Condomless sex in HIV-diagnosed men who have sex with men in the UK: prevalence, correlates, and implications for HIV transmission.

Objective: HIV transmission is ongoing among men who have sex with men (MSM) in the UK. Sex without a condom (condomless sex, CLS) is the main risk factor. We investigated the prevalence of and factors associated with types of CLS.

Methods: Cross-sectional questionnaire study in UK HIV clinics in 2011/2012 (ASTRA). MSM diagnosed with HIV for ≥3 months reported on anal and vaginal sex, CLS with HIV-serodifferent partners (CLS-D) and CLS with HIV-seroconcordant (CLS-C) partners in the previous 3 months. Mutually exclusive sexual behaviours were as follows: (1) Higher HIV risk CLS-D (not on antiretroviral therapy (ART) or clinic-recorded viral load(VL) >50 c/mL), (2) Other CLS-D, (3) CLS-C without CLS-D, (4) Condom-protected sex only and (5) No anal or vaginal sex. Associations were examined of sociodemographic, HIV-related, lifestyle, and other sexual measures with the five categories of sexual behaviour. We examined the prevalence of higher HIV risk CLS-D incorporating (in addition to ART and VL) time on ART, ART non-adherence, and recent sexually transmitted infections (STIs).

Results: Among 2189 HIV-diagnosed MSM (87% on ART), prevalence of any CLS in the past 3 months was 38.2% (95% CI 36.2% to 40.4%) and that of any CLS-D was 16.3% (14.8%-17.9%). The five-category classification was as follows: (1) Higher HIV risk CLS-D: 4.2% (3.5% to 5.2%), (2) Other CLS-D: 12.1% (10.8% to 13.5%), (3) CLS-C without CLS-D: 21.9% (20.2% to 23.7%), (4) Condom-protected sex only: 25.4% (23.6% to 27.3%) and (5) No anal or vaginal sex: 36.4% (34.3% to 38.4%). Compared with men who reported condom-protected sex only, MSM who reported any CLS in the past 3 months had higher prevalence of STIs, chemsex-associated drug use, group sex, higher partner numbers, and lifetime hepatitis C. Prevalence of higher HIV risk CLS-D ranged from 4.2% to 7.5% according to criteria included.

Conclusion: CLS was prevalent among HIV-diagnosed MSM, but CLS-D with higher HIV transmission risk was overall low. CLS-D is no longer the most appropriate measure of HIV transmission risk behaviour among people with diagnosed HIV; accounting for VL is important.