To determine the patient-dependent and virus-related factors that may predict sustained response to interferon-alpha therapy, we prospectively evaluated 60 consecutive patients with chronic hepatitis C who received a standardized treatment schedule of interferon-alpha. Twenty-eight patients achieved a long-term sustained remission, 14 patients had temporary responses during treatment but relapsed after completing it and 14 patients did not respond. Four patients dropped out because of severe side effects and were excluded from evaluation of efficacy. Twenty-one variables were chosen as possible predictors of sustained response and were analyzed by means of multivariate analysis. Variables related to the hepatitis C virus included genotype and concentration in serum before treatment. The former was determined by means of the polymerase chain reaction with genotype-specific primers for genotypes PT, K1, K2a and K2b, which were deduced from nonstructural region 5 of the hepatitis C virus genome. The latter was measured with a competitive polymerase chain reaction technique. Three variables were statistically significant (p < 0.05) on univariate analysis: viral genotype, pretreatment level of viremia and Knodell's fibrosis score. In multivariate analysis viral genotype and Knodell's fibrosis score were correlated independently with a sustained response (p < 0.05 for both). Incidence of sustained response was 40% and 91% in patients with genotypes K1 and K2a, respectively. We conclude that with the treatment schedule used, viral genotype was the most useful predictor of a sustained response to treatment with interferon-alpha.