Same-day discharge after endoscopic biliary sphincterotomy: observations from a prospective multicenter complication study

Background: Same-day discharge after endoscopic biliary sphincterotomy (ES) is a common clinical practice, but there have been few data to guide appro priate selection of patients. Using a prospective, multicenter database of complications, we examined outcomes after same-day discharge as it was prac ticed by a variety of endoscopists and evaluated the ability of a multivari ate risk factor analysis to predict which patients would require readmissio n for complications. Methods: A 150-variable database was prospectively collected at time of ES, before discharge and again at 30 days in consecutive patients undergoing E S at 17 centers. Complications were defined by consensus criteria and inclu ded all specific adverse events directly or indirectly related to ES requir ing more than 1 night of hospitalization. Results: Six hundred fourteen (26%) of 2347 patients undergoing ES were dis charged on the same day as the procedure, ranging from none at 6 centers to about 50% at 2 centers. After initial observation and release, readmission to the hospital for complications occurred in 35 (5.7%) of 614 same-day di scharge patients (20 pancreatitis and 15 other complications, 3 severe). Of the same-day discharge patients, readmission was required for 14 (12.2%) o f 115 who had at least one independently significant multivariate risk fact or for overall complications (suspected sphincter of Oddi dysfunction, cirr hosis, difficult bile duct cannulation, precut sphincterotomy, or combined percutaneous-endoscopic procedure) versus 21 (4.2%) of 499 without a risk f actor (odds ratio 3.1: 95% confidence interval [1.6, 6.3], p < 0.001). Of c omplications presenting within 24 hours after ES, only 44% presented within the first 2 hours, but 79% presented within 6 hours. Conclusions: Same-day discharge is widely utilized and relatively safe but results in a significant number of readmissions for complications. For pati ents at higher risk of complications, as indicated by the presence of at le ast one of five independent predictors, observation for 6 hours or overnigh t may reduce the need for readmission.