Two hundred and thirty patients with histologically proven chronic hepatitis C were randomized to receive one of the following treatment protocols: (a) 3 million units of interferon alfa-2b thrice weekly for 6 months, (b) 5 million units thrice weekly for 6 months, or (c) 3 million units thrice weekly for 2 years. The short-term response to treatment was defined by normal alanine aminotransferase for at least 3 months and until the end of treatment, and was confirmed by loss of hepatitis C viraemia in 42 (91%) of 46 cases as determined by reverse transcription-polymerase chain reaction. Short-term response to interferon alfa-2b was independent of the incremental dose, being 64% for 5 million units and 58% for 3 million units. Long-term response to interferon alfa-2b was defined by continued normality of alanine aminotransferase levels for at least 6 months after treatment withdrawal. The long-term response rates among responders treated for 6 months and those treated for 2 years were 29% and 54%, respectively (p < 0.001). Among all 18 patients tested, serum HCV-RNA was negative at both 6 and 12 months of follow-up in all long-term responders, and none have subsequently relapsed. Improvement in hepatic necroinflammatory changes was confirmed by quantitative histology (Scheuer score) in responders at the end of interferon alfa-2b treatment. The changes were significantly greater among those who had been treated for 2 years compared with those treated for 6 months (p < 0.05 and p < 0.02, respectively, for portal and lobular inflammation scores). Several pretreatment characteristics could be correlated with a favourable response to interferon alfa-2b. Thus, absence of cirrhosis was associated with a short-term response of 75%, while only 42% of patients with cirrhosis had a short-term response (p < 0.001). The frequency of short-term response to interferon alfa-2b also differed according to mode of disease acquisition, being best for injecting drug use (71%), less favourable for blood transfusion (56%) and worst for sporadic cases (43%) (p < 0.01). This observed difference, however, was not independent of histology on multivariate analysis. In summary, a 5 million unit dose of interferon alfa-2b failed to improve the short-term or long-term response to interferon alfa-2b treatment, but prolongation of interferon alfa-2b treatment to 2 years resulted in substantially improved long-term response rate.