Previous studies from our groups have demonstrated improvements in biochemical markers of liver function when cystic fibrosis patients with associated liver disease were administered oral ursodeoxycholic acid. The magnitude of the response was somewhat less than that found when comparable doses (10 to 15 mg/kg body wt/day) of ursodeoxycholic acid are given to other liver disease patients; this may be explained by the bile acid malabsorption that is characteristic of the disease. For this reason a dose-response study was carried out in nine cystic fibrosis patients with liver disease to establish whether improved efficacy could be obtained with higher doses. Ursodeoxycholic acid in doses of 5,lO and 15 mgkg body wt/day was given orally for consecutive 2-mo periods in a replicated Latinsquare design. After this, all patients received 20 mg/kg body wt/day. Liver function, individual serum bile acids and biliary bile acid composition were determined at entry and at the end of each treatment period. Our data demonstrate that the magnitude of the biochemical improvement in serum liver enzymes was sisnificantly greater with higher doses of ursodeoxycholic acid; at 20 rn& body wt/day it was similar to that reported for patients with other liver diseases administeredlower doses. Biliary ursodeoxycholic acid enrichment increased with increasing doses, attaining 42% +-6% of the total biliary bile acids with the highest dose. Fasting serum ursodeoxycholic acid concentrations increased during ursodeoxycholic acid administration but were variable and correlated poorly with the dose of ursodeoxycholic acid administered, whereas no correlation was found between serum ursodeoxycholic acid concentration and the proportion of d e o x y c h o l i c acid in bile. Our data indicate a need for doses of ursodeoxycholic acid of 20 mgkg body wt/day or greater to obtain optimal improvement in biochemical indices of liver function and