Appropriate patient selection in the management of common bile duct stones: when not to do ERCP

Background. Magnetic resonance cholangiopancreatography (MRCP) is noninvasive and accurate for diagnosing intra common bile duct stones (ICSs). However, given limited access, routine utilisation for investigating all patients with gallstone disease is neither practical nor cost-effective. Conversely, many individuals proceed directly to endoscopic retrograde cholangiopancreatography (ERCP), an invasive test with appreciable complications. Aim. Identify factors associated with ICS in order to improve risk-stratification for MRCP/ERCP. Methods. All patients having undergone cholecystectomy between November 2007 and October 2008 were reviewed. High-risk features for ICS were predefined, and their true presence confirmed by ERCP or intraoperative cholangiogram. Multivariate logistic regression was performed on candidate risk features. Results. Of 231 patients, 10.4% had ICS. Defining a high-risk group with "both" biochemical and ultrasound risk factors predicted ICS with 92% specificity and also bore strong association (OR 8.88). However, isolated hyperbilirubinaemia, ultrasound impression of CBD stones, and clinical risk factors did not (OR 1.10, 0.97, and 1.26). Normal liver biochemistry and normal ultrasound had a NPV of 99.5% for ICS. Conclusions. Ultrasound impression of CBD calculi without ductal dilatation is not predictive of ICS. Patients with normal liver biochemistry and normal CBD diameter on ultrasound are unlikely to have ICS and should not proceed to ERCP.