The effect of zidovudine on disease progression in asymptomatic human immunodeficiency virus type 1 (HIV-1)-infected men (n = 52) in relation to CD4 T cell numbers, T cell reactivity, and HIV-1 biologic phenotype was studied in a double-blind randomized trial over 2 years. CD4+ cell numbers and T cell reactivity did not differ significantly between the zidovudine- and placebo-treated groups, except for a transient improvement of both parameters in the zidovudine-treated group during the first 9 months. A marked differential efficacy of zidovudine was observed depending on the HIV-1 phenotype present. Zidovudine did not prevent the emergence of high-replicating syncytium-inducing (SI) variants, and clinical progression was observed in persons with SI variants despite zidovudine treatment. In contrast to nontreated HIV-1-infected asymptomatic persons, zidovudine-treated men who did not develop SI variants did not progress to AIDS. The beneficial effect of zidovudine during the asymptomatic phase may be mainly limited to persons who do not develop SI variants in the course of HIV-1 infection.