endoscopic prophylactic treatment for gastric variceal bleeding:
rebleeding. J Hepatol 2014;61:1014–19. a randomized controlled trial. J Hepatol 2012;56:1025–32.
156 Li L, Yu C, Li Y. Endoscopic band ligation versus pharmacological therapy for variceal 183 Kahloon A, Chalasani N, Dewitt J, et al. Endoscopic therapy with 2-octyl-
bleeding in cirrhosis: a meta-analysis. Can J Gastroenterol 2011;25:147–55. cyanoacrylate for the treatment of gastric varices. Dig Dis Sci 2014;59:
157 Cheung J, Zeman M, van Zanten SV, et al. Systematic review: secondary 2178–83.
prevention with band ligation, pharmacotherapy or combination therapy 184 Kanagawa H, Mima S, Kouyama H, et al. Treatment of gastric fundal varices by
after bleeding from oesophageal varices. Aliment Pharmacol Ther balloon-occluded retrograde transvenous obliteration. J Gastroenterol Hepatol
2009;30:577–88. 1996;11:51–8.
158 Ding SH, Liu J, Wang JP. Efficacy of beta-adrenergic blocker plus 5-isosorbide 185 Lee YT, Chan FK, Ng EK, et al. EUS-guided injection of cyanoacrylate for bleeding
mononitrate and endoscopic band ligation for prophylaxis of esophageal variceal gastric varices. Gastrointest Endosc 2000;52:168–74.
rebleeding: a meta-analysis. World J Gastroenterol 2009;15:2151–5. 186 Lo GH, Lai KH, Cheng JS, et al. A prospective, randomized trial of butyl
159 Ravipati M, Katragadda S, Swaminathan PD, et al. Pharmacotherapy plus cyanoacrylate injection versus band ligation in the management of bleeding gastric
endoscopic intervention is more effective than pharmacotherapy or endoscopy varices. Hepatology 2001;33:1060–4.
alone in the secondary prevention of esophageal variceal bleeding: a meta-analysis 187 Mishra SR, Chander SB, Kumar A, et al. Endoscopic cyanoacrylate injection versus
of randomized, controlled trials. Gastrointest Endosc 2009;70:658–64. beta-blocker for secondary prophylaxis of gastric variceal bleed: a randomised
160 Gonzalez R, Zamora J, Gomez-Camarero J, et al. Meta-analysis: combination controlled trial. Gut 2010;59:729–35.
endoscopic and drug therapy to prevent variceal rebleeding in cirrhosis. Ann Intern 188 Oho K, Iwao T, Sumino M, et al. Ethanolamine oleate versus butyl
Med 2008;149:109–22. cyanoacrylate for bleeding gastric varices: a nonrandomized study. Endoscopy
161 Funakoshi N, Segalas-Largey F, Duny Y, et al. Benefit of combination beta-blocker 1995;27:349–54.
and endoscopic treatment to prevent variceal rebleeding: a meta-analysis. World J 189 Rajoriya N, Forrest EH, Gray J, et al. Long-term follow-up of endoscopic histoacryl
Gastroenterol 2010;16:5982–92. glue injection for the management of gastric variceal bleeding. QJM
162 Thiele M, Krag A, Rohde U, et al. Meta-analysis: banding ligation and medical 2011;104:41–7.
interventions for the prevention of rebleeding from oesophageal varices. Aliment 190 Ramond MJ, Valla D, Mosnier JF, et al. Successful endoscopic obturation of gastric
Pharmacol Ther 2012;35:1155–65. varices with butyl cyanoacrylate. Hepatology 1989;10:488–93.
163 Puente A, Hernández-Gea V, Graupera I, et al. Drugs plus ligation to prevent 191 Romero-Castro R, Ellrichmann M, Ortiz-Moyano C, et al. EUS-guided coil versus
rebleeding in cirrhosis: an updated systematic review. Liver Int 2014;34:823–33. cyanoacrylate therapy for the treatment of gastric varices: a multicenter study (with
164 Luca A, D’Amico G, La Galla R, et al. TIPS for prevention of recurrent bleeding in videos). Gastrointest Endosc 2013;78:711–21.
patients with cirrhosis: meta-analysis of randomized clinical trials. Radiology 192 Soehendra N, Grimm H, Nam VC, et al. N-butyl-2-cyanoacrylate: a supplement to
1999;212:411–21. endoscopic sclerotherapy. Endoscopy 1987;19:221–4.
165 Papatheodoridis GV, Goulis J, Leandro G, et al. Transjugular intrahepatic 193 Tan PC, Hou MC, Lin HC, et al. A randomized trial of endoscopic treatment of
portosystemic shunt compared with endoscopic treatment for prevention of variceal acute gastric variceal hemorrhage: N-butyl-2-cyanoacrylate injection versus band
rebleeding: a meta-analysis. Hepatology 1999;30:612–22. ligation. Hepatology 2006;43:690–7.
166 Zheng M, Chen Y, Bai J, et al. Transjugular intrahepatic portosystemic shunt versus 194 Tantau M, Crisan D, Popa D, et al. Band ligation vs. N-butyl-2-cyanoacrylate
endoscopic therapy in the secondary prophylaxis of variceal rebleeding in cirrhotic injection in acute gastric variceal bleeding: a prospective follow-up study. Ann
patients: meta-analysis update. J Clin Gastroenterol 2008;42:507–16. Hepatol 2013;13:75–83.
167 Garcia-Villarreal L, Martinez-Lagares F, Sierra A, et al. Transjugular intrahepatic 195 Williams SG, Peters RA, Westaby D. Thrombin–an effective treatment for gastric
portosystemic shunt versus endoscopic sclerotherapy for the prevention of variceal variceal haemorrhage. Gut 1994;35:1287–9.
rebleeding after recent variceal hemorrhage. Hepatology 1999;29:27–32. 196 Przemioslo RT, McNair A, Williams R. Thrombin is effective in arresting bleeding
168 Qi X, Liu L, Bai M, et al. Transjugular intrahepatic portosystemic shunt in from gastric variceal hemorrhage. Dig Dis Sci 1999;44:778–81.
combination with or without variceal embolization for the prevention of variceal 197 Ramesh J, Limdi JK, Sharma V, et al. The use of thrombin injections in the
rebleeding: a meta-analysis. J Gastroenterol Hepatol 2014;29:688–96. management of bleeding gastric varices: a single-center experience. Gastrointest
169 Orloff MJ, Isenberg JI, Wheeler HO, et al. Liver transplantation in a randomized Endosc 2008;68:877–82.
controlled trial of emergency treatment of acutely bleeding esophageal varices in 198 McAvoy NC, Plevris JN, Hayes PC. Human thrombin for the treatment of gastric
cirrhosis. Transplant Proc 2010;42:4101–8. and ectopic varices. World J Gastroenterol 2012;18:5912–17.
170 Henderson JM, Boyer TD, Kutner MH, et al. Distal splenorenal shunt versus 199 Smith MR, Tidswell R, Tripathi D. Outcomes of endoscopic human thrombin
transjugular intrahepatic portal systematic shunt for variceal bleeding: injection in the management of gastric varices. Eur J Gastroenterol Hepatol
a randomized trial. Gastroenterology 2006;130:1643–51. 2014;26:846–52.
171 Boyer TD, Henderson JM, Heerey AM, et al. Cost of preventing variceal rebleeding 200 Holster IL, Poley JW, Kuipers EJ, et al. Controlling gastric variceal bleeding with
with transjugular intrahepatic portal systemic shunt and distal splenorenal shunt. endoscopically applied hemostatic powder (Hemospray). J Hepatol
J Hepatol 2008;48:407–14. 2012;57:1397–8.
Gut: first published as 10.1136/gutjnl-2015-309262 on 17 April 2015. Downloaded from http://gut.bmj.com/ on June 3, 2020 by guest. Protected by copyright.
Gastroenterol Clin Biol 1993;17:431–4. J Roentgenol 2007;189:W365–72.
203 Sanyal AJ, Freedman AM, Luketic VA, et al. Transjugular intrahepatic portosystemic 216 Tanihata H, Minamiguchi H, Sato M, et al. Changes in portal systemic pressure
shunts for patients with active variceal hemorrhage unresponsive to sclerotherapy. gradient after balloon-occluded retrograde transvenous obliteration of gastric varices
Gastroenterology 1996;111:138–46. and aggravation of esophageal varices. Cardiovasc Intervent Radiol 2009;32:1209–16.
204 Stanley AJ, Jalan R, Ireland HM, et al. A comparison between gastric and 217 Sabri SS, Abi-Jaoudeh N, Swee W, et al. Short-term rebleeding rates for isolated gastric
oesophageal variceal haemorrhage treated with transjugular intrahepatic varices managed by transjugular intrahepatic portosystemic shunt versus balloon-
portosystemic stent shunt (TIPSS). Aliment Pharmacol Ther 1997;11:171–6. occluded retrograde transvenous obliteration. J Vasc Interv Radiol 2014;25:355–61.
205 Barange K, Peron JM, Imani K, et al. Transjugular intrahepatic portosystemic shunt 218 Saad WE. The history and evolution of balloon-occluded retrograde transvenous
in the treatment of refractory bleeding from ruptured gastric varices. Hepatology obliteration (BRTO): from the United States to Japan and back. Semin Intervent
1999;30:1139–43. Radiol 2011;28:283–7.
206 Chau TN, Patch D, Chan YW, et al. “Salvage” transjugular intrahepatic 219 Wang J, Tian XG, Li Y, et al. Comparison of modified percutaneous transhepatic
portosystemic shunts: gastric fundal compared with esophageal variceal bleeding. variceal embolization and endoscopic cyanoacrylate injection for gastric variceal
Gastroenterology 1998;114:981–7. rebleeding. World J Gastroenterol 2013;19:706–14.
207 Lo GH, Liang HL, Chen WC, et al. A prospective, randomized controlled trial of 220 Orozco H, Mercado MA. The evolution of portal hypertension surgery: lessons
transjugular intrahepatic portosystemic shunt versus cyanoacrylate injection in the from 1000 operations and 50 years’ experience. Arch Surg 2000;135:1389–93.
prevention of gastric variceal rebleeding. Endoscopy 2007;39:679–85. 221 Jaroszewski DE, Schlinkert RT, Gray RJ. Laparoscopic splenectomy for the treatment
208 Procaccini NJ, Al Osaimi AM, Northup P, et al. Endoscopic cyanoacrylate versus of gastric varices secondary to sinistral portal hypertension. Surg Endosc
transjugular intrahepatic portosystemic shunt for gastric variceal bleeding: a single- 2000;14:87.
center US analysis. Gastrointest Endosc 2009;70:881–7. 222 Stone PA, Phang D, Richmond B, et al. Splenic artery embolization for the
209 Mahadeva S, Bellamy MC, Kessel D, et al. Cost-effectiveness of N-butyl-2- treatment of bleeding gastric varices secondary to splenic vein thrombosis.
cyanoacrylate (histoacryl) glue injections versus transjugular intrahepatic Ann Vasc Surg 2014;28:737–11.
portosystemic shunt in the management of acute gastric variceal bleeding. Am J 223 Greig JD, Garden OJ, Anderson JR, et al. Management of gastric variceal
Gastroenterol 2003;98:2688–93. haemorrhage. Br J Surg 1990;77:297–9.
210 Koito K, Namieno T, Nagakawa T, et al. Balloon-occluded retrograde transvenous 224 Riemenschneider T, Bermel RE, Hirner A. [Results of devascularization surgery of
obliteration for gastric varices with gastrorenal or gastrocaval collaterals. AJR Am J the gastroesophageal junction in recurrent hemorrhage of esophageal and fundus
Roentgenol 1996;167:1317–20. varices]. Zentralbl Chir 1994;119:291–7.
211 Choi YH, Yoon CJ, Park JH, et al. Balloon-occluded retrograde transvenous 225 Thomas PG, D’Cruz AJ. Distal splenorenal shunting for bleeding gastric varices.
obliteration for gastric variceal bleeding: its feasibility compared with transjugular Br J Surg 1994;81:241–4.
intrahepatic portosystemic shunt. Korean J Radiol 2003;4:109–16. 226 Mishra SR, Sharma BC, Kumar A, et al. Primary prophylaxis of gastric variceal
212 Hong CH, Kim HJ, Park JH, et al. Treatment of patients with gastric bleeding comparing cyanoacrylate injection and beta-blockers: a randomized
variceal hemorrhage: endoscopic N-butyl-2-cyanoacrylate injection versus balloon- controlled trial. J Hepatol 2011;54:1161–7.
occluded retrograde transvenous obliteration. J Gastroenterol Hepatol 227 Kang EJ, Jeong SW, Jang JY, et al. Long-term result of endoscopic Histoacryl
2009;24:372–8. (N-butyl-2-cyanoacrylate) injection for treatment of gastric varices. World J
213 Jang SY, Kim GH, Park SY, et al. Clinical outcomes of balloon-occluded retrograde Gastroenterol 2011;17:1494–500.
transvenous obliteration for the treatment of gastric variceal hemorrhage in Korean 228 Katoh K, Sone M, Hirose A, et al. Balloon-occluded retrograde transvenous
patients with liver cirrhosis: a retrospective multicenter study. Clin Mol Hepatol obliteration for gastric varices: the relationship between the clinical outcome and
2012;18:368–74. gastrorenal shunt occlusion. BMC Med Imaging 2010;10:2.