We wished to evaluate the efficacy and safety of a low and an intermediate daily dose of interferon-alpha2b (IFN-alpha2b) with didanosine in patients with acquired immunodeficiency syndrome (AIDS)-associated Kaposi's sarcoma (KS). HIV-seropositive subjects with biopsy-confirmed cutaneous KS were randomized to receive either a low (1 million IU) or an intermediate (10 million IU) dose of IFN-alpha2b once daily with twice daily doses of didanosine. Treatment assignment was stratified by CD4 count. Response, toxicity, changes in CD4 counts, and survival were evaluated. Sixty-eight eligible subjects were accrued, 35 to low-dose and 33 to intermediate-dose IFN-alpha2b. The response rate was 40% in the low-dose group (95% CI, 24-58) and 55% in the intermediate-dose group (95% CI, 36-72) (p = 0.338). The median response duration was approximately 110 weeks in both groups. Intermediate-dose IFN induced grade 3/4 neutropenia more often (21% vs. 3%, p = 0.048) and grade 3/4 toxicity faster (p = 0.0231) and necessitated treatment discontinuation earlier for drug-related toxicities (p = 0.0416) than low-dose IFN. There were no significant differences in survival between the treatment groups. Baseline CD4 count was the only significant factor predicting response. Once-daily low-dose and intermediate-dose IFN-alpha2b induced similar response rates, which were achieved without optimal antiretroviral therapy. The slightly higher response rate in the intermediate-dose group was offset by its significantly poorer tolerance. These findings justify the use of lower, well-tolerated IFN doses for treatment of KS with currently used antiretroviral regimens.