Factors associated with Kaposi's sarcoma in a cohort of homosexual and bisexual men.

The relationship between the occurrence of Kaposi's sarcoma (KS) and a range of factors was studied in a cohort of homosexual and bisexual men. Overall, 134 men had AIDS, including 46 for whom KS was the first AIDS-defining illness. The proportion of men diagnosed with AIDS whose first AIDS-defining illness was KS declined from 54 to 24% between 1984-85 and 1989-90 (p = 0.03 for linear trend). Kaposi's sarcoma was not associated with a history of any specific communicable disease, nor with the reported use of selected nontherapeutic drugs, at the 0.05 level of significance. No single sexual practice was significantly associated with KS, and only the relative risks for insertive oroanal contact (1.5, 95% confidence limits 0.73-2.4) and receptive fisting (1.3, 0.61-2.6) exceeded unity. Kaposi's sarcoma occurred significantly less frequently in men who reported orogenital intercourse than in those who did not (relative risk 0.47, 0.23-0.98 for receptive orogenital intercourse). History of sexual contact in the United States was associated with KS, but this association was not statistically significant. While this study provides some support for the hypothesis that Kaposi's sarcoma in people with AIDS may be caused by a sexually transmissible infectious agent, no specific sexual practice could be implicated. In particular, there was little evidence that the postulated KS agent was transmitted by fecal-oral contact.