Geographical variations in HIV prevalence have been observed between less-developed and more-developed countries, as well as within regions of similar socioeconomic development [1–5]. The epidemiology of HIV and other infectious diseases have been associated with both religious practices and male circumcision [1–19]. Religious beliefs and practices dictate many societal and sexual behaviors that influence transmission of sexually-transmitted infections (STIs) [20]. Male circumcision has been more common among populations with lower rates of HIV, cervical cancer, and other STIs [3, 10–13], and shown in one randomized trial to reduce HIV transmission [21]. Although religious affiliation is a major determinant of male circumcision status [12], many analyses have not controlled for religion when examining relationships between male circumcision and infectious diseases.

This study builds upon and further expands our previously reported analyses of variables associated with country-specific HIV prevalence and cervical cancer incidence [5, 22]. In our extensive analysis of HIV co-factors, among 81 variables male circumcision had the strongest association with HIV prevalence [5]. We now present a more thorough examination of the association between male circumcision and HIV prevalence by better adjusting for religion, by separately analyzing the sub-Saharan African region, and by conducting separate analyses between countries with sexual versus non-sexual primary modes of HIV transmission. Our previous ecological analysis of cervical cancer utilized 54 country-level variables, but did not include the important determinant of male circumcision [13, 22]. We complete our previous analysis by describing the relationships between male circumcision and cervical cancer, and by including male circumcision in the previously reported multivariate model. In addition, we further expand on our previous studies by describing the epidemiology of male circumcision among developing countries and by describing associations between male circumcision and five other infectious diseases.

We conducted an ecological study of country-level variables among developing countries. The United Nations Development Programme (UNDP) 'Human Development Report 2004' provided the Human Development Index, which determines each country's development status on the basis of life expectancy, educational attainment, and adjusted real income [23]. Countries classified as high human development ("developed") countries were excluded from the analyses under the assumption that they have greater capacity to sustain national treatment and prevention programs, and have different epidemiological infectious disease profiles. Therefore, subsequent country-specific data were collected for 122 low and medium human development ("developing") countries.

This ecological study of 118 developing countries expands results from our previous analyses [5, 22] by elucidating patterns and associations between male circumcision, religion, and infectious diseases, particularly HIV. Male circumcision, which is routinely practiced in the Middle East, northern and western Africa, and western Asia, was associated with lower rates of certain STIs, HIV and cervical cancer (a proxy for HPV), but not with infections transmitted by non-sexual routes. In general, more male circumcision was strongly associated with lower cervical cancer rates and fewer HIV cases, independent of religion. Furthermore, male circumcision was independently associated with HIV among countries with primarily heterosexual HIV transmission, and not among countries with primarily homosexual or injection drug use HIV transmission. These findings all suggest that male circumcision is a true protective factor that reduces the sexual transmission of HIV and possibly HPV, independent of Muslim and Christian religions.

An ecologic study of this type has limitations, as we have previously acknowledged [5, 22]. In brief, ecological analyses cannot measure correlates of risk at the individual-level. Second, the temporal sequence of events in individuals is undetermined. Since the approximate age at circumcision varies by country, the results could be affected if a significant proportion delayed male circumcision until after initiation of sexual activity. Third, the validity of country-level data undoubtedly varies, and some data, such as HSV-2 prevalence, were not available for many countries. Although mean HSV-2 prevalence was lower among countries with more male circumcision, the limited number of countries with HSV-2 data (23 countries) gave relatively low power for examining the significance of this association. Fourth, and perhaps most importantly, statistics on the distribution and variation within countries were not complete. Some countries, such as Kenya, have widely varying regional prevalences of male circumcision and HIV [32]. For example, western Kenya, where less than 20% of males are circumcised, has a much higher HIV prevalence than regions of the country where nearly all men are circumcised [3, 32]. The 2003 Kenya survey also found uncircumcised males had over three-fold higher HIV prevalence as compared to circumcised males, and HIV prevalence in circumcised men was nearly identical among the major religious groups (2.6% among Catholics, 3.0% in Protestant/other Christians, and 2.9% in Muslims) [32]. Finally, not all population-level measures that may impact infectious disease transmission, such as patterns of risk behaviors, condom availability and utilization, and injection drug use, were included in this analysis. Despite these limitations, findings from this ecological analysis support a biological relationship between male circumcision and certain STIs.

Previous studies, including a recent Cochrane review [16], have also found male circumcision to be associated with a reduced risk of HPV detection in men [14], cervical cancer in female partners [13], and HIV infection [3, 11, 12, 15, 17], while associations between male circumcision and HSV-2 and gonorrhea are less clear [15, 19, 43]. However, these studies were generally conducted among geographically-limited populations and many were unable to adequately control for religion. By comparison, our study described the global distribution of male circumcision prevalence, examined a large number of geographically-diverse countries and several infectious diseases, separately analyzed countries by primarily heterosexual and non-heterosexual transmission of HIV, and included relatively complete and recent surveillance data. Our study supports results of other studies by demonstrating independent associations of male circumcision with reduced cervical cancer rates and HIV cases, after stratifying the countries for Muslim and Christian religions, among a large sample of developing countries.

Others have described biologically plausible mechanisms by which male circumcision may reduce the transmission of certain STIs [18, 19, 44, 45]. First, circumcised males may have less difficulty maintaining penile hygiene, which may reduce the acquisition of STIs by decreasing inflammation and the carriage time of pathogens in the foreskin [19, 45]. Secondly, the non-keratinized epidermis of the prepuce in uncircumcised males may provide an easier portal of entry to STIs [19, 44, 45]. Third, the inner mucosal surface of the prepuce, which has a high density of HIV target cells (CD4+ T-cells, Langerhans cells, macrophages), has been shown to become more easily infected with HIV as compared to outer foreskin tissue [45]. Given these biological differences, it is entirely plausible that uncircumcised males are at a greater risk of acquiring some STIs, and of transmitting them to their sexual partners [12].

According to UNAIDS, heterosexual transmission was the primary mode of HIV transmission in Africa, the Middle East, the Caribbean, and South and South-East Asia, and homosexual transmission was the primary mode of HIV transmission in Latin America [37]. In our analysis, male circumcision was strongly associated with HIV among developing countries with heterosexual contact as the primary mode of HIV transmission, and not among developing countries whose primary mode of HIV transmission was not heterosexual contact (Eastern Europe, Central and Eastern Asia, Latin America, and the Pacific). These results further support the biological role of male circumcision as a protective factor in HIV transmission. 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 [24].

This study demonstrates that male circumcision was strongly associated with lower cervical cancer rates and fewer HIV cases among countries with heterosexual contact as the primary mode of HIV transmission, independent of religion. One randomized controlled trial has demonstrated that male circumcision is highly protective of HIV acquisition [21]. Therefore, while HIV and cervical cancer are impacted by a complex set of biological, social, and public health influences, male circumcision appears to play a prominent role in decreasing transmission of certain STIs. Although male circumcision must not substitute for other HIV and STI prevention strategies [46], the international public health and medical community should consider the implications and practicalities of integrating safe, voluntary male circumcision services with existing HIV prevention programs, particularly in countries with low prevalence of male circumcision and high prevalence of sexually-transmitted HIV.