Abstract Serum bile acid levels and distributions were studied every 6 mo in patients with primary biliary cirrhosis who were randomly assigned to receive ursodeoxycholic acid (13 to 15 mg/kg/day) (n = 73) or a placebo (n = 73) over a 2-yr period. In the ursodeoxycholic acid group, ursodeoxycholic acid was the predominant serum bile acid at 6 mo and throughout the 2-yr treatment period. The total concentration of endogenous bile acids decreased with a reduction in cholic acid (in the ursodeoxycholic acid group and the placebo group, respectively [mean ± S.E.]: 13.0 ± 2.2 and 12.6 ± 2.5 μmol/L at entry vs. 3.5 ± 0.6 and 9.0 ± 2.2 μmol/L at 2 yr; p < 0.002), chenodeoxycholic acid (in the ursodeoxycholic acid group and the placebo group, respectively: 12.1 ± 1.7 and 12.7 ± 2.3 μmol/L at entry vs. 5.8 ± 0.8 and 10.7 ± 2.2 μmol/L at 2 yr; p < 0.02) and 3β-hydroxy-Δ5-cholenoic acid. The concentration of deoxycholic acid did not change, whereas that of lithocholic acid increased significantly (in the ursodeoxycholic acid group and the placebo group, respectively: 0.63 ± 0.06 and 0.81 ± 0.12 μmol/L at entry vs. 1.26 ± 0.12 and 0.90 ± 0.15 μmol/L at 2 yr; p < 0.001). These changes were independent of the histological stage of the disease. Thus during ursodeoxycholic acid administration the liver was exposed to a lower level of endogenous bile acids and to an increased concentration of ursodeoxycholic acid. These changes in circulating bile acids and their beneficial effects could be explained by the following mechanisms: (a) initially, ursodeoxycholic acid would directly increase hepatocyte excretion of bile acids; (b) their return to the liver would then be limited by an inhibitory effect of ursodeoxycholic acid on their intestinal reabsorption; and (c) liver cell metabolism and functions would be improved by the sustained reduction in the concentration of endogenous bile acids and the predominant presence of ursodeoxycholic acid.