Daniel Halperin, PhD Senior Research Scientist Harvard University School of Public Health Center for Population & Development Studies daniel_halperin@harvard.edu (617) 496-7019; cell (617) 335-5347

Nineteen HIV and reproductive health experts -nearly all of them Africans or people who've worked in Africa for many years -have posted an online response regarding this "new study". The response was posted by Daniel Halperin.

Text follows:



Original Article Size Matters: The Number of Prostitutes and the Global HIV/AIDS Pandemic

Male Circumcision Matters (as One Part of an Integrated HIV Prevention Response)

Posted by DanielHalperin on 27 Jun 2007 at 13:34 GMT

Dear Editors,

The recent paper by John Talbott(1) has attracted considerable attention, largely due to widespread dissemination of his press release announcing that male circumcision is overstated as a prevention tool against HIV-AIDS.(2) We feel compelled to respond to this erroneous conclusion from his paper, which in fact contains no data on male circumcision. The length of this response is due to the fact that a number of us felt it was important to respond to both the issues raised in the paper and some related concerns.

Whilst applauding one implied conclusion of Talbott’s paper -- that providing sex workers with quality prevention services should be an important component of global HIV prevention(3) -- we disagree with other principal conclusions of the paper. Firstly, we believe that the epidemiological correlation between male circumcision and heterosexual HIV transmission is very clear, and especially with three randomized controlled trials now showing similar protective effects,(4-6) this certainly cannot be explained solely by an association with Muslim religion. Secondly, we disagree that the main explanation for the higher levels of HIV in Africa is because vastly greater numbers of African women are “prostitutes.”

Surprisingly, while this paper claims to refute the association (now widely accepted by most scientists and leading international health institutions) between male circumcision and heterosexual HIV infection, it actually contains no data on male circumcision. In the author’s model, he relies on the percentage of Muslims in countries as a proxy for the percentage of men who are circumcised. His approach is flawed because, while almost all Muslims are circumcised, not all non-Muslims are uncircumcised.(7,8)

Although the author’s critique of the circumcision data mainly focuses on a 2004 paper by Drain et al, a more recent ecological study by Drain et al published in 2006,(9) which was not cited, is more relevant. In the 2006 study, a multiple regression analysis was specifically conducted to model the correlation of both male circumcision and religion (Muslim vs. Christian) with HIV prevalence. The study abstract's conclusion reads (emphasis added):

"Male circumcision was significantly associated with lower cervical cancer incidence and lower HIV prevalence in sub-Saharan Africa, independent of Muslim and Christian religion. As predicted, male circumcision was also strongly associated with lower HIV prevalence among countries with primarily heterosexual HIV transmission, but not among countries with primarily homosexual or injection drug use HIV transmission. These findings strengthen the reported biological link between MC and some sexually transmitted infectious diseases, including HIV and cervical cancer."

That 2006 paper also noted:

"Furthermore, the fact that HIV prevalence in many predominantly Christian countries that practice male circumcision, such as the Philippines, Benin, Ghana, Equatorial Guinea, and Gabon, is similarly low as in predominantly Muslim countries in the same regions, suggests that the biological effect of male circumcision may be at least as important as religion in determining HIV prevalence."

Although the HIV pandemic is by far the most severe in southern Africa, accounting for some 2% of the world’s population yet nearly half of all HIV cases globally, this region may have been largely overlooked in Talbott's analysis, which was mainly based on a 2006 study by Vandepitte et al.(10) That study contained data from only one high HIV prevalence, southern Africa country: Zambia. In this region, the only two countries with relatively low HIV prevalence are Angola (under 3%) and Madagascar (under 1%), and these are also the only countries in the region that have high male circumcision prevalence (e.g., the circumcision rate in the most recent Madagascar Demographic and Health Survey was over 98%). In these countries, male circumcision clearly cannot be a marker of being Muslim. In fact, less than 1% of the population in Angola and fewer than 10% in Madagascar are Muslims.(11) (And according to the same Vandepitte et al study that Talbott cites, the highest percentage of "prostitutes" in the entire world was found in Madagascar, reportedly up to 12% of the adult female population.(10)

The situation is similar in southeast Asia, where male circumcision is very common in the Philippines (over 90% circumcision prevalence(12), and HIV prevalence is still extremely low.(13) Similarly, the Philippines is an overwhelmingly Christian (Catholic) country, where prostitution is quite common (just as it is in Angola and Madagascar).

Talbott has also questioned why, in our earlier (2004) study, we "weighted" our ecological analysis by population size. Our decision to weight countries by population size was due to several factors, but mainly from an assumption that a small country like Kosovo probably should not carry the same epidemiological weight as a huge one like China in an ecological study. However, we agree that this matter is open to debate, and it is true that some ecological studies carry out this type of analysis differently.

The main point is that all ecological studies, including the one by Talbott, represent a lower level of epidemiological proof, and randomized controlled trials remain the gold standard for evidence-based medicine. That said, we decided to conduct and publish the 2004 study, and our subsequent ecological analysis (the 2006 paper), because with some things -- and male circumcision represents an intriguing example -- the ecological evidence provides a kind of natural experiment for how such a factor plays out over time in the real world. That decision has now been supported by the unprecedented findings (for prevention of sexual acquisition of HIV) from three randomized clinical trials in Africa as well as several rigorous biological studies, and some 40 other published epidemiological studies showing how male circumcision has impacted on many different populations and countries (including those without predominantly Muslim populations).(14-19) Incidentally, all three of the clinical trials were terminated early by their ethics boards due to a strong, highly statistically significant effect.(20)

In fact, it is this type of rigorous data, particularly from the randomized controlled trials, which has convinced many international organizations, including the World Health Organization and other United Nations bodies, such as UNAIDS, of the protective effect of safe adult male circumcision for heterosexual HIV transmission.(21) After much internal and external debate, they have concluded that an intervention which has been proven beyond any reasonable doubt to be about 60% effective at preventing HIV infection in men (and which would also over time, even if only indirectly, have a considerable impact on infection rates in women(22) should be made available to those people who seek out the service. Of course, normative practices of respect for human rights, confidentiality and informed consent must be followed, as they must be for any surgical procedure or other public health service.

The data behind circumcision's efficacy is about as convincing evidence as one gets in public health. One wonders if we were talking about a partially protective vaccine or microbicide instead, would there even be this kind of "debate" regarding the health of the populations of Africa?

Finally, the paper claims that the reason that HIV prevalence is so much higher in Africa is that prostitution is much more common than elsewhere in the world, yet no convincing data was provided to support this claim, nor was "prostitution" or its measure clearly defined. The notion that there are vastly more prostitutes in Africa than elsewhere is simply not supported by the epidemiological and other relevant literature. Much of the confusion probably stems from a misunderstanding of what "prostitution" means. Many longer-term, regular relationships in Africa involve an important "transactional" element (exchange of gifts, etc.), which is often construed as "prostitution," although it is fundamentally different.(23,24) Thus, while in many places -- most famously Thailand -- effective "100% condom" programs have been successfully implemented in the context of brothels, such an approach would by definition be very different -- and extremely more difficult to implement --among people in Africa who consider themselves "lovers", even if there is a transactional element to their romantic relationships.

With regard to regional differences, according to the Vandepitte et al paper cited by Talbott, the highest percentage of sex workers in the female population in Africa was 4.3% (in Burkina Faso, a country with only 1.5% adult HIV prevalence), while in Latin America (where HIV prevalence is usually even lower) up to 7.4% of women are "prostitutes," according to the same study. Meanwhile, in the mining center of Ndola, Zambia, where HIV prevalence has been upwards of 30%, just 2.4% of women were considered "prostitutes," about similar to the percentage of prostitutes in several other -- and vastly lower HIV prevalence -- developing countries included in the Vandepitte et al study, e.g. Cambodia.(10)

In fact, a more useful epidemiologic indicator than the percentage of the female population who are "prostitutes," would be the percentage of men who report paying for sex. In the Demographic and Health Surveys conducted in numerous developing nations, many more men report visiting sex workers or paying for sex in various Asian and Latin American countries, than do men in the high HIV prevalence countries of southern Africa. Talbott or others may cite some anecdotal information to the contrary, but this is what most surveys (and also the Vandepitte et al study upon which his paper is mainly based) have consistently shown. For example, in the 2002 Behavioral Surveillance Surveys conducted by Family Health International in Swaziland and Lesotho, the world's highest and 3rd highest HIV prevalence countries, respectively, among "high risk" men (soldiers, miners, truck drivers, policemen, etc.) less than 3% of them reported paying for sex in the previous year, while up to 80% of these (mostly married) men reported having had "casual sex" in the past year.(25)

Furthermore, the widely cited "4-city" study conducted by UNAIDS et al also examined the prevalence of women engaging in sex work as one of the many possible variables that could help explain the long-standing heterogeneity of HIV across different regions of Africa, yet this factor did not prove to be predictive.(26) Male circumcision, however, was found to be the strongest predictor of HIV in that seminal study. There is increasing evidence that it is a particular combination, or “lethal cocktail,” of risky sexual behavior – i.e., a pattern of multiple concurrent partnerships among both genders, as is common in much of Africa, including southern Africa – along with low male circumcision prevalence, which largely explains the worst HIV epidemics in Africa.(27)

Perhaps the resistance to the matter of voluntary adult male circumcision for HIV prevention is not surprising. This lingering skepticism may be due, at least in part, to the fact that most international HIV and public health organizations declined to disseminate much information on the topic until fairly recently, when the findings of multiple randomized trials convincingly proved that circumcision directly reduces the risk of heterosexual HIV infection. Thus, even though WHO, UNAIDS, and others have now fully endorsed safe adult male circumcision for HIV prevention, various postings on the internet and even this peer-reviewed study continue to question the large body of existing evidence proving a significant protective effect of circumcision. Since many people may have only recently begun to hear of the link between circumcision and HIV (and other health problems(28), it is understandable that some are initially skeptical that male circumcision could have value for public health. Clearly the idea that a minor surgery could help prevent an infectious disease will take time to be widely accepted. As is the case with many other prevention methods (including those currently being developed and tested, such as pre-exposure prophylaxis with ARVs, microbicides, and vaccines), male circumcision is not a perfect intervention, especially given that it only provides partial protection, and so there would be the potential for behavioral “risk compensation” with any of these approaches.(29)

In conclusion, we commend Talbott for the reminder that addressing the sex work domain -- especially in the more concentrated epidemics outside of southern and eastern Africa, where in fact sex work accounts for a proportionally much larger share of total HIV transmission(25,30) -- must continue to be a focus of international HIV prevention efforts, along with other important approaches such as partner reduction and making safe male circumcision/male reproductive health services more available.(3,31) However, for prevention of HIV infection among high risk groups like sex workers, we would argue that a more effective public health approach would be the kinds of condom promotion and other risk reduction and risk avoidance strategies that have been successfully employed in places like Thailand and Abidjan,(32-34) rather than some of the more pejorative language(35 )and strategies that Talbott advocates, such as mandatory testing and treatment and legal sanctions against sex workers.

Daniel Halperin, PhD, MS (Harvard University), Helen Weiss, PhD (London School of Hygiene and Tropical Medicine), Paul Drain, MD, MPH (Stanford University), James Hughes, PhD (University of Washington), Bertran Auvert, PhD (University of Paris), Saifuddin Ahmed, PhD (Johns Hopkins University), David Serwadda, MD (Makerere University, Uganda), Jesse Kagimba, MD (Uganda), Kawango Agot, PhD, MPH (University of Nairobi), Emmanuel Oladipo Otolorin, MD (JHPIEGO/Johns Hopkins University, Nigeria), Helen Epstein, PhD (Princeton University), Godfrey Woelk, MD, MPH (University of Zimbabwe Medical School), Antonio de Moya, PhD (Federal University of the Dominican Republic), Quarraisha Abdool Karim, PhD (University of Natal, South Africa), Neil Martinson, MD (University of the Witwatersrand, South Africa), John Bongaarts, PhD (Population Council), Sharif Sawires, MA (University of California, Los Angeles), Dean Peacock, MA (Gender Justice, South Africa), Malcolm Potts, MD (University of California, Berkeley)

1. Talbott J. Size matters: the number of prostitutes and the global HIV/AIDS pandemic. PloS One 2007 (http://www.plosone.org/article/fetchArticle.action?articleURI=info:doi/10.1371/journal.pone.0000543).
2. Than K. Debate flares over whether circumcision curbs HIV. Live Science, June 26, 2007
   (http://www.livescience.com/health/070626_prostitute_hiv.html).
3. Halperin DT, Steiner M, Cassell M, Green EC, Hearst N, Kirby D, Gayle H, Cates W [149 signers in total]. The time has come for common ground on preventing
   sexual transmission of HIV. Lancet 2004; 364: 1913-1915
   (www.kaisernetwork.org/daily_reports/rep_index.cfm?hint=1&DR_ID=26931).
4. Bailey RC, Moses S, Parker CB, et al. Male circumcision for HIV prevention in young men in Kisumu, Kenya: a randomised controlled trial. Lancet 2007; 369: 643–56.
5. Gray RH, Kigozi G, Serwadda D, et al. Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet 2007; 369: 657–66.
6. Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, Puren A. Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial. PLoS Med 2005; 2: e298.
7. WHO/UNAIDS. Male circumcision: global trends and determinants of prevalence, safety and acceptability, 2007 (see http://www.unaids.org/en/MediaCentre/PressMaterials/FeatureStory/20070226_MC_pt1.asp).
8. Halperin DT, Weiss H, Hayes R, et al. Comments on male circumcision and HIV acquisition and transmission in Rakai, Uganda. AIDS 2002; 16 :810-12.
9. Drain PK, Halperin DT, Hughes JP, Klausner J, Bailey RC. Male circumcision, religion, and infectious diseases: an ecologic analysis of 118 developing countries. Bio-Med-Central Infect Dis. 2006; 6: 172 (http://www.biomedcentral.com/1471-2334/6/172),
10. Vandepitte J, Lyerla R, Dallabetta G, Crabbé F, Alary M, Buvé A. Estimates of the number of female sex workers in different regions of the world. Sex Transm Infect 2006; 82: Suppl III: iii18–iii25.
11. International Religious Freedom Report, 2004 (http://www.state.gov/g/drl/rls/irf/2004/c12778.htm).
12. Castellsague X, Bosch FX, Munoz N, et al. Male circumcision, penile human papillomavirus infection, and cervical cancer in female partners. N Engl J Med 2002;346:1105-1112 (http://content.nejm.org/cgi/content/short/346/15/1105?query=TOC).
13. UNAIDS. Geneva: Report on the Global AIDS Epidemic 2006.
14. Bailey RC, Plummer FA, Moses S. Male circumcision and HIV prevention: current knowledge and future research directions. Lancet Infect Dis 2001; 1: 223-231 (http://www.ingentaconnect.com/content/els/14733099/2001/00000001/00000004/art00117).
15. Weiss HA, Quigley MA, Hayes RJ. Male circumcision and risk of HIV infection in sub-Saharan Africa: a systematic review and meta-analysis. AIDS 2000; 14: 2361-2370.
16. Halperin DT. Male circumcision: A potentially important new addition to HIV prevention. Contact (‘HIV Prevention: Current Issues and New Technologies’) 2006; 82: 32-36, World Council of Churches (http://www.wcccoe.org/wcc/news/con-182.pdf).
17. Donoval BA, Landay AL, Moses S, et al. HIV-1 target cells in foreskins of African men with varying histories of sexually transmitted infections. Am J Clin Pathol 2006; 125: 386-391.
18. 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: 921-929.
19. Westercamp N, Bailey RC. Acceptability of male circumcision for prevention of HIV/AIDS in sub-Saharan Africa: a review. AIDS and Behav 2006 (DOI:10.1007/s10461-006-9169-4).
20. McNeil D. Circumcision’s anti-AIDS effect found greater than first thought. New York Times, February 23, 2007 (http://www.nytimes.com/2007/02/23/science/23hiv.html).
21. McNeil D. W.H.O. urges circumcision to reduce spread of AIDS. New York Times, March 29, 2007 (http://www.nytimes.com/2007/03/29/health/29hiv.html?ei=5070&en=
    4f962a3432644116&ex=1176091200&adxnnl=1&adxnnlx=11759938).
22. Williams BG, Lloyd-Smith J O, Gouws E, et al. The potential impact of male circumcision on HIV in sub-Saharan Africa. PLoS Med 2006; 3: e262 (http://medicine.plosjournals.org/perlserv/?request=get-document&doi=10.1371/journal.pmed.0030262).
23. Leclerc-Madlala S. Transactional sex and the pursuit of modernity. Social Dynamics 2003; 29: 1-21.
24. Epstein H. The fidelity fix. New York Times Magazine 2004; 13 June: 54-59.
25. Halperin DT. Evidence-based behavior change HIV prevention approaches for Sub-Saharan Africa. Presentation at Harvard Medical School, 17 January 2007 (http://www.globalhealth.harvard.edu/HUPASeminar_Halperin.html; see slides #26-27).
26. Ferry B, Caraël M, Buvé A, Auvert B, Laourou M, Kanhonou L, et al. Comparison of key parameters of sexual behavior in four African urban populations with different levels of HIV infection. AIDS 2001; 15: suppl S41–S50
    (http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=Retrieve&db=PubMed&list_uids=
    11686462&dopt=AbstractPlus&holding=f1000%2Cf1000m%2Cisrctn).
27. Halperin DT, Epstein H. Why is HIV prevalence so severe in southern Africa? The role of multiple concurrent partnerships and lack of male circumcision. Southern African Journal of HIV Med 2007; 26: 19-25.
28. Bailis SA, Halperin DT. Male circumcision: time to re-examine the evidence. student British Medical Journal 2006; 14:179-80. (http://www.studentbmj.com/issues/06/05/editorials/179.php).
29. Cassell MM, Halperin DT, Shelton JD, Stanton D. Risk compensation: The Achilles’ heel of emerging innovations in HIV prevention? British Medical Journal 2006; 332: 605-07.
30. Cote AM, Sobela F, Dzokoto A, et al. Transactional sex is the driving force in the dynamics of HIV in Accra, Ghana. AIDS 2004; 18: 917–25.
31. Sawires SR, Dworkin SL, Fiamma A, Peacock D, Szekeres G, Coates TJ. Male circumcision and HIV/AIDS: challenges and opportunities. Lancet 2007; 369: 708–13.
32. Hearst N, Chen S. Condom promotion for AIDS prevention in the developing world: is it working? Stud Fam Plann 2004; 35: 39-47.
33. Jana S, Bandyopadhyay N, Mukherjee S, Dutta N, Basu I, Saha A. STD/HIV intervention with sex workers in West Bengal, India. AIDS 1998; 12 (suppl B): S101-108.
34. Ghys PD, Diallo MO, Ettiegne-Traore V, et al. Increase in condom use and decline in HIV and sexually transmitted diseases among female sex workers in Abidjan, Cote d'Ivoire, 1991-1998. AIDS 2002; 16: 251-258.
35. Talbott J. Africans Against Aids, Inc. website (http://www.africansagainstaids.org/; for example, see ppt. presentation, slides 37-38).

Author John Talbott Responds to Halperin
johntalbs replied to DanielHalperin on 01 Jul 2007 at 02:03 GMT

I wrote this paper because I uncovered a fairly strong correlation between both the number of prostitutes and their infection rates and HIV prevalence in a country. This correlation survived even after intoducing many other variables to the analysis that had been found to be important.

It also was true that a country's percentage of Muslims became insignificant in predicting HIV levels by country once the number of prostitutes was introducd to the study. This Muslim percentage is a very close proxy to circumcision percentages and has been used as such in many of the pro-circumcision papers to date. The paper's evidence was suggesting that it was not circumcision that was an important correlate with AIDS across countries of the world, but the degree of prostituion.

As a side note, I also discovered that two of the most cited papers supporting circumcision as an effective deterrent to AIDS had made a very serious error in their methodology. (1)(2) They, for some unknown reason, had weighted their regression results by the population of each country in their studies. Such an obvious error makes the reported results of the two most cited papers in support of male circumcision not only suspect, but worthless as scientific evidence.

Upon publication of this paper I now find that circumcision supporters are not happy that I uncovered new scientific evidence in support of a possible cause of the AIDS pandemic and are defintely not happy that I exposed their most cited research as fatally flawed, they insist that I review all the papers in the field in support of circumcision. While outside the scope of my paper, let me take a minute and do that.

Halperin refers to his 2006 paper in which he creates a rough circumcision index and allows it to compete alongside a Muslim variable in a regression explaining HIV across countries. Here he has mae another fundamental error in methodology. Not only does this paper suffer the same omitted variable bias of his prior work by not including some measure of prostitution, his inclusion of circumcision alongside Muslim religion is a serious error in that the terms are very highly correlated (p<.001) (3). Again, a fundamental error of this magnitude makes any results he reports not only suspect but fairly worthless.

Finally, the pro-circumcision cabal is upset that I did not address the three recent randomized clinical trials done in Africa. (4)(5)(6). While outside the scope of my paper I would like to comment on them here because the opposition is using them to try to argue that somehow their flawed cross country regression papers got to the right conclusion even though they were frought with statistical methodology errors.

First, it is impossible to run double blind trials involving major surgery like circumcision and no one can predict what the behavior modifications might be of either the circumcised group or the control group. All volunteers wanted to eventually be circumcized, and without knowing their reasons, one can only guess at how they might modify their behavior either after, or in anticipation of the surgery.

Second, even if you believe the reported 50% to 60% declines in incidence rates from these studies, they are far from being of such a magnitude to be able to help in the fight against the AIDS pandemic in Africa. An average .8% delta in annual incidence rates in these studies suggests that if implemented in a mixed population of men and women such a 100% circumcision approach in men would result in a .4% decline in incidence in the combined male and female population. Optimistically, if 25% of all men in a country get circumcised this translates into a .1% reduction in incidence. So, if the pro-circumcision crowd and their flawed research is to be believed and implemented broadly, we might see incidence rates drop from say, 2.5% to 2.4%, almost immaterial and surely not large enough to suggest circumcision as the solution the AIDS pandemic.

1. Drain, PK, Smith JS, Hughes JP, Halperin DT and Holmes, KK (2004) Correlates of National HIV Seroprevalence. An Ecologic Analysis of 122 Developing Countries. J Acquir Immune Defic Syndr April 1, 35,4.
2. Halperin DT, Bailey RC (1999) Male circumcision and HIV infection: 10 years
   and counting. Lancet 354:1813–1815.
3. Drain PK, Halperin DT, Hughes JP, Klausner J, Bailey RC. Male circumcision, religion, and infectious diseases: an ecologic analysis of 118 developing countries. Bio-Med-Central Infect Dis. 2006; 6: 172 (http://www.biomedcentral.com/1471-2334/6/172),
4. Bailey RC, Moses S, Parker CB, et al. Male circumcision for HIV prevention in young men in Kisumu, Kenya: a randomised controlled trial. Lancet 2007; 369: 643–56.
5. Gray RH, Kigozi G, Serwadda D, et al. Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet 2007; 369: 657–66.
6. Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, Puren A. Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial. PLoS Med 2005; 2: e298.