We performed a 24-week, placebo-controlled, comparative trial of hydroxyurea (HU) monotherapy, didanosine(ddI) monotherapy, and the combination of ddI plus HU administered as 1000 mg qd or 1500 mg qd in antiretroviral-naive and experienced subjects with CD4+ lymphocyte counts of 200-700 cells/mm3. Enrollment included 134 subjects. HU enhanced the antiviral activity of ddI by 1.0 log10 copies/ml after 8 weeks of therapy, with sustained responses over 24 weeks. HU alone over 4 weeks had no effect. Lamivudine resistance had little impact on antiretroviral activity when examined across treatment arms. Increases in absolute CD4+ T cell counts, but not CD4+ T cell percentages, were less in subjects who received HU compared to ddI monotherapy, and lymphoproliferative responses to antigenic and mitogenic stimuli were not altered. Subjects who received HU 1500 mg were more likely to experience dose-limiting hematological toxicities compared to those who received 1000 mg, without any additional antiviral benefit. HU may continue to have a role as a component of HIV therapy.