Unprotected anal intercourse, risk reduction behaviours, and subsequent HIV infection in a cohort of homosexual men.

OBJECTIVE: A range of risk reduction behaviours in which homosexual men practise unprotected anal intercourse (UAI) has been described. We aimed to assess the extent of any reduction in HIV risk associated with these behaviours.
DESIGN: A prospective cohort study of HIV-negative homosexual men in Sydney, Australia.
METHODS: Men were followed up with 6-monthly detailed behavioural interviews and annual testing for HIV. The four risk reduction behaviours (behaviourally defined) examined were serosorting, negotiated safety, strategic positioning, and withdrawal during receptive UAI (UAI-R).
RESULTS: In 88% of follow-up periods in which UAI was reported, it occurred in the context of consistent risk reduction behaviours. Compared with those who reported no UAI, the risk of HIV infection was not raised in negotiated safety [hazard ratio = 1.67, 95% confidence interval (CI) 0.59-4.76] and strategic positioning (hazard ratio = 1.54, 95% CI 0.45-5.26). Serosorting outside negotiated safety was associated with an intermediate rate of HIV infection (hazard ratio = 3.11, 95% CI 1.09-8.88). Withdrawal was associated with a higher risk than no UAI (hazard ratio = 5.00, 95% CI 1.94-12.92). Patterns of UAI differed greatly according to partner's serostatus. Men who reported serosorting were less likely to report either strategic positioning or withdrawal.
CONCLUSION: Each behaviour examined was associated with an intermediate HIV incidence between the lowest and highest risk sexual behaviours. The inverse association between individual behaviours suggests that men who practise serosorting rely on this protection. The high prevalence of these behaviours demands that researchers address the contexts and risks associated with specific types of UAI.