Objectives: Transabdominal ultrasonography (US) usually reveals diagnosis of biliary acute pancreatitis (AP). Guidelines suggest repeating US in AP patients without cause at first examination. This approach has been poorly investigated, as well as the accuracy of repeated US as compared with that of magnetic resonance cholangiopancreatography. This study aims at evaluating the diagnostic accuracy of repeated US for biliary AP.
Methods: The accuracy of each test for diagnosis of biliary AP was evaluated according to the final diagnosis. Comparison between tests was obtained by examining the areas under the receiver operating characteristic curves.
Results: Among 155 patients, the etiology was biliary in 52% and alcoholic in 20%. The accuracy of the first US alone and of the 2 combined examinations for a biliary etiology were 66% and 83%, respectively. Comparison of receiver operating characteristic curves showed a better performance of repeated US (difference between areas under the curve, 0.135; 95% confidence interval, 0.02-0.24; P = 0.021). Magnetic resonance cholangiopancreatography had high specificity (93%) but low sensitivity (62%), with 76% accuracy. The accuracy of the combination of the 2 US examinations and of elevated alanine transferase was 87%.
Conclusions: Repeated US is effective for biliary AP diagnosis. The combination of repeated US examinations and biochemical tests seems an effective approach, whereas magnetic resonance cholangiopancreatography might be restricted to selected cases.