Background:
Many MSM adopt different sexual behaviors based on the perceived HIV serostatus of their sex partners.

Objective:
To assess the association between of HIV infection and reported sexual practices with partners of different HIV status categories among MSM.

Method:
From 2001-2005, clinicians asked MSM STD clinic patients how often they used condoms for insertive and receptive anal sex with partners who were HIV positive, HIV negative, and of unknown HIV status. We categorized sexual behaviors hierarchically based on magnitude of transmission risk, and attributed HIV positivity to subjects' riskiest behavior.

Result:
3028 MSM HIV tested 4867 times; 135 tested HIV positive. HIV positivity varied significantly based on patients' riskiest behavior, from 10.2% among men who reported unprotected receptive anal intercourse with an HIV positive partner to 1.0% among men who reported no unprotected anal intercourse (UAI). Among men with newly diagnosed HIV, 29% reported UAI with only HIV negative partners, and 17% reported no UAI in the preceding year; these percentages did not change when we restricted analysis to persons who reported testing HIV negative in the preceding year. On multivariate analysis, newly diagnosed HIV was significantly associated with UAI with an HIV positive partner or partner of unknown HIV status, methamphetamine use, longer time since last HIV test, older age and being uncircumcised. The median time since last HIV test among persons testing HIV positive was 361 days.

Conclusion:
Serosorting is protective against HIV, but many new HIV cases occur among MSM who report no UAI with HIV positive or unknown HIV status partners.

Implications:
Public health messages should indicate that although serosorting reduces the risk of HIV acquisition, it's an imperfect prevention strategy. Prevention programs should attempt to shorten the interval between HIV tests among MSM as a means to decrease transmission from men who incorrectly disclose to partners that they HIV negative.