We have recently shown that ERCP is the most useful technique for detecting a biliary origin of acute pancreatitis and can be done without side effects. We now report on a second series of 50 patients with acute pancreatitis in whom ERCP, computed tomography (CT), ultrasound (US), and clinical and laboratory assessment were performed within the first 24 to 48 hours of hospitalization. A score for ERCP, CT and US was used to assess the severity of the disease. Patients were followed up until discharge or death and their condition classified according to outcome as mild (less than or equal to 1 complication), severe (greater than 1 complication) or fatal. ERCP was superior in detecting choledochal stones (ERCP 100%, US 25%, CT 50%) and dilated intrahepatic ducts (ERCP 75%, US 75%, CT 37%) but not gallbladder stones (ERCP 70%, US 100%, CT 60%). When the ERCP severity score was calculated there was no relevant difference between patients thereafter having a mild course (0.66 +/- 0.91, range 0-3), a severe course (1.3 +/- 0.80, range 0-3), or a fatal outcome (1.0 +/- 1.1, range 0-3). In contrast, the CT score was different in all three groups (mild: 3.0 +/- 1.9; severe: 5.3 +/- 3.2; lethal: 6.3 +/- 3.1) as was the US score (mild: 1.5 +/- 1.3; severe: 3.2 +/- 2.3; lethal: 4.4 +/- 1.4). It is concluded from these results that ERCP is of value in defining the origin of acute pancreatitis. When a biliary origin is detected this can lead to immediate treatment using endoscopic sphincterotomy.(ABSTRACT TRUNCATED AT 250 WORDS)