The Risk of HIV Infection Among Brazilian Men Is Twice That Among Women

C. Coren

First published online:

Among Brazilians, injecting cocaine, having an HIV-positive partner and having male-male sexual relations may each significantly increase the risk for HIV infection. These risk factors were identified in a cross-sectional study of voluntarily tested adolescents and adults in which HIV was most prevalent among those who had syphilis or an HIV-infected partner or who injected cocaine or shared injection equipment.1 The researchers of the study note that their results may be particularly useful because "risk factors for HIV infection have been scarcely studied in developing countries."

The study was conducted in Porto Alegre, one of the most populous cities in southern Brazil, between April and November 1996. The study enrolled persons visiting one of three participating counseling and testing centers to be tested, anonymously and free of charge, for HIV infection, presumably because they suspected possible infection. Participants completed a detailed questionnaire before undergoing testing.

Participants were considered to be infected with HIV if they had a positive result on two enzyme-linked immunosorbent assays and on indirect immunofluorescent assay. (Persons with an initial positive test result unconfirmed by positive fluorescence were excluded from analysis.) The researchers assessed the prevalence of HIV infection in various subgroups. Their analyses examined numerous independent variables, relating to socioeconomic and demographic characteristics, social behavior, sexual behavior, drug use and characteristics of sexual partners. All independent variables were self-reported on the study questionnaire, except for syphilis, which was determined by laboratory analysis.

Of the 3,045 participants, 48% were women and 52% were men; most considered themselves white (78%). The median age was 25 years; persons younger than 20 composed 16% of the sample. Most participants (74%) resided in Porto Alegre; the rest lived in the surrounding metropolitan area or elsewhere in the Brazilian state of Rio Grande do Sul. Nearly all (99%) had had at least one sexual encounter.

Twelve percent of the overall sample tested positive for HIV. The proportion of participants with HIV infection was higher among men than among women (15% vs. 8%), and it was higher among nonwhites than among whites (14% vs. 11%). Prevalence tended to increase with age and to decrease with level of education.

In univariate analyses, several social and sexual behaviors were associated with HIV infection. Study participants who had been in prison and those who had tattoos were more likely than others to be infected. Prevalence tended to increase with decreasing age at first intercourse and with increasing number of lifetime partners. Participants who were homosexual or bisexual men and those who had ever had anal intercourse were more likely than other participants to be HIV-positive, as were those who had exchanged sex for money, gifts, drugs or food. A lifetime history of gonorrhea also was associated with HIV-positive status, as was current infection with syphilis.

The univariate analyses also showed associations between drug use or sexual partner characteristics and HIV-positive status. For example, men who had ever had sex with a transvestite or male prostitute were more likely than other male participants to be HIV-positive. Moreover, the researchers observed a relatively high prevalence of infection among participants who knew or suspected that a partner had used injection drugs or had HIV infection. High proportions of infection were also seen among persons who themselves used cocaine and persons who shared needles or syringes.

After adjustment for effects of potentially confounding variables in a regression analysis, several socioeconomic and demographic variables emerged as independent risk factors for HIV infection. The odds of infection were elevated among men (odds ratio, 2.1), participants aged 20 to 39 years (2.1-2.7) and those with 1-10 years of schooling (2.2-2.8).

Sexual and social behaviors independently associated with higher odds of HIV infection were male homosexuality (3.9), previous juvenile or prison detention (1.5), age of 11-12 years at first intercourse (1.9) and a positive result on the syphilis test (3.5). Of the variables related to sexual partner, several were associated with elevated odds of infection: HIV-positive status was more likely in participants who had had sex with a male prostitute (2.6) or with a partner known to have HIV infection (3.5) or to inject illicit drugs (1.9). Finally, using cocaine--by snorting (2.4) or by injecting (4.5)--and sharing injection equipment (2.6) were each independently associated with a higher likelihood of infection.

The authors acknowledge that their self-selected study group composed a biased sample, but they nonetheless believe that their findings may be useful for identifying at-risk groups and planning targeted interventions. They note that among the many variables associated in the univariate analyses with HIV infection, relatively few proved to have an independent association with HIV infection. They assert, therefore, that "some putative risk factors for HIV infection," such as tattoos or number of lifetime partners, "may just be proxies of the true risk factors."--C. Coren


1. Barcellos NT, Fuchs SC and Fuchs FD, Prevalence of and risk factors for HIV infection in individuals testing for HIV at counseling centers in Brazil, Sexually Transmitted Diseases, 2003, 30(2):166-73.