Of 157 consecutive patients who had histologically diagnosed hepatocellular carcinoma, 52 with a good performance score, bilirubin less than 2 mg/dl, and absence of bloody ascites were treated with a 75-mg/m2 dose of doxorubicin (unadjusted for hepatic function) every 3 weeks in a prospective trial. Forty-six patients were treated in West Africa and six in Southern California. No complete responses were seen and only six patients (11%) achieved partial objective responses. Plasma concentrations of doxorubicin and doxorubicinol (adriamycinol) were determined at four selected time points for up to 72 hours, corresponding to the terminal phase of disposition in eight African patients. The African patient results were compared to those seen in North American patients with hepatocellular carcinoma and other malignancies. In African patients with hepatocellular carcinoma, terminal half-life of doxorubicin was prolonged at 39.8 +/- 15.9 hours. The ratios of the corresponding concentration X time values of doxorubicinol to doxorubicin, which reflect the overall metabolite to parent drug ratio, ranged from 0.7 to 4.6, with a mean ratio of 2.03 +/- 1.20 in the African patients with hepatocellular carcinoma compared to a mean ratio of 0.76 +/- 0.31 in North American patients with other malignancies. Pharmacokinetic findings in hepatocellular carcinoma patients in North America and Africa were similar, reflecting elevation and prolongation of doxorubicinol metabolite relative to doxorubicin in the plasma of patients with this disease from both areas. We conclude that: (a) doxorubicinol disposition is altered in African as well as North American patients with hepatocellular carcinoma; and (b) even when given in full dose to patients with favorable prognostic features, iv doxorubicin has only limited activity against hepatocellular carcinoma.