3800 Pancreatic duct leak in necrotizing pancreatitis role of diagnostic and therapeutic ercp as part of a multi-disciplinary approach.

Abstract

Although PD disruption has been documented in some individuals with severe pancreatitis, to date there are no studies which have systematically reviewed ductal integrity and the consequences of treating a disruption, if found. Materials and Methods: Medical records of 100 patients (pts) with severe necrotizing pancreatitis (Balthazar CT score ≥ 6) who presented to VMMC between 6/90 and 4/99. 80 of whom were referred by other medical institutions, were reviewed. PD leak was defined as demonstrable duct disruption at ERCP, injection of the PD through a percutaneous or surgically placed drain, and/or a persistent high volume, high amylase (≥2500 IU) output from a JP drain. ERCP findings and treatment results were reviewed and placed into the perspective of overall therapy. Results: 78 pts (78%), 54 M/24 F had a documented PD leak. 56 of these pts (72%) had 62 PD leaks documented by ERCP: head 11, genu 15, body 23, tail 12, and side branches 1; whereas an additional 22 pts had no demonstrable leak. Concomitant ERCP findings included CBD stricture 28, PD stricture 20, PD stone 6, and CBD/PD fistula 6. Endoscopic treatment included placement of PD stent in 43 pts, with leak resolution in 32/40 (80%), CBD stent placement in 29, CBD and/or PD sphincterotomy 46, cystgastrostomy 11, NP drain placement 5, cystduodenostomy 3, and NB drain 2. Possible acute complications occurred in 9 pts and included post-ERCP pain exacerbation 5, bleeding 1, cholangitis 1, and positive bacterial cultures noted in JP drain sites 4 (9/89, 10%). Additional therapies in the 100 pts included CT drainage 73, and surgeries 55 (cholecystectomy 26, debridement 15, cystgastrostomy 2, Peustow 1, pancreatic resection 36, revision of pancreaticojejunostomy 4). There were a mean of 1.85 ± 0.2 hospitalizations and 79/100 pts had a mean hospital stay of 25 ± 3 days, although 2 pts had a LOS >4 months. There were 7 deaths in the series including 4 with PD leak and 3 without. Conclusions: 1) Severe pancreatic necrosis was associated with PD leak in 78% of pts in the series and may be etiologic in the severity of the disease process. 2) Although not risk-free, ERCP proved useful in both the diagnostic (72%) and successful resolution (80%) of these leaks. 3) The relatively low-mortality rate (7%) in this series may be a consequence of leak resolution and/or the multi-disciplinary approach to pancreatic necrosis practiced in our institution.