Male circumcision provides long-term indirect protection to women by reducing the risk of heterosexual men becoming infected with HIV. In this Review, we summarise the evidence for a direct effect of male circumcision on the risk of women becoming infected with HIV. We identified 19 epidemiological analyses, from 11 study populations, of the association of male circumcision and HIV risk in women. A random-effects meta-analysis of data from the one randomised controlled trial and six longitudinal analyses showed little evidence that male circumcision directly reduces risk of HIV in women (summary relative risk 0.80, 95% CI 0.53-1.36). Definitive data would come from a further randomised controlled trial of circumcision among men infected with HIV in serodiscordant heterosexual relationships, but this would involve enrolling about 10 000 couples and is likely to be logistically unfeasible. As circumcision services for HIV prevention are scaled-up in high HIV prevalence settings, rapid integration with existing prevention strategies would maximise benefits for both men and women. Rigorous monitoring is essential to ensure that any adverse effects on women are detected and minimised.