Abstract
BACKGROUND: Reports on the sexual behavior of people on antiretroviral therapy (ART) are inconsistent. We selected 14 articles that compared the sexual behavior of people with and without ART for this analysis. METHODS: We included both cross-sectional studies that compared different ART-naïve and ART-experienced participants and longitudinal studies examining the behavior of the same individuals pre- and post-ART start. Meta-analyses were performed both stratified by type of study and combined. Outcome variables assessed for association with ART experience were any sexual activity, unprotected sex and having multiple sexual partners. Random-effect models were applied to determine the overall odds ratios. Sub-group analyses and meta-regression analyses were performed to examine sources of heterogeneity among the studies. Sensitivity analysis was also conducted to evaluate the stability of the overall odds ratio in the presence of outliers. RESULTS: The meta-analysis failed to show a statistically significant association of any sexual activity with ART experience. It did, however, show an overall statistically significant reduction of any unprotected sex, having multiple sexual partners and unprotected sex with HIV negative or unknown HIV status with ART experience. Meta-regression showed no interaction between duration of ART use or recall period of sexual behavior with the sexual activity variables. However, there was an association between the percentage of married or cohabiting participants included in a study and reductions in the practice of unprotected sex with ART. CONCLUSION: In general, this meta-analysis demonstrated a significant reduction in risky sexual behavior among people on ART in sub-Saharan Africa. Future studies should investigate the reproducibility and continuity of the observed positive behavioural changes as the duration of ART lasts a decade or more.