Endoscopic ultrasound (EUS) is an essential tool for identification, staging, sampling
and follow-up of benign and malignant bilio-pancreatic lesions. Indeed, EUS achieves
high performance in choledocholithiasis diagnosis [1] and avoids two-thirds of unnecessary endoscopic retrograde cholangiopancreatography
(ERCP) procedures [2]. EUS also has a higher yield than magnetic resonance cholangiopancreatography in
diagnostic workup of idiopathic acute pancreatitis assessment [3], and is also the mainstay of exploration of pancreatic solid and cystic masses;
finally, more invasive techniques such as EUS-guided fine-needle aspiration/fine-needle
biopsy and interventional EUS have expanded considerably recently.
Diagnostic endosonography has traditionally been associated with a very low rate of
complications (1 % – 2 %) [4], with perforation, mostly cervical, at less than 0.05 % and associated with risk
factors such as esophageal strictures. However, most reports date back from early
experiences in which duodenal intubation was much less common than it is today. In
addition, many young endoscopists willing to learn EUS have no experience with duodenal
intubation with a side or oblique-viewing endoscope and are not necessarily coached
by a senior endoscopist as thoroughly as during ERCP training, due to the low expected
occurrence of iatrogenic side effects.
Widespread use of EUS makes knowledge and management of these complications mandatory,
even during diagnostic procedures.
Perforation remains one of the most common complications in endoscopy and in EUS. Although
surgery allows simultaneous treatment of perforation and of an associated biliary
or pancreatic disorder, it is invasive, carrying its own morbidity and mortality,
especially when undertaken in the context of an emergency. We wanted to assess the
feasibility, efficacy, and safety of immediate endoscopic repair of EUS-induced duodenal
perforations with over-the-scope clips (OTSC) and report our experience to determine
whether this procedure can be recommended in such cases.
Patients and method
We performed a retrospective study of patients with EUS-related iatrogenic perforation
observed in our unit from 2011 to August 2018.
We included all consecutive patients with EUS-related perforation with immediate diagnosis
who received conservative endoscopic management. Patients with primary surgical management
and conservative, non-interventional management, as well as perforation resulting
from an endoscopic intervention (eg endoscopic sphincterotomy, cyst fenestration,
etc), were excluded.
Preinterventional clinical data included age, sex, procedure indication, endoscopic
report data, perforation location and size. All perforations were diagnosed during
the EUS procedure. Once a perforation was diagnosed, air insufflation was shifted
to CO2, the EUS scope was retrieved, and a standard (10-mm diameter, 3.2-mm working channel)
forward-viewing endoscope was used to identify and examine the perforation site. The
gastroscope was removed, equipped with an OTSC (11/6 mm, type t, Ovesco Endoscopy
AG, Tübingen, Germany), and reintroduced and advanced up to the perforation site.
Tight closure was assessed endoscopically and by contrast injection ([Fig.1]). Once the defect was closed, patients were observed overnight with no oral intake,
continuous low aspiration tube, and intravenous antibiotics.
Fig. 1 Clipping procedure after EUS-related duodenal perforation. a, b Retropneumoperitoneum. c, d Endoscopic view of the perforation. e Perforation closure and injection of contrast showing no leak. f Endoscopic view of the clip positioning.
Outcome data included technical success, clinical success, length of hospital stay,
necessity of transfer to an intensive care unit, adverse events (AE), morbidity and
mortality. Technical success was defined as endoscopic perforation management allowing
secure positioning of the Ovesco clip with tight sealing of the defect and no visible
leak on massive (50 mL) contrast injection in front of the breach. Clinical success
was defined as absence of any post-procedural AE and confirmation of perforation closure
on post-op computed tomography (CT) with digestive contrast swallowing, or on repeat
endoscopic evaluation.
Results
Thirteen EUS-induced duodenal perforations occurred during the 7-year study period,
among 8504 EUS procedures, or a ratio of 0.15 %. All happened during a diagnostic
EUS, whether alone or preceding an ERCP. Two patients were excluded, one for a large
duodenal tear requiring immediate surgery, the other one for misdiagnosis of perforation
leading to early discharge and readmission 24 hours later with peritonitis and emergency
surgery. All but two of 11 included EUS had been initiated by trainees.
[Table 1] shows patient characteristics. All 11 patients were women, mean age of 75 years
(range 68 – 88). Eight of 11 perforations (72.7 %) were due to a radial probe EUS-scope
Olympus GF160 (Olympus America, Center Valley, Pennsylvania, United States) and the
other three (27.3 %) were due to a curvilinear probe EUS-scope Olympus GF180. Indications
for EUS were as follows: assessment of a pancreatic mass in five patients (45 %),
suspected choledocholithiasis in two (18 %), exploration of cholestasis in two (18 %),
staging of an ampullary tumor in one (9 %), and of a cholangiocarcinoma the last patient
(9 %).
Table 1
Study population characteristics.
|
Number
|
|
Age
|
75 (range 68 – 88)
|
|
Sex
|
|
|
11 (100 %)
|
|
Type endoscope
|
|
|
8 (76.7 %)
|
|
|
3 (27.3 %)
|
|
EUS indication
|
|
|
11 (100 %)
|
|
|
5 (45 %)
|
|
|
2 (18 %)
|
|
|
2 (18 %)
|
|
|
1 (9 %)
|
|
|
1 (9 %)
|
|
Perforation location
|
|
|
9 (81 %)
|
|
|
1 (9 %)
|
|
|
1 (9 %)
|
|
Risk factor
|
|
|
3 (27.3 %)
|
|
|
2 (18 %)
|
|
|
1 (9 %)
|
EUS, endoscopic ultrasound; UDF, upper duodenal flexure; LDF, lower duodenal flexure
Perforations were located in the superior part and superior flexure of the duodenum
in nine of 11 (81 %), in the descending part of the duodenum in one of 11(9 %), or
in the inferior duodenal flexure in one of 11 (9 %). Defect size as described in the
endoscopic report ranged from 10 to 15 mm.
All clipping procedures resulted in technical as well as a clinical success as defined
above. However only nine of 11 had CT evaluation after the procedure, with five having
remaining pneumoperitoneum, three a liquid collection of less than 20mm, and one a
suspected persistent contrast leak which was not confirmed thereafter but led to upholding
fasting for this patient; no duodenal defect was seen when the patient underwent cholecystectomy
4 days later.
One patient had a second perforation 6 days later due to an ERCP procedure, which
was managed conservatively by fasting and antibiotics. One patient had a successful
ERCP during the same procedure as EUS immediately before perforation closure with
an OTSC. Three of 11 patients (27 %) stayed in an ICU for less than 72 hours and total
hospital stay ranged from 3 to 22 days ([Table 2]).
Table 2
Perforation clipping details and follow-up.
|
Estimated defect size
|
Technical success
|
ICU stay
|
Surgery
|
CT1
|
CT2
|
Follow-up endoscopy
|
Delayed adverse event
|
Total LOS
|
|
10 – 12 mm
|
Yes
|
No
|
No
|
24 h: PNP, small collection, no leakage
|
7 days: Normal
|
No
|
No
|
12 J
|
|
–
|
Yes
|
48 h
|
No
|
6 days: small collection
|
No
|
1 month: Clip still in place
|
No
|
22 J
|
|
15 mm
|
Yes
|
|
No
|
No
|
No
|
|
No
|
|
|
15 mm
|
Minimal leak
|
72 h
|
No
|
72 h: leak next to common bile duct, PNP
|
No
|
No
|
No
|
|
|
8 mm
|
Yes
|
No
|
No
|
24 h: PNP + retro-PNP No leak
|
72 h: Normal
|
45 days: Clip still in place
|
No
|
7 J
|
|
15 mm
|
Yes
|
No
|
No
|
No
|
No
|
No
|
No
|
7 J
|
|
10 mm
|
Yes
|
|
No
|
72 h: PNP, no leakage
|
No
|
No
|
No
|
|
|
10 – 12 mm
|
Yes
|
24 h
|
No
|
96 h: small collection, no leak
|
No
|
Second perforation after ERCP
|
Yes: Second perforation after ERCP
|
8 J
|
|
15 mm
|
Yes
|
No
|
No
|
24 h: PNP, no leakage
|
No
|
No
|
No
|
3 J
|
|
10 mm
|
Yes
|
No
|
No
|
48 h: normal
|
No
|
1 month: clip still in place
|
No
|
4 J
|
|
10 – 12 mm
|
Yes
|
No
|
No
|
No
|
No
|
No
|
No
|
3 J
|
ICU, intensive care unit; CT, computed tomography; LOS, length of stay; PNP, pneumoperitoneum;
ERCP endoscopic retrograde cholangiopancreatography
Discussion
Gastrointestinal perforation during EUS is rare. Duodenal perforation occurs in 0.022 %
of procedures, according to a German and an international survey conducted from 1982
to 1992 [4]
[5]. More recently another study reported a 0.86 % duodenal perforation rate [6] in patients with suspected pancreatic cancer. As a consequence of broader EUS indications
and of more trainees involved, perforation may have a higher incidence, especially
in high-volume academic units.
Perforation may be due to EUS scope design, since EUS scopes are larger than standard
gastroscopes at 12.4 to 14.6mm vs. 9 to 11 mm and they are also less flexible and
more difficult to maneuver through the proximal duodenum with a side-viewing system,
and a longer rigid distal part containing the ultrasound transducer. All these properties
lead to partially blind advancement in the digestive tract to reach the second duodenal
part, which may cause traumatic lesions, especially in parts with tight angulations.
This explains why the majority of perforations occurred in the upper duodenal flexure,
right after or at the distal end of the bulb, and they were relatively limited in
size, rarely exceeding that of the scope head, or 10 to 15 mm.
As seen in some cases, perforation may also be due to anatomic particularities or
alteration, like an ulcer, a tumoral compression or an inflammatory stenosis. In these
cases, completing EUS and attempts to pass beyond the stenosis may require difficult
maneuvers and induce tension on the duodenal wall, causing the perforation, with risks
of a large tear. Surprisingly in this series and despite studies showing EUS safety
in elderly patients [7]
[8], all perforations occurred in elderly women. That may suggest that aging and female
sex are possible risk factors for EUS-induced perforation.
Perforation remains the most feared of AEs in endoscopy, because its occurrence can
bear serious clinical consequences, especially when it is overlooked or misdiagnosed.
Perforation management may require surgery, infers additional costs and raises liability
issues with legal consequences. Untreated duodenal perforation may progress to acute
peritonitis and septic shock with a high mortality rate. A nationwide survey and a
single-center report documented a total of 14 fatal complications due to duodenal
tears after EUS [9]
[10].
Nevertheless, development of endoscopic devices and techniques for endoscopic management
of perforations by the endoscopist has completely transformed the way we consider
such complications and has allowed more daring and sometimes risk-taking procedures
[11].
However, when duodenal perforation occurs during a diagnostic procedure, endoscopic
management comes with several caveats attached. First, because delayed management
worsens prognosis, the closure procedure must be done immediately, ie within the same
anesthesia, preferably after airway intubation and with the aid of a senior endoscopist
experienced in endoscopic defect closure [12]
[13]. Second, air insufflation must be immediately switched to low-CO2 flow to reduce peritoneal air insufflation. Rapid management limits air inflation
through the defect and switching to CO2 inflation prevents postoperative pain as CO2 is rapidly absorbed. Moreover, rapid defect closure prevents spillage outside the
gastrointestinal tract, which would cause infection and peritonitis. Third, standard
measures of conservative management must be taken post-op, with patient fasting and
broad-spectrum antibiotics covering gram-negative, beta-lactamase-producing and anaerobic
germs. Perforation closure with through-the-scope clips (TTS) has been reported in
case reports and small case series [14]
[15] with good outcomes for small defects in the duodenum. However, for large defects
where many TTS clips are necessary, the procedure is challenging and the outcome uncertain
due to limited space and difficult placement of clips in the duodenum; moreover, TTS
endoclips do not grip the duodenum deeply and can easily fall out long before definitive
sealing of the defect.
Fully-covered metallic stents (FCSEMS) could also be an alternative to clipping and
prevent leakage inside peritoneum. Yet stent radial expansion can stretch the tear.
Moreover, when used for benign conditions, FCSEMS have a high migration rate 62.5 %
[16], making stenting a less reliable method.
Ovesco clip (Tubingen, Germany) is an OTSC made from nitinol with self-memory shape;
provide a tighter, deeper, and larger grasp of tissue than TTS clips. Ovescoclips
have been approved for closure of mural defects up to 18 mm in size [17] and are considered an effective tool for closure of iatrogenic perforations and
their use in the duodenum has previously been reported, although not specifically
for EUS-related duodenal punches and perforations [18]. In our study, all procedures resulted in technical and clinical success. Nevertheless,
clipping procedure might be tricky, and require some expertise. Hands-on training
may help increase familiarity with OTSC clipping. Some specific precautions are mandatory:
a good assessment of defect size and careful introduction of the clip mounted at the
tip of the scope through the upper esophagus, to avoid clip migration or disinsertion
as well as esophageal tears. Before releasing the clip, the twin grasper forceps need
to be completely inside the clip cap to avoid forceps incarceration [19]. After releasing the clip, it is important to assess complete defect closure by
injecting a high volume of contrast and then place an aspiration tube. Patients can
stay in a standard care unit and, if asymptomatic, be rapidly discharged.
ERCP was performed in one patient during the same procedure as EUS and defect closure
was achieved without an additional AE, but carrying on with the planned interventional
procedure risks enlargement of defect size and a more difficult closure, with continuing
insufflation worsening the pneumoperitoneum. A better alternative, apart from a life-threatening
biliary emergency, is to continue the procedure after clipping or even better, a few
days later [18]. Such was our choice in most patients. The second procedure should indeed be gentle
and avoid any tension during scope advancement to prevent tearing around the clip.
Conclusion
Although bearing a low rate of specific complications, diagnostic EUS is not AE -free
and observing proper indications is necessary, especially in elderly patients. Duodenal
perforation is a potentially serious AE, but conservative endoscopic treatment with
OTSC represents a feasible, efficient, and safe treatment that can prevent surgery
in most instances.
