Measurement of serum HCV-RNA is a useful index for evaluating the antiviral effect of interferon therapy in chronic hepatitis C. In the present study, we investigated whether the detection of hepatic HCV-RNA after interferon treatment, using a polymerase chain reaction assay, predicted long-term response to therapy in patients with chronic hepatitis C. Thirty-three patients underwent liver biopsies before and after interferon therapy. Histology and clinical courses were compared after treatment. Before therapy, serum and hepatic HCV-RNA was detected in specimens from 32 (97%) and 33 (100%) patients, respectively. Serum HCV-RNA became undetectable in samples from 22 (67%) patients; however, in 10 of these patients (45%), serum HCV-RNA levels relapsed after therapy. Hepatic HCV-RNA became undetectable in 14 patients after therapy and the serum aminotransferase concentration remained within normal limits during and following (24-92 weeks) therapy in 12 of these patients (86%). All 11 patients with detectable hepatic HCV-RNA also had serum HCV-RNA and elevated aminotransferase concentrations refractory to therapy. The absence of hepatic HCV-RNA at the end of interferon treatment thus predicted a long-term complete response to therapy with a sensitivity of 100%, a specificity of 90% and an accuracy of 94%. We conclude that hepatic rather than serum HCV-RNA is a more useful index for the prediction of the long-term efficacy of interferon therapy.