To evaluate the incidence, severity, and course of ketoconazole-associated hepatic injury, 211 patients with onychomycosis were randomized by a ratio of 2:1 to receive either ketoconazole (137 patients) or griseofulvin (74 patients). All of them were seronegative for hepatitis B surface antigen (HBsAg) and anti-hepatitis C virus (HCV) and had no biochemical abnormality before therapy. The two groups were comparable in age, sex, and pretherapy liver biochemical tests. Liver biochemical tests were followed up biweekly for 3 months, and then at monthly intervals during the remaining course of therapy. No biochemical abnormality or hepatic injury was found in patients during griseofulvin treatment. Of the patients treated with ketoconazole, 24 (17.5%; 95% confidence interval [CI], 11.1% to 23.9%) showed asymptomatic transaminase elevation. Ketoconazole was discontinued immediately after overt hepatitis developed in another 4 patients (2.9%; 95% CI, 0.1% to 5.7%) who did not succumb to hepatic decompensation. The abnormal biochemical changes in patients with overt hepatitis returned to normal after discontinuing ketoconazole. Elderly patients were more prone to develop overt hepatitis. In patients with asymptomatic liver injury, the abnormal biochemical changes gradually returned to normal despite continuing ketoconazole therapy. The results of this cohort study suggest that ketoconazole-induced overt hepatitis is more common than previously believed and that transient subclinical injury is much more common than overt hepatitis. Therapy with ketoconazole may be continued with caution in the absence of symptoms and/or hyperbilirubinemia, but should be discontinued if overt hepatitis occurs.