Abstracts
2000 Digestive Disease Week

# 2310 Determinants of Survival After Surgery for Pancreatic Necrosis Complicating Acute Pancreatitis.
Garth C Beattie, James J. Powell, David Swan, Krishnakumar Madhavan, Ajith K. Siriwardena, Edinburgh, United Kingdom

Introduction: Some specialist centers report mortality rates in the order of 10% in patients undergoing pancreatic necrosectomy for severe acute pancreatitis. The indications for, and the optimum timing of surgery are now better defined. However, these excellent results may not be readily translatable to wider practice and publication bias may result in a false perception of current outcomes after pancreatic necrosectomy. Aims: The aim of this study was to examine outcome in patients undergoing pancreatic necrosectomy and to identify factors influencing survival. Methods: During the 8 year period January 1991 to March 1999, 382 patients were admitted to the Royal Infirmary of Edinburgh with a diagnosis of acute pancreatitis. Fifty four(14%) underwent surgical necrosectomy. Median age (range) was 55 years (19-74). Thirty three (61%) were tertiary referrals with the median delay to referral being 1 day (0-74). Most frequent etiologic agents were gallstones in 28 (52%), alcochol in 14 (26%). Prophylactic antibiotic therapy was prescribed in 32(59%). Median delay to necrosectomy was 26 days (2-83). Results: Multiple factor stepwise logistic regression analysis of the following variables (age, admission APACHE II score, admission multiple organ dysfunction score (MODS), pre-operative MODS, post-operative MODS, use of prophylactic antibiotics, delay to referral, presence of infected necrosis, presence of polymicrobial infection, use of post-operative irrigation, duration of post-operative irrigation and delay to necrosectomy revealed that delay to operation independently predicted adverse outcome (p=0.008). Operation was associated with an initial adverse effect on MOD score [admission MODS 3 (0-12), pre-operative MODS 3 (0-14) and 24 hours postoperative MODS 6 (0-16)]. There were 23 deaths (43% mortality). Ten of these deaths were attributed to refractory multiple organ failure, 5 to sepsis, 4 to exsanguinating hemorrhage and 1 each to stroke, respiratory failure, cardiac failure and bacterial endocarditis. Mortality rate in the first 4 years was 35% (7/20) and 47% (16/34) in the second period (p=NS). Median duration of hospital stay was 84 days (4-307) with no differnce in stay between survivors and those succumbing to disease. Conclusions: Although it is probable that many inter-related factors combine to influence outcome, delay to surgery is an independent predictor of adverse outcome. The excellent results reported from some centers represent one aspect of a spectrum of outcomes.