No shortcuts in circumcision

Snippety, snip, snip. Could that be the sound of AIDS in retreat in Africa?

The auditory metaphor and its effect come to mind if you spend any time reflecting on a much-publicized South African study tying a dramatic decline in HIV infection rates to circumcision. But well might you think just the opposite if you were apprised of the sometimes putrid public health politics underlying publication of said study.

The facts seem straightforward on the surface. Some 3,000 young men – hardly any married – were selected from a semi-rural area near Johannesburg. Half were put in a group that got circumcisions, half in a non-circumcised group. After more than a year, 20 of the circumcised men had become HIV positive versus 49 of the uncircumcised men, this even though the circumcised men had more sexual encounters.

The scientists were jubilant. "The result is equivalent to saying that during the period [of the study] the intervention prevented six out of 10 potential infections," wrote the French and South African researchers who conducted the research, adding, "this provides a degree of protection against acquiring HIV infection equivalent to what a vaccine of high efficiency would have achieved."

The finding was so striking, the researchers stopped the trial before it was supposed to finish because the "protective effect of male circumcision was so high it would have been unethical to continue." That is to say, they thought there was no doubt that circumcision protected against HIV infection.

Finally, the effectiveness of circumcision led the scientists to argue that decision makers should herewith consider making circumcision of all African men a public health priority.

Wonderful, except for all the ethical and methodological mud splattered on the findings.

As a background you should know the project grew out of more than two decades of observation that places in Africa where circumcision was the norm had less AIDS than the uncut places. We are not, by the way, simply talking about Muslim Africa but also tribes where male circumcision is part of a rite of passage into adulthood.

While scientists have been able to come up with numbers of physical explanations for why what has sometimes been snidely called the "cut cure" works, nobody has been certain about the connection. This is because no previous studies have controlled for confounding factors – age at circumcision, number of sexual partners, safe sex practices – that could distort results. So the Johannesburg study was vital in translating anecdote and intimation into believable science.

However, there was a fundamental ethical problem with its methodology. The men were given HIV tests before the project began and 146 were found to be HIV positive, but – underline the following in lipstick red – they weren't told about their status as researchers "considered it unethical to inform participants of their HIV status without their permission."

Gasp. You don't tell people with a life-threatening, highly infectious disease they are both sick and dangerous to others because knowing somehow violates their sense of privacy?

To justify this position, French and South African scientists involved argued that they were just respecting a fear of AIDS stigma so intense that "many of these people prefer to be dead than rejected by their communities."

Gasp again. This is a public health position so obviously crazy – think in the Canadian context of not telling people with SARS they had it because they would be stigmatized and quarantined – it led the British journal The Lancet to reject the publishing of the AIDS paper on the grounds it was ethically flawed.

And there may be a worse confusion to come. Not everyone believes the HIV infection numbers in men who hadn't been circumcised were so conclusive they justified the trial being shut down early. Part of the concern was caused by two recent papers that suggest that clinical trials claiming huge, big, early effects from drugs or other treatments as often as not turn out to be statistical blips and not true results.

Could that be the case here?

"My sense is that the circumcision study may have been stopped too early and that there is a real danger we may be subjecting hundreds of thousands or millions of men to having circumcisions that may not have the benefit we assume," Jeremy Grimshaw, director of the Clinical Epidemiology Program at the University of Ottawa, warns me about the South African study.

Gasp a third time. So why stop a study when the number of people who had become infected wasn't even half as large as the number who had the disease to start with and weren't told they had it? My guess is a guilty conscience. The doctors wanted an excuse to tell all the infected of their condition, no matter the stigmatization, and the early, positive statistics gave them just such an out.

My justification for this charge is that immediately after they closed down the trial, the researchers changed the rules so they could inform people of their disease, even if the people initially said they didn't want to know.

Maybe other, still ongoing trials will support the cut cure, but for the time being my faith in this one has gone snippety, snip, snip.