Circumcision and HIV

The most important, most current, and perhaps best supported claim for benefits from circumcision is the partial protection it may offer against men becoming infected sexually by HIV.  This page should be read in together with General Notes on the Medical Evidence.

Three studies performed in Africa would appear to give good support to the claim.  We do not seek to refute these studies, but would like to offer some comments on their interpretation.

The authors of the three studies were engaged in a long campaign to gain acceptance of male circumcision as an intervention to prevent men from becoming infected with HIV.  A Cochrane Systematic Review of Male circumcision for prevention of homosexual acquisition of HIV in men initially found insufficient evidence to support an interventional effect of male circumcision on HIV acquisition in heterosexual men.  The authors stated that the results of randomised trials would be required and would need to be carefully considered before circumcision could be implemented as a public health intervention for prevention of sexually transmitted HIV.

The three African studies were implemented in response to that need for randomised studies before circumcision could be implemented as a public health intervention for prevention of sexually transmitted HIV.

In recent years there have been concerns raised in relation to clinical trials in so far as industry-funded trials which do not favour the product being studied may not be published.  This has lead to the efficacy and safety of some drugs being overstated.  To deal with this there is a growing requirement for clinical trials to be registered in advance.  Trial registration should take place before the trials recruit subjects for study and should give details of the statistical methods to be used and any criteria for early termination of the study.  This makes it possible to trace unpublished studies and to confirm that the the methods used in the research were those originally intended by the researchers.  Medical journals will often refuse publication to trials that are not registered.  There is a campaign for all trials to be be registered and all trial data to be publicly available.  You can sign their petition.

The Auvert study was registered on 19 July 2005 the study having started in July 2002
and completed by July 2005.  No trial data are given for this registration.

The Bailey study was registered on 23 April 2003, having already started in February 2002.  The study completed in December 2006.  No trial data are given for this registration.

The Gray study was registered on 23 January 2007 having started in August 2002
and completed in December 2006.   No trial data are given for this registration.

As such none of the three studies was registered before recruitment of subjects, none provided prospective information about the statistical methods to be used or the criteria for early termination.

All three of the studies were stopped early “for benefit”.  The implication is that, given overwhelming evidence that the intervention was effective, it would be unethical to continue the study since it is generally regarded as unethical to deny effective treatment.  The decision for these studies to terminate early is not transparent since the membership of the Data Monitoring and Safety Review Boards is not known, nor are their terms of reference and the criteria for early termination were not registered in advance.

Concern has been raised that clinical trials terminated early for benefit overstate the benefits of treatment.  RCTs stopped early for benefit are becoming more common, often fail to adequately report relevant information about the decision to stop early, and show implausibly large treatment effects, particularly when the number of events is small. Montori suggests clinicians should view the results of such trials with skepticism.  Investigators have reported instances in which study results exceeded stopping boundaries, but because the results yielded implausibly large treatment effects, the DSMB and the investigator agreed to continue the trial and ultimately found no significant benefit.

The three circumcision and HIV studies having been published, the Cochrane Systematic Review was updated to disregard the observational studies which the authors said were limited by confounding which is unlikely to be fully adjusted for and to rely solely on the results from the RCTs.

The updated Cochrane Review concluded that “there is strong evidence that medical male circumcision reduces the acquisition of HIV by heterosexual men by between 38% and 66% over 24 months.  This has since been widely reported as “60% protection” and even hyped to “73% protection”.

So what does “60% protection” mean?

It is fair to say that there is a common deception in published clinical trials for drugs.  It relies on mixing two types of risk: absolute risk and relative risk.

  • Relative risk is usually quoted as a percentage – that the treatment arm might cause a reduction of, say, 50% in the risk of, say, death compared to those in the control (no treatment) group.
  • Absolute risk looks at the actual numbers of people protected from, or exposed to, a harmful outcome in relation to a treatment.  Thus if 1% of the treatment group died where 2% of the control group died, the absolute risk protected against by the treatment is 2%-1%=1%

The two cases above are identical in so far as a 1% absolute risk averted in the study also amounts to a 50% relative risk.

The deception is to present the benefits of the drug in terms of relative risk, perhaps a 50% reduction in the risk of fatal heart attack, but to present the risks of side-effects in terms of absolute risk.  So if there were side effects due to treatment in 2% the treatment has harmed more people than it has helped.  A recent paper by Diamond described how this deception created the appearance that statins are safe and effective in the prevention of heart disease.

In the case of circumcision and HIV, the three African studies do indeed imply a reduction in relative risk for the circumcised men of 60% and they also imply a reduction in absolute risk of 1.3% for catching HIV during the 2 years of the study.  The studies reported complications of circumcision in up to 3% of those in the study.  One study reported an adverse event resulting in erectile dysfunction.

So is a circumcised man 60% less likely to get HIV than an non-circumcised man?

In a word, no.  The circumcised man may be 60% less likely to get HIV from each time he has unprotected sex, but if he is promiscuous and has unprotected sex, and if he lives somewhere with a high prevalence of HIV, the odds are he will succumb to HIV eventually  The non-circumcised man who is not promiscuous and only has protected sex is unlikely ever to become infected.

So, given the constraints of the study, men who had a circumcision were 60% less likely to get HIV from female partners in sex during the first 2 years of a 5 year study than those who did not have a circumcision.  This finding is likely to have been biased by, amongst other things, the fact that circumcised men in the study were counseled to have no sex for 6 weeks and the circumcised men reported a higher level of condom use.

So does this mean that circumcision protects the population from HIV?

In a word, no.  The authors of the updated Cochrane Review were at pains to remove the observational evidence and rely solely on the randomised studies.  This is in keeping with the belief among doctors that RCTs are “Gold Standard” and that they provide direct proof of cause and effect. But this attitude is over-simplistic.

A meta-analysis by Garrenne of thirteen countries where Demographic and Health Survey data were available compared male HIV prevalence in circumcised and non-circumcised groups was compared and found no difference between the two groups (combined risk ratio [RR] = 0.99, 95% CI = 0.94–1.05). Eight of those countries showed no significant difference in HIV seroprevalence in circumcised and uncircumcised groups, while two countries (Kenya and Uganda) showed lower HIV prevalence among circumcised groups, and three countries (Cameroon, Lesotho and Malawi) had higher HIV prevalence among circumcised groups.

The WHO has embarked on a mission to promote circumcision in Africa for the prevention of HIV and scarcely seem to have carefully considered the RCTs before circumcision was “rolled out”, and later “scaled up” as a public health intervention for prevention of sexually transmitted HIV.  Now they are promoting Voluntary Medical Male Circumcision (VMMC) programmes which are targeting infants without any awareness of the irony this implies.

Even given the high prevalence of HIV in the countries concerned, more men may get a complication of circumcision than will be protected from HIV.  In countries with a lower prevalence of HIV the risk to benefit ratio will be even less favourable.  This makes it all the more surprising that the US CDC is promoting circumcision to prevent HIV on its own turf.

Our position is that adult men have the right to choose circumcision if they believe it is worthwhile to protect them from HIV, but they have a right to full information about the risks and harms and a realistic idea of how likely they are to get protection.  Children should be left to choose when they are adults.

If you know of any misconduct in the African circumcision and HIV trials, you can leak it confidentially.

Further reading

Demographic Health Survey Data

A surgical scandal that will promote HIV infection

“Cold Spots” of Male Circumcision and lower incidence of female HIV in Tanzania?

No correlation between circumcision and HIV in Ethiopia’s Regions

Information on avoiding circumcision.

[gravityform id=”5″ title=”false” description=”false”]


About the Author: