To evaluate clinical efficacy and toxicity of low-dose oral natural human interferon-α (nHuIFNα) on CD4+ lymphocyte counts and clinical symptoms in patients with HIV-1 infection. Double-blind, randomized, placebo-controlled trial with crossover. Private practice specializing in the treatment of patients with AIDS. Only patients with HIV-1 infection and CD4+ lymphocyte counts between 200 and 500 × 106/l were included for study. Thirty out of thirty-one patients at study entry completed treatment with placebo, and 29 completed nHuIFNα treatment. Mean patient age was 36 years (range, 25–58 years). The 30 patients included 26 men, of whom 22 were homosexual, and four women; five were drug users and none were currently on zidovudine therapy, although three had been previously. Patients were randomly assigned to cohorts of 10 to receive either 200 IU nHuIFNα once daily orally absorbed or placebo with crossover after 6 weeks. Every 2 weeks, a detailed history, physical examination, and laboratory tests, including CD4+ and CD8+ lymphocyte counts, were conducted. There was only a slight, transient increase in mean CD4+ lymphocyte counts after 4 weeks of treatment with nHuIFNα, compared with a slight decline when placebo was administered. This effect reached statistical significance in a subgroup of patients only and was not sustained after 6 weeks. There were no significant changes in weight and clinical symptoms. All patients remained HIV-1-antibody-positive. Treatment-related adverse reactions were not observed. Our double-blind, randomized, placebo-controlled clinical trial did not confirm a previous report of efficiency of oral nHuIFNα. Although non-toxic, our data do not justify the widespread use of low-dose oral nHuIFNα in HIV-infected patients outside controlled clinical trials.