Outcomes and predictors of success of endoscopic retrograde pancreatic drainage for pancreatic leaks and fistulas: a multicenter retrospective study

Pancreatic duct leaks and fistulas are challenging complications of pancreatitis and pancreatic surgery. Although endoscopic retrograde pancreatic duct drainage (ERPD) is widely used, optimal drainage strategies remain incompletely defined. This study aimed to evaluate outcomes and identify predictors of clinical success of ERPD in patients with pancreatic leaks and fistulas. We conducted a multicenter retrospective study of consecutive patients with pancreatic duct leaks or fistulas refractory to conservative therapy who underwent ERPD at four tertiary centers between July 2007 and August 2023. Fistulas were classified using an endoscopy-oriented system. Drainage strategies were categorized as type A (stent placed across the leak), type B (stent placed adjacent to the leak), or type C (stent placed into the collection cavity). The primary outcome was clinical success within 3 months. Independent predictors of clinical success were assessed using multivariable logistic regression. A total of 106 patients were included. Technical success was achieved in 98.2% of procedures, and overall clinical success was 79.2%. Clinical success rates were similar between pancreatitis-related leaks and postoperative or traumatic leaks (78.9% vs. 80.0%, P = 0.893). Leak-bridging drainage (type A) was associated with a significantly higher clinical success rate than non–leak-bridging drainage (94.4% vs. 71.4%). Type IIC fistulas had the lowest success rate (36.4%). On multivariable analysis, leak-bridging drainage independently predicted clinical success (OR 6.5, 95% CI 1.3–31.2; P = 0.020), whereas type IIC fistulas were associated with reduced odds of success (OR 0.114, 95% CI 0.027–0.476; P = 0.003). When leak bridging was not feasible, type B drainage achieved higher success than type C drainage (83.3% vs. 53.6%, P = 0.007). ERPD is a safe and effective treatment for pancreatic duct leaks and fistulas. Leak-bridging drainage is strongly associated with clinical success and should be prioritized whenever feasible. When bridging cannot be achieved, stent placement adjacent to the leak appears more effective than intracavitary drainage. Type IIC fistulas remain challenging and may require alternative or adjunctive interventions.