REVIEW ARTICLE
A comprehensive approach to the management of acute endoscopic
perforations (with videos)
Todd H. Baron, MD, FASGE,1 Louis M. Wong Kee Song, MD,1 Martin D. Zielinski, MD, FACS,2
Fabian Emura, MD, PhD, FASGE,3 Mehran Fotoohi, MD,4 Richard A. Kozarek, MD, FASGE4
Rochester, Minnesota, USA
The only method to prevent iatrogenic luminal perfo-
ration at the time of endoscopy is the avoidance of endo-
scopic procedures. Luminal perforation is among the most
feared adverse events of GI endoscopy, and the rationale
for this is multifactorial: (1) it may carry significant mor-
bidity and mortality; (2) perforation may not be remedia-
ble without a surgical procedure; (3) management of per-
foration requires a multidisciplinary approach, often
beyond the endoscopist’s control; (4) perforation usually
requires hospitalization, thus adding to the cost of care;
and (5) perforation has medicolegal implications, includ-
ing liability not only for the endoscopist but other provid-
ers involved with the case.1 In addition, the mere mention
of the word perforation triggers alarm and panic among
faculty and trainees alike, often with the assumption that
all perforations require surgical management. In some
cases, there may even be disagreement among involved
specialties as to the approach toward perforation.
With the development of endoscopic devices and tech-
niques, such as natural orifice transluminal endoscopic
Abbreviations: ESD, endoscopic submucosal dissection; FCSEMS, fully
covered self-expandable metal stent; FDA, U.S. Food and Drug Admin-
istration; NOTES, natural orifice transluminal endoscopic surgery;
OTSC, over-the-scope clip; PCSEMS, partially covered self-expandable
metal stent; PEG, percutaneous endoscopic gastrostomy; TTS, through-
the-scope.
DISCLOSURES: T. Baron has received travel and research support from
Boston Scientific and is a consultant and speaker for Cook Endoscopy. L.
Wong Kee Song and F. Emura are consultants for Boston Scientific. R.
Kazarek has received research support from Boston Scientific and hon-
oraria from Cook Endoscopy. No other financial relationships relevant to
this publication were disclosed.
Use your mobile device to scan this
QR code and watch the author in-
terview. Download a free QR code
scanner by searching ‘QR Scanner’
in your mobile device’s app store.
See CME section; p. 835.
Copyright © 2012 by the American Society for Gastrointestinal Endoscopy
0016-5107/$36.00
http://dx.doi.org/10.1016/j.gie.2012.04.476
838 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
surgery (NOTES), luminal perforation is a component of
the procedure that can be readily managed endoscopi-
cally.2 On the other hand, surgical exploration and repair
are generally required in patients in whom endoscopic
measures are unsuccessful or technically not feasible in
closing the perforation or in whom recognition of perfo-
ration is delayed. The challenge lies in determining which
patients can undergo nonsurgical treatment alternatives
and knowing when to realize whether the management
chosen is successful or not.
Despite the extensive, peer-reviewed literature on endo-
scopic perforations, data from controlled clinical trials do not
exist to provide the clinician with consensus management
algorithms regarding the care of the patient with an endo-
scopically induced perforation. In this review, we outline
the general principles regarding the approach to endo-
scopic perforations, discuss the anatomic peculiarities
of perforations throughout the GI tract, and provide a
framework for management of luminal endoscopic
perforations.
GENERAL PRINCIPLES
Table 1 summarizes the general principles related to
endoscopic perforation. We consider these the “Ten Com-
mandments of Endoscopic Perforation.” An understanding
of these basic concepts guides the establishment of proper
management strategies.
(1) Prompt recognition of endoscopic perforation
is essential to improvement in outcome
Early recognition of perforation, preferably during the pro-
cedure, allows prompt intervention that is critical for a successful
outcome. Identification of an iatrogenic perforation during the
procedure allows one not only to limit the accumulation of
extraluminal air by minimizing insufflation but, more impor-
tantly, to prevent the egress of luminal contents by rapid closure
of the perforation, diversion, or immediate referral for surgery,
as appropriate. Delayed identification of perforation typically
results in egress of luminal contents that lead to infection and a
worse outcome.3-5
(2) The presence of extraluminal air does not
automatically mean the need for surgery
Extraluminal air is the hallmark of endoscopic perfora-
tion and may be incidental (asymptomatic) or symptom-
www.giejournal.org
Baron et al
TABLE 1. The ten commandments of endoscopic
perforation
1. Prompt recognition of endoscopic perforation is
essential to improvement in outcome.
2. The presence of extraluminal air does not
automatically mean the need for surgery.
3. The volume of extraluminal air is not necessarily
proportional to the size of the perforation.
4. Extraluminal air per se is not infectious.
5. Extraluminal air under pressure is a medical
emergency.
6. Extraluminal air can dissect into distant spaces.
7. Residual extraluminal air may persist without clinical
significance.
8. Perforations tend to close after drainage or diversion
of luminal contents.
9. Oral, rectal, or injected contrast material extravasation
should elicit prompt intervention.
10. Failed endoscopic closure of a perforation generally
requires surgical intervention.
atic. For certain procedures, such as percutaneous endo-
scopic gastrostomy (PEG) tube placement, the presence of
extraluminal air is expected and inconsequential.6 Inci-
dental extraluminal air is generally identified when radio-
graphs, such as abdominal radiographs and CT, are ob-
tained after uncomplicated endoscopy.7,8 Typically, the
patient undergoes a prescheduled imaging study pursuant
to the problem already being investigated. In the absence
of concerning symptoms or signs, specific treatment is
usually not necessary, although serial examination and
monitoring are recommended.
A small volume of extraluminal air without any endo-
scopically visible sign of perforation may occur in some
endoscopic submucosal dissection (ESD) procedures as-
sociated with large dissections, long procedure times, and
after-exposure or minimal damage of the muscularis pro-
pria. During gastric ESD, for example, localized extralumi-
nal air close to the gastric wall may be visible on CT but
not on plain radiographs, and this has been defined as a
transmural air leak.9,10 In some ERCP procedures, the use
of compressed air to maintain luminal patency may result
in asymptomatic retroperitoneal air, which does not re-
quire surgical intervention.8,11. The use of carbon dioxide
rather than air for insufflation may minimize the amount of
extraluminal air because the former is rapidly absorbed.12
This may prevent the development of overt symptoms in
the presence of a small perforation or transmural air leak
or diminish the severity of presentation after a perforation,
although data are lacking to support this.
www.giejournal.org
Acute endoscopic perforations
(3) The volume of extraluminal air is not neces-
sarily proportional to the size of the perforation
The volume of extraluminal air present is not deter-
mined primarily by the size of the perforation but by the
degree of air insufflation after perforation. A very small
perforation may be associated with a large volume of free
air. Conversely, a large hole that is recognized immedi-
ately and followed by prompt removal of the endoscope
may not generate any free air at all. The volume of ex-
traluminal air is directly related to the duration of the
procedure and degree of insufflation performed after the
perforation has occurred11,13 and whether or not the per-
foration is contained by surrounding tissues.
(4) Extraluminal air per se is not infectious
Free air in and of itself is not infectious. Extraluminal air
will resolve spontaneously as long as the perforation
closes, and enteric contents do not egress.6-8 The spillage
of luminal contents outside the GI tract accounts for the
inflammatory/infectious response and will cause symp-
toms due to contamination of the sterile extraluminal
spaces by oral and gut flora. Patients with signs or symp-
toms suggestive of perforation after endoscopy including
pain, abdominal distention, fever, and tachycardia should
undergo urgent imaging, such as abdominal radiography,
CT, and/or a contrast upper GI series. Confirmation of free
air, along with the earlier-mentioned signs and symptoms,
generally requires surgical intervention.14,15
(5) Extraluminal air under pressure is a medical
emergency
Although extraluminal air does not directly introduce
infectious elements to sterile planes, free air under pres-
sure can result in a medical emergency. Examples include
tension pneumothorax, tension pneumomediastinum, ten-
sion pneumopericardium, and abdominal compartment
syndrome.16-18 Treatment to decompress the cavity under
pressure should be carried out immediately (see the
following).
(6) Extraluminal air can dissect into distant
spaces
Air can remain trapped in a single compartmental space
or localized in a specific soft tissue, but it also can diffuse
rapidly through diverse and distant anatomical planes,
potentially resulting in a life-threatening complication.19
Colorectal perforations, for example, can be manifested by
subcutaneous emphysema in the neck (Video 1, available
online at www.giejournal.org).16,17 The degree and reach
of air dissection are related to ongoing air insufflation and
long procedure times.
(7) Residual extraluminal air may persist without
clinical significance
Free air can persist and continue to disperse into
tissues for a variable amount of time after closure of the
perforation (Fig. 1). Postoperative pneumoperitoneum
usually resolves within a week20 but may persist for 10
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 839
Acute endoscopic perforations
Figure 1. (A) Abdominal CT obtained for follow-up of seemingly un-
complicated transmural drainage of pancreatic fluid collection 3 weeks
earlier shows massive intraperitoneal air. Patient was asymptomatic and
remained well.
to 24 days or, rarely, several weeks without any clinical
consequence.6,20
(8) Perforations tend to close after drainage or
diversion of luminal contents
With prompt recognition of a perforation, egress of
luminal contents can be prevented by closure of the per-
foration and/or diversion of luminal contents away from
the perforated site, depending on the site (stomach versus
duodenum, eg) and size of perforation. Diversion is more
readily achieved in the esophagus, stomach, and duode-
num by using a properly positioned nasoenteric tube
placed to suction. Colon contents cannot be effectively
diverted nonsurgically. Thus, nonsurgical management of
colon perforations is limited to techniques that enable
closure of the perforation.
(9) Free oral or injected contrast material extrav-
asation should elicit prompt intervention
The demonstration of free oral contrast material extrav-
asation on imaging studies correlates with easy spillage of
luminal contents into the extraluminal space. Urgent clo-
sure, diversion of luminal contents (if possible), or oper-
ative intervention is required.21
(10) Failed endoscopic closure of a perforation
generally requires surgical intervention
Surgical exploration and repair of perforations are typ-
ically required when endoscopic closure or luminal con-
tent diversion cannot be performed because of operator
inexperience, technical difficulty, or lack of endoscopic
accessories. In selected cases, nonoperative management
can be successful by using percutaneous drain placement.
However, if nonoperative management fails, the outcome
may be worse than those cases treated initially with oper-
ative repair.21
840 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
TABLE 2. Early indicators of endoscopic perforation
Subcutaneous air/crepitus
Chest pain
Abdominal distention
Pneumothorax
Endoscopically difficult to maintain endoluminal air/
luminal collapse
Hemodynamic instability (hypotension, tachycardia)
Respiratory distress (oxygen desaturation, tachypnea)
TABLE 3. Delayed signs and symptoms of perforation
(>24 hours after procedure)
Systemic inflammatory response (fever, leucocytosis,
tachycardia, tachypnea)
Acute abdominal pain (peritoneal irritation)
Nausea/vomiting
Back pain/flank pain
Unexpected localized swelling (neck, scrotum)
Abdominal distention
Pneumothorax
Severe chest pain
Inability to handle oral secretions
Shortness of breath
Hypotension
Mental confusion
DIAGNOSIS AND INITIAL MANAGEMENT
Although an iatrogenic perforation is readily evident at
the time of the procedure in most cases, a high index of
suspicion is required in some patients in whom an
endoscopic-related perforation has occurred. This allows
for early diagnosis and subsequent optimal management
and, thus, better prognosis. Symptoms may be masked
initially in certain circumstances, such as in the overly
sedated and elderly patient with multiple comorbidities,
making the diagnosis more difficult. Early diagnosis in-
creases the likelihood that the patient can be managed
nonoperatively, which may translate into a shorter hospital
stay and fewer or less harmful adverse events (Table 2).
Clinical suspicion for a perforation that is not readily ap-
parent initially should be heightened in the presence of
ongoing abdominal distention/pain, chest pain, shortness
of breath, or subcutaneous emphysema.17 In cases of clin-
www.giejournal.org
Baron et al
Figure 2. A, Water-soluble contrast study after attempted clip closure of
large, proximal, esophageal perforation. No extravasation is seen. B,
Chest CT obtained later the same day shows contrast material outside the
esophageal lumen consistent with an active leak (arrow).
ical deterioration hours after an endoscopic procedure,
delayed perforation should be considered (Table 3).
An initial imaging assessment should include a chest
and a flat/upright abdominal radiograph, which may dem-
onstrate pneumothorax, pneumomediastinum, subcutane-
ous emphysema, and free or retroperitoneal air.22 If plain
films are unrevealing in the setting of suspected perfora-
tion, water-soluble contrast studies (administered orally,
via nasogastric or nasoduodenal tube or per rectum) may
show contained or free contrast material extravasation,
although chest/abdominal CT scans may be more sensitive
at detecting free air and small or contained leaks (Fig. 2).23
Once a perforation is diagnosed, endoscopic closure
should be attempted if feasible. In cases of successful
closure or where conservative treatment alone is consid-
ered, concomitant measures include cessation of oral in-
take; administration of intravenous fluid and broad spec-
trum antibiotics; monitoring of vital signs, temperature,
and inflammatory markers (eg, leukocyte count); and se-
rial clinical examination with special attention to inflam-
matory response and signs of mediastinitis24 or peritonitis.
Placement of a nasogastric tube or nasoduodenal tube is
recommended in patients with gastric, duodenal, or prox-
www.giejournal.org
Acute endoscopic perforations
TABLE 4. Management of a suspected/known
perforation
Endoluminal closure (metallic clips, stents) for immediate
or early postprocedure recognition
Nasogastric or nasoduodenal decompression
Abdominal/chest decompression if there is air under
tension
Radiographic evaluation
Intravenous administration of broad-spectrum antibiotics
Fluid resuscitation
Pain medication
Nothing by mouth
Close clinical observation
Percutaneous or surgical intervention
imal small-bowel perforations to reduce the volume and
egress of GI contents. Blind insertion of a nasogastric tube
is not recommended in cases of esophageal perforation
because it may exit the perforated site (Table 4). A multi-
disciplinary approach is essential, with periodic assess-
ment of the need for surgical treatment in the event of
clinical deterioration or percutaneous drainage of symp-
tomatic fluid collections.25
Management of air under tension
Decompression should be performed emergently when
extraluminal air is under tension. A tension pneumothorax
requires immediate needle decompression of the affected
site with a needle catheter inserted along the midclavicular
line in the second intercostal space.26,27 A chest tube
should be placed after relief of the tension pneumothorax.
Moderate, nontension pneumothorax does not necessitate
needle decompression but requires chest tube or percuta-
neous catheter placement for drainage. A small asymptom-
atic pneumothorax, defined as a radiographic rim of air
around the lung, can be managed conservatively.27
Although subcutaneous emphysema is not clinically
hazardous and usually resolves in a few days, massive air
tracking into soft tissues of the neck can result in airway
obstruction and may require endotracheal intubation. Al-
teration in phonation should alert the clinician to this
possibility. Because it cannot be evacuated endoscopi-
cally, asymptomatic pneumomediastinum after esopha-
geal or hypopharyngeal perforation should be managed
conservatively. When the perforation is successfully man-
aged, it usually resolves, along with the subcutaneous
emphysema in the neck.
Abdominal compartment syndrome can develop as a re-
sult of tension pneumoperitoneum, causing unstable vital
signs, tense abdomen, and respiratory distress. Insertion of
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 841
Acute endoscopic perforations
an 18 or 20 – gauge trocar needle (eg, standard PEG tube kit
trocar) into the abdomen with the patient in the supine
position will result in rapid release of air and improvement in
hemodynamics.28,29 After skin cleansing with a topical anti-
septic agent, the trocar needle should be inserted in either
lower abdominal quadrants just at or inferior to the umbilicus
to minimize injury to the urinary bladder and other solid
organs. The needle should be removed while the plastic
sheath is left in situ to allow for continuous decompression of
the peritoneal cavity as the procedure resumes and endo-
scopic closure is attempted (Video 2, available online at
www.giejournal.org). This maneuver avoids the need for
decompressive celiotomy and enables the use of endoscopic
air insufflation for visualization and performance of thera-
pies, such as endoscopic clip placement.
MANAGEMENT OF PERFORATION
Endoscopic
The principal determinant for a successful nonsurgical
outcome is early recognition of the perforation and insti-
tution of measures to prevent egress of intraluminal con-
tents as soon as possible. Several endoscopic devices have
been used successfully for closure of perforations, and
some are approved for use by the U.S. Food and Drug
Administration (FDA) for this purpose. These include en-
doscopic clips, stents, and endoscopic suturing devices.
Endoscopic clip placement is currently the standard
method for closing luminal perforations.30 Several through-
the-scope (TTS) clips are available, including the QuickClip2
(Olympus Inc, Center Valley, Pa), the Resolution clip (Boston
Scientific Inc, Natick, Mass), and the Tri-Clip and Instinct clip
(Cook Medical, Winston-Salem, NC). As yet, there are no
comparative human studies establishing the superiority of
one type of clip over another for closure of perforations, and
the selection of a particular type of clipping device should be
based on characteristics of the defect, device availability,
experience, cost, and operator’s preference. Both the oper-
ator and assistant should be familiar with the use and han-
dling of a particular device for a successful outcome. Maneu-
vers to enhance successful clip closure include (1) gentle air
suction to approximate the borders of the perforation and
capture of more tissue within the prongs of the opened clip,
(2) initiating clip closure of the edges at the far end of the
defect and proceeding toward the proximal border of the
defect, and (3) placement of as many clips as needed in a
zipper fashion to ensure adequate closure. For larger gastric
defects, one can place TTS clips around the circumference of
the perforation and lasso them together with a detachable
plastic snare (Endo-loop; Olympus).31 Another method for
closing large perforations is clip closure of an omental patch
brought through the perforation to the border of the defect,
similar to the Graham’s patch used in surgical procedures.32
The over-the-scope clip (OTSC; Ovesco Endoscopy AG,
Tubingen, Germany) is significantly different in design com-
pared with TTS clips, with higher compression force and
842 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
capacity to capture a larger volume of tissue. It is FDA
approved for closure of luminal perforations ⬍20 mm in size.
Set-up and deployment of the OTSC is similar to that of a
variceal band ligator. Dedicated TTS grasping or anchoring
devices can be used to approximate tissue margins and
retract more tissue into the cap before clip release. Animal
studies have shown superior outcomes during closure of
NOTES defects and colonic perforations by using the OTSC
as compared with TTS clips.33,34 Ex vivo studies have shown
that full-thickness gastric defects 5 to 20 mm in size and colon
defects 10 to 30 mm in size can be closed with a single
OTSC.35 In recent clinical studies, surgery was avoided in
approximately 90% of cases of endoscopic perforations man-
aged with the OTSC.36,37
Stents, including partially covered self-expandable metal
stents (PCSEMS), fully covered self-expandable metal stents
(FCSEMS), and self-expandable plastic stents, have been
used to seal esophageal and other upper GI tract perfora-
tions38 (including anastomotic perforations after endoscopic
dilation) (Video 3, available online at www.giejournal.org),
although they are not FDA approved for this purpose in the
setting of benign disease. The selection of a particular type of
stent is influenced by several factors, including device avail-
ability, operator preference, lesion features (location, diam-
eter of defect, concomitant stricture), and risk for stent mi-
gration. The optimal stent dwell time to facilitate sealing of
perforations has not been established and ranges from 2 to 12
weeks among various studies. Our practice is to leave the
stent in place a minimum of 2 weeks and up to 1 month if the
underlying tissue is otherwise healthy (ie, no radiation dam-
age or malignancy).
Several endoscopic suturing devices have been devel-
oped for NOTES and antireflux and bariatric procedures,
including T-tags (Ethicon Endo-Surgery, Cincinnati, OH,
USA),30 pursed-string modified T-tags (Cook Endoscopy,
Winston-Salem, NC, USA),39 the Overstitch suturing sys-
tem (Apollo Endosurgery, Austin, TX, USA),40 the pursed-
string-suturing device (LSI Solutions, Victor, New York,
USA), the flexible Endostitch (Covidien, Mansfield, MA,
USA),41 the NDO Plicator (NDO Surgical Inc, Mansfield,
MA, USA),42 and the flexible stapler (Power Medical Inter-
ventions, Langhorne, PA, USA).43 Some of these devices
may prove suitable for closure of defects throughout the
GI tract, but experience regarding their use for closure of
acute perforations remains limited at this time.44
When an endoscopic procedure is performed with fluo-
roscopic assistance, water-soluble contrast material pre-
loaded into a 60-mL syringe can be injected rapidly through
the working channel of the endoscope to assess for a leak
before and after attempted endoscopic closure. Contrast ma-
terial extravasation is diagnostic for an active leak. It is im-
portant to note, however, that if a CT scan is obtained shortly
after such a maneuver, the radiologist performing the CT
must be aware of prior contrast material use because contrast
material still present from the endoscopic procedure may be
www.giejournal.org
Baron et al Acute endoscopic perforations

mistaken for active extravasation when oral or rectal contrast

material is administered for the CT.

Interventional radiologic
Once the source of the perforation has been addressed,
continued monitoring of the patient can ensure early de-
tection of complications of perforation. If the patient de-
velops worsening symptoms including fever, chills, ab-
dominal pain, or leukocytosis, imaging of the site is
indicated. CT has a high level of sensitivity for detection of
early fluid collections in patients with perforation.23 If a
symptomatic fluid collection is diagnosed, percutaneous
drainage is an effective way to prevent further deteriora-
tion of clinical symptoms.25 The size of the catheter (8F-
28F) will depend on the viscosity and debris within the
fluid collection. The number of catheters inserted is de-
pendent on the number and size of the fluid collections.
The catheters will need to be flushed with 10 to 15 mL of
saline solution frequently (every 8 hours) to prevent oc-
clusion. If the patient’s symptoms do not improve, or they
worsen, a repeat CT scan of the site with intravenous and
oral contrast material as well as fluoroscopic catheter in-
jection need to be performed. This will reveal the presence
of any undrained or new fluid collections. Injection of
contrast material into the catheter under fluoroscopic
guidance can reveal any fistula to the perforated site. If a
fistula is present, it can be addressed by endoscopic means
or by downsizing the catheter and sequentially pulling the
catheter back from the site of the leak over a period of 2
to 3 weeks. This has to be done with close monitoring of
the patient’s symptoms because downsizing or pulling
back the catheter can cause reaccumulation of the fluid
collection and recurrence of symptoms.45,46

Surgical
As previously mentioned, enteric contents that escape
the lumen after endoscopic perforation are a source of
infection for areas that are normally sterile. In symptomatic
patients, surgical intervention is generally required when
endoscopic closure and/or stent diversion is not possible
or is unsuccessful, and percutaneous drainage has failed.
Surgical management, however, will be influenced by se-
verity of illness, comorbid conditions, available local ex-
pertise, and the presence or absence of an ongoing leak.
Surgical management includes perforation closure or di-
version, drainage and lavage of collections, and partial/
complete removal of the compromised organ.
MANAGEMENT OF SPECIFIC TYPES OF
PERFORATIONS
Hypopharyngeal/proximal esophageal
perforations
Perforation of the hypopharynx and most proximal
esophagus is a rare complication of upper endoscopy,47
often not recognized during the procedure, but rather after
Figure 3. A, Perforation after flexible endoscopic cricopharyngeal my-
otomy for Zenker’s diverticulum. CT scan after oral contrast shows
extensive subcutaneous air but no contrast material extravasation. B,
Water-soluble contrast study shows small contained extravasation. The
patient did well with medical management and was discharged home 4
days later.
the procedure by symptoms such as sore throat, dyspha-
gia, or subcutaneous emphysema. In most cases, this iat-
rogenic perforation occurs as a result of operator tech-
nique and interventional procedures, such as endoscopic
removal of ingested foreign bodies.48 Blind passage
of nasogastric tubes, duodenoscopes, echoendoscopes,49
and transesophageal echocardiography probes50 also can
result in perforation of the hypopharynx. Finally, inadver-
tent passage through a Zenker’s diverticulum51 or other

www.giejournal.org Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 843

Acute endoscopic perforations Baron et al

hypopharyngeal pouch and endoscopic cricopharyngeal

myotomy also can lead to perforation.52
Confirmation of a hypopharyngeal perforation requires
neck CT with administration of oral contrast material
and/or esophagogram with water-soluble contrast mate-
rial (Fig. 3). Patients with confirmed hypopharyngeal per-
forations are treated conservatively with nasogastric tube
feeding, suspension of oral intake, and parenteral antibi-
otics. Surprisingly, there are scant data to guide optimal
therapy for this type of perforation, and a review on sword
swallowing injuries suggests that conservative manage-
ment of a hypopharyngeal endoscopic perforation is con-
sidered a safe management strategy.53 In our series, all
patients with perforation could be managed without sur-
gery after Zenker’s diverticulum therapy.52 The manage-
ment of upper esophageal perforations, defined for the
purpose of this review as those occurring just below the
cricopharyngeus, also can be managed without surgery
with or without with endoscopic TTS clip closure (as
outlined in the following), although endotherapy may be
technically difficult because of the confined working
space. Stent placement is not usually technically feasible
because of the inability to bridge the perforation site.

Esophageal perforations beyond the proximal

esophagus
Esophageal perforation is an infrequent complication of
upper endoscopy and rarely occurs in the absence of
underlying pathology or endoscopic therapy, such as di-
lation. Other factors including operator technique, impac-
tion of the endoscope against large osteophytes in the
cervical esophagus,54 and severe retching during the pro-
cedure (Fig. 4 and Video 4, available online at www.
giejournal.org) may result in perforation. Passage of trans-
esophageal echocardiography probes also can result in
perforation. In the setting of underlying pathology, perfo-
ration may be caused by insertion of the endoscope (or
transesophageal echocardiography probe) into an unrec-
ognized obstruction, by application of excessive force
when negotiating benign or malignant strictures or dis-
lodging a food bolus, after dilation of strictures and
webs, and after pneumatic dilation for achalasia. Fi-
nally, endoscopic interventions such as foreign body
removal, polypectomy, endoscopic marsupialization of an
esophageal duplication cyst, EMR, and ESD may result in
perforation.
In cases of suspected esophageal perforation, a chest
radiograph often shows extraluminal air within the pleural
space, mediastinum, or subcutaneous tissues. To confirm a
suspected perforation, an esophagogram with the use of Figure 4. A, Contrast swallowing study for post-procedure chest pain
water-soluble contrast material or a chest CT with admin- shows free extravasation at the distal esophagus (arrow). B, Endoscopic
istration of oral contrast material are useful diagnostic view of perforation. C, Fully covered self-expandable metal stent placed
methods. These imaging studies can determine the site across perforation. A contrast swallowing study was obtained after stent
and size of the perforation and whether or not the perfo- removal 4 weeks later and was normal.
ration is contained. The presence of contrast material ex-
travasation and/or extraluminal fluid, as assessed by chest

844 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012 www.giejournal.org

Baron et al Acute endoscopic perforations

Figure 5. Endoscopic perforation induced by EUS insertion. A, Radiographic image taken during contrast material injection through the working
channel of the endoscope shows active extravasation. B, Endoscopic image of perforation centered within the cap of the over-the-scope clip. C,
Endoscopic image of deployed clip. D, Radiographic image shows no extravasation (arrow).

CT, confirms the perforation, and treatment should be
initiated immediately.27,55 Treatment options depend on
the clinical setting and severity of symptoms and include
conservative management, endoscopic repair, surgical in-
tervention, or a combination of these measures. Conser-
vative treatment consists of nothing by mouth, intravenous
fluids, and broad spectrum antibiotics, with or without
nasogastric tube placement. Conservative approaches
without surgery are being used more frequently with im-
proved outcomes.56
When an esophageal perforation is recognized during
the procedure or early (within hours) after the procedure,
the principles of closure and/or diversion are used. Perfo-
rations that are found during the procedure can be closed
by using TTS endoscopic clip placement57,58 with or with-
out temporary stent placement (Video 5, available online
at www.giejournal.org).59 Often, multiple TTS clips are
required, and more than one endoscopic session may
be needed. The recently available OTSCs are likely to be
more effective for closure of esophageal perforations and
have longer clip retention times (Fig. 5).36,37,60 Large,
esophageal perforations can be sealed by temporary
placement of self-expandable stents,61-68 although their
use in this setting remains off-label. Placement of FCSEMSs
or self-expandable plastic stents across the gastroesopha-
geal junction typically results in stent migration in the
stomach, particularly in the absence of a stricture or shelf
to anchor the stent. In this circumstance, placement of a
PCSEMS is preferable to minimize the risk for migration,
although subsequent removal of the stent may necessitate
the stent-in-stent technique.69 In patients with inoperable
malignant esophageal obstruction and iatrogenic perfora-

www.giejournal.org Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 845

Acute endoscopic perforations
Figure 6. Chest CT shows bilateral large pleural effusions with right
hydropneumothorax from same patient in Figure 2. A percutaneous drain
was placed in the right pleural cavity. Patient was successfully managed
nonoperatively. A, air; F, fluid.
tion, PCSEMSs or FCSEMSs are suitable both for sealing of
the perforation and palliation of dysphagia.70
Postendoscopy contrast studies are recommended to
document and confirm successful sealing or closure of the
perforation, either at the time of the procedure (Fig. 5D) or
within 24 hours after the procedure. There are little data to
guide timing of resumption of oral intake after successful
conservative/endoscopic treatment and satisfactory clini-
cal progress. In two studies, oral intake was initiated at
approximately 3 days after covered stent placement for
esophageal perforations.71,72 We believe oral intake of
liquids can be resumed on the fourth or fifth day. Surgical
intervention is mandatory if endoscopic closure or stent
diversion is not possible or is unsuccessful or in the setting
of clinical deterioration despite endoscopic and percuta-
neous (see the following) intervention.
Esophageal perforations that are recognized late (ⱖ12
hours) or fail to close with conservative therapy are gen-
erally symptomatic. The symptoms are associated with
fluid tracking into the mediastinum and/or pleural cavity.
In general, these patients should be considered for surgi-
cal intervention if it is confirmed by CT imaging that
extraluminal leakage of GI contents has occurred, which
cannot be managed easily percutaneously. The manage-
ment depends on the severity of illness of the patient,
available local expertise, and the presence or absence of
an ongoing leak. When symptoms and laboratory test
findings suggest only mild abnormalities, however, con-
servative treatment alone may be a reasonable initial ap-
proach, assuming that the CT scan shows minimal extralu-
minal fluid. In a recent study of 57 patients with
esophageal perforations, approximately two-thirds were
treated ⬍24 hours after perforation (early) and the remain-
der ⬎24 hours (late). Length of stay was lower in the early
treatment group, but there was no difference in adverse
events or mortality compared with the late-treatment
group. During the study period, the use of nonoperative
846 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
therapy increased from 0% to 75% and, interestingly, was
used more often in the late-treatment group presenting
⬎24 hours after perforation (46% vs 38%).56
Interventional radiologic considerations
Esophageal perforation can lead to formation of empy-
ema or fluid collections in the mediastinum (Fig. 6). In
some patients, percutaneous drainage can be a safe and
effective method of treatment.73
Care has to be taken in the initial approach to medias-
tinal fluid collections because there are many vital organs
that can make access to the collection difficult. After place-
ment of a percutaneous catheter, output of the catheter
needs to be monitored. Persistent output may indicate the
presence of a fistula, which may be confirmed by injection
of contrast material through the catheter under fluoro-
scopic guidance or by a radiographic swallowing study.
Repeat CT scan and catheter assessment should be per-
formed about 1 week after catheter placement to reveal
persistent or new collections as well as assess for catheter
dysfunction. In case of formation of an empyema, a chest
tube can be placed. These collections often can be com-
plex, with significant fibrinous content. This may be ad-
dressed by increasing the size of the chest tube and infu-
sion of thrombolytics through the chest tube.74 Persistent
collections not responding to percutaneous treatment may
require surgical intervention.
Surgical considerations
The presence or absence of septic shock and whether
the perforation is free or contained are principal determi-
nants for surgical intervention. Hemodynamically stable
patients without evidence of end-organ damage may be
treated by endoscopic and/or percutaneous techniques as
mentioned earlier, but frequent re-evaluations based on
vital signs, physical examination, and laboratory study
results (eg, white blood cell count) must be undertaken.
Those patients who are, or who become, unstable dur-
ing nonoperative management should undergo surgical
exploration.
Patients who demonstrate a noncontained or free per-
foration may undergo expedient endoscopic intervention
initially, but a follow-up study (upper GI or CT) that
continues to demonstrate free extravasation of oral con-
trast material is an indication for surgical exploration.72
The surgical approach is dependent on the anatomic
location of the perforation. The esophagus is unique
among GI organs in that it courses through the neck and
traverses the thoracic and abdominal cavities. Perforations
of the cervical esophagus will require cervical incisions on
the same side as the perforation. The approach to perfo-
rations within the thoracic cavity also is dependent on the
location of the perforation. Patients with evidence of a
leak in the proximal two-thirds of the esophagus should
undergo right thoracotomy, whereas those with perfora-
tions in the lower third should undergo left thoracotomy.
www.giejournal.org
Baron et al
An upper midline celiotomy or thoracoabdominal ap-
proach may be required for perforations at the gastro-
esophageal junction.
The type of surgical procedure performed is critically
related to the timing of the perforation and the quality of
the tissues. Generally, patients who present within 24
hours of the perforation have limited inflammatory reac-
tion. In this situation, primary repair with healthy tissue
flaps as a buttress may be performed. If, however, severe
mediastinitis is encountered, then debridement of devital-
ized tissue, esophageal diversion/resection combined with
creation of a healthy muscle flap, wide drainage, and
gastrostomy tube placement should be performed, regard-
less of timing of the perforation. For perforations associ-
ated with large esophageal carcinomas, end-stage achala-
sia, and strictures (peptic and caustic), primary repair
should be avoided. The distal obstruction conferred by
these conditions will inhibit healing and result in fistula
formation. Endoscopic stents may be attempted to bridge
the obstruction(s). If treatment is unsuccessful or the per-
foration is not contained, resection with immediate recon-
struction in stable patients should be performed. Patients
with septic shock should undergo diversion with feeding
tube placement (gastrostomy or jejunostomy).
Gastric perforation
Gastric perforation in the absence of underlying pathol-
ogy is rare. When it occurs, it is commonly due to operator
technique and impaction of the endoscope against the
anterior wall of the stomach, as occasionally occurs during
ERCP. Intentional perforations include PEG tube place-
ment and transmural endoscopic interventions, such as
transmural drainage of extramural collections (eg, pancre-
atic pseudocyst and necrosectomy), although these can be
complicated by “unintentional” perforation. Other causes
of perforation include standard snare polypectomy, EMR,
ESD, dilation of gastroenteric anastomotic strictures, and
overdistension by gas (barotrauma) during argon plasma
coagulation or cryotherapy. Gastric barotrauma may go
unrecognized if the treatment site was the esophagus.
Among all sites within the GI tract, the stomach is the
most robust and forgiving after a perforation, as evidenced
by rapid closure of PEG tracts after tube removal. An
otherwise healthy stomach can close with simple diversion
by using nasogastric suction (Fig. 7). With regard to risk
factors for gastric perforation, the thinner-walled proximal
stomach and the presence of an ulcer or unhealthy (eg,
irradiated) tissue are associated with a higher risk of per-
foration during gastric therapeutic procedures.75,76
When the perforation is recognized during the proce-
dure or soon (within hours) after the procedure, the prin-
ciples of closure and/or diversion are again used. For
small, endoscopically identifiable perforations identified
during the procedure, endoscopic clipping is an accepted
method of treatment (Video 6, available online at
www.giejournal.org). For larger perforations (eg, immedi-
www.giejournal.org
Acute endoscopic perforations
ately after ESD) and when the intra-abdominal omentum is
visible, the latter can be brought through the defect and
clipped to the edges of the defect (omental patch) (Fig. 8).
Perforations occurring after dilation of gastrojejunal anas-
tomotic or antral strictures can be managed by placement
of removable self-expandable plastic stents or FCSEMSs.
Placement of a nasogastric tube also is used to decompress
the stomach and should be placed to suction as soon as
possible. Placement under endoscopic view enables pre-
cise positioning of the tube at the site of the perforation.
After endoscopic management of the perforation, it is
advisable to obtain an imaging study, either abdominal CT or
UGI study using water-soluble contrast material, to confirm
the absence of an ongoing leak. Percutaneous drainage and
surgical intervention are the next steps when endoscopic
closure or stent diversion is not possible or is unsuccessful.
Asymptomatic gastric perforations that are recognized
late (ⱖ12 hours) by the presence of free air on plain films
but not recognizable during an endoscopic procedure,
such as ESD, usually can be managed conservatively with
nothing by mouth, gastric suction, and administration of
antibiotics. These “perforations” also are known as trans-
mural air leaks. In a recent report of 33 patients with
transmural air leaks, none of the patients showed clinical
evidence of peritonitis or required interventions beyond
conservative management. Twelve patients had fever
⬎37.5°C, probably because of the transmural inflamma-
tory effect (similar to postpolypectomy syndrome), and
were managed with antibiotic therapy alone.21 If signs of
peritonitis appear, the necessity for surgical management
is reassessed. If the patient clinically improves, oral intake
is allowed after resolution of fever.
Late recognition of gastric perforations in the presence
of symptoms of inflammatory response is generally asso-
ciated with peritonitis caused by luminal contents leaking
into the peritoneal cavity. Abdominal CT enables identifi-
cation of extraluminal air and/or intra-abdominal fluid.
These patients generally require surgical closure of the
perforation, lavage, and evacuation of the contaminants.
Symptomatic perforations after transgastric drainage of
pancreatic fluid collections may be related to aggressive
balloon dilation. If air alone is present (retroperitoneal
and/or intraperitoneal), conservative measures are war-
ranted if the site of drainage is the antrum or body, be-
cause the muscular gastric wall usually will close (Fig. 7).
Placement of a covered self-expandable metal stent to seal
the leak may be effective.77 The presence of extraluminal
fluid warrants percutaneous therapy or surgery.
Interventional radiologic considerations
In patients with gastric perforation in whom percutaneous
drainage is indicated, this is often performed under CT guid-
ance. The catheter needs to be manipulated to the leak site if
possible. The size and number of catheters are dictated by
the viscosity and size of the fluid collections. If the patient’s
symptoms do not improve, repeat CT scan and catheter
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 847
Acute endoscopic perforations Baron et al

Figure 7. Gastric perforation, mid-body, posterior wall after large-diameter dilation into PFC. A and B, 20-mm balloon was used to dilate the posterior
gastric wall over a guidewire. C, Endoscopic view of the catheter through the posterior wall outside of the collection. The patient developed tension
pneumoperitoneum during the procedure, which was managed with needle catheter decompression and nasogastric tube placement. D, CT scan with
water-soluble contrast material via a nasogastric tube shows extensive free air (A) but no extravasation. The patient did well without surgery, and the
PFC resolved. PFC, pancreatic fluid collection.

assessment may be needed. Catheter upsizing may be done
if the characteristics of the fluid change. The initial drained
fluid is often not very viscous. Persistent leakage can cause
necrosis of the perigastric fat and produce particulate debris
within the collection that often occludes the drainage cathe-
ter. The catheter should be upsized and flushed with 10 to 15
mL of normal saline solution every 8 hours to prevent cath-
eter obstruction. If the patient does not respond to percuta-
neous drainage and endoscopic therapy, surgical interven-
tion should be considered.
Surgical considerations
Clinical signs and symptoms may be sufficient to diag-
nose an intra-abdominal gastric perforation and the need
for emergent surgical exploration. Patients who present
with sepsis and peritonitis should undergo immediate ex-
ploration. In those patients without signs of sepsis or
peritonitis, a large amount of free intra-peritoneal fluid or
demonstration of free contrast material extravasation on
upper GI studies or CT despite endoscopic therapy should
prompt immediate surgical exploration. In the absence of
these signs, the patient may be observed safely with serial
examination.
Surgical exploration can be performed laparoscopically
or via celiotomy, dependent on the surgeon’s comfort with
either approach. Wedge resection can be performed if
anatomically feasible. Alternatively, resection with Billroth
I or II reconstruction can be performed if the tissue integ-
rity is acceptable based on the severity of inflammation.

848 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012 www.giejournal.org

Baron et al
Figure 8. A, Lesser curvature early gastric cancer. B, Gastric perforation
after endoscopic submucosal dissection, with visible omentum through
the defect. C, Omentum was brought into the gastric lumen and clipped
to the gastric mucosa (omental patch). Courtesy of Dr Hiroyuki Ono,
Shizuoka Cancer Center, Japan.
The perforation should be oversewn only if carcinoma has
been ruled out.
Duodenal perforation
The approach to duodenal perforation depends
on whether the perforation is periampullary, as may
occur during pancreaticobiliary interventions, or not
periampullary.
www.giejournal.org
Acute endoscopic perforations
Periampullary perforation. Periampullary perfora-
tions are rare, carry high mortality, and are usually ERCP-
related.13 Perforations vary in size and severity. Manage-
ment is variable, with some patients requiring immediate
surgery, although most do well with conservative treat-
ment only.78 Diagnosis requires a high index of clinical
suspicion to allow for early detection and optimal man-
agement. Although treatment increasingly involves non-
operative management, some practitioners still advocate
surgical repair of all duodenal perforations. Conservative
measures may include nasogastric or nasoduodenal suc-
tion, nothing by mouth, intravenous fluids, antibiotics, and
a somatostatin analog. In a recent review of 251 cases of
duodenal perforation, conservative management was used
in 62.2% and was successful in 92.9% of these patients.
Ten who failed conservative management required sal-
vage surgery. Six of those patients died, and 5 deaths
occurred in those with a delay in diagnosis/intervention
beyond 3 days.79
Four types of perforations are described after endoscopic
sphincterotomy (Fig. 9).80 Type I perforations involve the
lateral duodenal wall. Lateral wall perforations, including
Billroth II afferent limbs, tend to be large and usually are
caused by the tip of the endoscope. This may occur during
stone extraction in which a countercoup injury occurs (Fig.
10). Historically, these perforations required immediate sur-
gery because enteric contents leak into the retroperitoneal or
intraperitoneal space. However, recent technological ad-
vances with OTSCs have allowed large lateral wall perfora-
tions to be closed endoscopically.36 Type II or peri-vaterian
injuries vary in severity and are less likely to require surgery,
although if CT shows a large amount retroperitoneal fluid,
surgical or percutaneous intervention is required, especially
in the setting of severe inflammatory response syndrome
(Fig. 11). Endoscopic management of these types of perfo-
rations includes placement of a nasoduodenal tube and bil-
iary stent81 and/or TTS clips (Fig. 12).82 Until recently, these
biliary stents have been large-bore plastic stents; temporary
placement of FCSEMSs have been used, with excellent
results.83-85 Despite earlier reports of successful clip closure,82
medial wall perforations are difficult to manage with TTS
clips because they function poorly, if at all, through side-
viewing endoscopes. OTSCs cannot be effectively deployed
because of angulation and difficult access to the medial wall
and the potential for closing the ampullary orifice. If TTS clips
can be placed successfully and/or the perforation is small,
the patient can be managed conservatively with careful mon-
itoring, but surgery is indicated in the presence of systemic
inflammatory response. Diversion can be achieved with
placement of a nasoduodenal tube across the leak site. This
is accomplished by cutting the tip of a nasogastric tube to
allow an end hole for guidewire passage. A small-caliber
endoscope is passed transnasally into the duodenum, and a
guidewire is then advanced fluoroscopically into the fourth
duodenum. The nasogastric tube is passed over the wire and
positioned fluoroscopically across the leak site (Fig. 12D).
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 849
Acute endoscopic perforations
Figure 9. Types of duodenal wall perforations that can occur as a result of ERCP.
Figure 10. Intraoperative photograph taken during repair of large ERCP-
induced lateral duodenal wall perforation. A large stone (arrow) can be
seen impacted in the ampulla through the duodenal defect.
Type III injuries probably represent distal bile duct
injuries related to wire or basket instrumentation near an
obstructing entity. They are often small in size and usually
are of no clinical consequence. Perforation of the distal
bile duct often can be managed by placement of plastic
stents and, more recently, by temporary placement of
FCSEMSs.86 Retroperitoneal air alone (type IV) probably is
850 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
related to the use of insufflated air to maintain patency of
the lumen, although it can occur with small perforations
and can be associated with postprocedural pain (Fig. 13).
If incidentally identified, this is not considered a true
perforation or an adverse event and does not require
surgical intervention.
Duodenal perforation after endoscopic sphincterotomy
may be treated conservatively with success if identified
during endoscopic sphincterotomy or early after the pro-
cedure.87,88 Although ERCP-related lateral wall duodenal
wall perforations usually require immediate surgery, clips,
including the endoloop-endoclip method,89 and fibrin
glue have been used.90 OTSCs have been shown to be
useful for closure of large perforations in the jejunum in
altered anatomy.36,91 Snare resection of the major duode-
nal papilla can result in perforation, and small perforations
usually can be managed conservatively with92 or without93
biliary FCSEMS placement.
Patients with duodenal perforation and failed closure,
as determined endoscopically (visually and by contrast
material injection at the time of procedure), clinically, and
radiographically, and those with delayed diagnosis and/or
retroperitoneal fluid, require surgical intervention (Fig.
11). A nonsurgical option, especially in poor operative
www.giejournal.org
Baron et al
Figure 11. A, Abdominal CT obtained 24 hours after ERCP with needle-
knife precut biliary sphincterotomy complicated by type II perforation.
Extensive nondependent air and retroperitoneal fluid/necrosis are seen
(arrows). B, Upright abdominal radiograph of same patient after place-
ment of multiple percutaneous retroperitoneal drains and transhepatic
drain (upper right on image).
candidates, is placement of retroperitoneal drains (Fig.
11B and 14).
Duodenal perforations that are not periampul-
lary. Duodenal perforations not related to ERCP usually
are the result of endoscope trauma, duodenal polypec-
tomy, coagulation of bleeding ulcers, and attempts to pass
forcibly beyond a duodenal stricture. As in type I ERCP-
related duodenal perforations, these typically require sur-
gical intervention, although the use of TTS clips94 and
more recently OTSCs can be effective in closing duodenal
perforations in animal models95 and in humans.36,96 Cov-
ered SEMSs can be placed across duodenal perforations,
especially those associated with strictures,97 although the
www.giejournal.org
Acute endoscopic perforations
only available large-diameter covered SEMSs suitable for
such purposes in the United States are esophageal SEMSs.
The short-length delivery systems pose technical chal-
lenges to placement (Video 5) One recently available
FDA-approved fully covered esophageal SEMS passes
through the endoscope channel and may be useful in such
cases.
Interventional radiologic considerations
Duodenal injuries can lead to retroperitoneal collections,
which usually can be drained by a percutaneous approach,
as described earlier. However, in cases of a peritoneal per-
foration, the collection must be treated like an amylase-rich
or bile-rich collection that is often seen in patients with
pancreatitis, because they may contain solid debris and may
take longer to resolve. These collections need more frequent
monitoring to assess for formation of necrotic debris or fistula
to other hollow organs. The presence of amylase-rich fluid
and a catheter in the periduodenal region can lead to vascu-
lar injury, such as a pseudoaneurysm, which is a medical
emergency. Any evidence of internal bleeding, such as
formed clot within the catheter, needs to be investigated by
CT angiography. Suspicion of an arterial pseudoaneurysm
has to be confirmed by a conventional angiogram and
treated by endovascular means.98
Surgical considerations
Surgical exploration is required in a lesser proportion (29%)
of periampullary perforations.78 The majority of patients with
iatrogenic perforations who require exploration will have lateral
(free wall, type I) duodenal perforations (86%). As with gastric
and esophageal perforations, patients with hemodynamic insta-
bility and who fail endoscopic (progression of symptoms or
demonstration of ongoing free leakage of oral contrast material)
and/or percutaneous management should undergo emergent
operative exploration. Duodenal injuries generally can be re-
paired by modified Gram patch procedures. If severe inflam-
mation is present, wide drainage or pyloric exclusion proce-
dures may be required. Patients with severe sepsis may require
damage control laparotomy with wide drainage during the ini-
tial procedure.
Biliary tract perforation
Biliary tract perforations can take the form of guidewire-
induced perforations or from dilation of strictures either
within native bile ducts or at biliary-enteric anastomoses.
Guidewire perforations can occur during attempts to traverse
benign or malignant strictures or when an excessive amount
of wire is passed intrahepatically (Fig. 15). As long as the
perforation is recognized, and stent placement or dilation is
not performed through the tract and outside the biliary sys-
tem (Fig. 16), most patients do well with either proper posi-
tioning of the wire with stent placement across the perfora-
tion (Fig. 17) or with conservative management and
antibiotics. The latter is especially true if the perforation
occurred below an obstructing stricture, which is protective
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 851
Acute endoscopic perforations Baron et al

Figure 12. A, Endoscopic view of perforation (arrows) immediately after ampullectomy and biliary sphincterotomy. B, Radiographic image taken after
contrast material injection through the endoscope shows contrast material extravasation and a large amount of extraluminal air. C, Endoscopic image
after placement of biliary fully covered self-expandable metal stent. D, Radiographic image taken after placement of a nasoduodenal tube for suction,
biliary stent (black arrows denote stent ends) and pancreatic stent placement (white arrow denotes proximal end). Patient did well with conservative
management.

of a leak. In our experience, guidewire perforations occur
most commonly in patients with difficult strictures due to
primary sclerosing cholangitis. If conservative management
fails, percutaneous drainage of the biliary tree and/or sepa-
rate catheter drainage of a contained leak (biloma) or free
leak may be required.
It is extremely important to recognize that some bil-
iary procedures may create free air; intervention is not
required and, in fact, may cause undue harm. This can
occur in EUS-transgastric biliary interventions, with or
without a rendezvous approach, as air escapes between
the stomach and liver.99 Similarly, a percutaneous-
endoscopic rendezvous procedure allows air to escape
outside or around a percutaneous transhepatic puncture
site from the endoscopic procedure and/or biliary
sphincterotomy.100 Finally, after ERCP with biliary sp-
hincterotomy and/or stent placement for treatment of a
bile leak, air introduced during the procedure can es-
cape from the leak and into the abdomen.101 Abdominal
films often are obtained in such patients to evaluate
post-procedural pain. Although such pain often is related
to small amounts of bile leakage, the pain may be errone-
ously attributed to a perforated viscus and lead to unwar-
ranted surgical intervention.
Patients who undergo surgical exploration for hepati-
copancreatic biliary perforations may require advanced

852 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012 www.giejournal.org

Baron et al
Figure 13. Abdominal CT obtained for postprocedure pain after appar-
ent uncomplicated ERCP. A small amount of retroperitoneal air is seen.
The patient did well with conservative management.
Figure 14. A, Endoscopic view of lateral wall duodenal perforation. B,
Abdominal CT shows contrast extravasation despite attempted clip clo-
sure of the perforation. A percutaneous drain was placed.
www.giejournal.org
Acute endoscopic perforations
Figure 15. Intrahepatic guidewire–induced biliary perforation. A, During
dilation of a hilar stricture in a patient with primary sclerosing cholangitis,
the guidewire passed too far proximally into the biliary tree. B, Contrast
material injection shows extravasation outside the liver (arrow). C, A
biliary stent placed into the right intrahepatic system. The patient did well
clinically.
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 853
Acute endoscopic perforations
Figure 16. Abdominal CT shows the proximal end of a biliary stent
extending outside the liver (arrow) into the peritoneal cavity. Extralumi-
nal air and fluid can be seen.
surgical procedures, such as Roux-en-Y choledochojeju-
nostomy. If such treatment is considered, transfer to a
tertiary-care surgical center is recommended.
Enteroscopy-related perforation
Small-bowel enteroscopy by using the single-balloon or
double-balloon technique is associated with a higher com-
plication rate compared with standard upper endoscopic
procedures. A total of 11 (0.4%) perforations were re-
ported in a series of 2478 double-balloon endoscopy ex-
aminations performed at 9 centers. The perforation rate
was significantly higher in patients with altered surgical
anatomy and in those who underwent retrograde exami-
nations.102 Therapeutic procedures, particularly dilation of
strictures and polypectomy, carry higher perforation rates.
Although clip placement can be attempted, technical dif-
ficulties and limited accessibility to place TTS clips or
OTSCs to close small-bowel perforations often lead to
surgical management.
Interventional radiologic considerations
After diagnosis of a perforation, a contrast enhanced CT
scan should be performed regardless of endoscopic inter-
vention. If the perforation is contained, but a fluid collec-
tion is evident, the patient can be managed by percutane-
ous means. Drainage of contained perforation is similar to
drainage of an abdominal abscess, which can be per-
formed by CT or US guidance. As with other types of
perforation, close monitoring of the patient’s clinical status
as well as imaging findings is essential.
Surgical considerations
Unless the perforation is in the proximal jejunum and is
small and accessible to endoscopic clip closure, good
operative candidates should undergo emergent explora-
854 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
tion. Bowel resection is generally necessary, but if the
perforation is small and the tissues have limited inflamma-
tory changes oversewing can be performed. Primary anas-
tomosis is preferred to jejunostomy due to the severe fluid,
electrolyte, and nutritional difficulties associated with
proximal stomas, even if the tissues are severely inflamed.
Ileostomy may be performed if sufficient proximal bowel
absorptive surface is present.
Colorectal perforation
Perforation as a complication of colonoscopy is esti-
mated to occur in 0.01% to 0.3% of procedures, both in
referral and ambulatory settings. Management of colo-
rectal perforation differs depending on whether the
perforation is rectal or colonic. It is important to note
that although endoscopic (percutaneous endoscopic ce-
costomy) and medical diversion (transanal tube place-
ment) are theoretically possible, these options are rarely
used.
Rectal perforation
The rectum is located below the peritoneal reflection,
and, thus, when wall dehiscence occurs, intraluminal air
does not leak into the abdomen but is contained or dis-
sects through perirectal tissues. Because endoscopic per-
foration usually implies air leakage into a space, when it
occurs in the rectum, penetration to perirectal tissue might
be a more appropriate term. Penetration has been defined
as visual or radiographic evidence of unintended penetra-
tion beyond the mucosa or duct without perforation.1
Frequent causes of rectal penetration are endoscope ret-
roflexion, EMR, and ESD of polyps and tumors. When
penetration occurs during rectal ESD, subcutaneous em-
physema may appear at distant sites, such as the head,
neck, and upper limbs, due to the existence of a contin-
uum of fascial planes that connect cervical soft tissues with
the mediastinum and retroperitoneum, permitting aberrant
air in these areas. Treatment consists of nothing per mouth
and administration of broad-spectrum antibiotics. Subcu-
taneous emphysema usually resolves in a few days. Defect
closure by using endoscopic clips (TTS103,104 and OTC105)
can be effective.
Colonic perforation
Causes of colonic perforation differ between diagnostic
and therapeutic procedures. Perforations related to diag-
nostic colonoscopy are due to mechanical trauma and can
be caused by direct impact of the endoscope tip or by
antimesenteric impaction of the endoscope body (loop)
against the colon wall. Cecal perforation may occur from
barotrauma, usually in the setting of difficult colonoscopy
and overinflation. The sigmoid colon is not only the most
tortuous and challenging portion to deal with during
colonoscopy but also a common location for diverticula,
pedunculated polyps, and postoperative adhesions. At-
tempts to overcome the sigmoid colon are often related to
www.giejournal.org
Baron et al
Figure 17. A, Dilating balloon positioned across a focal common bile duct stricture. Note waist in mid-portion of balloon. B, Contrast extravasation
noted after removal of balloon dilator. C, Two 10F plastic biliary stents placed across the biliary perforation. The patient did well clinically.
forceful torque maneuvers, loop formation, lumen oblit-
eration, and overinsufflation, making the sigmoid colon
the most frequent site for colon perforation. Perforations
tend to occur most often in the sigmoid colon, particularly
in elderly women with a history of abdominal/pelvic sur-
gery and/or diverticular disease.
In most cases, the perforation is unrecognized until
pneumoperitoneum and abdominal distention develop.
Needle decompression for tension pneumoperitoneum is
undertaken, and management depends on the size of the
perforation, the patient’s clinical condition, underlying co-
morbidities, the degree of colon preparation, the presence
of extraluminal fecal soiling, and elapsed time after the
perforation.106 Because diversion cannot be achieved, im-
mediate closure by using TTS clips107 (Fig. 18) or OTSCs
(Fig. 19)36,108 should be attempted when the perforation is
recognized during or immediately after the procedure and
when the size of the perforation is ⱕ2 cm. Band ligation
can be useful for closing colon perforations when TTS clip
placement has failed.109
The mechanism of colon perforation differs when ther-
apeutic procedures are performed. Transmural electrocau-
tery injury is responsible for perforations when removal of
colon polyps or early cancers is attempted by standard
polypectomy, EMR, or ESD. When a polyp is being re-
sected en bloc during EMR, the identification of a “target
sign” on the underside of the specimen and a mirror target
at the resection site is potentially indicative of a full-
thickness resection (perforation) and should be promptly
treated with endoscopic closure.110 Due to the thin wall
and relative complexity in dealing with proximal lesions,
the cecum and right side of the colon are by far the most
www.giejournal.org
Acute endoscopic perforations
frequently reported sites of colon perforation after thera-
peutic colonoscopy. Endoscopic clip management can be
undertaken as mentioned earlier (Video 7, available online
at www.giejournal.org). However, close monitoring for
clinical decompensation and need for surgery are neces-
sary. As in other locations, abdominal CT with rectally
administered contrast material is useful to detect ongoing
leakage and spillage of enteric contents into the extralu-
minal space, which will require surgical or, occasionally,
percutaneous management. The latter interventions also
are needed when endoscopic therapy is not feasible or
recognition of perforation is delayed.
Interventional radiologic considerations
As with other types of perforations, in the presence of
a fluid collection, the most important first step is to divert
leaked contents from the abdomen by a percutaneous
catheter. In case of colon perforations, the consistency of
the drain output has to be monitored closely. If the output
of the catheter is consistent with fecal material, the cath-
eter needs to be upsized. Small catheters are notorious for
occlusion in the presence of colon fistulas. The duration of
drainage is often longer with colon fistulas compared with
other enteric fistulas. Although some fistulas may not com-
pletely resolve by percutaneous means, percutaneous
drainage can improve symptoms so that the patient may
be a better candidate for surgical treatment.
Surgical considerations
Anterior and lateral rectal wall perforations in the upper
two-thirds of the rectum can result in intraperitoneal sepsis. If
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 855
Acute endoscopic perforations
Figure 18. A, Rectal penetration after ESD. B, Clip closure of
penetration.
this is present, surgical exploration via celiotomy or laparos-
copy with washout and fecal diversion is recommended.
Perforation isolated to the extraperitoneal space generally
can be treated by nonoperative means. Percutaneous drain-
age may be required if an abscess develops. If this abscess is
inaccessible by percutaneous means, then presacral drainage
may be performed. Routine operative presacral drainage,
however, should be avoided. Transanal repair can be con-
sidered in low rectal perforations.
Until recently, most patients who present with a
colonoscopic perforation required operative management.
Patients who present within 24 hours of the perforation
event have significantly improved outcomes compared
with those whose presentation is delayed.107 Late presen-
tations are associated with a greater rate of feculent peri-
tonitis and, therefore, inflammatory changes. Primary
repair is possible if the tissues appear healthy, and lapa-
roscopic closure is possible.107 If not, fecal diversion in the
form of colostomy or ileostomy is necessary.
856 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
Figure 19. A, Perforation of colon diverticulum with tip of colonoscope.
B, Successful over-the-scope clip closure of perforated diverticulum.
Images courtesy of Dr Andrew Ross, Virginia Mason Medical Center,
Seattle, Washington.
CONCLUSIONS
Endoscopic perforation can provoke an array of emo-
tional responses from both patients and caregivers alike.
Prompt recognition and management are key determinants
for lessening the significant morbidity and mortality associ-
ated with this complication. A multidisciplinary approach is
essential for a successful outcome. In selected patients, en-
doscopic therapy plays a dominant role via diversion of
enteric contents and/or mechanical closure of the perfora-
tion, with or without placement of percutaneous drains. Sur-
gery is generally reserved for patients in whom endoscopic
therapy is unfeasible or unsuccessful, recognition of perfo-
ration is delayed, or clinical deterioration occurs despite en-
doscopic and percutaneous intervention.
REFERENCES
1. Cotton PB, Eisen GM, Aabakken L, et al. A lexicon for endoscopic ad-
verse events: report of an ASGE workshop. Gastrointest Endosc 2010;
71:446-54.
2. Watson RR, Thompson CC. NOTES spin-off for the therapeutic gastro-
enterologist: natural orifice surgery. Minerva Gastroenterol Dietol
2011;57:177-91.
3. La Torre M, Velluti F, Giuliani G, et al. Promptness of diagnosis is the
main prognostic factor after colonoscopic perforation. Colorectal Dis
2012;14:e23-6.
www.giejournal.org
Baron et al
4. Iqbal CW, Cullinane DC, Schiller HJ, et al. Surgical management and
outcomes of 165 colonoscopic perforations from a single center insti-
tution. Arch Surg 2008;143:701-6; discussion 706-7.
5. Vallböhmer D, Hölscher AH, Hölscher M, et al. Options in the manage-
ment of esophageal perforation: analysis over a 12-year period. Dis
Esophagus 2010;23:185-90.
6. Wiesen AJ, Sideridis K, Fernandes A, et al. True incidence and clinical
significance of pneumoperitoneum after PEG placement: a prospec-
tive study. Gastrointest Endosc 2006;64:886-9.
7. Tamiya Y, Nakahara K, Kominato K, et al. Pneumomediastinum is a
frequent but minor complication during esophageal endoscopic sub-
mucosal dissection. Endoscopy 2010;42:8-14.
8. Genzlinger JL, McPhee MS, Fisher JK, et al. Significance of retroperito-
neal air after endoscopic retrograde cholangiopancreatography with
sphincterotomy. Am J Gastroenterol 1999;94:1267-70.
9. Onogi F, Araki H, Ibuka T, et al. “Transmural air leak”: a computed to-
mographic finding following endoscopic submucosal dissection of
gastric tumors. Endoscopy 2010;42:441-7.
10. Coriat R, Leblanc S, Pommaret E, et al. Transmural air leak following
endoscopic submucosal dissection: a non-useful computed tomogra-
phy finding. Endoscopy 2010;42:1117; author reply 1118.
11. Enns R, Eloubeidi MA, Mergener K, et al. ERCP-related perforations: risk
factors and management. Endoscopy 2002;34:293-8.
12. Maeda Y, Hirasawa D, Fujita N, et al. A pilot study to assess mediastinal
emphysema after esophageal endoscopic submucosal dissection with
carbon dioxide insufflation. Endoscopy. Epub 2012 Mar 9.
13. Fujii L, Lau A, Fleischer DE, et al. Successful nonsurgical treatment of
pneumomediastinum, pneumothorax, pneumoperitoneum, pneu-
moretroperitoneum, and subcutaneous emphysema following ERCP.
Gastroenterol Res Pract 2010;2010:289135.
14. Assalia A, Suissa A, IIivizki A, et al. Validity of clinical criteria in the man-
agement of endoscopic retrograde cholangiopancreatography-
related duodenal perforations. Arch Surg 2007;142:1059-64.
15. Maunder RJ, Pierson DJ, Hudson LD. Subcutaneous and mediastinal
emphysema. Pathophysiology, diagnosis, and management. Arch In-
tern Med 1984;144:1447-53.
16. Ignjatović M, Jović J. Tension pneumothorax, pneumoretroperito-
neum, and subcutaneous emphysema after colonoscopic polypecto-
my: a case report and review of the literature. Langenbecks Arch Surg
2009;394:185-9.
17. Tam WY, Bertholini D. Tension pneumoperitoneum, pneumomedias-
tinum, subcutaneous emphysema and cardiorespiratory collapse fol-
lowing gastroscopy. Anaesth Intensive Care 2007;35:307-9.
18. Lin BW, Thanassi W. Tension pneumoperitoneum. J Emerg Med 2010;
38:57-9.
19. Mai CM, Wen CC, Wen SH, et al. Iatrogenic colonic perforation by
colonoscopy: a fatal complication for patients with a high anesthetic
risk. Int J Colorectal Dis 2010;25:449-54.
20. Gayer G, Hertz M, Zissin R. Postoperative pneumoperitoneum: preva-
lence, duration, and possible significance. Semin Ultrasound CT MR
2004;25:286-9.
21. Merchea A, Cullinane DC, Sawyer MD, et al. Esophagogastroduodenoscopy-
associated gastrointestinal perforations: a single-center experience.
Surgery 2010;148:876-80; discussion 881-2.
22. Zissin R, Shapiro-Feinberg M, Oscadchy A, et al. Retroperitoneal perfo-
ration during endoscopic sphincterotomy: imaging findings. Abdom
Imaging 2000;25:279-82.
23. Sherck J, Shatney C, Sensaki K, et al. The accuracy of computed tomog-
raphy in the diagnosis of blunt small-bowel perforation. Am J Surg
1994;168:670-5.
24. Vallböhmer D, Hölscher AH, Hölscher M, et al. Options in the manage-
ment of esophageal perforation: analysis over a 12-year period. Dis
Esophagus 2010;23:185-90.
25. Flancbaum L, Nosher JL, Brolin RE. Percutaneous catheter drainage of
abdominal abscesses associated with perforated viscus. Am Surg 1990;
56:52-6.
www.giejournal.org
Acute endoscopic perforations
26. Zengerink I, Brink PR, Laupland KB, et al. Needle thoracostomy in the
treatment of a tension pneumothorax in trauma patients: what size
needle? J Trauma 2008;64:111-4.
27. Henry M, Arnold T, Harvey J; Pleural Diseases Group, Standards of Care
Committee, British Thoracic Society. BTS guidelines for the manage-
ment of spontaneous pneumothorax. Thorax 2003;58(suppl 2):ii39-52.
28. Fu K, Ishikawa T, Yamamoto T, et al. Paracentesis for successful treat-
ment of tension pneumoperitoneum related to endoscopic submuco-
sal dissection. Endoscopy 2009;41(suppl 2):E245.
29. Chiapponi C, Stocker U, Korner M, et al. Emergency percutaneous nee-
dle decompression for tension pneumoperitoneum. BMC Gastroen-
terol 2011;5:11:48.
30. Mangiavillano B, Viaggi P, Masci E. Endoscopic closure of acute iatro-
genic perforations during diagnostic and therapeutic endoscopy in
the gastrointestinal tract using metallic clips: a literature review. J Dig
Dis 2010;11:12-8.
31. Matsuda T, Fujii T, Emura F, et al. Complete closure of a large defect
after EMR of a lateral spreading colorectal tumor when using a two-
channel colonoscope. Gastrointest Endosc 2004;60:836-8.
32. Desilets DJ, Romanelli J, Surti VC, et al. The ties that bind: durable,
transmural, purse-string-like gastrotomy closure using a novel device
[abstract]. Gastrointest Endosc 2007;65:AB292.
33. von Renteln D, Schmidt A, Vassiliou MC, et al. Endoscopic closure of
large colonic perforations using an over-the-scope clip: a randomized
controlled porcine study. Endoscopy 2009;41:481-6.
34. von Renteln D, Vassiliou MC, Rothstein RI. Randomized controlled trial
comparing endoscopic clips and over-the-scope clips for closure of
natural orifice transluminal endoscopic surgery gastrotomies. Endos-
copy 2009;41:1056-61.
35. Matthes K, Jung Y, Kato M, et al. Efficacy of full-thickness GI perforation
closure with a novel over-the-scope clip application device: an animal
study. Gastrointest Endosc 2011;74:1369-75.
36. Voermans RP, Le Moine O, von Renteln D, et al. CLIPPER study-group
efficacy of endoscopic closure of acute perforations of the gastrointes-
tinal tract. Clin Gastroenterol Hepatol. Epub 2012 Feb 20.
37. Baron TH, Wong Kee Song LM. Ross A, et al. Use of an over-the-scope
clipping device: multicenter retrospective results of the first U.S. expe-
rience. Gastrointest Endosc 2012. In press.
38. Swinnen J, Eisendrath P, Rigaux J, et al. Self-expandable metal stents
for the treatment of benign upper GI leaks and perforations. Gastroin-
test Endosc 2011;73:890-9.
39. Pham BV, Raju GS, Ahmed I, et al. Immediate endoscopic closure of
colon perforation by using a prototype endoscopic suturing device:
feasibility and outcome in a porcine model (with video). Gastrointest
Endosc 2006;64:113-9.
40. Ryou M, Pai RD, Sauer JS, et al. Evaluating an optimal gastric closure
method for transgastric surgery. Surg Endosc 2007;21:677-80.
41. Voermans RP, Worm AM, van Berge Henegouwen MI, et al. In vitro
comparison and evaluation of seven gastric closure modalities for nat-
ural orifice transluminal endoscopic surgery (NOTES). Endoscopy
2008;40:595-601.
42. McGee MF, Marks JM, Onders RP, et al. Complete endoscopic closure of
gastrotomy after natural orifice translumenal endoscopic surgery us-
ing the NDO Plicator. Surg Endosc 2008;22:214-20.
43. Magno P, Giday SA, Dray X, et al. A new stapler-based full-thickness
transgastric access closure: results from an animal pilot trial. Endos-
copy 2007;39:876-80.
44. Raju GS. Endoscopic closure of gastrointestinal leaks. Am J Gastroen-
terol 2009;104:1315-20.
45. Bakal CW, Sacks D, Burke DR, et al; for the Society of Interventional
Radiology Standards of Practice Committee. Quality improvement
guidelines for adult percutaneous abscess and fluid drainage. J Vasc
Interv Radiol 2003;14:S223-5.
46. vanSonnenberg E, Wing VW, Casola G, et al. Temporizing effect of per-
cutaneous drainage of complicated abscesses in critically ill patients.
Am J Roentgenol 1984;142:821-6.
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 857
Acute endoscopic perforations
47. Minami S, Gotoda T, Ono H, et al. Complete endoscopic closure of
gastric perforation induced by endoscopic resection of early gastric
cancer using endoclips can prevent surgery (with video). Gastrointest
Endosc 2006;63:596-601.
48. Geraci G, Pisello F, Modica G, et al. Complications of elective esophago-
gastro-duodenoscopy (EGDS). Personal experience and literature re-
view. G Chir 2009;30:502-6.
49. Sharma N, Jindal M, Mahon B, et al. Hypopharyngeal perforation dur-
ing endoscopic ultrasound treated by primary repair. Head Neck 2011;
33:756-8.
50. Min JK, Spencer KT, Furlong KT, et al. Clinical features of complications
from transesophageal echocardiography: a single-center case series of
10,000 consecutive examinations. J Am Soc Echocardiogr 2005;
18:925-9.
51. Sobrino MA, Kozarek R, Low DE. Primary endoscopic management of
esophageal perforation following transesophageal echocardiogram.
J Clin Gastroenterol 2004;38:581-5.
52. Case DJ, Baron TH. Flexible endoscopic management of Zenker diver-
ticulum: the Mayo Clinic experience. Mayo Clin Proc 2010;85:719-22.
53. Moberly AC, Fritsch MH, Mosier KM. Management of sword-swallower
injuries. J Laryngol Otol 2011;125:217-9.
54. Rathinam S, Makarawo T, Norton R, et al. Thoracic osteophyte: rare
cause of esophageal perforation. Dis Esophagus 2010;23:E5-8.
55. Søreide JA, Viste A. Esophageal perforation: diagnostic work-up and
clinical decision-making in the first 24 hours. Scand J Trauma Resusc
Emerg Med 2011;19:66.
56. Kuppusamy MK, Hubka M, Felisky CD, et al. Evolving management
strategies in esophageal perforation: surgeons using nonoperative
techniques to improve outcomes. J Am Coll Surg 2011;213:164-71; dis-
cussion 171-2.
57. Mangiavillano B, Viaggi P, Masci E. Endoscopic closure of acute iatro-
genic perforations during diagnostic and therapeutic endoscopy in
the gastrointestinal tract using metallic clips: a literature review. J Dig
Dis 2010;11:12-8.
58. Qadeer MA, Dumot JA, Vargo JJ, et al. Endoscopic clips for closing
esophageal perforations: case report and pooled analysis. Gastrointest
Endosc 2007;66:605-11.
59. Kuppusamy MK, Felisky C, Kozarek RA, et al. Impact of endoscopic as-
sessment and treatment on operative and non-operative manage-
ment of acute oesophageal perforation. Br J Surg 2011;98:818-24.
60. Seebach L, Bauerfeind P, Gubler C. “Sparing the surgeon”: clinical ex-
perience with over-the-scope clips for gastrointestinal perforation. En-
doscopy 2010;42:1108-11.
61. Tang SJ, Singh S, Wait MA, et al. Endotherapy for a 5-cm mid-
esophageal perforation with tandem stenting above the lower esoph-
ageal sphincter (with videos). Surg Endosc 2009;23:2836-41.
62. Siersema PD, Homs MY, Haringsma J, et al. Use of large-diameter me-
tallic stents to seal traumatic nonmalignant perforations of the esoph-
agus. Gastrointest Endosc 2003;58:356-61.
63. Bakken JC, Wong Kee Song LM, de Groen PC, et al. Use of a fully covered
self-expandable metal stent for the treatment of benign esophageal
diseases. Gastrointest Endosc 2010;72:712-20.
64. van Heel NC, Haringsma J, Spaander MC, et al. Short-term esophageal
stenting in the management of benign perforations. Am J Gastroen-
terol 2010;105:1515-20.
65. Fischer A, Thomusch O, Benz S, et al. Nonoperative treatment of 15
benign esophageal perforations with self-expandable covered metal
stents. Ann Thorac Surg 2006;81:467-72.
66. Johnsson E, Lundell L, Liedman B. Sealing of esophageal perforation or
ruptures with expandable metallic stents: a prospective controlled
study on treatment efficacy and limitations. Dis Esophagus 2005;18:
262-6.
67. Gelbmann CM, Ratiu NL, Rath HC, et al. Use of self-expandable plastic
stents for the treatment of esophageal perforations and symptomatic
anastomotic leaks. Endoscopy 2004;36:695-9.
858 GASTROINTESTINAL ENDOSCOPY Volume 76, No. 4 : 2012
Baron et al
68. van Heel NC, Haringsma J, Spaander MC, et al. Short-term esophageal
stenting in the management of benign perforations. Am J Gastroen-
terol 2010;105:1515-20.
69. Hirdes MM, Siersema PD, Houben MH, et al. Stent-in-stent technique
for removal of embedded esophageal self-expanding metal stents.
Am J Gastroenterol 2011;106:286-93.
70. White RE, Mungatana C, Topazian M. Expandable stents for iatrogenic
perforation of esophageal malignancies. J Gastrointest Surg 2003;6:
715-9.
71. Dai Y, Chopra SS, Kneif S, et al. Management of esophageal anasto-
motic leaks, perforations, and fistulae with self-expanding plastic
stents. J Thorac Cardiovasc Surg 2011;141:1213-7.
72. Ben-David K, Lopes J, Hochwald S, et al. Minimally invasive treatment
of esophageal perforation using a multidisciplinary treatment algo-
rithm: a case series. Endoscopy 2011;43:160-2.
73. vanSonnenberg E, Wittich GR, Goodacre BW, et al. Percutaneous drain-
age of thoracic collections. J Thorac Imaging 1998;13:74-82.
74. Rahman NM, Maskell NA, West A, et al. Intrapleural use of tissue plas-
minogen activator and DNase in pleural infection. N Engl J Med 2011;
365:518-26.
75. Imagawa A, Okada H, Kawahara Y, et al. Endoscopic submucosal dis-
section for early gastric cancer: results and degrees of technical diffi-
culty as well as success. Endoscopy 2006;38:987-90.
76. Toyokawa T, Inaba T, Omote S, et al. Risk factors for perforation and
delayed bleeding associated with endoscopic submucosal dissection
for early gastric neoplasms; analysis of 1123 lesions. J Gastroenterol
Hepatol 2012;27:907-12.
77. Iwashita T, Lee JG, Nakai Y, et al. Successful management of perforation
during cystogastrostomy with an esophageal fully covered metallic
stent placement. Gastrointest Endosc. Epub 2011 Aug 31.
78. Fatima J, Baron TH, Topazian MD, et al. Pancreaticobiliary and duode-
nal perforations after periampullary endoscopic procedures: diagnosis
and management. Arch Surg 2007;142:448-54; discussion 454-5.
79. Machado NO. Management of duodenal perforation post-endoscopic
retrograde cholangiopancreatography. When and whom to operate
and what factors determine the outcome? A review article. JOP 2012;
13:18-25.
80. Stapfer M, Selby RR, Stain SC, et al. Management of duodenal perfora-
tion after endoscopic retrograde cholangiopancreatography and
sphincterotomy. Ann Surg 2000;232:191-8.
81. Howard TJ. Re: Stapfer M, et al. Management of duodenal perforation
after endoscopic retrograde cholangiopancreatography and sphinc-
terotomy. Ann Surg 2001;234:132-3.
82. Baron TH, Gostout CJ, Herman L. Hemoclip repair of a sphincterotomy-
induced duodenal perforation. Gastrointest Endosc 2000;52:566-8.
83. Baron TH. Covered self-expandable metal stents for benign biliary tract
diseases. Curr Opin Gastroenterol 2011;27:262-7.
84. Vezakis A, Fragulidis G, Nastos C. Closure of a persistent
sphincterotomy-related duodenal perforation by placement of a cov-
ered self-expandable metallic biliary stent. World J Gastroenterol 2011;
17:4539-41.
85. Akbar A, Irani S, Baron TH, et al. Use of covered self-expandable metal
stents for endoscopic management of non-stricture–related benign
biliary diseases. Gastrointest Endosc 2012. In press.
86. Jeon HJ, Han JH, Park S, et al. Endoscopic sphincterotomy-related per-
foration in the common bile duct successfully treated by placement of
a covered metal stent. Endoscopy. 2011;43 Suppl 2 UCTN:E295-6.
87. Baron TH. Covered self-expandable metal stents for benign biliary tract
diseases. Curr Opin Gastroenterol 2011;27:262-7.
88. Vezakis A, Fragulidis G, Nastos C. Closure of a persistent
sphincterotomy-related duodenal perforation by placement of a cov-
ered self-expandable metallic biliary stent. World J Gastroenterol
201128;17:4539-41.
89. Nakagawa Y, Nagai T, Soma W, et al. Endoscopic closure of a large
ERCP-related lateral duodenal perforation by using endoloops and en-
doclips. Gastrointest Endosc 2010;72:216-7.
www.giejournal.org
Baron et al
90. Mutignani M, Iacopini F, Dokas S, et al. Successful endoscopic closure
of a lateral duodenal perforation at ERCP with fibrin glue. Gastrointest
Endosc 2006;63:725-7.
91. Buffoli F, Grassia R, Iiritano E, et al. Endoscopic “retroperitoneal fat-
pexy” of a large ERCP-related jejunal perforation by using a new over-
the-scope clip device in Billroth II anatomy (with video). Gastrointest
Endosc. Epub 2011 Aug 5.
92. Ali A, Irani S, Baron TH, et al. Use of Covered Self-Expandable Metal
Stents for Endoscopic Management of Non-Stricture- Related Benign
Biliary Diseases. Gastrointest Endosc 2012. In press.
93. Norton ID, Gostout CJ, Baron TH, et al. Safety and outcome of endo-
scopic snare excision of the major duodenal papilla. Gastrointest En-
dosc 2002;56:239-43.
94. Haider S, Kahaleh M. The use of endoscopic clipping devices in the
treatment of iatrogenic duodenal perforation. Gastroenterol Hepatol
(NY) 2010;6:660-1.
95. von Renteln D, Rudolph HU, Schmidt A, et al. Endoscopic closure of
duodenal perforations by using an over-the-scope clip: a randomized,
controlled porcine study. Gastrointest Endosc 2010;71:131-8.
96. Parodi A, Repici A, Pedroni A, et al. Endoscopic management of GI
perforations with a new over-the-scope clip device (with videos). Gas-
trointest Endosc 2010;72:881-6.
97. Small AJ, Petersen BT, Baron TH. Closure of a duodenal stent-induced
perforation by endoscopic stent removal and covered self-expandable
metal stent placement (with video). Gastrointest Endosc 2007;66:
1063-5.
98. Fotoohi M, Traverso LW. Management of severe pancreatic necrosis.
Curr Treat Options Gastroenterol 2007;10:341-6.
99. Park do H, Jang JW, Lee SS, et al. EUS-guided biliary drainage with
transluminal stenting after failed ERCP: predictors of adverse events
and long-term results. Gastrointest Endosc 2011;74:1276-84.
100. Hui YT, Lam WM, Lam TW, et al. Benign pneumoperitoneum devel-
oped after endoscopic biliary metallic stent placement with the ren-
dezvous procedure. Gastrointest Endosc 2008;67:179-80.
101. Bar-Meir S, Lang A, Shemesh E, et al. Pneumoperitoneum after inser-
tion of endoscopic biliary stent for post-cholecystectomy biliary leak.
Gastrointest Endosc 1993;39:818-20.
GIE on Facebook
GIE now has a Facebook page. Fans will receive news, updates, and links to
author interviews, podcasts, articles, and tables of contents. Search on Facebook
for “GIE: Gastrointestinal Endoscopy” and become a fan.
www.giejournal.org
Acute endoscopic perforations
102. Gerson LB, Tokar J, Chiorean M, et al. Complications associated with
double balloon enteroscopy at nine US centers. Clin Gastroenterol
Hepatol 2009;7:1177-82, 1182.e1-3.
103. Ahlawat SK, Charabaty A, Benjamin S. Rectal perforation caused by
retroflexion maneuver during colonoscopy: closure with endoscopic
clips. Gastrointest Endosc 2008;67:771-3.
104. Katsinelos P, Kountouras J, Chatzimavroudis G, et al. Endoscopic clo-
sure of a large iatrogenic rectal perforation using endoloop/clips tech-
nique. Acta Gastroenterol Belg 2009;72:357-9.
105. Coriat R, Leblanc S, Pommaret E, et al. Endoscopic management of
endoscopic submucosal dissection perforations: a new over-the-
scope clip device. Gastrointest Endosc 2011;73:1067-9.
106. Raju GS, Saito Y, Matsuda T, et al. Endoscopic management of colono-
scopic perforations (with videos). Gastrointest Endosc 2011;74:1380-8.
107. Cho SB, Lee WS, Joo YE. Therapeutic options for iatrogenic colon per-
foration: feasibility of endoscopic clip closure and predictors of the
need for early surgery. Surg Endosc. Epub 2011 Sep 23.
108. Seebach L, Bauerfeind P, Gubler C. “Sparing the surgeon”: clinical ex-
perience with over-the-scope clips for gastrointestinal perforation. En-
doscopy 2010;42:1108-11.
109. Han JH, Park S, Youn S. Endoscopic closure of colon perforation with
band ligation; salvage technique after endoclip failure. Clin Gastroen-
terol Hepatol 2011;9:e54-5.
110. Swan MP, Bourke MJ, Moss A, et al. The target sign: an endoscopic
marker for the resection of the muscularis propria and potential perfo-
ration during colonic endoscopic mucosal resection. Gastrointest En-
dosc 2011;73:79-85.
Received March 15, 2012. Accepted April 29, 2012.
Current affiliations: Division of Gastroenterology and Hepatology (1), De-
partment of Surgery (2), Mayo Clinic, Rochester, Minnesota; Emura Center
(3), Latino America, Universidad de La Sabana, Bogota DC, Colombia; Inter-
ventional Radiology and the Digestive Disease Institute at Virginia Mason
(4), Virginia Mason Medical Center, Seattle, Washington, USA.
Reprint requests: Todd H. Baron, MD, Division of Gastroenterology and
Hepatology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905.
Volume 76, No. 4 : 2012 GASTROINTESTINAL ENDOSCOPY 859