Fifty consecutive patients with acute pancreatitis were assessed with respect to a biliary origin of the disease. Endoscopic retrograde cholangiopancreaticography, surgery, and autopsy were used to define biliary pancreatitis. Ultrasound, computed tomography, and several laboratory tests (SGOT, SGPT, alkaline phosphatase, and bilirubin) were analyzed for their ability to detect a biliary origin of the disease. Ultrasound and computed tomography could not reliably make the diagnosis in the 10 patients found to have biliary disease. Receiver-operator-characteristic curves revealed that none of the laboratory tests assessed had sufficient sensitivity and specificity to determine the diagnosis, although all tests showed higher mean values in biliary pancreatitis. SGPT gave the best discrimination (positive predictive value 53%, negative predictive value 94%, cut off 40 units/liter). Therefore, initial ERCP is suggested for a reliable diagnosis of biliary origin of acute pancreatitis.

Endoscopic retrograde cholangiopancreatography (ERCP) was attempted in 32 patients and successfully performed in 31 patients with acute pancreatitis during the initial 12 h of hospitalization. 29% of the patients were found to have biliary pancreatitis, and in 4 patients common bile duct stones were removed. Abnormalities of the main pancreatic duct and the side branches were found in 89% of the patients with nonbiliary pancreatitis. In particular, defects in the filling of the pancreatic ducts, suggesting protein plugs and increased viscosity of the pancreatic secretion, were observed in 11 patients. In all but 1 patient a decrease of serum amylase was observed 1 day after ERCP. This preliminary study leads us to conclude that ERCP in acute pancreatitis is: (i) not harmful; (ii) detects unsuspected biliary pancreatitis better than other methods, and (iii) reveals abnormalities of the pancreatic ducts which may be of importance for the pathogenesis of nonbiliary pancreatitis.

Fifty consecutive patients suffering from acute pancreatitis were studied prospectively with regard to the role of ultrasound (US) in diagnosis, in detection of a biliary origin, and in initial assessment of prognosis. In six patients the pancreas could not be visualised, whereas in 19 only partial examination was possible. In 34% no diagnostic abnormalities were found. US was superior to computed tomography (CT) in respect to the detection of small amounts of ascites, but less suitable than CT for the detection of necrosis. Compared to endoscopic retrograde cholangiography US was of little help in detecting bile duct stones. Neither a lethal outcome nor a severe course could be predicted with sufficient accuracy. The positive predictive value for the presence of necrosis was 33% and the negative predictive value 67%. The data demonstrate a limited role of US in diagnosis and staging in acute pancreatitis.