To test whether hepatotoxicity occurring in National Cooperative Gallstone Study patients was caused by a toxic effect of chenodiol per se or of lithocholate caused by defective sulfation, bile samples were analyzed using a new high-performance liquid chromatography method that measures the proportions of the four individual lithocholate amidates (sulfated and unsulfated lithocholylglycine and lithocholyltaurine) and all common bile acid amidates. Samples were obtained from National Cooperative Gallstone Study patients (n = 17) with abnormal light microscopic liver biopsy results or major aminotransferase elevations and from a matched control group of patients (n = 14) who received similar chenodiol doses but had no evidence of liver injury. Bile samples from 45 healthy subjects were also analyzed. The analytical method was validated by showing that the percentage of chenodiol and cholic and deoxycholic acid obtained by high-performance liquid chromatography correlated highly (r greater than 0.94) with previous gas-liquid chromatography analyses of these samples by the National Cooperative Gallstone Study Reference Laboratory. No significant differences were seen between gallstone patients with and without evidence of liver injury for percent total lithocholate amidates, percent sulfated or unsulfated lithocholate amidates or percent chenodiol amidates. Lithocholate was partially sulfated in all bile samples (52% +/- 17% [mean +/- S.D., n = 50]), but the extent of sulfation varied widely between and within patients during the course of therapy. Mean values of healthy subjects were similar and also showed a wide range in the extent of lithocholate sulfation. It is concluded that (a) liver injury caused by these doses of chenodiol could not be attributed to the accumulation of unsulfated lithocholate per se in circulating bile acids; (b) liver injury appeared to be, directly or indirectly, caused by enrichment in circulating bile acids with chenodiol or chenodiol together with lithocholate, suggesting that the hepatocytes of those patients with hepatotoxicity were injured by the change induced in bile-acid metabolism by the feeding of chenodiol; and (c) about half of lithocholate amidates in bile samples were sulfated, but the extent of sulfation was highly variable both in gallstone patients and healthy subjects.