duodenal 900

Things are not always what they seem: A rare and serious complication in a child on
veno-venous ECMO
ABSTRACT:
Introduction: ECMO-related complications may conceal serious underlying problems.
Case report: A 22 month-old male patient was put on veno-venous ECMO for respiratory
failure using a 19 French dual-lumen cannula, with an additional cannula in the iliac vein to
optimize drainage. Following the development of abdominal distention, a peritoneal tap
returned fresh venous blood. Iliac vein perforation by the cannula was suspected. Urgent
laparotomy revealed against all odds, a perforated duodenum, which was primarily repaired.
Discussion: The management of vessel perforation by the ECMO cannula is a nightmare:
Limited cannulation sites; hemodynamic instability; Cessation of anticoagulation; massive
transfusions; transfer to radiology for a CT scan. These concerns may divert the attention from
other serious underlying complications.
Conclusion: With increasing use of ECMO, more attention should be given to rare
complications related to the underlying diseases. This is the first reported case of a duodenal
perforation in a child on ECMO.

Introduction:
Extracorporeal membrane oxygenation (ECMO) is a method of supporting the heart and or the
lungs while native organ function recovers. During the last 2 decades, the role of veno-venous
(VV) ECMO for refractory respiratory failure in the pediatric population has been established as
a lifesaving technique (1). However, despite extraordinary technical progress, ECMO-related
complications remain a real concern in 40% of the cases (1,2). Common complications include
bleeding, infection, and thromboembolic events (2,3). Less common complications include the
dislodgment of a cannula, or the perforation of a cannulated vessel (2-4). Critically ill children
may also present unique complications rarely encountered in the current era; a
perforated bleeding duodenal ulcer is a rare occurrence in modern day medicine (5).This
potentially fatal complication may present insidiously, especially when all fingers point at the
cannulas in a patient on ECMO. To the best of our knowledge, this is the first reported case of
a duodenal perforation in a child on ECMO.
Case:
A 22 month-old boy was transferred to our institution after 17 days in ICU for acute
respiratory failure secondary to H1N1 infection. V-V ECMO was initiated using a 19 French
dual lumen cannula. Suboptimal venous drainage necessitated the insertion of a second
drainage cannula, ideally in the inferior vena cava. The only available short 14 French cannula
didn’t reach the vena cava, and was positioned in the iliac vein. Despite appropriate support
with ECMO, the patient remained critical with borderline vital signs and urine output. There
were no signs of infection; he was on broad spectrum antibiotic therapy, antiviral treatment
with oseltamivir, total parenteral nutrition, and gastric protection with ezomeprazole. Laboratory

work including chemistries, coagulation parameters, and urinalysis were all within normal
limits. All ttempts at enteral nutrition were not tolerated.
On ECMO day 5, the patient started developing increasing abdominal distension, with
significant ascitis on ultrasound. A drop in hemoglobin from 9.8 to 8.0 g/dl was attributed to
circuit hemolysis. Ultrasound guided peritoneal tap returned 900 ml of fresh venous blood, and
continued draining fresh blood. Injury to the iliac vein by the inappropriate venous cannula was
immediately suspected. Although the patient was hemodynamically stable, it was deemed risky
to waste time performing a CT scan, and he was taken directly to the operating room.
The femoral cannula was removed but no perforation was identified. Tracking the
source of bleeding lead to the right upper abdominal quadrant. Hepatic vein injury from the tip
of the dual lumen cannula was then suspected, but no perforation was found.
Further dissection revealed some bile efflux from an area involving the transverse colon
and the duodenum. The exploration of this inflammatory and fibrous area exposed a 1cm
perforation (figure 1A) involving the first portion of the duodenum. The aspect of the
surrounding necrotic tissues was suggestive of chronicity. The perforation was closed and re-
enforced with an omental patch. The post-operative course was uneventful, and ECMO
support was continued using the initially inserted dual-lumen cannula for 34 days. The child is
currently going through rehabilitation physical therapy.
Because of the chronic aspect of the perforation, we retrieved the abdominal x-rays
taken upon arrival to the hospital, and discovered a large pneumoperitoneum (figure 1B).
Discussion:
Even though survival rates are improving, the journey while on ECMO is fraught with
complications. The most common complications for pediatric patients on VV ECMO include

cannula-related bleeding 16%, other cannula-related problems 15%, culture-proven infections
10.3%, cerebral hemorrhage 7%, and gastro-intestinal hemorrhage 4.9% (1,2).Few studies
discussed cannula related perforations, reporting percentages reaching 2% (3,4,6). In addition
to being an emergency, it is a dreaded complication and its management is a nightmare:
Limited alternative cannulation sites; hemodynamic instability; Cessation of anticoagulation;
massive transfusions needs, adding to lung injury; transfer of a critically ill, bleeding patient on
ECMO for a CT scan; and the need for urgent surgery. We decided to avoid this series of
challenges by operating before hemodynamic instability, with the presumed diagnosis of
vessel perforation. Our diagnosis proved to be incorrect, but all intensivists and surgeons with
ECMO expertise didn’t regret this decision. Senior intensivists and non-cardiac pediatric
surgeons considered for a moment a gastrointestinal cause for the bleeding, but were
overwhelmed by the prevailing diagnosis of vessel perforation and its potential consequences.
Perforated peptic ulcer in children is a rare cause of acute abdomen today, due to the
decreased prevalence of Helicobacter Pylori, and the improved efficacy of proton pump
inhibitors (5). While the literature about this disease is outdated in western countries, it is still
reported in developing nations, most cases being secondary to an underlying etiology including
steroid therapy, severe illness ( meningitis, malaria, and lymphoma), or increased intracranial
pressure (7,8). On the other hand, several rare gastrointestinal complications have been
reported in children on ECMO, including hemorrhage, emphysematous gastritis and gastric
perforation (9). In our patient, duodenal perforation occurred probably in the referring hospital,
during the first 2 weeks, when he was battling a severe respiratory infection with H1N1,
because the pneumoperitoneum was present upon arrival; had the perforation been diagnosed
early enough, the patient’s condition would have probably improved without ECMO. In the
adult population, 1 case of fatal perforated peptic ulcer was reported following ECMO removal,
and not during the ECMO run (10).

Conclusion
With the increasing indications for ECMO in critical care, it remains important to spread
awareness about the less common complications related to the underlying diseases.