MAJOR HEMORRHAGE FROM ENDOSCOPIC SPHINCTEROTOMY - RISK FACTOR-ANALYSIS

We carried out a retrospective cohort study on all patients undergoing sphincterotomy at our institution over a 4-year period. Major hemorrh age occurred in 10 of 189 patients (5.3%). Onset was usually delayed ( mean, 3.0 days; range, 0-9 days). Six potential risk factors for posts phincterotomy hemorrhage were assessed by univariate and multivariate analysis. Three factors predicted postsphincterotomy hemorrhage: hemod ialysis (relative risk, 8.4; 95% confidence interval, CI, 2.7-26.4), a prothrombin time prolonged at least 2 s above control (relative risk, 7.8; 95% CI, 2.4-25.6), and endoscopically observed bleeding at the t ime of sphincterotomy (relative risk, 5.9; 95% CI, 1.7-20.1). Features not independently associated with hemorrhage were sphincter of Oddi d ysfunction, aspirin or nonsteroidal anti-inflammatory drug (NSAID) use within 1 week prior to sphincterotomy, and sphincterotomy length. Whe n differentiated from endoscopically observed bleeding, clinically sig nificant hemorrhage was usually a delayed complication, primarily in p atients with hemostatic defects.