EUS-directed transgastric interventions in Roux-en-Y gastric bypass anatomy: a multicenter experience.

ABSTRACT

BACKGROUND AND AIMS: Placement of a lumen-apposing metal stent (LAMS) between the gastric pouch and the excluded stomach allows for EUS-guided transgastric interventions (EDGIs) in patients with Roux-en-Y gastric bypass (RYGB). Although EUS-guided transgastric ERCP (EDGE) outcomes have been reported, data are scant on other endoscopic interventions. We aimed to evaluate the outcomes and safety of EDGIs. METHODS: This retrospective study involved 9 centers (United States, 8; Europe, 1) and included patients with RYGB who underwent EDGIs between June 2015 and September 2021. The primary outcome was the technical success of EDGIs. Secondary outcomes were adverse events (AEs), length of hospital stay, and fistula follow-up and management. RESULTS: Fifty-four EDGI procedures were performed in 47 patients (mean age, 61 years; 72% women), most commonly for the evaluation of a pancreatic mass (n = 16) and management of pancreatic fluid collections (n = 10). A 20-mm LAMS was used in 26 patients and a 15-mm LAMS in 21, creating a gastrogastrostomy in 37 patients and jejunogastrostomy in 10. Most patients (n = 30, 64%) underwent a dual-session EDGI, with a median interval of 17 days between the 2 procedures. Single-session EDGI was performed in 17 patients, of whom 10 (59%) had anchoring of the LAMS. The most common interventions were diagnostic EUS (with or without FNA or fine-needle biopsy sampling; n = 28) and EUS-guided cystgastrostomy (n = 8). The mean procedural time was 97.6 ± 78.9 minutes. Technical success was achieved in 52 patients (96%). AEs occurred in 5 patients (10.6%), of which only 1 AE (2.1%) was graded as severe. Intraprocedural LAMS migration was the most common AE, occurring in 3 patients (6.4%), whereas delayed spontaneous LAMS migration occurred in 2 (4.3%). Four of the 5 LAMS migration events were managed endoscopically, and 1 required surgical repair. LAMS anchoring was found to be protective against LAMS migration (P = .001). The median duration of hospital stay was 2.1 ± 3.7 days. Of the 17 patients who underwent objective fistula assessment endoscopically or radiologically after LAMS removal, 2 (11.7%) were found to have persistent fistulas. In 1 case the fistula was intentionally left open to assist with weight gain. The other fistula was successfully closed endoscopically. CONCLUSIONS: EDGI is effective and safe for the diagnosis and management of pancreaticobiliary and foregut disorders in RYGB patients. It is associated with high rates of technical success and low rates of severe AEs. LAMS migration is the most common AE with evidence that anchoring can be protective against its occurrence. Persistent fistulas may occur, but endoscopic closure seems to be effective.