Backgrounds/aims: Endoscopic retrograde cholangiopancreatography (ERCP) remains the primary treatment for a subset of patients with pancreatic fistulae. The objective of this study was reporting outcomes of ERCP and predictors of resolution in patients with pancreatic fistulae refractory to conservative therapy.
Methods: Retrospective review of patients who underwent ERCP and pancreatic stent placement for pancreatic fistula not responding to medical therapy was performed. Clinical features, laboratory parameters, radiological features and pancreatogram findings were noted. Clinical resolution of fistula was the primary outcome measure.
Results: Sixty-eight patients underwent ERCP for high-output pancreatic fistula (Mean age 34.1 years, 91.1% males, 35/68 chronic pancreatitis, 52.9% alcohol etiology). Internal fistulae (pancreatic ascites, pleural effusion, or pericardial effusion) were seen in 55 (80.9%) patients and external fistula in 13 (19.1%) patients. Technical success for ERCP was 92.6% (63/68). Leak was seen in 98.4% (62/63). The most common leak site was body (69.8%). Multiple leak sites were seen in 23.1%. Pancreatic stricture was found in 36.5%. In 44 (69.4%) patients, stent was placed beyond the site of the leak. Resolution at six weeks was achieved in 76.4% (52/68). On univariate and multivariate analyses, placement of stent beyond site of leak was significantly associated with resolution of high-output fistulae (3/41 [7.3%] vs. 5/19 [26.3%], p = 0.03; odds ratio: 6.5, 95% confidence interval: 1.211-34.94).
Conclusions: In our experience, ERCP was successful in 76% of patients with pancreatic fistulae refractory to conservative therapy. Stent placement beyond the site of leak was associated with higher resolution of fistulae.
Keywords: Acute pancreatitis; Chronic pancreatitis; Endoscopic retrograde cholangiopancreatography; Pancreatic fistula.