The fasting concentrations of ursodeoxycholic acid were determined in peripheral and portal venous serum of untreated (n = 12) and ursodeoxycholic acid-treated (n = 7) patients undergoing cholecystectomy. The levels of ursodeoxycholic acid were also determined in peripheral venous serum of 9 healthy subjects before and during treatment with ursodeoxycholic acid. Ursodeoxycholic acid, as well as cholic, chenodeoxycholic, and deoxycholic acids, were analyzed by a highly specific method based on isotope dilution-mass spectrometry. The fasting peripheral venous serum concentration of total (unconjugated plus conjugated) ursodeoxycholic acid averaged 0.14 mumol/L in the untreated gallstone patients and 0.19 mumol/L in the healthy subjects. The corresponding value in portal venous serum was 0.44 mumol/L. Treatment with ursodeoxycholic acid raised the level of this bile acid about 25-fold in portal as well as in peripheral venous serum. The proportion of unconjugated ursodeoxycholic acid was 34% in portal and 49% in peripheral venous serum of treated subjects. The mean hepatic uptake of ursodeoxycholic acid was calculated to be about 60% both in untreated and treated subjects. This uptake was significantly lower than that of cholic acid (83%). The hepatic uptake of ursodeoxycholic acid also tended to be lower than that of chenodeoxycholic acid (68%). This was mainly due to a lower hepatic uptake of unconjugated ursodeoxycholic acid (34%) compared with unconjugated chenodeoxycholic acid (49%). The relatively low hepatic uptake of unconjugated ursodeoxycholic acid explains why serum levels of the administered bile acid are higher during treatment with ursodeoxycholic acid than during treatment with chenodeoxycholic acid. Our results also give evidence that the hepatic uptake of ursodeoxycholic acid cannot be saturated under physiologic conditions.