Therapy may explain recent deficits in AIDS incidence.

Since the middle of 1987, fewer consistently defined AIDS cases have been reported than expected among homosexual and bisexual men in the United States. This "AIDS deficit" was greater among homosexual and bisexual men in New York City, San Francisco, and Los Angeles, but was also striking among all homosexual and bisexual men in the United States. Deficits were virtually absent among intravenous drug users (IVDUs) in the United States. Three independent sources of data--placebo-controlled trials, pharmaceutical company reports, and the San Francisco Men's Health Study--were used to demonstrate that the amounts of zidovudine (AZT) given prophylactically to those at highest risk of AIDS since March 1987 have been sufficient to account for most of the observed AIDS deficits. Other advances in the medical care of pre-AIDS patients may have combined with AZT to produce the deficits. Other hypothesized explanations were examined and found insufficient to account for the observed AIDS deficits, including: (a) a sudden halt in new human immunodeficiency virus (HIV) infections during the early or mid-1980s; (b) misspecification of the distribution of AIDS incubation times following HIV infection; (c) increasing delays in the reporting of AIDS cases; (d) changes in the surveillance definition of AIDS in 1987; and (e) evolution of attenuated HIV strains. The hypothesis that therapy is affecting national AIDS rates has important implications. Failure to take the effects of therapy into account can lead to serious underestimates by back-calculation of the cumulative numbers infected with HIV and of AIDS incidence over the longer term. Moreover, it appears that AIDS incidence could be retarded in underserved groups, such as IVDUs, by making AZT and other state-of-the-art treatments readily available to AIDS-free patients with advanced immunodeficiency.