Individual and poly-substance use and condomless sex among HIV-uninfected adults reporting heterosexual sex in a multi-site cohort

Abstract

Abstract Background We analyzed the association between substance use (SU) and condomless sex (CS) among HIV-negative adults reporting heterosexual sex in the Seek, Test, Treat, and Retain (STTR) consortium. We describe the impact of SU as well as person/partner and context-related factors on CS, identifying combinations of factors that indicate the highest likelihood of CS. Methods We analyzed data from four US-based STTR studies to examine the effect of SU on CS using two SU exposures: 1) recent SU (within 3 months) and 2) SU before/during sex. Behavioral data were collected via 1:1 or self-administered computerized interviews. Adjusted individual-study, multivariable relative risk regression was used to examine the relationship between CS and SU. We also examined interactions with type of sex and partner HIV status. Pooled effect estimates were calculated using traditional fixed-effects meta-analysis. We analyzed data for recent SU (n = 6781; 82% men, median age = 33 years) and SU before/during sex (n = 2915; 69% men, median age = 40 years). Results For both exposure classifications, any SU other than cannabis increased the likelihood of CS relative to non-SU (8–16%, p-values< 0.001). In the recent SU group, however, polysubstance use did not increase the likelihood of CS compared to single-substance use. Cannabis use did not increase the likelihood of CS, regardless of frequency of use. Type of sex was associated with CS; those reporting vaginal and anal sex had a higher likelihood of CS compared to vaginal sex only for both exposure classifications (18–21%, p < 0.001). Recent SU increased likelihood of CS among those reporting vaginal sex only (9–10%, p < 0.001); results were similar for those reporting vaginal and anal sex (5–8%, p < 0.01). SU before/during sex increased the likelihood of CS among those reporting vaginal sex only (20%; p < 0.001) and among those reporting vaginal and anal sex (7%; p = 0.002). Single- and poly-SU before/during sex increased the likelihood of CS for those with exclusively HIV-negative partners (7–8%, p ≤ 0.02), and for those reporting HIV-negative and HIV-status unknown partners (9–13%, p ≤ 0.03). Conclusion Except for cannabis, any SU increased the likelihood of CS. CS was associated with having perceived HIV-negative partners and with having had both anal/vaginal sex.