Risk factors for post-ERCP pancreatitis: a prospective multicenter study.

OBJECTIVES

Pancreatitis is the most common and serious complication of diagnostic and therapeutic ERCP. The aim of this study is to examine the potential patient- and procedure-related risk factors for post-ERCP pancreatitis in a prospective multicenter study.

METHODS

A 160-variable database was prospectively collected by a defined protocol on patients undergoing diagnostic or therapeutic ERCP at 15 centers in the Midwest Pancreaticobiliary Group and participating in a randomized controlled study evaluating whether prophylactic corticosteroids will reduce the incidence of post-ERCP pancreatitis. Data were collected prior to the procedure, at the time of procedure, and 24-72 h after discharge. Post-ERCP pancreatitis was diagnosed and its severity graded according to consensus criteria.

RESULTS

Of the 1,115 patients enrolled, diagnostic ERCP with or without sphincter of Oddi manometry (SOM) was performed in 536 (48.1%) and therapeutic ERCP in 579 (51.9%). Suspected sphincter of Oddi dysfunction (SOD) was the indication for the ERCP in 378 patients (33.9%). Pancreatitis developed in 168 patients (15.1%) and was graded mild in 112 (10%), moderate in 45 (4%), and severe in 11(1%). There was no difference in the incidence of pancreatitis or the frequency of investigated potential pancreatitis risk factors between the corticosteroid and placebo groups. By univariate analysis, the incidence of post-ERCP pancreatitis was significantly higher in 19 of 30 investigated variables. In the multivariate risk model, significant risk factors with adjusted odds ratios (OR) were: minor papilla sphincterotomy (OR: 3.8), suspected SOD (OR: 2.6), history of post-ERCP pancreatitis (OR: 2.0), age <60 yr (OR: 1.6), > or =2 contrast injections into the pancreatic duct (OR: 1.5), and trainee involvement (OR: 1.5). Female gender, history of recurrent idiopathic pancreatitis, pancreas divisum, SOM, difficult cannulation, and major papilla sphincterotomy (either biliary or pancreatic) were not multivariate risk factors for post-ERCP pancreatitis.

CONCLUSION

This study emphasizes the role of patient factors (age, SOD, prior history of post-ERCP pancreatitis) and technical factors (number of PD injections, minor papilla sphincterotomy, and operator experience) as the determining high-risk predictors for post-ERCP pancreatitis.