Because some patients with cirrhosis have serum transaminase levels within the normal range, a prospective study was undertaken to determine whether the concentration of individual serum bile acids would be a sensitive indicator of development of cirrhosis. The choline-deficient rat has been used as a model for study of these changes. Using high-performance liquid chromatography, we measured the concentrations of individual serum bile acids at 3, 6, 10, 20 and 30 wk of dietary intake. Serum levels of total glycine- and taurine-conjugated bile acids were elevated at all stages tested as compared with levels in control groups (choline supplemented). Similarly, unconjugated bile acids and, particularly, cholic acid showed significantly higher levels at all stages except with the occurrence of cirrhosis at 30 wk, at which time there was a significantly lower level for unconjugated bile acids (0.48 ± 0.11 vs. 1.40 ± 0.36 in controls) and for cholic acid (0.17 ± 0.05 vs. 0.91 ± 0.39 in controls). The ratio of serum cholic acid to serum chenodeoxycholic acid changed in temporal relationship to progression in the histological lesions in livers of these rats. The ratio was at its highest at 78 ± 3 at 3 wk (no histological change) and decreased with increasing time and changes in histological appearance until 30 wk, at which time it was down to 1.6 ± 0.6. The routinely used markers of liver injury (serum ALT, alkaline phosphatase and bilirubin), however, did not match the progression of hepatic histological changes. The relationship of the increase in serum bile acids to the cirrhotic process is supported by qualitatively similar findings in a second rat model. Thus the hepatic pathological changes in cirrhotic rats were reflected by the changes in the individual serum bile acids together with the ratio of serum cholic acid to serum chenodeoxycholic acid, but not with other standard tests of liver function. (HEPATOLOGY 1993;18:1224-1231).