Delayed application of condoms is a risk factor for human immunodeficiency virus infection among homosexual and bisexual men.

The objective of this analysis was to identify risk factors for recent human immunodeficiency virus (HIV) infection among homosexual and bisexual men in Ontario, Canada, during 1998-2001. Participants were recruited through the provincial HIV diagnostic laboratory and through physicians and community organizations. HIV test results were used to identify recent seroconverters (cases). A subsample of 183 men (62 cases and 121 controls) enrolled in the Polaris HIV Seroconversion Study as of June 2001 was analyzed. This analysis focused on sexual behaviors with partners who were HIV-positive or whose HIV status was unknown. In multiple logistic regression, independent risk factors were identified. Rates of unprotected receptive oral, insertive anal, and receptive anal sex and delayed application of condoms during receptive anal sex (RAS) were significantly higher among cases (97%, 41%, 53%, and 32%, respectively) than among controls (73%, 19%, 14%, and 2%). Independent risk factors for HIV infection were RAS without condoms (odds ratio = 4.4, p = 0.0004) and delayed application of condoms during RAS (odds ratio = 5.8, p = 0.01). There was an association with condom failure (breakage or slippage) during RAS that approached significance (odds ratio = 2.9, p = 0.09). Delayed application of condoms for RAS may result in contact with preejaculatory fluid. This behavior, which to date has received little attention, may pose as much risk for HIV infection as fully unprotected RAS.