Beneficial effects of ursodeoxycholic acid in chronic cholestatic liver diseases have been attributed to displacement of hydrophobic bile acids from the endogenous bile acid pool. To test this hypothesis, we determined pool sizes, fractional turnover rates, synthesis/input rates and serum levels of deoxycholic acid and chenodeoxycholic acid before and 1 mo after the start of treatment with ursodeoxycholic acid (13 to 15 mg/kg body wt/day) in four healthy volunteers and five patients with chronic cholestatic liver diseases (three with primary biliary cirrhosis and two with primary sclerosing cholangitis). Bile acid kinetics were determined by combined capillary gas chromatography-isotope ratio mass spectrometry in serum samples after administration of [2H4] deoxycholic acid and [13C]chenodeoxycholic acid. In healthy volunteers, deoxycholic acid pool sizes decreased during administration of ursodeoxycholic acid by 72%. In patients with cholestatic liver diseases, deoxycholic acid pool sizes before ursodeoxycholic acid treatment were only 13% of those in healthy volunteers and were unaffected by ursodeoxycholic acid treatment. Chenodeoxycholic acid pool sizes were not different in healthy volunteers and in patients with cholestatic liver disease, and were not altered by ursodeoxycholic acid treatment. In both healthy volunteers and patients with cholestatic liver disease, synthesis/input rates and serum levels of deoxycholic acid and chenodeoxycholic acid were not altered by ursodeoxycholic acid treatment. Because in our patients improvement of serum liver tests during short-term ursodeoxycholic acid treatment was noted without a decrease of the pool sizes of the major hydrophobic bile acids, we conclude that displacement of hydrophobic endogenous bile acids is not the mechanism of action of ursodeoxycholic acid in chronic cholestatic liver disease.