The Sunday Mail
This is the second instalment on the circumcision debate in which Dr Sebastian Ndlovu is arguing against the practice. From the last instalment, readers complained about the use of medical jargon and in as much as efforts are made to simplify the language to match everyday use, some of the terms remain purely medical and cannot be simplified any further.
********
Dr Sebastian Ndlovu, MBChB
In last week’s article, we tackled the basis of medical male circumcision and this week we take a closer look at the nitty gritties of that subject.
The studies on circumcision were performed in Kenya, South Africa and Uganda.
While the “gold standard” for medical trials is the randomised, double-blind, placebo-controlled trial, these studies were not properly randomised and were neither placebo —controlled nor double-blind.
To add insult to injury, they had other serious flaws including selection bias, attrition, as well as supportive bias (circumcised men received additional counselling sessions), lead time and duration bias.
Within just a mere two years from commencement, all these studies were stopped, the reason being that circumcision appeared to be able to confer a clear benefit towards preventing the contraction of the virus.
Therefore, it would seem unethical to further expose participants to the risk of HIV when a clear benefit could be ascertained so early in the trial.
According to me, that was the major flaw of all the three studies. How on planet earth could complete bio-statistics specialists derive an astronomical conclusion that is supposed to define a lifetime risk and I repeat LIFETIME risk of contracting HIV basing on a study of less than two years only? This was definitely great for the researchers who clearly had arrived at that conclusion way before they even thought about any studies. A fortuitous end and flagrant duration bias!
This is key because for statistical reasons, effectiveness of a treatment declines with the passage of time. Cutting the experiment short gives a falsely optimistic outcome. A recent meta-analysis of truncated trials (studies stopped prematurely) shows that the practice more than doubles the estimated benefit of any intervention under investigation. This suggests that the three HIV-circumcision RCTs would have showed much less benefit or none if they had not been truncated.
The selection process for all the three studies was not properly randomised. It was based on a person not only volunteering to be part of the study but also agreeing to stay in one area for the whole duration of the trial, and agreeing to have an HIV test whose result had to be negative for inclusion. It is, therefore, arguable that probably the men that volunteered to be part of the studies had a predilection towards being cautious, hygienic, and thoughtful.
Such men fitted the criteria of the cautious males in society who naturally did not have a proclivity towards risky behaviour. The risk-takers, those who tested positive had been effectively selected out. In short, the sample that was taken was not representative of the whole population. This selected out itinerants such as truck-drivers, who are at higher risk of HIV because of their greater variety of partners. Unlike in a real-world setting, study participants were provided free condoms, extensive education and counselling, and they were paid, making any comparisons to the general population egregious and highly dishonest.
Both the people that were being investigated and researchers could not be blinded (ideally both should not know which group is under investigation) due to the nature of the procedure which was under investigation. This made a significant difference between the experimental (cut) and control (intact) group. The control group could have been given a placebo operation, or another kind of placebo, and the same instructions. Even then, since the test could not be made double blind (neither experimenters nor subjects knowing who is circumcised), errors would still occur.
This made the trials subject to what is known as the Hawthorne Effect — all the subjects knew which group they were in, and what effect this was supposed to have. The Hawthorne Effect could have affected their sexual behaviour, making the circumcised men more aware of safer sexual practices, and perhaps more likely to implement them. The fact that the experimental group received more support and was in contact with the researchers more often also creates a major discrepancy between the groups.
HIV is rare in Cuba, where circumcision is also rare, and common in Lesotho, where circumcision is common.
It is obvious that the researchers told the circumcised group that they should continue using condoms, in essence, therefore, the results were probably investigating a protective effect of some form of behavioural change. Consequentially, they concluded that the effect was from circumcision when in this case circumcision alone was a giant, fictitious conjecture.
In all three studies, there was a patent, wrong assumption that infections resulted from heterosexual transmission, without any effort whatsoever to determine the actual sources of the infection. In all the three trials only 89 of 205 infections (about 43 percent) were supposedly “sexually” transmitted. The rest of the infections, a majority of 57 percent, the source is unknown or assumed to be “non-sexual”.
There was no attempt to find out what was responsible for the transmission among a statistically significant number of the non-circumcised group. How WHO in its sanity would move to unanimously recommend the application of a procedure based on results by studies that failed to account for the sources of infection is mischievous at best and outrageous at worst. It was very much necessary to explain the source of all the infections, especially in those who were not circumcised.
For instance in the South African trial, one third (23 of 69) of the HIV infections occurred in men who reported no unprotected sex during the period from their last negative test to their first positive test. In Uganda, 16 of 67 new infections occurred in men who reported no sex partners (6 infections) or 100 percent condom use (10 infections).
None of the studies reported on injections or on any other blood exposures during follow-up. In the Kenyan trial, four men became HIV-positive a month after circumcision. The circumcision itself might have infected them, but the study did not mention that possibility. It is appalling to note that the researchers claimed that the transmission was not sexually related and made no effort to find out the source.
Let’s suppose we choose to believe their hubris, how then on planet earth could circumcision have prevented that very kind of infection in the other group?
There has also been a lot of hullaballoo, both informed and uninformed, around the famous figure — 60 percent! Using such a big number to generate interest in a statistical model is mischievous at times, especially when you want to use it to formulate public policies. I suppose the researchers needed to create media hype, which is understandable because we had not had a lot of very positive news on HIV/Aids in a long time.
But do people know what the “60 percent” statistic is actually referring to? Allow me to shed some light. Across all three trials, of the 5 411 men subjected to male circumcision, 64 (1,18 percent) became HIV-positive and among the 5 497 controls, 137 (2,49 percent).
Therefore 60 percent is the relative reduction in infection rates, found by comparing two very small percentages. Relative risk is simply a ratio of incidence of the outcome in the exposed divided by the incidence of the outcome in the unexposed. According to Hill and Boyle in this instance, “it’s a bit of arithmetic that generates a big-seeming number, yet one which – without also reporting the absolute risk reduction alongside – arguably misrepresents the results of the study”. The absolute decrease in HIV infection between the treatment and control groups in these experiments was just 1,31 percent, which some experts argue is likely to have no appreciable effect at the demographic level.
In other words, if the 60 percent allegation was true you would have to circumcise 56 men to prevent one of them from contracting HIV in one year. However,when this 60 percent relative risk protection is corrected for lead time bias, that is the time the experimental group was charged not to partake in sexual activities post-circumcision when the control group was out there “getting busy”, this goes down to a 49 percent. A further correction for truncation, the figure effectively is reduced to about 20 percent “protection” and this does not augur well for the spirited alacrity that these studies have created, because that effect is already negligible without even factoring in other biases.
Two recent studies examining African circumcision rates and HIV prevalence found that circumcision status was not significantly associated with HIV. One study examined data from 13 sub-Saharan countries found no association, and another found that circumcision made no difference in HIV rates in South Africa. A 2007 study concluded that once commercial sex-worker patterns are factored in, male circumcision is not significantly associated with lower HIV.
Another 2009 publication also shows that circumcision status is not correlated with lower HIV prevalence rate. These correlations require highly selective use of statistics. For instance, HIV is rare in Cuba, where circumcision is also rare, and common in Lesotho, where circumcision is common, and common among both the Zulu of South Africa who do not circumcise, and the Xhosa, who do.
Meticulous verification of methodologies and discrepancies in the RCTs suggest that the benefit, if any, from the programme is not from the act of removing the foreskin but rather the significant contact with health professionals which may foster behavioural changes.
That benefit is virtually negligible or none existent if confounding factors are accounted for. Given that there are more effective, less invasive, less coercive alternatives such as condom-usage, it is inconsistent with biomedical ethics to endorse the risky genital cutting of anyone toward the same ostensible end.
With all due respect, the creation of a hodge-podge of measures to tackle the scourge based on methods that lack solid evidence is anachronistic with current practice. Yes, we may need an assortment of measures but they should not be redundant, should have cumulative effectiveness and above all must be proven worthy. Only then will we be able to dream about an Aids-free generation!
Dr Sebastian Ndlovu is a qualified medical doctor. He writes in his personal capacity and opinions expressed here are his personal opinions, not of any other corporate body or institution. He writes in an attempt to create public discourse on health matters and thereby improve public health policies. You can contact him on his email address: [email protected]
