Among men who have sex with men, the level of protection against HIV associated with serosorting—the practice of considering a partner's serostatus in making one's decision to have sex or use a condom with that partner—may differ by race and ethnicity.1 According to an analysis of data collected during visits to a Seattle STD clinic by men who have sex with men, equal proportions of visits by whites, blacks and Hispanics who had not previously tested positive for HIV included reports of serosorting behavior (29–30%); the proportion of visits during which clients who underwent testing received an HIV diagnosis was significantly lower among whites than among blacks or Hispanics (3% vs. 5% and 4%, respectively). In univariate analyses stratified by race and ethnicity, serosorting was associated with reduced odds of testing positive for HIV among whites (odds ratio, 0.5), but not among blacks or Hispanics.
Data for the analysis came from electronic medical records of men who had visited the STD clinic for a new problem between October 1, 2001, and December 31, 2010, and reported having had a male sexual partner in the previous year. During each visit, men were asked about their race and ethnicity; their HIV testing history; their and their partners’ HIV status; their sexual behavior, including the types of anal sex they had participated in during the prior two months (insertive, receptive, both or none); and their condom use. To explore the link between serosorting and HIV risk, researchers classified clinic visits into four groups by men's reports of protective behavior: unprotected anal sex only with partners of the same HIV status (serosorting), unprotected anal sex with partners of discordant or unknown HIV status, only protected anal sex and no anal sex. In addition, clinicians recommended testing for all men who reported not previously having tested positive for HIV. Generalized estimating equations were used to examine the association between serosorting behavior and testing positive for HIV by race and ethnicity.
Overall, the sample included 22,370 clinic visits by 10,620 white, black and Hispanic men who had sex with men. Some 20,735 visits had complete serosorting data; of these, 17,972 were by men who had not previously tested positive for HIV, and 2,763 were by men who had. During 13,657 visits, men who had not previously tested positive for HIV underwent HIV testing.
Sixty percent of clinic visits were by men who reported having had both insertive and receptive anal sex in the previous two months, 20% were by those who had had insertive anal sex only, 9% were by those who had had receptive anal sex only, and 12% were by those who had had no anal sex during the period. Compared with the proportion of visits by white men, a significantly greater proportion of visits by black men included reports of insertive anal sex only (33% vs. 19%), and a significantly smaller proportion included reports of both receptive and insertive anal sex (49% vs. 60%) or receptive anal sex only (6% vs. 9%). Similar proportions of visits by whites, blacks and Hispanics included reports of serosorting (29–30%), protected anal sex only (32–36%) and unprotected anal sex with partners of discordant or unknown HIV status (25–27%); a smaller proportion of visits by Hispanic men than of those by white men included reports of no anal sex (7% vs. 12%). Among the visits by men who underwent HIV testing, the proportion associated with a new HIV diagnosis was lower among whites than among blacks or Hispanics (3% vs. 5% and 4%, respectively).
In univariate analyses among whites, men who reported serosorting and men who reported protected anal intercourse only were less likely than those who reported unprotected anal sex with partners of discordant or unknown serostatus to test positive for HIV (odd ratios, 0.5 and 0.4, respectively). Among blacks and Hispanics, however, no associations were found between serosorting and HIV diagnosis; Hispanic men who had had protected anal intercourse only had a reduced likelihood of testing positive for HIV (0.4).
The authors highlight several limitations of their study, including low analytic power, and the possibility of social desirability bias and of variability in how clinicians surveyed clients. Even so, they conclude that serosorting "may not be protective against HIV" among black men who have sex with men, and suggest that their findings "prompt additional caution related to the promotion of serosorting among [black men who have sex with men] and highlight the need to increase HIV testing coverage and frequency in this critically important population."—J. Rosenberg