Impact of male circumcision on risk of HIV infection in men in a changing epidemic context - systematic review and meta-analysis

Introduction WHO/UNAIDS recommended Voluntary Medical Male Circumcision in 2007 based on systematic review of observational studies prior to 1999 and three randomized controlled trials (RCTs). To inform updated WHO guidance, we conducted a systematic review and meta-analysis of impact of circumcision on the risk of HIV infection among heterosexual men. Methods Studies in PubMed of HIV incidence and changes in prevalence in heterosexual men by circumcision status were identified. Pooled incidence rate ratios were computed using fixed- and random-effects meta-analysis and risk of bias was assessed using the ROBINS-I tool. Results and Discussion In three RCTs, the pooled incidence ratio was 0.41 (95% CI 0.30 to 0.56), with risk difference 10 (8 to 12) fewer infections per 1000 person-years (py). Pooled incidence ratios were 0.34 (0.24 to 0.49) in two post-RCT follow-up studies, 0.29 (0.19 to 0.43) in men at high HIV risk (five cohorts), 0.48 (0.33 to 0.70) in four community-based cohorts before circumcision scale-up, and 0.56 (0.49 to 0.64) (7 [6 to 8] fewer per 1000 py) in six community-based cohorts during circumcision and antiretroviral treatment scale-up. Heterogeneity between studies was low except in men at high HIV risk. We estimated 520,000 (425,000 to 605,000) fewer infections occurred in men by end of 2018 from 22.7 million circumcisions performed since 2008 and increasing by 155,000 (125,000 to 180,000) annually if epidemic conditions remain similar. After exclusion of studies with high risk of bias and those conducted outside Africa, pooled incidence ratios were similar. There was no evidence of confounding nor changes in risk behaviour following circumcision. In post-hoc exploratory analyses we observed a trend of decreasing effectiveness of circumcision in cohorts with lower HIV incidence. Conclusions Efficacy of medical male circumcision on HIV incidence from randomized controlled trials was supported by effectiveness from observational studies in populations with diverse HIV risk and changing epidemic contexts. Voluntary Medical Male Circumcision remains an important evidence-based intervention for control of generalized HIV epidemics.