Science Fact or Science Fiction:
Could Circumcision Really Prevent AIDS?
by Norm Cohen, Director, NOCIRC of Michigan
AIDS has been a major fear of sexually active people worldwide. The claim that circumcision might somehow prevent it has been widely circulated in the popular media and in medical journals. This claim may seem quite surprising to Americans, who have lived with circumcision far longer than they have lived with AIDS. In South Africa, however, with an HIV infection rate 31 times that of the United States, this proposal is being taken seriously.
1
In a Class All By Itself
As a preventative, circumcision is in a class all by itself. Nowhere else in medicine has universal surgery been recommended as a means of preventing disease. Nowhere else is surgery on a healthy organ considered an alternative to proper hygiene.
The extraordinary history of circumcision as a medical panacea, and as ancient religious and tribal rites, strongly suggests that latter-day claims in favor of circumcision should be regarded with a high degree of skepticism. For 140 years, circumcision has been proposed as a solution to the most frightening diseases of the times. Over 60 different diseases have been supposedly prevented or cured by circumcision. These diseases have included .masturbation insanity,. syphilis, gonorrhea, penile cancer, cervical cancer, urinary tract infections, and now AIDS.
Each time, evidence was subsequently produced (but not widely publicized) that negated the claim. It is from this historical context that calls for
circumcization (mass circumcisions) to prevent AIDS in Africa should be examined.
Whenever a lack of scientific understanding of ills associated with the penis is combined with the urgency of fear, the circumcision proposition finds fertile ground. Since there has not been much success in reducing the AIDS epidemic in Africa, it is very tempting to believe in quick fixes and miracle cures. Ironically, this proposal may negate any protective effects claimed by accepting risky sexual behavior as the status quo and by encouraging circumcisions in unsterile conditions.
Here are the main arguments that explain why circumcision to prevent AIDS is not good public health policy and cannot be taken seriously:
A Study in Contradictions
When presented with a proposed solution, a test of the solution should be made in populations other than those used to promote the solution to check for consistent results. Proposals that fail this test are inevitably bad science, no matter how fancy a study appears to be.
January 21, 2007 1 of 18 NOCIRC of Michigan
The proposal that circumcision prevents HIV infection fails this test. Of the 35 observational studies included in a stringent review, 16 gave inconsistent results for the general population.
2 At least 20 published observational studies do not show support for circumcision in preventing HIV infection.
All of the randomized controlled trials of circumcision were limited to three countries in southern Africa. Researchers detected an annual HIV incidence rate in
circumcised men over 6 times higher than the annual incidence rate for African-American men in the United States.3 Due to this .success., they stopped the trials early.
The probable mode of transmission for HIV (e.g. heterosexual sex) is often reported when a new infection is reported. The World Health Organization estimates that heterosexual sex has accounted for 75% of the HIV infections in adults worldwide.
4 Heterosexual intercourse has been the dominant route of transmission in Africa, Asia, South America, Central America and the Caribbean. In Western Europe, more than half were acquired during heterosexual intercourse.5 In the United States, one-third of HIV infections are transmitted through heterosexual contacts.6
The circumcision status of a man is not normally reported when a new infection is reported. Therefore, it is difficult to make estimates of the numbers of circumcised men who acquired the infection from heterosexual sex.
For the purpose of a worldwide estimate, and lacking other data, we will use the claim by researchers that there is roughly a 50% reduction in the incidence of HIV infection among circumcised men.
7 The estimates of adult males who are already circumcised is 62% in Africa8, 75% in the United States9, and 20% in the rest of the world10.
So even if circumcision reduced the risk of HIV infection by 50%, over
3.5 million circumcised men worldwide are living with an HIV infection that they acquired through heterosexual sex. This hardly qualifies circumcision as an AIDS vaccine!
The rate of circumcision and the rate of HIV infection in the United States are the highest among all developed nations. Over 500,000 circumcised American men have been infected with HIV from sex since the epidemic began.
11 Regional differences in American circumcision rates don.t match up with regional differences in HIV infections. No studies of men in the United States have been able to demonstrate a correlation between heterosexually acquired HIV infection and the presence of a foreskin.
The Xhosa people of South Africa circumcise, while the Zulus do not. However, both tribes have the same HIV infection rate. In Ethiopia, 93% of the men are circumcised.
12 In Uganda, 25% of the men are circumcised.13 However, both countries have the same rate of HIV infection. In Cote d’Ivoire and Gabon, 93% of the men are circumcised, but the HIV infection rate is even higher in those countries than in Ethiopia.
January 21, 2007 2 of 18 NOCIRC of Michigan January 21, 2007 3 of 18 NOCIRC of Michigan
In Lesotho, the HIV infection rate is substantially higher among circumcised men (23%) than among males who are not circumcised (15%).
14
The lowest rates for all sexually transmitted diseases, including AIDS, remains the Scandinavian nations, where circumcision is virtually unknown.
On the graph that follows there are several examples of countries that contradict the claim that circumcision prevents HIV infection.gan
Making Fiction from Science
The claim that circumcision prevents AIDS was made on the basis of observational studies of men already circumcised and randomized controlled trials where men underwent a circumcision at the start of the study.
All of the studies attempted to predict what happened on a microscopic level by studying conditions on a macro level, which is far less precise. Researchers were not able to observe exactly when, where, or how each individual got infected. Therefore, a fundamental assumption was made that it is possible to draw conclusions about the
mode of transmission of HIV by enumerating the success of transmission in specific populations.
This assumption could produce misleading conclusions about the role of the foreskin in HIV infection. The rules of evidence in medicine are much lower than the rules of evidence in our legal system. The exact
cause of changes in HIV infection rates did not have to be proved by the studies, and was not. In fact, the cause cannot be proved statistically.
Circumcised status and HIV infection are the two variables that researchers compared statistically. The researchers claim that there is a
correlation between the rate of HIV infection and circumcision status. Correlation is the interdependence between two variables. It does not mean that a change in the rate of HIV infection is necessarily caused by a change in circumcision status. Correlation is not the same thing as causation, but the researchers also claimed that the correlation was caused by circumcision status.
The researchers maintain that the difference in HIV infection rates is
not due to many other, confounding factors that might also influence infection rates. If these infection rates do actually depend on other confounding factors, then the results will be reinforced by these other influences.
In an attempt to prevent this, researchers try to .control for. these other factors. In an ideal experiment, all of the confounding factors that could possibly affect the experiment are kept constant so as to eliminate their effects on the outcome. These controlled factors are never supposed to change.
Researchers often claim to have adjusted for .potentially confounding factors,. but this can.t be done completely because all of these factors are simply not known. In practice, researchers can only control for known factors, and only then if they are measured without errors. Even if an attempt is made to control for these known factors, in practice, the control group is not always unpolluted by them.
When a researcher is able to control for the actual factors for the spread of a disease in a population, the hypothetical reason may be found to be wrong. Alternately, if a researcher
So convenient a thing it is to be a reasonable creature, since it enables one to find or make a reason for everything one has a mind to do.
Benjamin Franklin
January 21, 2007 5 of 18 NOCIRC of Michigan
did not control for the actual factors in the spread of a disease because it was not measured or is not known, then the conclusions drawn from the study will be wrong. If all the variables controlled for are not independent of each other, then the results of the study will be misleading and unreliable.
Since circumcision is practiced for religious and cultural reasons, it is not chosen or practiced at random. Foreskins are found among men who have risk factors that actually spread HIV without any help from their foreskins. Significant factors such as sexual practices, number of partners, limited healthcare, poor hygiene, and drug use all have some association with having or not having a foreskin. These and other confounding factors were often not controlled for in the studies and make meaningful comparisons impossible.
15
Any study may also introduce confounding factors by virtue of its existence. For example, the men studied were those who had easy access to a health clinic and willingly showed up there to be interviewed.
Observational studies comparing HIV infection rates in circumcised versus uncircumcised men were poorly designed, inconsistent, and misleading. The observational data varied widely, with no definitive support for circumcision. The studies often contradicted each other in their findings.
When several observational studies were refined with an analysis for confounding factors, the perceived advantage of circumcision disappeared.
16, 17 Controlling for confounding factors eliminated the difference between HIV infection rates between the circumcised and intact groups, indicating that circumcision was not the real reason for a decreased HIV infection rate.
If confounding factors are controlled for, the researcher has to somehow isolate the research subjects having those factors. This often divides the target population into smaller subgroups, thereby reducing the statistical power of the results. A positive correlation could be due merely to the over or under sampling of a subgroup.
|
The confounding factors listed below have all been associated with an increase or decrease in HIV infection. Some of these confounding factors have probably created a spurious correlation between circumcision status and HIV infection rate by operating in favor of the results.
age at circumcision |
non-sterile medical procedures |
|
age at first intercourse |
number of partners |
|
anal sex |
other infections |
|
condom use |
other sexually transmitted infections |
|
drug and alcohol use |
periodic abstinence |
|
.dry. sex |
post-intercourse hygiene |
|
duration of intercourse |
retractability of the foreskin |
|
fidelity |
severity of circumcision |
|
frequency of intercourse |
sex during menses |