Sheppard Haynes W.; Ascher, Michael S.; Winkelstein, Warren Jr.; Vittinghoff, Eric; Osmond, Dennis; Moss, Andrew R.; Shiboski, StephenJournal of Acquired Immune Deficiency Syndromes: March 1993 EPIDEMIOLOGY: PDF Only Free Abstract Infection with HIV-I and < 200 CD4+ lymphocytes/mm3 has been proposed as an AIDS-defining condition. We have evaluated the effects of using this and other T-cell subset measurements, in the diagnosis of AIDS in two cohorts of homosexual/bisexual men in San Francisco. Among 762 HIV-I infected men, 200 CD4 + lymphocytes/mm3 corresponded to 13 percent CD4 + lymphocytes and a CD4 + /CD8 + ratio of 0.23. If these AIDS-defining criteria had been implemented in mid-1991, the number of living AIDS cases would have increased by 106 (212%), 133 (266%), and 136 (272%), respectively. When these criteria were first met, either before or in the absence of a clinical AIDS diagnosis, about half of the subjects were asymptomatic and the median clinically AIDS-free interval was ∼2 years. Using two consecutive tests or pairwise combinations of criteria reduced the number of cases identified by testing error or transient biological variation, but the number of living AIDS cases would still be increased more than twofold. Finally, any AIDS case definition using a specific T-cell subset value will be compromised by the inherent variability in these measurements and the substantial overlap in the results for those with and without clinical manifestations of HIV infection. © Lippincott-Raven Publishers.