To distinguish responders from non-responders early in interferon (IFN) treatment would be beneficial in patients with chronic hepatitis C. Those patients unlikely to respond would be spared the cost and hazard of prolonged treatment. Forty-three chronic hepatitis C patients who had received IFN-alpha therapy (6-9 MU; six times weekly for 2 weeks followed by thrice weekly for 22 additional weeks) were randomly enrolled into the present study. Serially obtained sera were retrospectively tested for HCV-RNA by reverse transcription-polymerase chain reaction (Amplicor HCV) with a low limit detection of approximately 10(2) copies/mL. Genotypes were determined by type-specific primers. Sixteen subjects were defined as sustained responders (SR), who showed sustained loss of viraemia with normalized alanine aminotransferase values for at least 6 months of follow-up after completion of therapy. The other 27 subjects were non-responders (NR), whose viraemia persisted during follow-up. Pretreatment serum HCV-RNA levels (P < 0.0001) and the genotype (P < 0.01) were significant predictors for sustained response to IFN therapy. Hepatitis C virus RNA was detectable in only one (6%) SR and in 23 (85%) NR at the second week of therapy (P < 0.0001) and was detected in none of the SR subjects and in 18 (67%) NR at the fourth week of therapy (P < 0.0001). Pretreatment viral load was correlated with the time until loss of viraemia. Multivariate analysis revealed that loss of viraemia at the second week of therapy was the strongest predictor for a long-term response, followed by the initial viral load and loss of viraemia at the fourth week of therapy. These findings suggest that it is possible to predict a long-term response to IFN as early as at the second and fourth weeks after the start of therapy by identifying the presence or absence of HCV-RNA with a sensitive assay.