Abstract

OBJECTIVE--To determine whether certain sexually transmitted diseases are independent risk factors for HIV transmission in a cohort of homosexual men. METHODS--Eligible cases were identified as those who had seroconverted between November 1982 and November 1990. Two persistently HIV-seronegative control participants were randomly selected for each case from all participants who remained seronegative in November 1990. For cases, risk factor data were taken from an index visit which was defined as the first seropositive visit, while for controls these data were obtained from a matched visit which occurred within two months of the index visit for the corresponding case. Mantel-Haenszel methods and logistic regression were used to compare differences in risk factors for seroconversion between cases and controls. RESULTS--A total of 125 cases and 250 controls were eligible for this study. Cases were significantly more likely to have had reported any gonorrhoea (17% versus 6%; OR = 2.94; 95% CI: 1.51-5.73) or syphilis (7% versus 2%; OR = 3.78; 95% CI: 1.33-10.79) than controls during the seroconversion period. Multivariate logistic regression revealed rectal gonorrhoea to be independently associated with risk of seroconversion (odds ratio = 3.18; p = 0.044), whereas urethral gonorrhoea (p = 0.479) and pharyngeal gonorrhoea (p = 0.434) were not after inclusion of rectal gonorrhoea. In addition, the following variables were also shown to exert an independent effect on seroconversion: frequency of anal intercourse, use of illicit drugs, number of male sexual partners, and lack of a post-secondary education. CONCLUSIONS--In this observational study, rectal gonorrhoea was found to be associated with HIV seroconversion after adjustment for a number of HIV risk factors. We cannot rule out that rectal gonorrhoea was not directly associated with HIV infection but rather with other residual lifestyle factors not fully adjusted for in the analysis. However, the relationship with gonococcal involvement of a specific anatomic site lends support to a biological association between gonorrhoea and HIV infection, rather than to alternative non-biologic explanations. Our findings are consistent with previous studies reporting an association between HIV infection and non-ulcerative sexually transmitted diseases. Such a direct association might be explained by postulating that gonorrhoea results in inflamed rectal mucosa and compromised epithelial integrity, thereby predisposing an individual to subsequent HIV infection.