Circumcision and HIV

For simplicity, this page may now be cited as www.circumstitions.com/hiv.

Circumcision and HIV

A lie will be halfway around the world before the truth has got its pants on.

- Rev. C. H. Spurgeon, 1855
who called it an old proverb

"Scientists have power by virtue of the respect commanded by the discipline. We may therefore be sorely tempted to misuse that power in furthering a personal prejudice or social goal -- why not provide that extra oomph by extending the umbrella of science over a personal preference in ethics or politics? But we cannot, lest we lose the very respect that tempted us in the first place."

- Stephen Jay Gould
Bully for Brontosaurus, pp 429-30

(But some, it seems, are willing to take that risk.)

Contents

Summary
"Therefore Carthage must be destroyed"
Flawed studies
the Random Clinical Tests
A misreported study
Contrary studies - A Cochrane Review
Where circumcision doesn't prevent AIDS
Between Correlation and Recommendation
A Vaccine? Hardly!
A Solution Looking For A Problem
Ethics
"Dry Sex"
Alcohol
Female Genital Mutilation
Sexual Selection
Wife Inheritance
Heterosexual transmission - Europe vs the United States
A voice of sanity from UNAIDS
A voice of sanity from the Terrence Higgins Trust
A UK survey of gay men that found more circumcised men with HIV

The hazards of unblinded trials
Other studies that show no correlation or a negative correlation
between intactness and HIV/AIDS

"Therefore Carthage must be destroyed"

(The Roman senator Cato was in the habit of ending every speech on any subject with those words - it helped that "must be destroyed" was a single splendid Latin word, "delenda". Eventually the Senate agreed to destroy Carthage, with disastrous consequences for Rome.)

Recently, several studies have been published, most from Africa, one from India, claiming to show a link between having an intact penis and a higher risk of HIV infection. They tend to have two things in common - flawed work, and a passage near the end saying "Therefore, universal male circumcision should be considered as a preventive measure against HIV infection" or words to that effect.

This advice is folly.

Any link between circumcision and HIV is statistically quite slight, so the protection would be quite inefficient compared to education in safe-sex practices and a culture of protected or otherwise safe(r) sex.
The studies are trumpetted by their Relative Risk Reduction (RRR), currently running about 50-60%, but the corresponding Number Needed to Treat (NNT) is much less impressive. (In the latest Uganda study it is 56 circumcisions to prevent one HIV infection per year. That corresponds to 380 circumcisions/infection/year in the US, where AIDS is less prevalent.)
As each new study corrects the errors of its predecessors, the protection claimed is less. When all the errors are corrected, what effect will be left?

In the case of randomised controlled trials (RCTs),

While large numbers of men enter a trial, only a very small number are infected, making random errors high.
The men were randomly assigned to be circumcised or left intact, but they were not a random sample of the population.
- They were all HIV-negative, meaning they were more likely to have any natural immunity than the rest of the population
- None were circumcised, meaning certain tribal groups had been selected out.
- All were willing to be circumcised
- They were significantly rewarded for taking part, skewing the socio-economic status of the sample
  It may have been impossible to correct for these (since humans are not lab rats), but they are issues none the less.
Significant numbers of men dropped out of the trials (were "lost to study"). Only those who stay the distance should be counted.
Those who are circumcised and contract HIV will be more likely to drop out than the others because
1. they got what they came for but
2. circumcision didn't protect them, so they would be disillusioned with the trial.
The three RCTs were cut short: this has reduced their accuracy.
- The control groups were then offered circumcision, making long-term follow-up impossible.
- Ethical approval for better studies will be harder to get, making these studies the last word.
The gold standard of medical testing is the double blind random controlled trial. Circumcision can not be concealed from the experimenter or the subject. The control groups were not given a placebo operation.
The after-effects of the operation are likely to alter sexual behaviour.
In an experimental environment, the subjects got counselling and safe-sex advice that would not be available in a mass circumcision campaign.
The circumcised group had specific instructions to abstain from sex and use condoms that the intact control group does not.
- Experimenter and circumcision advocate Robert Bailey has admitted that "repeated study visits and intensive behavioural counselling" of the circumcised men were needed to reduce risk behaviours.

If these results are acted on, with mass circumcision campaigns:

Protection, if any, would be extended to a population, but it would be impossible to convince the average man that circumcision did not confer significant protection on him personally.
The temptation would be irresistible - especially if he had submitted to a painful operation in adulthood - for a man to say "I'm circumcised, I'm safe".
- He would be more likely to lean on partners for unprotected sex
- This disempowers women
Transmission from man to woman is easier than from woman to man. Circumcision has not been shown to protect women directly.
Since circumcision desensitises the glans, men circumcised in adulthood would be less willing to use condoms than before.
If they combine mass circumcision with Abstinence, Be faithful, Condoms, campaigns, as they propose, it will be impossible to tell what is responsible for the outcome: circumcision will be given the credit for any reduction, but will not have to take the blame for any lack of reduction.

The latest study (2006) is the most careful so far to avoid the mistakes of its predecessors, but it still falls far short of justifying mass circumcision campagns of men in Africa, let alone Routine Infant Circumcision. It claims to have found a less protective effect than the one before it.

Flawed Studies

But each of these studies is flawed in one way or another.

In the study of Kenyan truck drivers,
- 95 had intact penises, and of those, eleven men contracted HIV-1 in a 20 month period, compared to 32 of the 651 circumcised men in a 21 month period. That is to say, six more intact men contracted HIV-1 than the 5 out of 95 than the aggregated rate of 3.34 per hundred per year would predict. While this might look like a big difference, it is far too few, outside a laboratory, to draw any meaningful conclusions. "The law of small numbers" applies. Those six might have just been unlucky. Applying high-powered statistical methods to such a small sample as this, and with so many unknown variables, is using a sledgehammer to crack a nut.
- There were significant unexplained numerical discrepancies between two different publications of this study.
- The study's authors admit that circumcision is so closely tied to ethnicity that it was not possible to assess the effects of circumcision independently from those of ethnic origin.
- Nor were the effects assessed of ethnic origin or religion on other practices that might influence HIV-1 transmission, such as
  - anal sex. An analysis of same-sex activity by the truck drivers, and how that is affected by ethnicity and religion, might cast a completely different light on the results. In the nature of things, membership of a tribe or ethnic group correlates with a variety of different customs, including sexual practices, and it may be those, rather than circumcision, that is responsible for any difference in HIV transmission.
  - "dry sex": the use by women of herbal and other astringents to dry their vaginas (to increase men's pleasure, though it decreases their own). This causes micro-tears which can facilitatte HIV transmission.
  - Female Genital Mutilation, which is practised only where male circumcision is also (with one exception, the Pokot of Kenya, and they used to circumcise males, but have given it up).
A study released in Nairobi compares quite different populations of men. According to CBS: "The study focused on Benin's capital Cotonou and Cameroon's capital Yaounde, where circumcision is a widespread cultural practice, as well as the Zambian town of Ndola and the Kenyan town of Kisumu, where it is not." That is, the men compared lived in different countries, as much as 2300 miles (3600 km) apart!
The Rakai study in Uganda showed no circumcised men contracting HIV during its 30 month course, and this fact has been made much of by the likes of Szabo and Short. They fail to mention that more than a third of the circumcised men were infected before the study started, and hence were not admitted to it. Thus all the men in the study had been selected in advance for less than average susceptibility to HIV.
However that selection could in turn be affected by circumcision status. Adolescent circumcision may delay the age of onset of intercourse (in societies where women won't have intercourse with intact men, unlike women in non-circumcising societies) which in turn would affect their chance of contracting HIV and being excluded from the study before it began. The lower HIV rate may have merely been a result of circumcised men having taken risks for longer than the intact men, and hence being more likely to have some immunity to HIV when they entered the study.
A study published in Scientific American used nationality as a marker for circumcision status, yet African national boundaries are an historical accident arising frm the 19th Century "carve up of Africa".
Other studies often rely on self-reporting of circumcision status. Where a man was circumcised in infancy, he may very well imagine he is intact because he looks like all his peers, and studies have shown a misapprehension of one's own status of as much as 33%.
Only one of the African studies claims to have sufficiently corrected for the fact that circumcision in central Africa is largely a Muslim rite, and Islam requires
- ritual washing before prayer
- abstinence from alcohol
- periodic abstinence from sex, and
- marital fidelity
- all factors affecting HIV transmission. Islam allows polygamy, which makes extramarital sex less likely, just by exhaustion, and encourages female seclusion, which of course makes HIV transmission less likely.

That one study, confined to Christians in Kenya, compared men belonging to churches that encouraged circumcision with those that discouraged it. It used physical examination to determine circumcision status, and confined itself to churches that had similar views on polygamy and widow inheritance (of their late husbands' brothers as second husbands).
With those precautions, the correlation between circumcision status and HIV acquisition fell to 1.5 (20% of circumcised men had HIV, vs 30% of intact). With "adequate" genital hygiene, the rate among the intact fell to 26%. The circumcised men were more likely to be married and to have more than one wife, less likely to have ever been with a sex worker, or with more than three sex workers. (Equal proportions of both groups, 10-11%, had been with one or two sex workers.) These factors could well account for the difference.
The study had a 27% non-participation rate. The authors maintain that "because participants did not know their HIV-1 status at the time of our visit, bias from this source would seem unlikely." But many would know their HIV-1 status because of AIDS symptoms. All would know their circumcision status. There are thus unknown ways in which men might non-randomly "include themselves out".
This study - unlike others - found no effect of age of circumcision on HIV acquistion, even if the circumcision took place after sexual activity began and after HIV was prevalent. This suggests that circumcision itself is not the key factor. An unexplored area is what else the churches advocated or required beside circumcision. Since the church circumcisions occur on the eighth day after birth, it seems likely they model themselves on Judaism: what other Jewish practices do they advocate, and what effect could those have on HIV acquisition?
As controls on "psychic" research are tightened, the effects found steadily diminish, and when control is complete, the effects vanish. We see a similar effect here. These results are certainly consistent with the null hypothesis, that circumcision has no effect on HIV acquisition: the confounding factors have just not all been found yet. Yet as usual, this study advocates that "male circumcision should be seriously considered as an intervention to slow the spread of HIV-1 in uncircumcised populations". It is hard to escape the conclusion that this line was written before the study began.
One study, of gay men who visited STD clinics in Seattle, relied on self-reporting, and also found a significant correlation between being circumcised and intra-venous drug use. This was not commented on (and the parallel conclusion, that circumcision should be discouraged in order to prevent IVDU, was of course not drawn). Again, only a small number of the men (thought they) were intact - 59 out of the 313 HIV+ men and 18 out of the 186 HIV- men.
A study of men visiting STD clinics (which in itself skews the sample) in Pune, India is a classic example of inadvertent sorting by religion. In India, only Muslim (and Jewish) men are circumcised.

Two studies claim to find a mechanism for the proposed correlation, involving the Langerhans cells of the foreskin. However they base their conclusions on diametrically opposed data:

A study much touted in early 2000, that of Szabo and Short, is based on a search of the literature (the other flawed studies just listed) plus a histological examination of the penises of 13 cadavers, all aged over 60, only 6 of them with foreskins. It found Langerhans cells on the inner mucosa and concluded that they facilitated HIV transmission.

That of arch-circumcisionist Gerald Weiss of seven years earlier examined the foreskins of a cohort of circumcised babies and found a deficiency of Langerhans cells, and concluded that their absence rendered the foreskin vulnerable to HIV transmission.

These studies are contrasted side by side on another page.

Another experimental study compares dead, excised foreskin tissue, with dead excised cervical tissue. It cites both the Short and Weiss studies without mentioning the contradiction between them. It cites the Quinn et al. (Rakai) study in Uganda that found 0 of 40 cut men seroconverting, compared to 40 out of 137 intact men - but ignores Quinn's reply to one of this paper's authors explaining why circumcision was not a factor.
It tests the uptake of HIV by foreskin tissue with that of uterine cervical tissue - rather than vaginal, labial or clitero-preputial mucosa, or the mucosa of the male glans for example - for no apparent reason. (One reason could be that cervical tissue was easier to obtain, via hysterectomy).
The dead foreskin and cervical tissue was subjected to an extraordinary amount of processing before it was even ready to be inoculated with HIV or HIV genes. The experimenters may answer that since the cervical and foreskin tissues were subjected to the same processing, any experimenter effects would be cancelled out - but, since they are different tissues, how does anyone know that?
They use skin from the outside of their sample foreskins as a surrogate for the shaft skin of circumcised men, but fail to take into account that only very low and tight circumcisions will result in a shaft covered only in skin: the traditional African method of drawing the foreskin forward on a block and slicing or chopping (as described by Nelson Mandela in excruciating detail in his autobiography) results in a circumcision that is low and loose, leaving plenty of mucosa.
Doubtless this study will now be cited again and again as proof that live HIV is more likely to infect live intact men than live circumcised men - even though no live penile tissue (and no circumcised penile tissue) was involved in the experiments.
The paper again proposes mass circumcision as an HIV preventative measure, considering only "acceptablity and operational feasibility," not ethics. It throws a sop to the false sense of security this would engender, recommending
"...counselling parents and men against increasing sexual risk behaviours in the belief that circumcision fully protects against HIV acquisition." In other words, they propose to persuade men to be circumcised because that will protect them, and simultaneously tell them not to have unsafe sex because it won't. A mixed message indeed!

A Misreported Study

Studies that claim to find a correlation between intactness and HIV transmission are not uncommonly misreported in a way that plays up the "protective effect". In one particularly glaring case, a study that found no statistical signficance was widely reported as finding a protective effect.

The study (Gray R et al. Male circumcision and the risks of female HIV and sexually transmitted infections acquisition in Rakai, Uganda. Thirteenth Conference on Retroviruses and Opportunistic Infections, Denver, abstract 128, 2006.) was an attempt to find whether circumcised men were less likely to infect their female partners with HIV.

299 couples where the man was intact were compared with 44 where the man was circumcised. After 30 months (if the pattern of the rest of the study was followed), infection rates were 7 per 100 person-years for the wives of circumcised men and 10 for the wives of intact men. This may look like a protective effect, but in statistical terms, p=0.22, meaning no statistical significance. In real terms, it can be back-calculated that 8 of the wives of circumcised men were infected. If 11 had been, the rate would be the same for both, and that difference of three infections in 30 months is too few to be considered significant.

But the study was widely reported (by Reuters) as showing that all 299 wives of intact men were infected, compared with only 44 wives of circumcised men, as if these were just the small (infected) samples of two much larger and equal samples. This makes the supposed protective effect look much greater.

See the garbled report and the relevant part of a more accurate report.

Why, one wonders, was the study ever published, and why in this very misleading form?

Contrary Studies

Circumcsion does not protect black South Africans

A total of 2585 males over the age of 15 were administered questionnaires and provided specimens for HIV testing. 916 (35.4%) of them said they were circumcised. HIV prevalence among circumcised males was 10.7% and among uncircumcised males was 12.1%, p = 0.9 [i.e. no statistical significance]. Blacks were less likely to be circumcised (28.8%) compared to other racial groups, 42.6%, p = 0.002. When the data was stratified by racial group, circumcised Blacks showed similar rates of HIV as uncircumcised Blacks, (OR: 0.8, p = 0.4) however other racial groups showed a strong protective effect, (OR: 0.3, p = 0.01) [or rather, a correlation]. When the data are further stratified by age of circumcision, there is a slight protective effect [correlation] between early circumcision and HIV among Blacks, OR: 0.7, p = 0.4.
Conclusion In general, circumcision offers slight protection. The effect is much stronger in other racial groups than in blacks. This racial difference cannot be explained by age of circumcision.

HIV and circumcision in South Africa
C.A. Connolly, O. Shisana, L. Simbayi, M. Colvin.
Poster at the XV AIDS Conference in Bangkok [MoPeC3491]

No protection among young South Africans

A 2001 study by Bertran Auvert et al (who also ran the 2005 Random Controlled Study) of HIV infection among youth in a South African mining town found it is associated with the Herpes simplex 2 virus

It was "a community-based, cross-sectional study" of a random sample of men (n = 723) and women (n = 784) living in a township in the Carletonville district of South Africa.
Risk factors associated with HIV were recorded by questionnaire and biological tests were performed on serum and urine.
It found that women were much more likely to have HIV (34%) than men (9%) and HSV-2 (53% vs 17%) Two thirds of the 24-year-old women had HIV. Of the men,

Circumcision status	n	HIV+	Odds ratio	95% Confidence interval
No	498 (89.1%)	11.2%	1
Yes	61 (10.9%)	16.4%	1.6	0.7-3.2

Thus, the circumcised men in the study were more likely to be HIV+, but the result was not significant (the 95% CI straddles 1.0 - in real terms, 10 of the 61 circumcised men had HIV, three more than would be expected if they had the same rate as the intact men) But it certainly casts doubts on the claim that circumcision protects against HIV infection. Typically, Auvert expresses this cautiously, in terms of the prevailing mythology: "No protective effect of circumcision on HIV prevalence was shown."

No protection to insertive gay men:

"Our finding that 17% of homosexual men with newly acquired HIV infection reported insertive UAI [unprotected anal intercourse] as their highest risk activity suggests that insertive UAI is an important means of HIV transmission in this population. However, we found no association between circumcision status and infection by insertive UAI. In addition, men who had seroconverted despite no reported event of UAI were also no more likely to be uncircumcised. These data strongly suggest that the foreskin is not the main source of HIV infection in homosexual men who become infected by insertive UAI, and that other sites, such as the distal urethra, must be important in HIV infection.

"Our data showing that there is no difference in the circumcision status of men infected by receptive or insertive UAI, in a population with a circumcision prevalence of approximately 75%, suggests that circumcision is not strongly protective against HIV infection in homosexual men. Larger studies, preferably of prospective design, are needed to confirm the absence of a relationship between circumcision and HIV infection risk in gay men. In the meantime, educational messages to homosexual men should continue to emphasize that insertive anal sex is a high-risk activity for HIV transmission whether or not the insertive partner is circumcised."

- Grulich AE, Hendry O, Clark E, Kippax S, Kaldor JM.
Circumcision and male-to-male sexual transmission of HIV.
AIDS 2001 Jun 15;15(9):1188-1189.

No protection to US Black and Latino men who have unprotected insertive sex with men

JAIDS
December 15, 2007

Circumcision Status and HIV Infection Among Black and Latino Men Who Have Sex With Men in 3 US Cities.

Millett, Gregorio A; Ding, Helen; Lauby, Jennifer; Flores, Stephen; Stueve, Ann; Bingham, Trista; Carballo-Dieguez, Alex; Murrill, Chris; Liu, Kai-Lih; Wheeler, Darrell; Liau, Adrian; Marks, Gary

Abstract:
Objective: To examine characteristics of circumcised and uncircumcised Latino and black men who have sex with men (MSM) in the United States and assess the association between circumcision and HIV infection.

Methods: Using respondent-driven sampling, 1154 black MSM and 1091 Latino MSM were recruited from New York City, Philadelphia, and Los Angeles. A 45-minute computer-assisted interview and a rapid oral fluid HIV antibody test (OraSure Technologies, Bethlehem, PA) were administered to participants.

Results: Circumcision prevalence was higher among black MSM than among Latino MSM (74% vs. 33%; P < 0.0001). Circumcised MSM in both racial/ethnic groups were more likely than uncircumcised MSM to be born in the United States or to have a US-born parent. Circumcision status was not associated with prevalent HIV infection among Latino MSM, black MSM, black bisexual men, or black or Latino men who reported being HIV-negative based on their last HIV test. Further, circumcision was not associated with a reduced likelihood of HIV infection among men who had engaged in unprotected insertive and not unprotected receptive anal sex.

Conclusions: In these cross-sectional data, there was no evidence that being circumcised was protective against HIV infection among black MSM or Latino MSM.

JAIDS Journal of Acquired Immune Deficiency Syndromes. 46(5):643-650, December 15, 2007.

No protection to women

Author: Turner AN | Morrison CS | Padian NS | Kaufman JS | Salata RA

Source: AIDS. 2007 Aug 20;21(13):1779-1789.

Abstract: The objective was to assess whether male circumcision of the primary sex partner is associated with women's risk of HIV. Data were analyzed from 4417 Ugandan and Zimbabwean women participating in a prospective study of hormonal contraception and HIV acquisition. Most were recruited from family planning clinics; some in Uganda were referred from higher-risk settings such as sexually transmitted disease clinics. Using Cox proportional hazards models, time to HIV acquisition was compared for women with circumcised or uncircumcised primary partners. Possible misclassification of male circumcision was assessed using sensitivity analysis. At baseline, 74% reported uncircumcised primary partners, 22% had circumcised partners and 4% had partners of unknown circumcision status. Median follow-up was 23 months, during which 210 women acquired HIV (167, 34, and 9 women whose primary partners were uncircumcised, circumcised, or of unknown circumcision status, respectively). Although unadjusted analyses indicated that women with circumcised partners had lower HIV risk than those with uncircumcised partners, the protective effect disappeared after adjustment for other risk factors [hazard ratio (HR), 1.03; 95% confidence interval (CI), 0.69-1.53]. Subgroup analyses suggested a non-significant protective effect of male circumcision on HIV acquisition among Ugandan women referred from higher-risk settings: adjusted HR 0.16 (95% CI, 0.02-1.25) but little effect in Ugandans (HR, 1.33; 95% CI, 0.72-2.47) or Zimbabweans (HR, 1.12; 95% CI, 0.65-1.91) from family planning clinics. After adjustment, male circumcision was not significantly associated with women's HIV risk. The potential protection offered by male circumcision for women recruited from high-risk settings warrants further investigation. (author's)

Date Posted: 3 September 2007

Greater risk to women whose partners are circumcised:

Int J Epidemiol. 1994 Apr;23(2):371-80.

Risk factors associated with prevalent HIV-1 infection among pregnant women in Rwanda.
National University of Rwanda-Johns Hopkins University AIDS Research Team.
Chao A, Bulterys M, Musanganire F, Habimana P, Nawrocki P, Taylor E, Dushimimana A, Saah A.
Department of Epidemiology, School of Hygiene and Public Health, Johns Hopkins University, Baltimore, MD 21205.

Abstract: This study evaluated risk factors associated with prevalent HIV-1 infection among pregnant women in a semi-rural but densely populated area surrounding the town of Butare in Rwanda. Overall seroprevalence was 9.3% in 5690 pregnant women who sought antenatal care at one of five health centres. Factors associated with higher seroprevalence of HIV-1 included history of multiple sexual partners, history of at least one sexually transmitted disease (STD), relatively high socioeconomic status (SES), being unmarried, young age at first pregnancy, and low gravidity. Women who had used oral contraceptives, smoked more than one cigarette per day, whose partners were circumcised, and had had sex to support themselves were also at higher risk of being infected. A history of blood transfusion in the past 5 years was not associated with HIV-1 infection. History of multiple sexual partners, history of STD, high household income, partner circumcision, and past oral contraceptive use remained strongly associated with HIV-1 infection even when simultaneously controlling for other covariates. Among legally married women who lacked sexual behaviour risk factors, history of STD, high SES, young age at first pregnancy, and low gravidity were significantly associated with HIV-1 seroprevalence.

PMID: 8082965 [PubMed - indexed for MEDLINE]

No correlation in a high-risk population

International AIDS Society

Prevalence of male circumcision and its association with HIV and sexually transmitted infections in a U.S. navy population
A G Thomas, L N Bakhireva, S K Brodine, R A Shaffer
Int Conf AIDS. 2004 Jul 11-16; 15: abstract no. TuPeC4861.

Background: Lack of male circumcision has been found to be a risk factor for HIV and sexually transmitted infection (STI) in several studies performed in developing countries. However, the few studies conducted in developed nations have yielded inconsistent results. Policy regarding circumcision of male infants as a prevention measure against HIV/STI remains a controversial topic. This study describes the prevalence of circumcision and its association with HIV and STI in a U.S. military population.

Methods: This is a case-control study of male HIV infected U.S. military personnel (n= 232) recruited from 7 military medical centers and male U.S. Navy controls (n=516) from a general aircraft carrier population. Cases and controls completed similar self-administered HIV behavioral risk surveys. Case circumcision status was abstracted from medical charts while control status was reported on the survey. Cases and controls were frequency matched on age. Multiple logistic regressions were constructed separately to evaluate the role of circumcision in the acquisition of HIV and STI.

Results: The proportion of circumcised men did not significantly differ between cases (84.9%) and controls (81.8%). Prevalence of circumcision among men born in the U.S. was higher (85.0%) than those born elsewhere (58.1%). After adjustment for demographic and behavioral risk factors lack of circumcision was not found to be a risk factor for HIV (OR = 0.9; 95% CI: 0.51, 1.7) or STI (OR = 1.08; 95% CI 0.52, 2.26). The odds of HIV infection were 2.6 higher for irregular condom users, 5 times as high for those reporting STI, 6.2 times higher for those reporting anal sex, 2.8-3.2 times higher for those with 2-7+ partners, nearly 3 times higher for Blacks, and 3.5 times as high for men who were single or divorced/separated.

Conclusions: Although there may be other medical or cultural reasons for male circumcision, it is not associated with HIV or STI prevention in this U.S. military population.

No protection by traditional circumcision

J Acquir Immune Defic Syndr. 2007 Aug 1;45(4):371-9.
The protective effect of circumcision on HIV incidence in rural low-risk men circumcised predominantly by traditional circumcisers in Kenya: two-year follow-up of the Kericho HIV Cohort Study.
Shaffer DN, Bautista CT, Sateren WB, Sawe FK, Kiplangat SC, Miruka AO, Renzullo PO, Scott PT, Robb ML, Michael NL, Birx DL.
US Army Medical Research Unit, Walter Reed Project HIV Program, Kericho, Kenya. dshaffer@wrp-kch.org

BACKGROUND: Three randomized controlled trials (RCTs) have demonstrated that male circumcision prevents female-to-male HIV transmission in sub-Saharan Africa. Data from prospective cohort studies are helpful in considering generalizability of RCT results to populations with unique epidemiologic/cultural characteristics.

METHODS: Prospective observational cohort sub-analysis. A total of 1378 men were evaluated after 2 years of follow-up. Baseline sociodemographic and behavioral/HIV risk characteristics were compared between 270 uncircumcised and 1108 circumcised men. HIV incidence rates (per 100 person-years) were calculated, and Cox proportional hazards regression analyses estimated hazard rate ratios (HRs).

RESULTS: Of the men included in this study, 80.4% were circumcised; 73.9% were circumcised by traditional circumcisers. Circumcision was associated with tribal affiliation, high school education, fewer marriages, and smaller age difference between spouses (P < 0.05). After 2 years of follow-up, there were 30 HIV incident cases (17 in circumcised and 13 in uncircumcised men). Two-year HIV incidence rates were 0.79 (95% confidence interval [CI]: 0.46 to 1.25) for circumcised men and 2.48 (95% CI: 1.33 to 4.21) for uncircumcised men corresponding to a HR = 0.31 (95% CI: 0.15 to 0.64). In one model controlling for sociodemographic factors, the HR increased and became non-significant (HR = 0.55; 95% CI: 0.20 to 1.49).

CONCLUSIONS: Circumcision by traditional circumcisers offers protection [That's not what "non-significant" means.] from HIV infection in adult men in rural Kenya. Data from well-designed prospective cohort studies in populations with unique cultural characteristics can supplement RCT data in recommending public health policy. PMID: 17558336 [PubMed - indexed for MEDLINE]

No protection to men

Declining Rates in Male Circumcision amidst Increasing Evidence of its Public Health Benefit

Mor Z, Kent CK, Kohn RP, Klausner JD (2007) Benefit. PLoS ONE 2(9):

The study objective was to describe male circumcision trends among men attending the San Francisco municipal STD clinic, and to correlate the findings with HIV, syphilis and sexual orientation.
Methods and Findings. A cross sectional study was performed by reviewing all electronic records of males attending the San Francisco municipal STD clinic between 1996 and 2005. The prevalence of circumcision over time and by subpopulation such as race/ethnicity and sexual orientation were measured. The findings were further correlated with the presence of syphilis and HIV infection. Circumcision status was determined by physical examination and disease status by clinical evaluation with laboratory confirmation.

Among 58,598 male patients, 32,613 (55.7%, 95% Confidence Interval (CI) 55.2–56.1) were circumcised. Male circumcision varied significantly by decade of birth (increasing between 1920 and 1950 and declining overall since the 1960’s), race/ethnicity (Black: 62.2%, 95% CI 61.2–63.2, White: 60.0%, 95% CI 59.46– 60.5, Asian Pacific Islander: 48.2%, 46.9–49.5 95% CI, and Hispanic: 42.2%, 95% CI 41.3–43.1), and sexual orientation (gay/ bisexual: 73.0%, 95% CI 72.6–73.4; heterosexual: 66.0%, 65.5–66.5).

Male circumcision may [or, equally, may not] have been modestly protective against syphilis in HIV-uninfected heterosexual men (PR 0.92, 95% C.I. 0.83–1.02, P = 0.06) . [No correlations were found between circumcision and HIV or syphilis in any of the groups of men studied, but the paper tries its best to make it look as if they were]

From the Results:
Table 2. Percent circumcised in those with and without syphilis infection by HIV status and sexual orientation, as determined during male patient visits, San Francisco municipal STD clinic, 1996-2005.
Sexual orientation
Syphilis infection
HIV-infected
HIV-uninfected

Circumcised %
(n/N)
PR*
(95% CI)
Circumcised %
(n/N)
PR
(95% CI)

Heterosexual
Yes
62.5
(10/16)
0.85
(0.40-1.56)
66.7
(384/576)
0.92
(0.83-1.02)

No
73.8
(1,050/1,423)
Ref.

72.4
(36,290/50,128)
Ref.

Gay/ bisexual
Yes
75.8
(214/282)
1.0
(0.87-1.15)
72.7
(384/528)
0.98
(0.88-1.08)

No
75.4
(15,910/21,090)
Ref.

74.6
(34,210/45,869)
Ref.

*PR = Prevalence ratio of circumcision status by syphilis infection (Yes/No)
Table 2 shows the proportion of visits by circumcised men at the San Francisco municipal STD clinic from 1996 through 2005 by sexual orientation, syphilis and HIV infection status. There was a trend towards a protective effect of circumcision for syphilis infection in heterosexual HIV-uninfected men and in a lesser extent in HIV-infected men. Among gay/bisexual men, no such protective effect was seen and also no association was found between circumcision status and HIV infection (71.1% circumcised versus 72.2%, PR = 0.97, 95% CI 0.90-1.0, P =0.52).

Conclusions. Male circumcision was common among men seeking STD services in San Francisco but has declined substantially in recent decades. Male circumcision rates differed by race/ethnicity and sexual orientation. Given recent studies suggesting the public health benefits of male circumcision, a reconsideration of national male circumcision policy is needed to respond to current trends.

[And therefore Carthage must be destroyed. The conclusion does not follow at all from the data.

"A trend towards a protective effect" is weasel wording for no correlation.

Class		Prevalence of circumcision ratio Syphilis : No Syphilis
Heterosexual	HIV-	0.92:1
Heterosexual	HIV+	0.85:1
Gay/Bisexual	HIV-	0.98:1
Gay/Bisexual	HIV+	1.00:1

However, none of the ratios is statistically significant.

Considering HIV, in every row except the first, the percentage on the right (circumcised men with HIV) is greater than the percentage on the left (intact men with HIV), and in the first row, there are only six intact (heterosexual) men with HIV (and syphilis). Here is a different presentation of the same data:

Class		Prevalence of circumcision ratio HIV+ : HIV-
Heterosexual	Syphilis	0.94:1
Heterosexual	No Syphilis	1.02:1
Gay/Bisexual	Syphilis	1.04:1
Gay/Bisexual	No Syphilis	1.01:1

In all classes except the first, men with HIV are very slightly more likely to be circumcised than men without HIV, but in no class does the difference reach statistical significance. (And in the first class - because only six of the men with HIV were intact - if one more HIV+ man had been circumcised, that ratio would also have been greater than 1:1.)

There are other problems with this paper. According to its Table 1 there were 15,515 intact men, while according to Table 2 intact men paid only 14,409 visits to the clinic.

A published response to the paper

Circumcision as a risk of HIV transmission

The Bagisu people of Eastern Uganda circumcise boys aged 12-18 years. The cultural practices associated with circumcision are a risk to HIV transmission. HIV transmission awareness programmes have been running in the local media but the message is mainly perceived by urban, literate people. The researchers found it is hard to change the attitude of the Bagisu towards their cultural circumcision practices despite the risks.

A. Kataami Moiti. Joint Clinical Research Centre, Kampala, Uganda
The Importance of education in addressing risk factors associated with cultural circumcision practices among Bagisu community, Uganda
Poster at the XV AIDS Conference in Bangkok, July 2004 [ThPeC7544]

A Cochrane Review of HIV-circumcision studies finds:

"Despite the positive results of a number of observational studies, there are not yet sufficient grounds to conclude that male circumcision, as a preventive strategy for HIV infection, does more good than harm."

"Circumcision itself may be a proxy measure of the knowledge and behaviour learnt during initiation, when young men are taught about traditional sexual practices, including monogamy and penile hygiene."

"Selection bias was problematic in all studies, and results were potentially confounded by other risk factors for transmission of HIV such as sexual behaviour and religion. Circumcised and uncircumcised groups (in cohort and cross-sectional studies) and HIV-positive and HIV-negative groups (in case-control studies) were seldom balanced for all or most of the 10 risk factors that we identified as potential confounders prior to quality assessment."

"Age
Sexual behaviour
Location of trial
Religion
Education, occupation, socio-economic status
Sexual behaviour – measured by age at first intercourse, number of sexual partners, contact with sex workers
Any sexually transmitted infections
Condom use
Migration status, travel to different countries
Other possible exposures, e.g. injection, blood transfusions"

"As HIV is related to sexual behaviour, which may in turn be partly determined by culture and religion, strong confounding factors in these studies seem likely."

"It is important to note that observational studies, unlike R[andom] C[ontrolled] T[rial]s, can only adjust for known confounders, and only then if they are measured without error. The effect of unknown confounders may well be operating in either direction within and across all of the included studies."

The Medical Research Council of South Africa has a good summary of it.

Other studies showing no correlation, or a negative correlation between intactness and HIV.

Where circumcision doesn't prevent AIDS

Country	% of men circumcised	% HIV prevalence in
Country	% of men circumcised	Adults	Circumcised men	Uncircumcised men
Burkina Faso	88	1.8	1.8	2.9
Cameroon	93	5.5	4.1	1.1
Cote d'Ivoire	96	4.7	2.8	3.8
Ethiopia	91	1.4	0.9	1.1
Ghana	95	2.2	1.6	1.4
Kenya	83	6.7	3.0	12.6
Lesotho	49	23.5	22.8	15.2
Malawi	20	11.8	13.2	9.5
Rwanda	9	3.0	3.5	2.1
Tanzania	69	7.0	6.5	5.6
Uganda	25	6.4	3.8	5.6
Source: National surveys, available at: www.measuredhs.com/countries/

Between Correlation and Recommendation

"Circumcision status should be viewed as a proxy for other aspects of human behavior. Unless one can control for these aspects, one cannot draw reliable conclusions about the causative status of the presence or absence of a foreskin on the course of medical disease processes."

- Dr Anne Laumann
in a letter to Archives of Dermatology

Several intermediate steps need to be taken, between the association shown in some (not all) studies and recommending general circumcision as a preventative measure. Married men in Africa have a higher rate of HIV infection than single men, but so far no one has called for the abolition of marriage. The proofs of links to circumcision and to marriage are similar, but first

the association needs to be clearly established (in the case of circumcision, some think it has been, but many think it has not). Then
a case for causality needs to be made (it falls short). Then
a cost-utility estimate needs to be done to see if it is feasible, and, if so,
under what conditions. Then
a randomised trial needs to be performed. Finally, if all of those items fall into place, only then can one reasonably
make the call for universal or selective circumcision.

A Vaccine? Hardly!

Australian Doctor
November, 2005

Circumcision equal to a vaccine for HIV

by Rebecca Jenkins

CIRCUMCISION offers the same level of protection against HIV infection in heterosexual men as a highly effective vaccine, according to a landmark study. In the first randomised controlled trial of its kind, researchers found circumcision provided 60% protection against the virus, confirming the results of a large body of observational studies.

A vaccine of high efficacy is expected to offer long-term protection of 95% or above. Smallpox was eradicated with such a highly efficient vaccine. If control of tetanus, measles, and poliomyelitis has been largely achieved in the world, it has been a result of high-efficacy vaccines. ... A 96%-efficient measles vaccine means that 96% of vaccinated persons exposed to measles are indeed protected against infection. Protection lasts for many years, and revaccination permits dealing with loss of immunity over time. What Auvert and colleagues show is ... a 60% reduction in disease incidence over an 18-month period among circumcised men compared with uncircumcised men with similar exposure. To our knowledge, this does not mean that those men are really "protected" against HIV, especially in the case of repeated exposure. It simply means "reduced risk," or reduced probability of contamination.

- Michel Garenne, Male Circumcision and HIV Control in Africa

In a text for upper division and/or graduate study of immunology, a table gives the percentage of reduction obtained by vaccines for the diseases modern societies associate with successful immunization programs.

Smallpox, diphtheria, and polio vaccinations resulted in 100% reduction of incidence. Vaccination against measles, Mumps, and rubella (German measles) resulted in >99% reduction of incidence. Tetanus (lockjaw) was reduced by more than 98%; Pertussis (whooping cough) by more than 87%.

(No vaccines that reduced incidence by as little as 70% were included in the table.)

- Kindt, Thomas J, Goldsby, R.A., and Osborne, B.A.
(Kuby) Immunology (6th Ed), New York: W.H. Freeman, 2007.
[Kindt - NIH, Goldsby - Amherst College, Osborne - UMass, Amherst]

A Solution Looking For A Problem

The question arises, why have so many studies been done apparently looking for this correlation (and prematurely making the recommendation)? For over a hundred years, circumcision has been a solution looking for a problem, and the problem has typically been the most frightening disease (or "disease") of the day -

"masturbation insanity" in the 19th century,
then tuberculosis,
Sexually Transmitted Diseases (then called Venereal Disease or VD) after World War I,
penile cancer in the 1930s, and
cervical cancer in the 1950s, when cancers were terrifyingly untreatable,
Urinary Tract Infections from 1982 onward,
and now HIV.

Today's calls are just the latest in a long series, and no better founded than those.

Ethics

As ethicist Dr Margaret Somerville (Gale professor of law and a professor in the faculty of medicine at the McGill Centre for Medicine, Ethics and Law) says:

"...even assuming that reducing the risk of HIV transmission could be a justification for infant male circumcision, this justification would not be available until it became at least more likely than not that circumcision would reduce the risk of HIV transmission.

"...even assuming that circumcision could help to protect against HIV infection, it would not be necessary to carry it out on unconsenting infants. One could wait until the person was about to become sexually active and could decide for himself.

"...one is ethically required to use the least harmful, least invasive means of achieving a good, the achievement of which involves harm. Consequently, a surgical intervention aimed at preventing the spread of HIV could only be justified if there were no other reasonable way to achieve this. And, even if circumcision helped to protect people in developing countries from the spread of HIV, we would not be justified in carrying this out for this purpose in developed countries, where other, better means of protection are much more readily available."

- The Gazette, Montreal, October 24, 1998, pB6

"Dry Sex"

Meanwhile, an explanation seems to be to hand: "dry sex" - the use by women of herbal and other astringents to dry their vaginas.

Wet, Dry, Man, Woman: Heterosexuals and Anal Sex

formerly at http://hivinsite.ucsf.edu/

Wet/Dry and Tight/Loose

DH: We ... encountered a notion of "dry sex" that appears to be shared in Haiti, the country with which the Dominican Republic shares an island, as well as various parts of Africa. It's complex and it varies from place to place, but the basic idea is that sex should be very tight and should be dry. In the Dominican Republic, I couldn't help but begin to think that maybe that was part of the appeal for anal sex, both bisexual male anal sex and heterosexual anal sex. Particularly if, as you say, women have given birth and so on. A lot of women there and in countries like Brazil will have operations to tighten the vaginal opening. There's actually a surgical procedure in the Dominican Republic that translates as "the cut that makes the husband happy." It's basically a tightening of the vagina after the woman has given birth.
Given this notion that sex should be tight, there's potentially an interaction with the foreskin there, because we seem to mainly find dry sex practices in areas where most men are not circumcised. One explanation may be that circumcised men don't have the lubricative mechanism of the foreskin rolling back and forth across the glans. Presumably, it would be quite painful and uncomfortable for most men to have dry sex if they are circumcised. But uncircumcised men in the Domincan Republic and in parts of Africa commonly report tearing and bleeding of the foreskin during dry sex. ...

"Dry sex practices appear to be primarily restricted to certain predominately non-male[-]circumcising regions of eastern and southern Africa, including many of the countries reporting the world's highest HIV seroprevalence (for example, Zimbabwe, Botswana, Zambia, Malawi). Presumably, such practices would appear to be less appealing to the drier (non-prepucial secreting) circumcised males of western Africa or other regions. Reportedly, very few men in the Dominican Republic or Haiti [where dry sex is also widely practised] have been circumcised . . . ."
Halperin, Daniel T. Dry sex practices and HIV infection in the Dominican Republic and Haiti. Sexually Transmitted Infections 1999; 75:445-446.

Alcohol

One confounding factor that the circumcisionists haven't noticed (because they weren't looking), is alcohol usage. An eight-year study in Uganda has shown a correlation between alcohol consumption and HIV infection (because people who have been drinking are less likely to practise safe sex). Islam prohibits alcohol and also prescribes circumcision. It is at least as reasonable that the prohibition as the prescription protects against HIV.

Circumcisionists are fond of claiming that their statistics have been "adjusted" to correct for this kind of confounding error, but Ted Goertzel argues that such "adjustments" are just an attempt to blind us with science.

Sexual selection

People don't have sex with just anyone, they tend to do so within their own social groups, so HIV stays within social groups. (The clearest case is that in the US, gay men have sex with gay men, heterosexual men with heterosexual women. So once it started with them, HIV would have spread mainly among gay men regardless of other factors.) So in Africa, if HIV first spread in societies where men were intact, it would continue to do so, and not in societies where men were cut.

Female Genital Mutilaton

Abstract: Female circumcision and HIV infection in Tanzania: for better or for worse? Stallings R.Y, Karugendo E. (PowerPoint)

Introduction: ...The authors sought to explain an unanticipated significant crude association of lower HIV risk among circumcised women [R{isk} R{atio}=0.51; 95% C{onfidence} I{nterval} 0.38,0.70] in a recent survey by examining other factors which might confound this crude association.

Methods: Capillary blood was collected ... from a nationally representative sample of women age 15 to 49 during the 2004 Tanzania Health Information Survey. Eighty-four percent of eligible women gave consent for their blood to be anonymously tested for HIV antibody. Interview data was linked ... to final test results for 5753 women. The chi-square test of association was used to examine the bivariate relationships between potential HIV risk factors with both circumcision and HIV status. Restricting further analyses to the 5297 women who had ever had sexual intercourse, logistic regression models were then used to adjust circumcision status for other factors found to be significant.

Results: By self-report, 17.7 percent of women were circumcised. Circumcision status varied significantly by region, household wealth, age, education, years resident, religion, years sexually active, union status, polygamy, number of recent and lifetime sex partners, recent injection or abnormal discharge, use of alcohol and ability to say no to sex. In the final logistic model, circumcision remained highly significant [O{dds} R{atio}=0.60; 95% C{onfidence} I{nterval} 0.41,0.88] while adjusted for region, household wealth, age, lifetime partners, union status, and recent ulcer.

Conclusions: A lowered risk of HIV infection among circumcised women was not attributable to confounding with another risk factor in these data. Anthropological insights on female circumcision as practiced in Tanzania may shed light on this conundrum.

Will there be Randomised Controlled Trials of 3000 HIV-negative women, where 1500 are circumcised and they see how many seroconvert - followed by calls for mass circumcision of women to prevent the spread of HIV? Of course not.

The correlation one way between FGM and MGM is 100%. That is, females are circumcised only if males are. So if FGM reduced the incidence of HIV, it could be mistaken for an effect of MGM.

(Jan 2002: One exception has been found, the Pokot tribe in Kenya - but they used to circumcise males.)

Wife Inheritance

Among the Luo people of Kenya (who do not practise circumcision), when a man dies, his wife is "inherited" by his brother. She is required to have intercouse with him, and that intercourse must be unprotected. Otherwise the husband's spirit is not free, and the wife is not free to remarry. The rate of HIV among people tested in that region was 2/3. One man said it makes no difference if they know the woman is HIV positive. They do not believe AIDS is caused by a virus: "If a man dies, it is because he has done something wrong."
There can be no doubt that wife-inheritance is a potent factor in HIV transmission - especially where the death rate from HIV is high: it's a vicious circle. Wife inheritance is seldom if ever mentioned as a confounding factor in studies of HIV transmission.
If there should be a correlation between intactness and wife-inheritance, or between circumcision and the shunning of wife-inheritance, that might go a long way toward explaining the supposed intactness-HIV link.
A BBC story 18 November 2003.
The Washington Post November 8, 1997
Christianity Today August 28, 2000
"The Luo people are often polygamous, and several widows may be inherited by a single family member. Another element of the tradition is the practice of holding a "cleansing" ritual in which the widow has sex with an outsider before being given to her brother-in-law or other family member."

Heterosexual transmission, Europe vs the United States

A common criticism of "Circumcision prevents HIV" is "But HIV is very common in the US, where circumcision is prevalent." A common reply from the pro-circumcision lobby is that HIV is primarily transmitted homosexually in the US, heterosexually in Africa, and anal receptivity of HIV is unaffected by circumcision. This can be countered by comparing the United States with Europe, where homosexual and heterosexual rates of transmission are comparable, but circumcision rates are very different. The US proves to have a much higher rate of HIV than Europe, and a disproportionate rate of male to female transmission.

Advocates of circumcision then have to put considerable spin on the statistics. For example, Bailey and Halperin write:

Remarkably, there is consistent evidence that female-to-male HIV transmission, compared with male-to-female transmission, is much higher in Europe than in the USA . . . Data from the European Multicenter Partners Study and comparable research from the USA suggest that the ratio of female-to-male transmission (compared with male to female transmission) is about 10 fold higher in Europe.³

[3 De Vincenzi I. Heterosexual transmission of HIV. JAMA 1992; 267: 1919.]

The implication is that intact European men are being infected with HIV at an alarming rate compared to their circumcised counterparts in the US.

This is assisted by the straightforward but false interpretion that the rate of female-to-male transmission is higher in Europe. Bailey and Halperin actually mean the ratio of the ratios of (female-to-male vs male-to-female in) Europe vs (female-to-male vs male-to-female in) the US.

Yet if the four sets of data are compared, standardising the US total to 100, M-to-F amounts to 95, F-to-M to 5, and in Europe, M-to-F 20 and F-to-M 10. So Halperin's extraordinary ratio is (10/20)/(5/95) = (1/2)/(1/19) = 9.5
(In exact figures,(10.10/20.20)/(4.76/95.24)=10.0)
Expressed pictorially:

Clearly, what needs to be explained is not a high female-to-male HIV transmission rate in Europe, but the high male-to-female rate in the US. Could the reason be the rougher action of dry, circumcised US penises, creating micro-tears on US women's vaginal walls? Perhaps not, perhaps it is is the different strains of HIV prevalent in the US and Europe, but this kind of difference between fact and interpretation illustrates that simple correlations do not necessarily translate into simple solutions.

"Russian Roulette with two bullets rather than three"

Male circumcision and HIV infection

For several years, researchers have been debating the relationship between male circumcision and HIV. Several studies have indicated that circumcised men are less likely to become infected with HIV than uncircumcised men. However, because circumcision is usually linked to culture or religion, it has been argued that the apparent protective effect of the procedure is likely to be related not to removal of the foreskin but to the behaviours prevalent in the ethnic or religious groups in which male circumcision is practised. In addition, some researchers have assumed that any association between circumcision and HIV must be complicated by the presence of other sexually transmitted infections, which have been found to be more common among uncircumcised men.

Clearly, the correlations are not straightforward. In the higher income countries, the rates of HIV infection among men who have sex with men do not vary greatly even though the circumcision rates do: few men in Europe and Japan but four-fifths of men in the United States are circumcised. In Africa, however, circumcision seems to confer some protection. A study in Nyanza Province, Kenya, among men from the same ethnic group, the Luo, found that one-quarter of uncircumcised men were infected with HIV, compared with just under one-tenth of circumcised men. The protective effect remained even after other factors, such as sexual behaviour and sexually transmitted infections, had been taken into account. A study of over 6800 men in rural Uganda has suggested that the timing of circumcision is important: HIV infection was found in 16% of men who were circumcised after the age of 21 and in only 7% of those circumcised before puberty. A recent review of 27 published studies on the association between HIV and male circumcision in Africa found that, on average, circumcised men were half as likely to be infected with HIV as uncircumcised men. When African men with similar socio-demographic, behavioural and other factors were compared, circumcised men were nearly 60% less likely than uncircumcised men to be infected with HIV.

Even though the weight of evidence increasingly suggests that circumcising men before they become sexually active does provide some protection against HIV, the practical implications for AIDS prevention are not obvious. Circumcision, where it is practised, usually has links to religious or ethnic identities and life-cycle ceremonies, and may customarily be done after puberty. If the same scalpel were used without sterilization on a number of boys, this could actually contribute to the transmission of HIV. Finally, if circumcision were promoted as a way of preventing HIV infection, people might abandon other safe sexual practices, such as condom use. This risk is far from negligible - already, rumours abound in some communities that circumcision acts as a "natural condom". A sex worker interviewed in the city of Kisumu in Kenya summed up this misconception, saying: "I can sleep with circumcised men without a condom because they don't carry a lot of dirt on their penis". While circumcision may reduce the likelihood of HIV infection, it does not eliminate it. In one study in South Africa, for example, two out of five circumcised men were infected with HIV, compared with three out of five uncircumcised men. Relying on circumcision for protection is, in these circumstances, a bit like playing Russian roulette with two bullets in the gun rather than three. [...assuming the gun has only five chambers - or, if it had the more usual six, 2.4 bullets rather than 3.6.]

- Report on the global HIV/AIDS epidemic
UNAIDS, June 2000
A large file, >275KB.

In the wake of three incomplete Random Controlled Tests of circumcision, the head of UNAIDS, Dr Peter Piot, has chosen to forget these wise words.

A British survey of gay men found slightly more of the circumcised men were HIV-positive.

Know
the
score

Findings from
the National
Gay Men’s
Sex Survey
2001

David Reid
Peter Weatherburn
Ford Hickson
Michael Stephens

...

Introduction and methods

1.1 CONTENT OF THE REPORT

This research report outlines the main findings of Vital Statistics 2001 – which was the fifth annual national Gay Men’s Sex Survey (henceforth GMSS). The survey was carried out during the summer of 2001 by Sigma Research in partnership with 73 health promotion agencies across England and Wales.

...

Chapter 2 gives a brief description of the sample of 14,616 men living in England and Wales who either had sex with another man in the last year or expected to have sex with a man in the future.

...

1.2 BACKGROUND TO THE FIFTH NATIONAL GAY MEN’S SEX SURVEY
The Gay Men’s Sex Survey uses a short self-completion questionnaire to collect a limited amount of information from a substantial number of men. ...

1.3 PRIDE EVENTS: RECRUITMENT DATES, EVENTS AND RETURNS
Recruitment occurred at seven community-based events in the summer of 2001. ...

4.5 CIRCUMCISION
It has been suspected for some time that when uninfected men are insertive in UAI with positive men, whether or not the uninfected man is circumcised has a bearing on the probability of HIV transmission occurring. The hypothesis is that the cells of the fore-skin are more susceptible to infection by HIV and therefore circumcision has a protective function.

Men were asked Are you circumcised? and were asked to tick No, Yes or Don’t Know. Overall, 0.9% said Don’t know by which we think they mean they do not know the word rather than not knowing whether they have a foreskin. Excluding this small group, 22.1% of men indicated that they were circumcised. The proportion rose with increasing age, from 16.1% among the under 20s, through 18.8% (in the 20s), 21.3% (in the 30s), 24.8% (in the 40s) and 40.2% among the over 50s.

Circumcision also significantly varied by ethnicity, being highest among Bangladeshi men (100%, 5/5), Pakistani men (97.5%, 39/40), other Asian men (77.3%, 68/88) and Black African men (76.1%, 35/46). Of all sixteen ethnic groups, White British men had the lowest level of circumcision (18.7%, 2201/11764).

If circumcised men are less likely to acquire HIV than men with foreskins, then we should expect fewer of the circumcised men to have tested positive than the men with a foreskin. However, more of the circumcised men had tested positive for HIV (6.1%) than had those with a foreskin (5.0%). This small but significant difference is in the opposite direction than predicted if foreskins are contributing to transmission, and was observed in all ethnic groups and across the age range.

The survey found no evidence to support the adoption of ‘the proportion of HIV uninfected men who are not circumcised’ as a population level target for HIV prevention programmes for gay and bisexual men. [... let alone evidence to support the promotion of circumcising anyone]

[The only possible confounder remaining is selection bias. The results would not reflect the actual position if circumcised men who have HIV (and know it), or intact men who don't, are more likely to take the survey than intact men who have HIV (and know it) or circumcised men who don't, but it is very hard to see why that might be.

Only a small proportion of these men with HIV would have been infected trans-penilely, compared to the proportion infected tran-anally, so the small surplus of those HIV-positive men who are circumcised should not be taken as suggesting that circumcision makes HIV-infection more likely.]

The hazards of unblinded studies

"Scientists must constantly be on guard against this sort of self-deception [picking and choosing data to agree with the preconception that electromagnetic fields, as from power lines, cause leukaemia]. Unless studies are carefully designed to avoid it, the biases of the epidemiologist have a way of creeping into the results. To minimize the opportunity for bias, scientists rely on double-blind studies. An independent researcher might be given a list including both the homes of victims of childhood leukemia and an equal number of addresses of nonvictim children matched in age, gender, race, family income, etc., but without any indication of which are which. Without knowing which were the homes of victims and which were "controls," the researcher would rate them by whatever criteria were used to estimate the field strength. Someone else would then apply the key after the judgments were made.

[Double-blinding a study involving circumcision is hardly practicable, but much more could have been done to make the circumcised experimental groups and the intact control groups equivalent.]

But even if the study had been double blind, a "risk ratio" of only three for a rare disease such as childhood leukemia would be regarded by many epidemiologists as barely credible. The risk ratio for lung cancer from smoking, for example, is well over thirty^ that is, a 3,000 percent increase in the incidence of lung cancer among smokers. Yet it took years of checking and rechecking the figures, as well as a highly plausible mechanism in terms of known carcinogens in tobacco smoke and, finally, confirming laboratory studies on animals before the cancer link was firmly nailed down."

- "Voodoo Science" by Robert Park, pp 150-1

"The estimated reduction in the relative risk of infection with HIV [between circumcised and intact men in the Kenyan and Ugandan trials] was 51% (unadjusted modified intention-to-treat analysis) to 55% (as-treated analysis)."

Editorial comment in The Lancet.

Related pages:

A call for circumcision by two Australian researchers
Contradictory evidence but never mind, circumcise!
The much-touted random clinical tests in Africa
- and some rebuttals
Rebuttal of a biased TV documentary
Psychology
Ethics

External links:

Doctors Opposing Circumcision has issued a statement critical of HIV-circumcision claims
A site devoted to the HIV-circumcision claims.
A summary from a different point of view
The Onion pokes the borax at the idea.

Back to the Intactivism index page.

The African AIDS Epidemic
J. C. Caldwell and Pat Caldwell
Scientific American, March 1996.

Effect of Circumcision on Incidence of Human Immunodeficiency Virus Type 1 and Other Sexually Transmitted Diseases: A Prospective Cohort Study of Trucking Company Employees in Kenya
Ludo Lavreys, Joel P. Rakwar, Mary Lou Thompson, et al.
The Journal of Infectious Diseases 1999;180:330-336

Risk of HIV-1 in rural Kenya: A comparison of circumcised and uncircumcised men
Agot KE, Ndinya-Achola JO, Kreiss JK, Weiss NS
Epidemiology 2004;15(2):157-63.

The Association between Circumcision Status and Human Immunodeficiency Virus Infection among Homosexual Men

Joan K. Kreiss and Sharon G. Hopkins

The Journal of Infectious Diseases 1993:168:1404-8 (medline abstract)

Male circumcision and HIV infection

Robert C. Bailey and Daniel T. Halperin

Lancet, Volume 355, Number 9207 (11 March, 2000): 926-934 (Reply to correspondence)

Viral load and heterosexual transmission of immunodeficiency virus type 1.

Quinn TC, Wawer MJ, Sewankambo N, et al., for the Rakai Project Study Group.

N Engl J Med, 2000;342:921-9.

The New England Journal of Medicine
August 3, 2000, Vol. 343, No. 5

A Study in Rural Uganda of Heterosexual Transmission of Human Immunodeficiency Virus

To the Editor:

Largely ignored in the report by Quinn et al. of their study of the heterosexual transmission of human immunodeficiency virus type 1 (HIV-1) in Uganda (March 30 issue) (1) is the finding that of 137 uncircumcised men who were negative for HIV-1, 40 seroconverted, whereas 0 of 50 circumcised men seroconverted. This finding suggests that male circumcision is at least as protective against female-to-male transmission of HIV-1 as low viral load in the female partner. Yet the authors do not consider male circumcision among their list of possible strategies for the prevention of HIV-1 infection.

There are now more than 30 epidemiologic studies from sub-Saharan Africa dating back to 1987 that report a significant protective effect of male circumcision against HIV-1 infection. (2) Is it not time for those in Rakai, Uganda (where Quinn et al. conducted their study), as well as others, to benefit from these studies? The feasibility of offering information on voluntary male circumcision and circumcision services to this community with a high prevalence of HIV-1 infection could at least be investigated. Justice and scientific evidence demand it.

Robert C. Bailey, Ph.D., M.P.H.
University of Illinois School of Public Health
Chicago, IL 60302

References

1. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med 2000;342:921-9.

2. Halperin DT, Bailey RC. Male circumcision and HIV infection: 10 years and counting. Lancet 1999;354:1813-5.

[...a letter on another topic here...]

The authors reply:

To the Editor:

Bailey comments on the association between circumcision and reduced rate of acquisition of HIV-1 in male subjects in our study of couples discordant for HIV-1 status. Although circumcision was strongly associated with reduced acquisition of HIV-1 in these highly exposed couples, additional analyses suggest that generalization to the whole population is complicated by confounding. (1) In our representative population in Rakai, we found that circumcision was associated with a reduced rate of HIV-1 acquisition; this was particularly true for circumcision performed before puberty. However, this effect was mainly due to the lower incidence of HIV-1 among Muslims, who constitute the largest group of circumcised males. Circumcision was not significantly protective among non-Muslim men or in couples in which both partners were HIV-1-negative. (1) The 30 African epidemiologic studies mentioned by Bailey are mainly cross-sectional investigations with inconsistent findings and inadequate control for potential confounding. These observational data are difficult to interpret, and clinical trials are needed before circumcision can be promoted as a means of preventing HIV infection.

[...]

Thomas C. Quinn, M.D.
National Institute of Allergy and Infectious Diseases
Bethesda, MD 21205

Maria J. Wawer, M.D.
Columbia University
New York, NY 10032

Nelson K. Sewankambo, M.B., Ch.B.
Makerere University
Kampala, Uganda

References

1. Gray RH, Kiwanuka N, Quinn TC, et al. Male circumcision and HIV acquisition and transmission: cohort studies in Rakai, Uganda. AIDS (in press).

[...]

Susceptibility to Human Immunodeficiency Virus-1 Infection of Human Foreskin and Cervical Tissue Grown in Explant Culture

Bruce K. Patterson, Alan Landay, Joan N. Siegel, Zareefa Flener, Dennis Pessis, Antonio Chaviano, and Robert C. Bailey

American Journal of Pathology, Vol. 161, No. 3, September 2002

Briefly, inner mucosal foreskin tissue samples as well as external foreskin tissue samples and cervical tissue samples were soaked in a concentrated antibiotic wash (20,000 U/ml penicillin/streptomycin, 250 g/ml Fungizone, and 120 U/ml Nystatin) for 10 minutes.

The tissues were then washed three times in Raft media to wash away any remaining antibiotics. A 4.0-mm Acupunch biopsy scalpel (Acuderm, Ft. Lauderdale, FL) was used to provide a number of contiguous samples from each tissue, which were then measured for thickness. Three 4.0-mm biopsies from the inner mucosal surface and three from the outer external surface were cultured and infected in parallel in the same 12-well plate. Tissue biopsies were placed with the epithelial side up on a 3.0-m membrane in the top chamber of a 12-well Transwell (Costar, Cambridge, MA). A 3% solution of agarose (SeaKem Agarose; FMC BioProducts, Rockland, ME) in Hanks' balanced salt solution (Life Technologies, Inc., Grand Island, NY) was added to the area surrounding the tissue in the top well exposing only the epithelium. After 1 day in culture, the foreskin biopsies were infected with either 1000 TCID50 of the CCR5-using (R5) HIV-1Bal or the CXCR4-using (X4) HIV-1Lai. One day after infection, the tissues were harvested and infectivity quantified using real-time quantitative polymerase chain reaction for HIV-1 pol DNA. A qualitative assessment of the cell types infected was performed using simultaneous immunophenotyping for CD4, CD68, and/or CD1a and UFISH for HIV-1 gag-pol mRNA.

The following studies either show no relationship with circumcision staus or a higher risk in circumcised men.

No relationship to circumcision status (16 studies):

1. Hira SK, Kamanga J< Mcuacua R, et al. Genital ulcers and male circumcision as risk factors for acquiring HIV-1 in Zambia. J Infect Dis 1990;161:584-5.

2. Pépin J, Quigley M, Todd J, et al. Association between HIV-2 Infection and genital ulcer diseases among male sexually transmitted disease patients in The Gambia. AIDS 1992;6:489-93.

3. Bollinger RC, Brookmeyer RS, Mehendale SM,l et al. Risk factors and clinical presentation of acute primary HIV infection in India. JAMA 1997; 278:2085-9.

4. Chiasson M, Stoneburner RL, Hildebrandt DS, et al. Heterosexual transmission of HIV-1 associated with use of smokable freebase cocaine (crack). AIDS 1991;5:1121.

5. Carael M, Van De Perre, PH, Lepage PH, et al. Human immunodeficiency virus transmission among heterosexual couples in Africa. AIDS 1988;2:201-5.

6. Moss GB, Clemerson D, D'Costa L, et al. Association of cervical ectopy with heterosexual transmission of human immunodeficency virus: results of a study of couples in Nairobi, Kenya. J Infect Dis 1991;164:588-91.

7. Allen S, Lindan C, Serufilira A, et al. Human immunodeficiency virus infection in urban Rwanda: demographic and behavioral correlate in a representative sample of childbearing women. JAMA 1991; 266:1657-63.

8. Seidlin M, Vogler M, Lee E, et al. Heterosexual transmission of HIV in a cohort of couples in New York City. AIDS 1993;7:1247-54.

9. Konde-Lule JK. Bergley SF, Downing R. Knowledge attitudes and practices concerning AIDS in Ugandans. AIDS 1989;3:513-18.

10. Van de Perre P, Clumeck N, Steens M, et al. Seroepidemiological study on sexully transmitted diseases and hepatitis B in African promiscuous heterosexuals in relation to HTLV-III infection. Eur J Epidemiol 1987;3:14-8.

11. Quigley M, Munguti K, Grosskurth H, et al. Sexual behavior patterns and other risk factors for HIV infection in rural Tanzania: a case control study. AIDS 1997;11:237-48.

12. Urassa M, Todd J, Boerma JT, et al. Male circumcision and susceptibility to HIV infection among men in Tanzania. AIDS 1997;11:73-80.[study 2]

13. Urassa M, Todd J, Boerma JT, et al. Male circumcision and susceptibility to HIV infection among men in Tanzania. AIDS 1997;11:73-80.[study 3]

14. Urassa M, Todd J, Boerma JT, et al. Male circumcision and susceptibility to HIV infection among men in Tanzania. AIDS 1997;11:73-80. [study 5]

15. Hudson CP, Hennis AJM, Kataaha P, et al. Risk factors for the spead of AIDS in rural Africa, hepatitis B and syphilis in southwestern Uganda AIDS 1988; 2: 255-60.

16. Laumann EO, Masi CM, Zuckerman EW. Circumcision in the United States: prevalence, prophylactic effects, and sexual practice. JAMA 1997;277:1052-7.

A higher risk in circumcised men (4 studies):

1. Barongo LR, Borgdorff W, Mosha FF, et al. The epidemiology of HIV-1 infection in rural areas, roadside settlements and rural villages in Mwanza Region, Tanzania. AIDS 1992;6:1521-8.

2. Grosskurth H, Mosha F, Todd J, et al. A community trial of the impact of improved sexually transmitted disease treatment on the HIV epidemic in rural Tanzania: 2. Baseline survey results. AIDS 1995;9:927-34.

3. Chao A, Bulterys M, Musanganire F, et al.Risk factors associated with prevalent HIV-1 infection among pregnant women in Rwanda. National University of Rwanda-Johns Hopkins University AIDS Research Team. Int J Epidemiol 1994;23:371-380.

4. Urassa M, Todd J, Boerra JT, et al. Male circumcision and susceptibility to HIV infection among men in Tanzania. AIDS 1997;11:73-80. [study 1]

http://ije.oupjournals.org/cgi/reprint/dyh127v1.

Int J Epidemiol. 2004 Mar 24 [Epub ahead of print]

Trends in antenatal human immunodeficiency virus prevalence in Western Kenya and Eastern Uganda: evidence of differences in health policies?

Moore DM, Hogg RS.

Department of Health Care and Epidemiology, University of British Columbia, Vancouver, BC, Canada.

OBJECTIVE: To observe recent trends in human immunodeficiency virus (HIV) prevalence in antenatal clinic attendees to determine if previously noted falls in HIV prevalence are occurring on both sides of the Kenyan-Ugandan border. Design An ecologic study was conducted at the district level comparing HIV prevalence rates over time using data available through reports published by the Kenyan and Ugandan Ministries of Health and UNAIDS.

METHODS: Sentinel sites were compared with respect to population, ethnicity, language group, and the prevalence of circumcision practice. The prevalence of HIV found at each sentinel site was recorded for the years 1990-2000 and analysed visually and by conducting bivariate correlations.

RESULTS: Ethnographic analysis revealed a wide mix of ethnic and language groups and circumcision rates on both sides of the border. All sentinel surveillance sites in Uganda showed trends towards decreasing HIV prevalence, with three of five sites showing statistically significant declines (r = -0.87, -0.85, -0.86, P < 0.05). In contrast, all of the surveillance sites in Kenya showed trends toward increasing HIV prevalence, with two of the five sites showing statistically significant increases (r = 0.62, 0.84, P < 0.05).

CONCLUSIONS: The declines in HIV prevalence occurring in Uganda are not being seen in geographically proximal districts of Kenya. No obvious differences in ethnic groupings or their associated prevalence of circumcision appeared to explain these differences. This suggests that decreasing HIV prevalence in Uganda is not due to the natural course of the epidemic but reflects real success in terms of HIV control policies.

PMID: 15044420 [PubMed - as supplied by publisher]

http://www.ias2007.org/pag/Abstracts.aspx?SID=55&AID=2465

Circumcision status and risk of HIV seroconversion in the HIM cohort of homosexual men in Sydney
Presented by David James Templeton, Australia.

Templeton D.J.1, Jin F.1, Prestage G.P.1, Donovan B.1, Imrie J.2, Kippax S.C.2, Kaldor J.M.1, Grulich A.E.1
1National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Darlinghurst, Australia, 2National Centre in HIV Social Research, University of New South Wales, Randwick, Australia

Objectives: Circumcision substantially lowers the risk of HIV acquisition among heterosexual African men, but there are few data addressing circumcision status as a risk factor for HIV among homosexual men. We examined circumcision status as an independent risk factor for HIV seroconversion in the community-based Health in Men (HIM) cohort of homosexual men in Sydney, Australia.
Methods: Between 2001 and 2004, 1,427 initially HIV-negative men were enrolled. Circumcision status was self-reported at baseline and was validated by clinical examination during study visits in a sub-sample of participants. All participants were tested annually for HIV and offered testing for other sexually transmitted infections (STIs). Demographic information was collected at baseline and detailed information on sexual risk behaviours was collected every 6 months. Results: At baseline, 66% of participants reported being circumcised; mostly as infants. There were 49 HIV seroconversions through 2006, an incidence of 0.80 per 100PY. On multivariate analysis controlling for non-concordant unprotected anal intercourse (UAI), anorectal STIs and age, being circumcised was not associated with HIV seroconversion (RR = 0.88, 95% CI 0.45-1.74). Among men who reported no receptive UAI, there were nine seroconversions, an incidence of 0.35 per 100PY. When analyses were restricted to this group, there was also no association with HIV seroconversion (RR = 0.99, 95% CI 0.25-3.96).
Conclusion: Circumcision status was not associated with HIV seroconversion in this cohort. Although statistical power was limited, among men who were more likely to acquire HIV by insertive UAI, there was also no relationship. As most HIV infections in homosexual men occur after receptive anal sex, circumcision is unlikely to be an effective HIV prevention intervention in Australian gay men. However, further research in populations where there is more separation into exclusively receptive or insertive sexual roles by homosexually active men is warranted.