| Excerpt from The WHO Reproductive Health Library | Published by Update Software Ltd. |
This review evaluates the evidence for an interventional effect of male circumcision in preventing HIV acquisition by men through heterosexual intercourse. The authors found no randomized controlled trials on this topic. Hence, the review is based on 34 observational studies, including 5 cohort, 4 case–control and 25 cross-sectional studies addressing this issue. Approximately half of the studies were conducted among the general population (n=16) and the other half among high-risk populations (n=18). All five cohort studies (1 in the general and 4 in the high-risk population) showed a protective effect, with 3 of these having statistically significant results. The results from the 15 cross-sectional and 1 case–control study in the general population were inconsistent (high heterogeneity). However, the use of adjusted results tended to show stronger evidence of an association. In the high-risk population all cross-sectional and case–control studies showed a protective effect for male circumcision, with the majority of studies (8 cross-sectional and 2 case–control) reporting statistically significant results. The authors concluded that the existing studies show “a strong epidemiological association” between male circumcision and HIV acquisition, but that methodological weaknesses of these observational studies result in insufficient evidence to support a causal effect. Therefore, there is an urgent need to conduct randomized controlled trials to evaluate this issue. Feasibility of male circumcision as an intervention was beyond the scope of the review.
The search strategies used for the review were comprehensive and appropriate: various databases were searched and no language restrictions were applied. The authors were unable to contact authors of many conference abstracts and thus publication bias is a possibility (negative studies may remain unpublished). Independent reviewers applied clearly specified inclusion and exclusion criteria for the selection of studies for the review and the use of two reviewers to extract data independently ensured accurate data collection.
The authors’ choices of study outcome and the general analytical strategy were appropriate. They decided not to conduct a meta-analysis owing to the heterogeneity of the study results. Instead, they conducted stratified analyses by study type (e.g. cohort, cross-sectional) and further by study population (general vs. high-risk). They also considered the impact of the method of circumcision and the use of adjusted rather than crude results. These various ways of considering the data led to important insights (e.g. a stronger protective effect of male circumcision among high-risk compared with the general population groups). Unfortunately, owing to limitations in the studies they could not report separately on the effects of male circumcision by: the acquisition of HIV-1 and HIV-2, the background HIV incidence/prevalence levels, or the publication status of results. A careful quality assessment of the various studies identifying potential sources of bias was conducted. Given this, it would have been helpful for the authors to present a stratified analysis of the effect of male circumcision by study quality within each study design.
The full RHL commentary also includes sections on:
Relevance