Analyses of the effects of prophylactic use of zidovudine (AZT) on progression to acquired immune deficiency syndrome (AIDS) in human immunodeficiency virus seropositive (HIV+) asymptomatic persons with T4 lymphocyte (CD4+) cell counts > or = 500/mm3 is reported for data obtained from two studies, the Australian European Group Collaborative Study, a multi-centered double-blind placebo-controlled clinical trial of the effects of AZT on progression to AIDS and other clinical endpoints, and the San Francisco Men's Health Study, an observational cohort. The analyses of the data of both studies demonstrate no benefit from AZT treatment in terms of progression to AIDS for those who are asymptomatic with CD4+ cell counts > or = 500/mm3. The analysis of the San Francisco study, performed with Kaplan-Meier survivorship estimates, indicates a heterogeneity in the efficacy of AZT between baseline CD4+ cell count strata, 200-499/mm3 and 500-800/mm3. Within the 200-499 stratum, 47% of those receiving AZT therapy and 62% of those not receiving AZT therapy progressed to AIDS during the study period. By contrast, within the 500-800 stratum 41% of those receiving AZT therapy and 27% of those not receiving AZT therapy progressed to AIDS during the same period. Application of the Cox proportional hazards survivorship regression model for the relative risk of progression to AIDS to these same data accounts for this heterogeneity. The model includes an interaction between AZT treatment and baseline CD4+ cell counts. The hematological toxicity of AZT, demonstrated in clinical studies and laboratory investigations, indicates a biological correlate for this interaction: the toxic effects of AZT on the more intact immune system of those with CD4+ cell counts in the 500-800/mm3 range [corrected].