Background: Ideally, patients should only be referred to endoscopic retrograde cholangiopancreatography (ERCP) if therapy is indicated. The aim of this study was to evaluate whether or not the 'Trondsen Discriminant Function' (DF) could be used for selecting patients directly for ERCP.
Methods: The DF was calculated in 163 patients referred for ERCP with the DF value being unknown to the endoscopist. Compared to the final diagnoses of the patients, the sensitivity and specificity of a positive DF value for predicting biliary obstruction and need of endoscopic therapy were calculated.
Results: Ninety-three (57%) patients had obstruction of the bile duct and 84 (52%) needed endoscopic therapy. A positive DF value had a sensitivity, specificity, positive predictive value and negative predictive value for predicting biliary obstruction of 81%, 72%, 79% and 73%, respectively. If only patients with a positive DF value had been examined by ERCP, 50 (31%) patients would have been saved from this investigation. Had a negative DF value stopped the patients from further diagnostic evaluation, 18 (11%) would have had undiagnosed pathological conditions.
Conclusion: A positive DF value is useful for selecting which patients should be referred directly for ERCP because of a high probability that they will need endoscopic therapy. A negative DF value cannot be used to stop the patient from further diagnostic evaluation.