Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery

BACKGROUND: Reoperations for complications of bariatric surgery are associated with high morbidity and mortality. It is not known whether endoscopic treatment may reduce reoperation rates. METHODS: Twenty-one patients underwent endoscopic treatment for persisting large anastomotic leaks before considering redo surgery. Eight patients had a gastric bypass, eight had a sleeve gastrectomy combined with a duodenal switch (SDS), four had a sleeve gastrectomy alone, and one had a Scopinaro procedure (biliopancreatic diversion). Fistulas were gastrocutaneous in 15 patients, duodenocutaneous in 2, gastroperitoneal in 3, and gastrobronchial in 1. Partially covered self-expanding metal stents (SEMSs) were used, followed by additional endoscopic procedures if the SEMS failed. SEMSs were removed by traction alone or by insertion of a self-expanding plastic stent (SEPS) followed by extraction of both stents together. RESULTS: SEMS insertion led to 62 % (13/21) primary closures. Complementary endoscopic treatment led to 4 secondary closures. Total success rate was 81 % (17/21). Three patients in whom SEMSs failed underwent reoperation but died during postoperative follow-up; one patient died from pulmonary embolism before SEMS extraction. The success rates of endotherapy were 100 % (8/8) in the gastric bypass group, 62.5 % (5/8) in the SDS group, 75 % (3/4) in the sleeve gastrectomy group, and 100 % (1/1) for the Scopinaro procedure. Gastrocutaneous fistulas on sleeve sutures were successfully treated in 60 % of cases (6/10), while other anastomotic fistulas were successfully treated in 100 % of cases (11/11) ( P = 0.0351). CONCLUSIONS: Endoscopic treatment using SEMSs for complications of bariatric surgery is feasible. Healing of severe leaks was obtained in 81 % (17/21) of patients, avoiding high-risk reintervention. Gastrocutaneous fistulas on a sleeve suture are the most difficult condition to treat.