Lipoprotein X is considered to be an indicator of cholestasis, but its clinical usefulness is unclear owing to different methods of measurement in use and varying definitions of cholestasis. We investigated the reliability of the qualitative determination of lipoprotein X in predicting cholestasis, defined here as the presence of stainable bile in tissue sections. Liver tissue and serum were collected from 42 patients. Patients were divided into those who did and those who did not have stainable bile in tissue sections. Lipoprotein X was demonstrated by electrophoresis and polyanion precipitation. Lipoprotein X in serum correlated with the presence of stainable bile (p = 0.0006). A normal range of alkaline phosphatase values correlated with the absence (p = 0.04), but elevated values did not correlate with the presence of stainable bile. In patients known to have elevated alkaline phosphatase, the presence of lipoprotein X still correlated with the presence of bile (p = 0.008). We conclude that lipoprotein X is a reliable indicator of cholestasis and that it is more useful than alkaline phosphatase in this application.