In this paper, French demographer and reproductive health expert Michel Garenne shows why circumcision as an HIV preventive is unlikely to be as effective in real world situations as it appears to be in the artificial conditions of those famous clinical trials.
Abstract
This paper reviews the demographic evidence for the relationship between male circumcision and HIV infection in national or sub-national African populations. A meta-analysis based on 18 countries, representing more than half of the population of sub-Saharan Africa, shows no relationship [standardized odds ratio=1.00; 95% CI: 0.96–1.05]. There were even more countries in which HIV prevalence was higher among circumcised persons than countries where it was lower. In only five countries, the odds ratio of HIV prevalence (circumcised/intact) was significantly different from 1.0; three countries where it was higher, and two countries where it was lower. The contrast between lack of demographic impact and results from clinical trial is striking, and can probably be explained by the low clinical efficacy in situations of intense and repeated exposure, and by the interactions with the many other determinants of HIV spread. This paper also addresses some ethical and political issues, and in particular raises the question of power abuse, which may lie in the practice of genital mutilations and relevant international recommendations.
Keywords: Male circumcision; HIV/AIDS; Clinical efficacy; Population effectiveness; Randomized controlled trials; Ecological studies; Sub-Saharan Africa
In the study of the impact of public health interventions, two types of evidence can be opposed: the measure of “clinical efficacy,” or “biological effect,” shown in general by clinical trials or epidemiologic studies, and the measure of “population effectiveness,” or “demographic impact,” shown by their effect in large populations or by comparison between populations. Clinical trials aim primarily at demonstrating a biological effect, but they do not guarantee a large demographic impact. Conversely, a demographic impact is sometimes observed even when clinical trials tend to indicate a low effect. When clinical efficacy is very high (say above 95%), the population impact is usually marked, and close to what can be expected from clinical trials. When clinical efficacy is moderate (say around 50%), the population impact is often small, when not negligible. In rare studies, one also finds cases of population impact despite low clinical efficacy.
Let us give a few examples in the field of vaccines. The measles vaccine has a very high clinical efficacy, considered to be above 95%. This has been shown repeatedly in numerous randomized clinical trials all over the world, and similar values are found in case control studies (Redd et al., 1999). The measles vaccine is also an “efficient” vaccine: when administered on a large scale, not only does it protect those who received it, but it also tends to reduce the transmission, and even to stop epidemics through herd immunity. Measles was virtually eradicated from the United States in the early 1980s by mass vaccination campaigns, and came back only because it was re-introduced from foreign countries (Wood and Brunell, 1995). The cholera vaccine has an efficacy of about 50% in clinical trials, but has virtually no demographic impact: it does not stop epidemics, and its protection is short lasting (Tacket and Sack, 2008). This is why it is not recommended in public health programs. The pertussis vaccine (whooping cough) is a vaccine that has a very low efficacy in clinical trials, and has been shown to provide no protection against infection by the germ (Bordetella pertussis). However, it has a very large population effect, stopping epidemics and protecting individuals against the severe forms of the diseases, and its population effectiveness is close to that of the measles vaccine (Edwards and Decker, 2008; Pollard, 1980). This is why it is used in most vaccination programs, and it is part of the Expanded Program on Immunization recommended by the World Health Organization.
Similar comparisons can be made in the field of contraception, an issue more closely related to the control of sexually transmitted infections (STIs). The hormonal contraceptives (pill, injectables, implants) or the barrier methods, such as the Intra-Uterine-Device (IUD), have a very high efficacy in clinical trials, usually above 99%. They also have very high population effectiveness: women who use them properly have no unwanted pregnancy, and populations who use them on a large scale have a low fertility, close to that desired by couples. On the contrary, the Rhythm Method (Knauss-Ogino Method), which is based on very sound biological evidence and has a moderate efficacy in clinical trials (around 50%), has virtually no population effect nor any individual effect: women who use it tend to become pregnant sooner or later, and no country was able to control its fertility only with this method (Labbok and Queenan, 1989). Its only visible effect is a lengthening of birth intervals, which does not permit to bring fertility from 7 or 8 children to 2 children, as do very effective methods. The withdrawal method (coitus interruptus) will never be investigated in clinical trials since it involves primarily a very personal behavior, but was shown empirically to have a major population impact since this was the most common method to control fertility in Europe before 1950. The case of efficacy and effectiveness of male circumcision for controlling HIV transmission bears some similarity with that of the rhythm method to control fertility.
In a series of recent clinical trials, male circumcision was found to have an average clinical efficacy against HIV transmission (Auvert et al., 2005; Bailey et al., 2007; Gray et al., 2007). Is this enough to make recommendations for general use? Beyond clinical efficacy and demographic impact, are there not other issues related with such a practice? In this paper, we will focus on a lack of evidence showing a demographic impact of male circumcision on the HIV epidemics in sub-Saharan Africa, the continent most hard hit by the disease. We also address briefly some of the ethical issues raised by the recommendation for male circumcision (WHO/UNAIDS, 2007), viewed from an international health perspective.
In order to investigate the population impact of male circumcision, we will use several comparisons:
There are serious caveats involved in these comparisons. In Africa, sub-groups practicing—or not practicing—male circumcision are usually defined by ethnicity or religion, sometimes associated with social status or urban residence, and may not be comparable in terms of exposure, since they have different value systems, different marriage patterns, and different sexual behaviors. The case is even worse for country comparisons, where many other factors could bias the comparisons, and we will see some examples later.
Nevertheless, when a public health intervention is very efficient, its demographic impact is largely independent of any social variable, such as social status, religion, or ethnicity. There are many examples in the literature, and they apply similarly to the fields of vaccination, contraception, or medical treatments. The effect of vaccines, hormonal contraceptives, antibiotics, anti-malarial drugs, etc., is basically the same in all countries in the world, and their population effect is simply proportional to the population coverage. Therefore, when there is no visible demographic impact, one can seriously question the usefulness of a public health intervention.
HIV Prevalence and Proportion Circumcised at National Level
The first comparison deals with HIV prevalence between circumcised and intact men in generalized epidemics, some 20–25 years after the onset of the HIV epidemic, that is the situation in years 2002–2007. It measures basically a net effect of circumcision, before changing behavior (condom use, reducing number of partners) really took off and changed the ecological correlations. The data are drawn from the Demographic and Health Surveys (DHS), which are large scale surveys based on representative samples of African populations [all available on the DHS website], and on a similar survey conducted in South Africa, not part of the DHS program (South Africa, 2002). For a statistician, this is the best scenario for testing a potential demographic impact, since the surveys display at the same time the HIV serologic status and the circumcision status for men aged 15–49 or 15–59 years. A first analysis was done a few years ago, and published in 2008 in the African Journal of AIDS Research (Garenne, 2006, 2008). The data were updated with more recent surveys (Congo, Zambia, South Africa), so that 18 countries’ surveys are now included, covering about 55% of the population of sub-Saharan Africa.
Table 4.1 HIV seroprevalence by circumcision status in African countries
| Percent | Percent HIV+ | Ratio circumcised/intact | |||
|---|---|---|---|---|---|
| Country | Circumcised | Circumcised | Intact | RR | Significance |
| Burkina Faso | 89.7 | 1.8 | 2.9 | 0.62 | NS |
| Cameroon | 91.8 | 4.1 | 1.1 | 3.73 | ∗ |
| Côte d’Ivoire | 96.0 | 2.8 | 3.8 | 0.74 | NS |
| Congo Kinshasa | 97.7 | 1.0 | 0.0 | 1.00 | NS |
| Ethiopia | 92.3 | 0.9 | 1.1 | 0.82 | NS |
| Ghana | 95.3 | 1.6 | 1.4 | 1.14 | NS |
| Kenya | 83.4 | 3.0 | 12.6 | 0.24 | ∗ |
| Lesotho | 48.6 | 22.8 | 15.2 | 1.50 | ∗ |
| Liberia | 97.8 | 1.1 | 0.0 | 1.00 | NS |
| Malawi | 20.7 | 13.2 | 9.5 | 1.39 | ∗ |
| Niger | 99.5 | 1.0 | 0.0 | 1.00 | NS |
| Rwanda | 11.1 | 3.5 | 2.1 | 1.67 | NS |
| South Africa | 35.3 | 12.3 | 12.0 | 1.03 | NS |
| Swaziland | 8.2 | 21.8 | 19.5 | 1.12 | NS |
| Tanzania | 69.7 | 6.5 | 5.6 | 1.16 | NS |
| Uganda | 24.9 | 3.8 | 5.6 | 0.68 | ∗ |
| Zambia | 12.5 | 10.8 | 12.5 | 0.86 | NS |
| Zimbabwe | 10.5 | 16.6 | 14.2 | 1.17 | NS |
| Meta-analysis | 1.00 | NS | |||
Results show no effect of male circumcision in national African populations: the standardized odds-ratio was 1.00 (95% CI: 0.96–1.05), which means that, on the average, circumcised and intact men have the same HIV prevalence (Table 4.1). Out of the 18 countries studied, 12 have an odds ratio equal or higher than one (more HIV among circumcised men), 3 of them statistically significant (Cameroon, Lesotho, Malawi), and 6 have an odds ratio lower than one (less HIV among circumcised men), 2 of them statistically significant (Kenya and Uganda). Note that these results are based on large numbers: some 73,800 men sampled, who are representative of the general population. This is the most reliable evidence that we have on the population impact of male circumcision in Africa.
Case Studies
Let us give a closer look at a few selected countries, for which, contrary to expectations, the risk ratio of HIV prevalence (circumcised/intact) is higher than 1. In Tanzania, HIV prevalence is moderate, with some 6.3% of men infected. The epidemic has been going on for about 25 years, since the western part of the country is close to the epicenter of the epidemic, located around Lake Victoria. There are some 110 ethnic groups recorded in Tanzania, some 70% practicing male circumcision. This is a quasi-experimental situation. In Tanzania, HIV prevalence is higher among the circumcised groups (6.5% versus 5.6%). This is due in part to a correlation with urbanization: urban areas are at the same time more circumcised and have more HIV. But even if one controls for urbanization, the HIV prevalence is the same in the two groups: 9.7 and 9.5% in urban areas; 5.2 and 4.6% in rural areas, none of these differences being significant (Tanzania, 2005).
Lesotho is a tiny country embedded in South Africa. The main feature of Lesotho is its ethnic homogeneity, all people belonging to the same group: the Southern Sotho. The country has been exposed for about 15 years to HIV, and prevalence is very high, with 19.3% of adult men infected. About half of the men are circumcised, which is again a quasi-experimental situation, almost ideal given the homogeneity of the ethnic composition. Here again, the profile of HIV prevalence is contrary to expectations: circumcised men are more infected by HIV: 22.8% versus 15.2% for intact men, and this is true in both urban (28.6% versus 17.3%) and rural areas (21.8% versus 14.5%), in the various ecological zones, and for various measures of social status (Lesotho, 2005).
Malawi is a country located in South-Eastern Africa. Malawi has a rather high prevalence, with about 10% men infected, and the epidemic has been going on for about 20 years. Malawi is characterized by a strong dichotomy between North and South, and the dozen of ethnic groups recorded have major differences in demographic profiles and sexual behavior. The North is less circumcised and has less HIV, whereas the South is more circumcised and has more HIV, again contrary to expectations. As was the case in other countries, controlling for urbanization does not change the main picture: more HIV among the circumcised men. Ironically, when stratified by region, the relationship between circumcision and HIV prevalence is inverted: the more circumcised in a region, the higher is the seroprevalence (Malawi, 2005).
A recent publication, based on a 2002 survey, also showed a similar pattern in South Africa, where about a third of the population is circumcised and HIV prevalence amongst the highest on record: there was no difference in HIV prevalence in 2002 between the two groups, even after controlling for a variety of factors (Connolly et al., 2008).
HIV Incidence, and Dynamics of the HIV Epidemic in South Africa
Some authors have argued that male circumcision could change the dynamics of the epidemic by strongly reducing incidence and the net reproduction rate (Ro). This is not the case in South Africa, the only country where one can pursue this type of analysis, because routine HIV prevalence is recorded every year and published over a long period of time. South Africa also has a useful feature for analysis: about one third of men are circumcised; circumcision is primarily ethnic specific, and provinces are also largely ethnic specific. So, by comparing the dynamics of the HIV epidemic by province, one may infer the effect (or lack thereof) of circumcision in the general population. The nine provinces were classified into three groups: low, medium and high level of circumcision. Two indicators of the dynamics of the epidemic were computed: the average incidence between 1994 and 2004, and an estimate of the net reproduction rate of the epidemic between 1994 and 2004. Results again show no clear relationship between the prevalence of male circumcision and the prevalence of HIV: differences in incidence were small: 2.0, 2.5, 2.1%, and differences in net reproduction rates were even in the opposite order: higher in provinces with high level of circumcision than in those with low level (see details in Garenne, 2008). If one compares two contrasted provinces: one with no circumcision, the North-West province, populated by Tswana, and one with widespread circumcision, the Eastern Cape province, populated by Xhosa, one finds no difference in the dynamics of the epidemic from 1994 to 2004, and levels of seroprevalence were basically the same in 2004.
In conclusion, large-scale demographic surveys, as well as routine seroprevalence surveys among pregnant women, do not show any consistent population impact of male circumcision on either HIV prevalence or HIV incidence. Male circumcision does not appear to be the “Magic Bullet” presented by other researchers and based on results from clinical trials.
Other Evidence of a Lack of Demographic Impact
These findings are not really new, and could have been anticipated. Robert Van Howe conducted a large scale meta-analysis of the effect of male circumcision on HIV and other STIs in many risk groups, and found no protective effect. He even found a small increased risk in his meta-analyses (Van Howe, 1999a, b). Comparing the United States, where male circumcision is widespread, to Europe where it is rare, also goes rather in the opposite direction: more HIV in the former, and even more HIV transmitted heterosexually if one excludes cases imported from Africa in Europe. Even in the USA, the African-American population is more circumcised than average, and it is also more affected by HIV (Siegfried et al., 2007).
Let us also remember that, in Africa, about 70% of men are already circumcised, probably the highest rate in any continent: this did not hamper Africa to host the largest epidemic in any continent. A quick comparison with Eastern Asia (China, Japan) where virtually no men are circumcised and where HIV prevalence is extremely low is illuminating: the correlation goes in the opposite direction— the less circumcision, the less HIV prevalence. In contrast, when one makes similar comparisons with vaccination coverage for diseases or with contraceptive use for fertility, one finds the expected correlations.
Controversy About Geographical Correlations
Several studies conducted in the 1980s argued that African countries more affected by HIV were also less circumcised (Bongaarts et al., 1989; Moses et al., 1990; Weiss et al., 2000). This has been confirmed by recent population based studies, but is obviously correlated with other important confounding factors affecting sexual behavior, in particular, religion: Islam recommends male circumcision, but is very much opposed to any form of premarital and extramarital intercourse for women, therefore strongly reducing the risk of sexual transmission of diseases.
In another paper, we argued that HIV was more closely related with marriage pattern and permissiveness, both being measured simultaneously by premarital fertility (having a birth prior to first marriage). The map displaying premarital fertility levels is in fact close to the map of HIV prevalence levels, revealing the other confounding factors (Zwang and Garenne, 2008). Therefore, the ecological correlation between HIV prevalence and male circumcision appears misleading, and seems to reflect primarily other determinants of HIV spread. However, one should note that no country for which we have reliable data where male circumcision is widespread (>85% circumcised), including non-Muslim countries, such as Congo-Kinshasa, had a high level of seroprevalence (>5% among adults 15–49). This fact has never been properly explained, and deserves more research.
In summary, whatever the correlations, one has to remember that even a country half circumcised, such as Lesotho, can have a very large epidemic, with levels of seroprevalence close to the highest on record. This seems to give a far better picture of what could be the potential impact of mass circumcision campaigns: basically negligible in generalized epidemics.
Population Impact Versus Clinical Efficacy
The difference between the effect of male circumcision in clinical trials and the lack of any significant demographic effect may seem puzzling at first glance. The main reason seems to be the low level of clinical efficacy: a 50% reduction in risk is likely to have only a small demographic effect. Indeed, under repeated exposure, any circumcised man will eventually become infected, as will intact men. Circumcision does not really provide any protection, but simply reduces the risk at each exposure. This may delay the time to infection, but will rarely change the ultimate outcome, a situation similar to the effect of the rhythm method for contraception. The potential effect of mass circumcision campaigns will lead only to a very small impact, which will be buried into the many other factors of HIV transmission. If male circumcision had a 99% protective efficacy, the situation would have been different.
Rationale for Making Public Health Policies
Let us ask now a simple question, based on the findings of the clinical trials: is a 50% reduction in risk enough to make a policy? The answer is clearly no. For instance, if one compares with the field of contraception: the rhythm method is also 50% efficient in clinical trials, but is not recommended, because there are much better strategies available for birth control. Likewise, the cholera vaccine is also 50% efficient in clinical trials, but is not recommended, because there are much better strategies available for controlling cholera. There are also better alternative strategies to control HIV, summarized under the acronym “ABC” (for Abstinence, Be faithful, Condom use), which have worked in Africa and elsewhere: changing risky behavior worked well, for instance, in Uganda (Low-Beer, 2002; Low-Beer and Stoneburner, 2003), and condom use worked extremely well in Thailand (Brown et al., 1994). We do not have a full account of what has been happening in Africa since year 2000, but in almost all countries for which data are available, prevalence and incidence among young adults have been going down over the past 10 years, as a result of changing behavior. These ABC strategies seem to be able to change the course of the HIV epidemics. On the other hand, no country has ever been able to control an STI with male circumcision only. Let us remember the case of Japan: this country has among the lowest rates of HIV and of any STI, and makes a very wide use of condoms. This seems to be a far better strategy for controlling sexually transmitted infections.
Ethical Considerations
This paper focuses on the lack of demographic impact to be expected from male circumcision. Of course, there are many other dangers associated with mass circumcision campaigns. Firstly, at population level, shifting from ABC strategies, which, from experience, are the most likely to change the course of the epidemic to a strategy that has no chance of doing so, seems to be a serious mistake. Secondly, at the individual level, giving the impression that male circumcision “protects” against HIV transmission may have adverse effects: by giving a false sense of protection, it may induce riskier behaviors, and ultimately increase transmission. In this case, it will also have negative effects on the confidence that individuals have in the health system and in health education messages.
Ultimate Rationale of Male Circumcision
Male circumcision is a form of genital mutilation with numerous implications, amply documented in this book and in the ten international symposia organized over the past 20 years. This in itself raises many ethical issues, widely documented elsewhere (Aggleton, 2007; Doctors Opposing (male) Circumcision, 2006; Clark, 2006). Beyond individual cases when it is recommended for medical reasons, at population level male circumcision appears as a form of power abuse, especially when made compulsory, or at least strongly recommended. It is especially questionable when used on infants, children, or adolescents since it violates their rights, but is also questionable when recommended for consenting adults.
Numerous studies have highlighted the stakes behind this practice. For traditional societies, where circumcision is compulsory for adolescents, the power abuse comes from the elders. This is best expressed by Margaret Mead (1949), who had such a powerful insight on male circumcision. In her famous book, Male and Female, she says:
in South America, in Africa and in the South Seas, there are tribes in which the old men’s antagonism to the springing sexuality of the young induces fears that are later reduced in pantomine, cruel initiatory rites in which the young men are circumcised, their teeth knocked out, and, in various ways they are reduced and modified and humbled, and then permitted to be men.
When religious leaders recommend circumcision for newborns or young boys, the power abuse comes from religious hierarchy and applies to the whole society: by requesting the parents to accept the circumcision of their sons, the religious establishment ensures its power over the whole family. Robert Darby (2005) in his book, A Surgical Temptation, showed that the development of male circumcision for newborn infants in Victorian England is also a form of power abuse, this time coming from the medical establishment over the families. This came at a time when the political power of physicians and surgeons in society increased dramatically, and a new form of “biopower” emerged.
The question that can be raised now is whether recommending mass circumcision campaigns for Africans, in the absence of clear evidence of any demographic impact, is not a new form of power abuse, this time from newly established groups: international institutions and lobbies? This question certainly deserves further comments and in-depth discussions.
Aggleton P. (2007) Just a snip? A social history of male circumcision. Reprod Health Matters. 15(29):15–21.
Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, Puren A. (2005) Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: The ANRS 1265 trial. PLOS Med. 2(11):e298, 1–111.
Bailey RC, Moses S, Parker CB, Agot K, MacLean I, Krieger JN, Willams CFM, Campbell RT, Nidnya-Achola JO. (2007) Male circumcision for HIV prevention in young men in Kisumu, Kenya: A randomised controlled trial. Lancet. 369:643–656.
Bongaarts J, Reining P, Way P, Conant F. (1989) The relationship between male circumcision and HIV infection in African populations. AIDS. 3(6):373–377.
Brown T, Sittitrai W, Vanichseni S, Thisyakorn U. (1994) The recent epidemiology of HIV and AIDS in Thailand. AIDS. 8(Suppl 2):S131–S141.
Clark PA. (2006) To circumcise or not to circumcise? Health Prog. 87(5):1–9.
Connolly C, Simbayi LC, Shanmugam R, Nqeketo A. (2008) Male circumcision and its relationship to HIV infection in South Africa: Results of a national survey in 2002. S Afr Med J. 98(10):789–794.
Darby R. (2005) A Surgical Temptation: The Demonization of the Foreskin and the Rise of Circumcision in Britain. Chicago, IL: University of Chicago Press.
Demographic and Health Surveys. Web site: www.measuredhs.com
Doctors opposing male circumcision. (2006) Medical Ethics and the Circumcision of Children. Report, 2006. Available on web site: http://www.doctorsopposingcircumcision.org/pdf/A4-medicalEthicsReport.pdf
Edwards K, Decker M. (2008) Cholera vaccine. In: Plotkin SA, Orenstein WA. (eds.) Vaccines. Philadelphia, PA: Elsevier-Saunders.
Garenne M. (2006) Male circumcision and HIV control in Africa. PLoS Med. 3(1):e78.[Letter]
Garenne M. (2008) Long-term population effect of male circumcision in generalized HIV epidemics in sub-Saharan Africa. Afr J AIDS Res. 7(1):1–8.
Gray RH, Kigozi G, Serwadda D, Makumbi F, Watya S, Nalugoda F, Kiwanuka N, Moulton LH, Chaudhary MA, ChenMZ, Sewankambo NK,Wabwire-Mangen F, BaconMC,Williams CFM, Opendi P, Reynolds SJ, Laeyendecker O, Quinn TC, Waver MJ. (2007) Male circumcision for HIV prevention in men in Rakai, Uganda: A randomised controlled trial. Lancet. 369:657–666.
Labbok MH, Queenan JT. (1989) The use of periodic abstinence for family planning. Clin Obstet Gynecol. 32(2):387–402.
Lesotho – Ministry of Health and SocialWelfare (MOHSW), Bureau of Statistics (BOS), and ORC Macro. (2005) Lesotho Demographic and Health Survey 2004. Calverton, NY:MOH, BOS, and ORC Macro.
Low-Beer D. (2002) HIV incidence and prevalence trends in Uganda. Lancet. 360(9347):1788.
Low-Beer D, Stoneburner RL. (2003) Behaviour and communication change in reducing HIV: Is Uganda unique? Afr J AIDS Res. 1(2):9–21.
Malawi – National Statistical Office (NSO), and ORC Macro. (2005) Malawi Demographic and Health Survey 2004. Calverton, NY: NSO and ORC Macro.
Mead M. (1949) Male and Female: A Study of the Sexes in a Changing World. New York, NY: [Reprinted by Harper-Collins, 2002].
Moses S, Bradley JE et al. (1990) Geographical patterns of male circumcision practices in Africa. Int J Epidemiol. 19(3):693–697.
Pollard R. (1980) Relation between vaccination and notification rates for whooping cough in England and Wales. Lancet. 1(8179):1180–1182. Redd SC, Markowitz LE, Katz SL. (1999) Measles vaccine. In: Plotkin SA, Orenstein WA. (eds.) Vaccines. Philadelphia, PA: W.B. Saunders, pp 222–266.
Tacket C, Sack D. (2008) Cholera vaccine. In: Plotkin SA, Orenstein WA. (eds.) Vaccines. Philadelphia, PA: Elsevier-Saunders.
Siegfried N, Muller M, Volmink J, Deeks J, Egger M, Low N, Weiss H, Wlaker S, Williamson P. (2007) Male circumcision for prevention of heterosexual acquisition of HIV in men (Review). The Cochrane Library.
South Africa. (2002) The Nelson Mandela/HSRC study of HIV/AIDS, 2002. Human Science Research Council, Cape Town, South Africa. [available on web site: www.hsrcpress.ac.za]
Tanzania – Commission for AIDS (TACAIDS), National Bureau of Statistics (NBS), and ORC Macro. (2005) Tanzania HIV/AIDS Indicator Survey 2003–2004. Calverton, NY: TACAIDS, NBS, and ORC Macro.
Van Howe RS. (1999a) Does circumcision influence sexually transmitted diseases? A literature review. BJU Int. 83(Supp 1):52–62.
Van Howe RS. (1999b) Circumcision and HIV infection: Review of the literature and metaanalysis. Int J STD AIDS. 10:8–16.
Weiss HA, Quigley MA, Hayes RK. (2000) Male circumcision and risk of HIV infection in sub-Saharan Africa: A systematic review and meta-analysis. AIDS. 174:2361–2370.
Wood DL, Brunell PA. (1995) Measles control in the United States: Problems of the past and challenges for the future. Clin Microbiol Rev. 8(2):260–267.
WHO/UNAIDS. (2007) Recommendations from expert consultation on male circumcision for HIV prevention. Available at: http://www.who.int/hiv/mediacentre/news68/en/index.html
Zwang J, Garenne M. (2008) Premarital fertility and HIV/AIDS in Africa. Afr J Reprod Health. 12(1):64–74.
Michel Garenne, Mass Campaigns of Male Circumcision for HIV Control in Africa: Clinical Efficacy, Population Effectiveness, Political Issues, in George C. Denniston, Frederick M. Hodges and Marilyn Fayre Milos (eds), Genital Autonomy: Protecting Personal Choice (Dordrecht, Heidelberg, London, New York: Springer, 2010; ISBN 978-90-481-9445-2 e-ISBN 978-90-481-9446-9)
Professor Garenne is a demographer and reproductive health expert with the IRD (French Institute for Research and Development) and Institut Pasteur, Paris, France; e-mail: mgarenne@hotmail.com
The book may be purchased from Amazon. Individual chapters may be bought and downloaded from the publishers, SpringerLink.