While prevention of sexually transmitted infections is irrelevant to non-sexually-active newborns and children, nonetheless, prevention of human immunodeficiency virus (HIV) has become one of the main ‘medical benefits’ rationales given for circumcision.
A decade ago, three randomized controlled trials (RCTs) done in sub-Saharan Africa appeared to show, during the study period, a 38-66% relative reduction for the circumcised subjects in the risk of heterosexual, female to male only, transmission of HIV.[85-87] All three studies were terminated early, due to their apparently clear results. However, Dowsett and Couch examined the results of the three RCTs, but found insufficient evidence to recommend circumcision to prevent HIV infection. Green et al. reviewed the evidence and also found “insufficient data” as well as contrary evidence.[89,90]
While RCTs are often considered the gold standard of medical trials, this only applies to RCTs with study designs that minimize bias. The three African RCTs were very similar in study design and contained multiple sources of bias, outlined below[89-92]:
- Researcher expectation bias − Many of the investigators had written papers advocating for male circumcision to prevent HIV infection prior to undertaking these RCTs. Siegfried et al. note, on the topic of research on circumcision to prevent HIV, that “researchers’ personal biases and the dominant circumcision practices of their respective countries may influence their interpretation of findings.”
- Participant expectation bias − The majority of participants were convinced that circumcision would reduce their risk of HIV infection.
- Lead time bias − Men randomized to the intervention arm of the trials (the group that was circumcised) were considered to be at risk for becoming infected from the time of the surgery, even though they were told to avoid sexual activity during the period of wound healing. Men in the control arm (the ones who were not circumcised) were able to be sexually active from the beginning of the study.
- Selection bias − Only men who were interested in a free circumcision were eligible to participate, and therefore may not have been representative of the general population.
- Attrition bias − For every man who became infected with HIV during the trials, 3.5−7.4 men were lost to follow-up. This is a serious methodological problem that could alter the statistical significance of the findings.
- Early termination bias − Studies that are terminated early are more likely to overestimate any treatment effect.[96,97]
- Duration bias − Because men who were not initially circumcised were circumcised at the end of the study, long-term comparison of the effects cannot be accurately extrapolated, as some modelers have proposed.
- Source of infection unknown − If the studies were designed to determine whether circumcision reduced the risk of heterosexually-transmitted HIV, the investigators should have confirmed that the infections were indeed transmitted through heterosexual sexual contact. They did not. Using the data reported, it is estimated that about half of the infections of the men in these studies were not sexually transmitted.
The cumulative treatment effect in these trials – which claimed a 38-66% relative risk reduction – was an absolute risk reduction of 1.3%. This is a very small effect, which could easily have resulted from the various forms of bias, rather than being a true treatment effect. The findings are not robust, given that all of the trials had nearly identical methodologies and nearly identical results.
In any case, it appears that these trials were unnecessary in the first place. Data released before the trials began found a number of African countries where the prevalence of HIV infection was greater in circumcised men than in intact men.[100,101]
Unfortunately, the results from the three RCTs provided the impetus for the WHO to bypass the usual step of performing pilot studies to determine if circumcision was effective outside of a research setting. Instead, it recommended programs to circumcise millions of men in sub-Saharan Africa as quickly as possible. (These programs measure success by the number of males circumcised rather than by their impact on HIV incidence. Since the mass circumcision campaigns began in Uganda and Kenya, the incidence of new cases of HIV in both countries has increased.[102-104]) The WHO recommendations included that circumcision programs should be voluntary, free of coercion, and targeted to areas where the HIV prevalence is high (>15% of the population) and circumcision rates are low (<20%). None of these criteria apply to the situation of newborn babies (who cannot voluntarily consent) in the epidemiological setting of the United States (low HIV prevalence, and already high circumcision rate, as well as a much higher standard of living than Africa).
As with other STIs, there is no evidence that circumcision has had any impact on lowering the incidence of HIV infection in the United States. Of the eight HIV studies in North American heterosexual men,[100,106-112] only one has found a significant association between circumcision and HIV infection risk: it actually found that circumcised men were at greater risk of HIV infection. Furthermore, the HIV epidemic in the United States is concentrated among men who have sex with men (MSM) and injecting drug users. A meta-analysis of the studies published on this topic by the Centers for Disease Control and Prevention (CDC) found that the risk for HIV infection in MSM is the same in intact and circumcised men. A subsequent study yielded similar results.
While some authors have theorized that Langerhans cells present in the foreskin are a significant portal for HIV infection, and have used this as a rationale for circumcision as a preventative, evidence now shows that these cells produce a protein, langerin, that is actually protective against the virus. The authors of the langerin study argue that “strategies to combat [HIV] infection must enhance, preserve or, at the very least, not interfere with langerin expression and function [in the Langerhans cells of the foreskin].”
RCTs carried out among adults in Africa are not relevant to children anywhere, since children are not sexually active and are therefore not at risk of HIV infection by sexual transmission. For adults, condoms are an effective means of preventing sexually transmitted infections, including HIV. Other preventative interventions, such as “treatment as prevention” and pre-exposure prophylaxis, are more effective, less expensive, and less injurious than circumcision.[118,119]