Clinical evaluation, ultrasound, cholescintigraphy, and endoscopic retrograde cholangiography in cholestasis. A prospective comparative clinical study.

The accuracy of clinical evaluation (CE), ultrasonography (US), and cholescintigraphy (CS) was prospectively compared in a group of 72 cholestatic patients, using the results of endoscopic retrograde cholangiography (ERC) as the true standard. Forty-six cases (63.9%) had a final diagnosis of extrahepatic biliary obstruction. Clinical evaluation had a sensitivity of 82.6% and a specificity of 46.2% for the identification of mechanical obstruction. Ultrasonography was the single most powerful noninvasive diagnostic procedure, with a sensitivity of 65.2%, a specificity of 92.3%, a positive likelihood ratio of 8.5, and a positive predictive value (PVpos) of 93.8%, which was superior to the diagnostic power of CS (0.05 less than p less than 0.1). Inconclusive results or technical failures were counted as diagnostic errors. Contingency table analyses in an extended group of 112 patients, complete for CE, US, and CS, revealed that concordant results of CE and US had a high probability of correctness (combined PVpos 92.6%, PVneg 98.0%). Contradictory ratings of these two tests were unlikely to be clarified reliably by additional CS. We conclude that the combined evaluation of CE and US allows an accurate differentiation between intrahepatic cholestasis and extrahepatic biliary obstruction. When both give contradictory results, direct cholangiography, rather than further noninvasive procedures, appears to be the diagnostic strategy of choice.