Early versus deferred zidovudine monotherapy: impact on AIDS-free time and survival in the multicenter AIDS cohort study.

The objective of this study was to compare the time to AIDS and to death between men receiving zidovudine therapy before or not before the diagnosis of AIDS. For the time to AIDS comparison, 821 men receiving zidovudine therapy before the diagnosis of AIDS were pair matched with men who did not receive zidovudine therapy until after diagnosis on their CD4 cell count (+/- 75 cells/mm3), haemoglobin level (+/- 0.75 g/dl), number of clinical symptoms and study visit at the time of initiation of zidovudine therapy and were monitored for a median of 2.08 years. For the time to death comparison, 186 men who received zidovudine therapy prior to AIDS diagnosis were pair matched on the same variables to men who received zidovudine therapy only after the AIDS diagnosis, and were monitored for a median of 2.88 years. Only men with < 350 CD4 cells/mm3 who received zidovudine therapy prior to AIDS diagnosis remained AIDS free significantly longer than their pair match who did not (P < 0.0001). The median extension of time to AIDS was 0.61 years for men with < 200 CD4 cells/mm3 and 1.13 years for men with 200-349 CD4 cells/mm3. Cox regression analysis showed a significantly increased time to AIDS for the men with < 350 CD4 cells/mm3, both before and after adjustment for the use of prophylactic drugs against Pneumocystis carinii pneumonia. No difference was seen in the time to death between men receiving zidovudine therapy before or only after AIDS diagnosis. Zidovudine treatment of asymptomatic HIV-1-infected men provides significant benefit to men with < 350 CD4 cells/mm3 by extending AIDS-free time, but does not extend survival. The analytical technique used is applicable to other observational studies of treatment.