Endoscopic versus surgical treatment for thoracopancreatic fistulas complicating chronic pancreatitis: A systematic review.

Background: Thoracopancreatic fistulae are a rare complication of chronic pancreatitis. The aim of the present study is to evaluate potential risk factors for endoscopic treatment failure and explore the safety of surgery when utilized either upfront or as a "bail-out" procedure after failed endoscopic treatment.

Method: A comprehensive literature search was conducted on the MedLine, Scopus, Embase, and Web of Knowledge databases for cases of thoracopancreatic fistulae. Data regarding patient demographics, fistula anatomy, and treatment interventions performed were extracted for further analysis.

Results: The study pool consisted of 75 case reports and 19 case series published between the years 1972 and 2020. Duct disruption in the pancreatic body was most commonly encountered (41.1%), and a left pleural effusion was the most common manifestation (46%). Endoscopic treatment was attempted for 104 patients with an overall success rate of 42.3% (n = 44). Predictive factors for eventual success of endoscopic treatment were the ability of endoscopic retrograde cholangiopancreatography to diagnose the thoracopancreatic leak (odds ratio 9.76, 95% confidence interval 2.71-35.09, P < .001), the use of pancreatic duct stents (odds ratio 22.1, 95% confidence interval 7.92-61.61, P < .001), and the use of sphincterotomy (odds ratio 7.96, 95% confidence interval 2.1-30.1, P < .001). Conversely, the presence of pancreatic duct calculi was associated with endoscopic treatment failure (odds ratio 0.34, 95% confidence interval 0.12-0.94, P = .03). Pooled results suggest that surgical outcomes were comparable between the primary and salvage surgery groups.

Conclusion: A step-up approach from endoscopic management to salvage surgery may be effectively employed in cases of thoracopancreatic fistulae refractory to endoscopic treatment.