EPIDEMIOLOGIC ANALYSIS OF KAPOSI-SARCOMA AS AN EARLY AND LATER AIDS OUTCOME IN HOMOSEXUAL MEN

The authors separately studied the epidemiology (risk and risk factors ) of Kaposi's sarcoma occurring as an initial acquired immunodeficienc y syndrome (AIDS) outcome (early Kaposi's sarcoma) and later after a d ifferent initial AIDS outcome (later Kaposi's sarcoma) in a cohort of 2,591 human immunodeficiency virus type 1-infected gay men of the Mult icenter Al DS Cohort Study between 1984 and 1992. Among 844 AIDS cases , 202 presented with early Kaposi's sarcoma, 101 subsequently develope d later Kaposi's sarcoma, and 541 were not diagnosed with Kaposi's sar coma. Overall, 37.4% of AIDS cases were diagnosed with Kaposi's sarcom a prior to death. Kaposi's sarcoma diagnosed on the skin was significa ntly more common with early Kaposi's sarcoma (77.3%) than with later K aposi's sarcoma (65.1%). Men presenting with an AIDS outcome other tha n Kaposi's sarcoma were at high risk for later Kaposi's sarcoma. Later Kaposi's sarcoma onset in men with a previous AIDS outcome was associ ated with the following characteristics: 1) lower immune status prior to AIDS and 2) longer post-AIDS survival. A Kaposi's sarcoma diagnosis in a man with a previous AIDS illness approximately doubled the risk (hazard) for death. Histories of urethral gonorrhea and scabies prior to study entry were more common in early Kaposi's sarcoma cases than i n later Kaposi's sarcoma cases. However, self-reported sexual activity at study entry and prior to AIDS onset was highest in the later Kapos i's sarcoma group. In this cohort, cigarette smoking had a protective association against all Kaposi's sarcoma in univariate and multivariat e models. Only 21.0% of the later Kaposi's sarcoma and 25.0% of the ea rly Kaposi's sarcoma men smoked at least one-half pack of cigarettes d aily at study entry compared with 33.8% of non-Kaposi's sarcoma and 35 .5% of seroprevalent men still AIDS free. The reasons for this surpris ing association are unclear. However, other evidence which documents t hat habitual smoking alters the immune system (and possibly cytokine l evels) in ways that could perhaps influence Kaposi's sarcoma pathogene sis should be considered.