The efficiency of bile acid conjugation before and during therapy with 600 mg/day of ursodeoxycholic acid was measured in seven adult patients with early chronic cholestatic liver disease (6 with primary biliary cirrhosis; 1 with primary sclerosing cholangitis). Duodenal bile samples were obtained by aspiration and the proportion of unconjugated bile acids was determined using lipophilic anion exchange chromatography to separate bile acid classes, followed by analysis of individual bile acids by gas chromatography-mass spectrometry. The proportion of conjugated bile acids was determined by high-performance liquid chromatography. Use of a (99m)Tc-HIDA recovery marker permitted the absolute mass of unconjugated bile acids in the gallbladder to be calculated. Unconjugated bile acids comprised 0.4% of total biliary bile acids before and 0.2% during ursodeoxycholic acid therapy, indicating highly efficient conjugation of bile acids. During therapy, percentage unconjugated ursodeoxycholic acid significantly increased from (mean +/- S.D.) 13 +/- 13% to 54 +/- 12%; P < 0.002. When the unconjugated and conjugated fractions of bile acids were compared, there was an enrichment in unconjugated fraction for cholic acid and ursodeoxycholic acid and a depletion for chenodeoxycholic acid both in basal condition and during ursodeoxycholic acid therapy, suggesting that hydrophilic bile acids were conjugated less efficiently. During therapy, the conjugation efficiency significantly increased for cholic acid and ursodeoxycholic acid. The pretreatment mass of total unconjugated bile acids in the gallbladder was (mean +/- S.D.) 4.4 +/- 3.2 mumol, and was not significantly changed by ursodeoxycholic acid therapy (6.2 +/- 3.5 mumol). However, ursodeoxycholic acid therapy caused a significant increase in the mass of unconjugated ursodeoxycholic acid. It is concluded that endogenous bile acids and exogenous ursodeoxycholic acid when given at the usual dose are efficiently conjugated in patients with early cholestatic liver disease. Despite showing increased biliary unconjugated ursodeoxycholic acid during its oral administration, our data do not lend support to the occurrence of hypercholeresis due to cholehepatic shunting of bile acids.