Su1258 “INNOCENT AS A LAMS.” DOES SPONTANEOUS GASTROGASTRIC/JEJUNOGASTRIC FISTULA CLOSURE (HEALING BY SECONDARY INTENTION), AFTER EUS-DIRECTED TRANSGASTRIC ERCP (EDGE) VIA 20-MM LAMS, CONFER AN INCREASED RISK OF PERSISTENT FISTULA AND UNINTENTIONAL WEIGHT GAIN?

Abstract

EUS-directed transgastric ERCP (EDGE) is an effective modality for treating pancreaticobiliary disorders in Roux-en-Y gastric bypass (RYGB) anatomy. EDGE is a 2-in-1 procedure that consists of EUS-directed gastrogastrostomy (or jejunogastrostomy) creation using a lumen-apposing metal stent (LAMS), followed by, transgastric ERCP. After completion of transgastric ERCP and maturation of the gastogastric fistula (GGF) or jejunogastric fistula (JGF), the LAMS is removed with the intention of re-excluding the bypassed stomach. Closure of the GGF/JGF can occur via primary intent (endoscopic closure) or secondary intent (spontaneous closure). Persistent fistula is a concerning sequela of EDGE, as it may functionally reverse the bypass. Prior EDGE studies, using 15-mm LAMS, report persistent fistula rates of 4 – 20%, without a clear association with weight gain. Our study measures the incidence of persistent fistula and its association with unintentional weight gain, among patients whose fistula was allowed to spontaneously heal, after EDGE using 20-mm LAMS. We conducted a single-center retrospective study of 16 RYGB patients who underwent EDGE using 20-mm by one advanced endoscopist between 3/2018 – 9/2019. After LAMS extraction, all GGFs/JGFs were allowed to heal spontaneously. Objective testing for persistent fistula occurred 5-months (median) after LAMS extraction, using upper gastrointestinal series or EGD. A GGF/JGF was considered “persistent” if detected at least 2-months after LAMS extraction. Change in total body weight (TBW) was calculated relative to pre-EDGE (baseline) TBW. All statistical analyses were performed on R (Version 3.6.1). Persistent fistula was identified in 50% of patients (n = 8) (Table 1). A longer LAMS dwell time (median 89-days) was observed in the persistent fistula group, compared to those without persistent fistula (median 34-days) (p = 0.0499). Weight gain of > 5% TBW occurred in 50% (n = 4) of patients with persistent fistula, compared to 25% (n = 2) of patients without persistent fistulas (p = 0.6084). A 50% persistent fistula rate is the highest reported incidence to date (compared to studies using 15-mm LAMS). We hypothesize that 20-mm LAMS leaves a wider transmural defect than 15-mm LAMS, and that healing by secondary intent may be slower than by primary intent. Persistent fistula was significantly associated with a longer LAMS dwell time, possibly due to more permanent tissue remodeling around the LAMS. Weight gain was not significantly associated with persistent fistula; thus, we question the utility of routine fistula closure (by primary intent) and suggest a personalized approach to follow-up. Multicenter collaboration is needed to better understand the natural history and optimal management of persistent fistula occurring after EDGE.