Background:  Serosorting, i.e. seeking to only engage in sexual activities with partners of concordant HIV serostatus, has been described as a risk reduction strategy increasingly used by men who have with men (MSM). If replacing condom use, the implications of this practice for the risk of HIV transmissions are controversial. Few data exist on the effect of serosorting on the incidence of bacterial sexually transmitted infections (STI).

Methods:  In Germany, large national cross-sectional surveys on “Gay Men and AIDS” have been conducted since 1987. In 2007, a self-administered questionnaire was distributed via magazines and Internet sites for MSM. Questions included risk reduction strategies, condom use, numbers of sexual partners, and communication about HIV test results. We analyzed 8170 questionnaires. Adjusted odds ratios were calculated to estimate the influence of serosorting on frequencies of bacterial STI and HIV.

Results:  A substantial minority of participants engaging in anal intercourse with casual partners reported their last intercourse was unprotected because of presumed seroconcordancy (self-reported HIV^– 16%; HIV⁺ 20%). Pure guessing accounted for 55% of serostatus estimates among HIV⁺, and 36% among HIV^– MSM. However, among MSM not diagnosed HIV⁺, 62% had a test result older than 18 months or were untested. Contrastingly, 3% of HIV^–, but 21% of HIV⁺ MSM reported serosorting as a main risk-reduction “strategy.” Compared with traditional safer sex strategies and adjusted for the number of sexual partners, serosorting was strongly and significantly associated with a recent history of bacterial STI among HIV⁺ MSM (OR 4.3, 95%CI 2.2 to 8.3); and with a recent HIV⁺ result among all participants with a recent HIV test (OR 5.0, 95%CI 2.7 to 9.4). Recent syphilis diagnoses were reported by 26% of MSM with a recent positive, and by 3% with a recent negative HIV test result, suggesting high rates of co-infection.

Conclusions:  Serosorting is not uncommon among German MSM, but seems to qualify as “strategic” only among MSM who are HIV⁺. However, serosorting is frequently based on assumptions instead of disclosure. (Effective) serosorting among HIV⁺ MSM increases incidence and prevalence of bacterial STI, and hence the per contact risk for HIV transmission. Serosorting among MSM who believe they are HIV^– is highly ineffective and therefore may further increase the risk for HIV transmission.