Objective: To evaluate the decline in CD4+ counts in relation to the incidence of AIDS in different cohorts of homosexual men and to quantify possible consequences of laboratory variation in CD4+ measurement.
Methods: Our study includes 403 men with well documented dates of HIV seroconversion originating from five cohort studies among homosexual men. Differences in time from HIV seroconversion to the first CD4+ count dropping < 500 or 200 x 10(6)/l and to AIDS were evaluated using Kaplan-Meier survival analyses.
Results: We found considerable differences between cohorts in CD4+ depletion, but not in the incidence of AIDS (1987 definition).
Conclusions: Variation in CD4+ depletion appears to be mainly the result of laboratory differences. Policy recommendations on a basis of CD4+ counts probably requires a calibration of measurement. The 1993 AIDS case definition leads to a site-specific shortening of the incubation time, which complicates the study of the natural history of HIV infection and of trends in the AIDS epidemic.