To evaluate response rates to 3, 5, or 10 million units (MU) of interferon alfa-2b, given thrice weekly, and to determine whether higher doses of interferon increase the likelihood or durability of the response, a multicenter, randomized trial was performed at nine academic medical centers in the United States. Two hundred forty eight patients with chronic hepatitis C were randomized to receive 3, 5, or 10 MU of interferon alfa-2b thrice weekly for 12 weeks. Based on the alanine aminotransferase (ALT) response at treatment-week 12, the patients were rerandomized to additional therapy at the same or at increased doses for an additional 12 to 36 weeks; in the case of no response to the highest dose, the patients were discontinued from the study. Serum ALT concentrations and liver histology were measured. The overall complete response rates to 3, 5, or 10 MU were not different at treatment-week 12 (31% vs. 42% vs. 40%, not significant). The majority of week-12 responders continued to respond during additional treatment. When the treatment was discontinued, 15.4% to 19.0% of patients maintained their response. Of the nonresponders to 3 MU at week 12, who were continued on 3 MU for an additional 12 weeks, none responded. However, response to additional therapy occurred in 12% of week-12 nonresponders, whose dose was escalated from 3 or 5 MU to 10 MU. The only baseline features associated with the treatment response were the absence of fibrosis or cirrhosis on the pretreatment liver biopsy and viral genotype. We conclude that the initial response to interferon in patients with chronic hepatitis C is not increased by treatment with higher doses of the drug. Patients who do not respond to 3 MU by treatment-week 12 will not respond with continued therapy at that dose; however, a proportion of patients who do not respond to 12 weeks of treatment with 3 or 5 MU may respond to higher doses. Although the long-term sustained response rates are marginally increased with interferon doses above 3 MU three times per week, the side effects are difficult to tolerate. The analysis of baseline factors in relation to response identified no single baseline factor associated with a low-enough response rate to warrant withholding interferon therapy from patients with chronic hepatitis C.