PWE-063 Evaluation Of a Previously Described Scoring System for Predicting Complications Post Ercp

Introduction Endoscopic retrograde cholangiopancreatography (ERCP) is widely performed for the management of biliary and pancreatic duct disorders. Despite this, ERCP is associated with a significant complication rate, with the literature reporting rates of post-procedural pancreatitis of 1–5%, cholangitis in 1–5% and haemorrhage in 1%. A previous study (Jeurnink et al 20111) described a prognostic model for predicting those patients at greater risk of developing post ERCP complications, and identifying those who may be safely discharged shortly after ERCP. The aim of this study was to validate this scoring system in an external cohort to assess whether it can be used in general clinical practise. Methods Details of all patients undergoing ERCP over the 22 month period from May 2010 to February 2012 were recorded on an institutionally approved database. Electronic records were subsequently accessed to identify post ERCP complications within 30 days of procedure. The predictive score as described was retrospectively calculated and applied to all patients, with a score > 3 being considered high risk. Sensitivity, specificity, negative and positive predictive values were then calculated. Results 697 patients (409 females, mean age 64, mean ASA grade 2.35) underwent ERCP during the study period. The overall complication rate was 9.0% (63/697); cholangitis 2.3% (n = 16), pancreatitis 2.1% (n = 15), bleeding 1.6% (n = 11), perforation 1.3% (n = 9) and miscellaneous in 1.7% (n = 12). The mortality rate was 0.4% in our cohort (n = 3). 681/697 (97.7%) had a predictive score < 4 but ERCP grade 1/2/3 was 531/149/17 respectively. Of those with a predictive score ≥4, 12.5% (n = 2/16) developed a post-ERCP complication (both severe pancreatitis) versus 8.4% (n = 57/681) with a score < 4 (p = ns). Using the predictive score gave a sensitivity of 3.4%, specificity of 97.8%, positive predictive value of 13% and a negative predictive value of 92%. Conclusion The predictive scoring system as previously described does not accurately stratify patients into high or low risk groups or predict post-ERCP complications in our cohort. This may be due to case mix in the original cohort leading to lack of generalisation. Further work is needed to formulate a clinically applicable scoring system which has higher accuracy. Disclosure of Interest None Declared. Reference Jeurnink SM, Siersema PD, Steyerberg EW, Dees J, Poley JW, Haringsma J, Kuipers EJ. Predictors of complications after