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

Background. Post-ERCP pancreatitis is poorly understood. The goal of this s tudy was to comprehensively evaluate potential procedure- and patient-relat ed risk factors for post-ERCP pancreatitis over a wide spectrum of centers. Methods: Consecutive ERCP procedures were prospectively studied at 11 cente rs (6 private, 5 university). Complications were assessed at 30 days by usi ng established consensus criteria. Results: Pancreatitis occurred after 131 (6.7%) of 1963 consecutive ERCP pr ocedures (mild 70, moderate 55, severe 6). By univariate analysis, 23 of 32 investigated variables were significant. Multivariate risk factors with ad justed odds ratios (OR) were prior ERCP-induced pancreatitis (OR 5.4), susp ected sphincter of Oddi dysfunction (OR 2.6), female gender (OR 2.5), norma l serum bilirubin (OR 1.9), absence of chronic pancreatitis (OR 1.9), bilia ry sphincter balloon dilation (OR 4.5), difficult cannulation (OR 3.4), pan creatic sphincterotomy (OR 3.1), and 1 or more injections of contrast into the pancreatic duct (OR 2.7). Small bile duct diameter, sphincter of Oddi m anometry, biliary sphincterotomy, and lower ERCP case volume were not multi variate risk factors for pancreatitis, although endoscopists performing on average more than 2 ERCPs per week had significantly greater success at bil e duct cannulation (96.5% versus 91.5%, p = 0.0001). Combinations of patien t characteristics including female gender, normal serum bilirubin, recurren t abdominal pain, and previous post-ERCP pancreatitis placed patients at in creasingly higher risk of pancreatitis, regardless of whether ERCP was diag nostic, manometric, or therapeutic. Conclusions: Patient-related factors are as important as procedure-related factors in determining risk for post-ERCP pancreatitis. These data emphasiz e the importance of careful patient selection as well as choice of techniqu e in the avoidance of post-ERCP pancreatitis.