While efficacy of zidovudine (ZDV) in the acquired immunodeficiency syndrome (AIDS) is well established, the issue of survival after early ZDV treatment is still controversial. To assess survival benefits of ZDV treatment prior to AIDS, as compared to treatment after the onset of AIDS, we used an observational analysis of infected individuals infected with human immunodeficiency virus treated with ZDV and/or prophylaxis against Pneumocystis carinii pneumonia prior to or after AIDS in comparison to patients never treated with ZDV Nine German AIDS treatment centers entered case reports dating from January 1988 to January 1992. A total of 1425 HIV-infected patients were included, mainly homo-/bisexuals: 1338 males and 87 females, with a mean age of 38.9 years. Of these, 262 had received ZDV prior to AIDS, 376 after AIDS, and 787 had never received ZDV Survival from a first CD4 lymphocyte count below 0.200 × 10⁹/1 (or below 0.500 × 10⁹/1) to death was assessed by means of Kaplan-Meier analysis. Survival did not differ significantly when the first CD4 count below 0.200 × 10⁹/1 was taken as baseline. The median survival of patients receiving ZDV prior to AIDS was 662 days as compared to 572 days in patients treated after AIDS. Patients with earlier therapy showed longer survival in a subset of patients who were observed from their first CD4 count below 0.500 × 10⁹/1. Additional PcP prophylaxis significantly improved survival in all groups. We conclude that survival from the first CD4 count below 0.200 × 10⁹/1 to death does not differ in patients receiving ZDV prior to or after AIDS. Additional PcP prophylaxis improves survival in ZDV-treated patients and patients without ZDV