Secondary Logo

Journal Logo

Original Article

Risk of HIV-1 in Rural Kenya

A Comparison of Circumcised and Uncircumcised Men

Agot, Kawango E.*†; Ndinya-Achola, Jeckoniah O.; Kreiss, Joan K.; Weiss, Noel S.*

Author Information
doi: 10.1097/01.ede.0000112220.16977.82



Click on the links below to access all the ArticlePlus for this article.

Please note that ArticlePlus files may launch a viewer application outside of your web browser.

By the end of 2002, approximately 42 million people in the world were estimated to be living with HIV/AIDS, 29.4 million (70%) of whom were from sub-Saharan Africa.1 Currently, HIV/AIDS is the fourth leading cause of mortality worldwide and the number one killer in sub-Saharan Africa.1,2 A number of studies have examined factors suspected of fueling the epidemic in the subcontinent, including the presence of other sexually transmitted diseases,3–6 concurrent and multiple sexual partners,7–9 low prevalence of condom use,4 and contact with sex workers.10–12

There is also growing evidence linking lack of male circumcision with elevated risk of HIV-1.6,11–18 Some have recommended male circumcision as one possible strategy to control the rapid spread of HIV in Africa.6,14,15,20 Others, however, have felt that available evidence is not sufficient to justify such a recommendation.21–23 The main reservation with available data is the possibility that confounding has influenced the direction or the strength of the observed associations. Of particular concern is that cultural or religious differences between circumcised and uncircumcised men may be associated with their sexual and genital hygienic practices such as multiple sexual partnerships or washing after sex, which may in turn affect their exposure to HIV-1 and other sexually transmitted infections.21,24 Furthermore, most studies have been among high-risk populations such as long-distance truck drivers, sexually transmitted infections clinic patients, and clients of commercial sex workers, and the majority of these studies have been carried out in urban or periurban settings.6,11,12,25,26

In Kenya, approximately 13% of the population is currently infected with HIV-1, with the highest prevalence being in the west among the Luo ethnic community.1,2,6,19 Luo is the third largest community in Kenya, with a population of over 2 million. Many Luo men are Christians and members of African-instituted churches (churches or denominations founded in Africa, by Africans, as distinguished from those introduced by missionaries). Some, but not all, of these churches advocate circumcision as part of their teachings. Among the minority of Luo men who are circumcised, the practice is often linked to membership in specific denominations.

We conducted a study of circumcision in relation to the prevalence of HIV-1 among Luo members of African-instituted churches. This setting provided us with an opportunity to compare uncircumcised and circumcised men from the same ethnic community and similar religious backgrounds. Among these men, the prevalence of other risk factors for HIV-1 infection might be expected to be rather similar.


We conducted a cross-sectional study of the prevalence of HIV-1 among men in the Luo ethnic community in Kenya in relation to penile circumcision. Ethical approvals and research permits in Kenya were obtained from Kenyatta National Hospital, the Nyanza Provincial Medical Office, the Office of the President, and the Nyanza Provincial Administration. Approval was also obtained from the Human Subjects Review Committee of the University of Washington, Seattle.

Subject Selection and Enrollment Procedure

Between April and September 1999, we visited both the African and East African head offices of an organization that brings together Christian denominations founded by Africans, the Organization of African-Instituted Churches (OAIC), to obtain lists of member and nonmember denominations. We also identified and visited several leaders whose denominations were not on the lists provided by the OAIC. Altogether, approximately 80 such denominations were identified within the Luo community. We visited leaders of 68 of them, and leaders of 66 denominations gave us permission to carry out the study in their churches. Nine of the denominations advocated circumcision among male members, whereas 57 did not. We also obtained information on the teachings of the 66 denominations regarding the 2 main cultural practices believed to be placing the Luo community at disproportionate risk for HIV infection: polygamy and widow inheritance (a practice whereby the brother of a deceased husband assumes conjugal rights and other responsibilities over a widow). All 9 circumcising denominations strongly encouraged widow inheritance and polygamy, and were explicit in their promotion of the 2 practices. We used these 2 practices as the initial criteria to match circumcising and noncircumcising denominations, as described subsequently.

We ranked the 57 noncircumcising denominations in terms of their support for widow inheritance (yes/neutral/no) and polygamy (yes/neutral/no). Thirty-one denominations supported inheritance, 19 were neutral, whereas 7 denounced it; and 18 supported polygamy, 25 were neutral, and 14 opposed it. The 57 denominations did not support or denounce the practice of male circumcision explicitly; they assumed the status quo of noncircumcision was the norm. Because all the 9 circumcising denominations strongly supported both widow inheritance and polygamy, we selected for comparison 16 of the 57 noncircumcising denominations that also approved of both practices.

Leaders of the 9 circumcising and the 16 noncircumcising denominations were asked to compile lists of churches under them and their geographic locations (districts, divisions, locations, and sublocations). (Districts are the third largest politico-administrative units in Kenya, with a population of 100,000–300,000, whereas sublocations are the smallest units, with 5–10,000 people.27) A total of 1237 churches were listed in the 8 districts occupied by the Luo ethnic community, 448 from the 9 circumcising denominations and 789 from the 16 noncircumcising denominations. We then selected 9 of the 16 noncircumcising denominations that had churches located at closest geographic proximity to those of circumcising denominations. Altogether, the selected 9 noncircumcising denominations had 505 churches.

Between October 1999 and May 2000, we visited and enrolled men from 123 circumcising and 92 noncircumcising churches at sublocation level, and 118 circumcising and 66 noncircumcising churches at the location level. We did not proceed to the division and district levels because we had attained our targeted sample of 890 men. We were unable to obtain lists of male members of the churches, so we recruited all eligible men who were present at the time of a research team visit and who provided informed consent to join the study. We ultimately visited more churches from circumcising denominations (241 of 448, compared with 158 of 505 from noncircumcising denominations) because 23% of the 489 male participants from circumcising denominations were in fact not circumcised, compared with only 6% (n = 20) circumcised men from noncircumcising denominations.

Data Collection

To be eligible for this study, men needed to be Luo by ethnicity, members of an African-independent church, age 18–49 years, sexually active, and unaware of their HIV status. Eligible participants were taken through the informed consent process in Dholuo, the local language. Those who consented to participate by signing the form were issued confidential identification numbers before proceeding to the interview. The interview included questions on age, education, occupation, income, marital status, number of wives, history of spousal separation, sexual history (age at sexual debut, sexual partnerships, contact with sex workers, casual sex, sex during partner's menses, and condom use), and alcohol consumption (history and frequency of taking alcohol and of having sex while intoxicated).

Participants then proceeded to a male nurse-counselor who asked them questions regarding their medical history such as history of sexually transmitted infections (genital ulcer disease, urethral pain or discharge) and how those reporting a history of sexually transmitted infection managed their condition (whether they sought therapy, notified their partner, took action to prevent transmission to their partner). Men were then given pretest counseling, including information on when and where to learn about their test results, before being asked to give 3–4 mL of venous blood. Circumcision status was physically verified and recorded as follows: uncircumcised if the glans penis was covered completely by the foreskin on pulling down and releasing, partially circumcised if the glans penis was partly covered, and circumcised if there was no foreskin at all. Genital hygiene was assessed by asking several open-ended questions about the usual procedure and frequency of genital washing. In addition, the male nurse examined and classified the level of cleanliness under the foreskin as “good” if the glans penis had no evidence of smegma and as “poor” if smegma was evident.

Data Analysis

HIV-1 serologic testing was conducted using Detect HIV-ELISA test (Biochem Immunosystems, Inc., Montreal, Canada) for screening and Recombigen HIV-ELISA test (Cambridge Biotech, Worcester, MA) for confirmation. We compared demographic and selected HIV-1 risk profiles, as well as the seroprevalence of HIV-1, between (1) all circumcised men and all uncircumcised men in the study, regardless of their denominational background, and (2) circumcised and uncircumcised men only from denominations that advocate circumcision. The association between HIV-1 seroprevalence and circumcision status was assessed using prevalence difference, prevalence ratio (PR), and prevalence odds ratio (POR); 95% confidence intervals (CIs) were computed for the ratios. The prevalence difference was obtained by first computing the prevalence of HIV-1 in uncircumcised men, standardized for each risk factor individually using the risk factor distribution of circumcised men as the standard. We then calculated the difference between the 2 proportions.

We computed Mantel-Haenszel estimates of the adjusted prevalence ratios and performed logistic regression to obtain crude and adjusted prevalence odds ratios using covariates listed on the Table, as well as subjects’ denomination, to run stepwise incremental models. The final model included demographic characteristics (income, age, education, and marital status), sexual partnerships (extramarital and premarital sex partners, contact with commercial sex workers, polygamy, and widow inheritance), condom use, sexually transmitted infections (history of genital ulcer disease, urethral pain or discharge), denominational affiliation, and miscellaneous other risk profile (ever lived in town, resides with wife, age at sexual debut, alcohol consumption, and sex during partner's menses).

Distribution of Circumcised and Uncircumcised Men, by Demographic Profile and Selected HIV-1 Risk Factors


Altogether, 1217 participants were eligible—706 (58%) from denominations that advocate circumcision and 511 (42%) from denominations that do not advocate the practice. Thirty-four men (3%) did not give consent to join the study and were not interviewed, whereas 138 (11%) did not participate in the clinical procedures as a result of time constraint. Finally, 156 (13%) declined to give a blood sample or to have their circumcision status verified. In all, 889 men (73%) completed all aspects of the study—525 (74%) of the 706 who were members of denominations that advocate circumcision and 364 (71%) of the 511 who were members of denominations that do not advocate circumcision.

Our results are based only on those whose circumcision status was confirmed by physical examination rather than from self-report (Twenty-six of those who had reported being circumcised were in fact not circumcised on examination, and 17 of those who reported not being circumcised were circumcised.) Among the study participants, 447 were observed to be fully uncircumcised (336 from noncircumcising denominations and 111 from circumcising denominations), 398 were fully circumcised (378 from circumcising denominations and 20 from noncircumcising denominations), and 37 were partially circumcised. Seven participants had discordant results on HIV-1 serologic testing and were excluded from the analysis, as were the 37 participants who were partially circumcised. Our results are based on the 398 participants who were physically verified as fully circumcised and the 447 as fully uncircumcised.

The Table presents selected background characteristics and the HIV-1 risk profile of the participants, divided into uncircumcised men from denominations that advocate circumcision and those that do not, and circumcised men. As a result of the small number of circumcised men from noncircumcising denominations (n = 20), it was not feasible to analyze them separately from other circumcised men. Demographically, the circumcised men and the 2 groups of uncircumcised men were broadly similar, with the circumcised men being slightly older and more likely to be married, to be residing together with their wives, to have 2 or more wives, and to report no history of genital ulcers. Circumcised and uncircumcised men did not differ to any substantial degree with respect to sexual history, prior occurrence of sexually transmitted infections other than genital ulcers, or alcohol consumption.

Overall, we found 26% HIV-1 seropositivity—30% in uncircumcised men and 20% in circumcised men. Among members of denominations that advocate circumcision, 31% who had not been circumcised were seropositive for HIV-1. The crude prevalence ratio for HIV-1 associated with not being circumcised was approximately 1.5, irrespective of denominational affiliation. Overall, the crude prevalence ratio was 1.5 (95% CI = 1.2–2.0); after adjustment, as discussed previously in the “Methods” section, the result was similar (data not shown). The crude prevalence difference was 10–11 per 100 men. When we adjusted for each of the covariates in the Table, the prevalence ratios ranged from 1.5 to 1.7, whereas the difference ranged from 8.9 to 14.5 per 100 (data not shown). A logistic regression analysis that included the variables discussed previously in the “Methods” section gave a POR associated with not being circumcised of 2.1 (95% CI = 1.5–3.0); this adjusted POR was slightly different from the crude POR of 1.9 (CI = 1.3–2.6). Because of the high frequency of HIV-1 in the unexposed in this population (20%), the POR has overestimated the RR, hence, the discrepancy between the PR and the POR.

With one exception, the size of the association between lack of circumcision and HIV-1 prevalence was similar across all characteristics listed in the Table. That exception was condom use: the difference in HIV-1 prevalence between uncircumcised and circumcised men who always used a condom during intercourse during the prior year was only 0.6%, whereas the corresponding difference in men who sometimes used a condom was 18% and for those who did not use them at all the difference was 11%.

Among men who were circumcised, the prevalence of HIV-1 seropositivity was similar regardless of whether the circumcision was performed before or after the onset of sexual activity (19% in 214 men circumcised before and 21% in 184 circumcised after sexual debut). Altogether, 38% of the circumcisions were done on day 8 after birth (following the religious teachings of circumcising churches), whereas 61% had been performed by age 15. Also, we found that 26% (n = 257) of uncircumcised men with “adequate” genital hygiene (reporting retracting foreskin while washing in addition to having no visible smegma on the glans penis on physical examination) were HIV-1-seropositive, compared with 34% (n = 61) of those with “less than adequate” hygiene (PR = 1.3; CI = 0.9–2.0). There was no difference in HIV-1 seroprevalence between reported versus verified circumcision status (both 26%). However, using denominational policy on circumcision as an ecologic marker for the practice, HIV-1 prevalence in members of circumcising denominations was 22% compared with 30% in noncircumcising denominations (PR = 1.3; CI = 1.0–1.6).


We observed a higher prevalence of HIV-1 in Luo men who were not circumcised compared with those who were circumcised (30% vs. 20%). To minimize the likelihood that this difference in the prevalence of HIV-1 infection was confounded by unmeasured differences in sexual behavior, we also compared uncircumcised and circumcised men within denominations that advocate circumcision and obtained virtually identical results.

We also found that among uncircumcised men HIV-1 seroprevalence was somewhat higher in those with “less than adequate” genital hygiene (34%) compared with those with “adequate” hygiene (26%). However, even when genital hygiene among those who were not circumcised was deemed adequate, they still had a higher HIV-1 prevalence compared with circumcised men (26% vs. 20%). Although there was no difference in HIV-1 seroprevalence between reported and verified circumcision status (both 26%), relying on denominational policy on circumcision to designate circumcision status of individuals would have led to a small overestimation of HIV-1 prevalence in circumcised men (22% instead of the true prevalence of 20%) but identical results in uncircumcised men (both 30%).

Previous studies have found an association between age of circumcision and HIV-1 prevalence. A study in Uganda reported that, relative to uncircumcised men, men who were circumcised before the age of 13 years had a 61% reduction in HIV-1 seroprevalence, those circumcised between ages 13 and 20 years had a 54% reduction, but those circumcised after age 20 years had no difference.16 We found no difference in HIV-1 seroprevalence between men who were circumcised before their first sexual activity and those who were circumcised after (19% vs. 21%). Among men who were circumcised after the onset of sexual activity, the prevalence of HIV-1 infection varied little whether the procedure took place before the reported onset of high HIV-1 prevalence in this community (approximately 1990) or afterward (21% vs. 19%, respectively).

Our study had some limitations. First, 328 (27%) of the 1217 participants who were eligible either declined to participate (3%) or did not complete the study (24%). However, because participants did not know their HIV-1 status at the time of our visit, bias from this source would seem unlikely, especially because comparable proportions of men from denominations that do and do not advocate circumcision completed the study. As a result of imprecise measures of some variables such as condom use and genital hygiene, data pertaining to subgroups defined by these variables need to be interpreted with caution.

These shortcomings notwithstanding, our study addresses a number of limitations that have been identified in previous studies. By confining the study to one ethnic community, we addressed the concern that the association between lack of circumcision and HIV-1 reported in previous studies may have been confounded by ethnicity.6,21 Circumcised and uncircumcised men in this population differed little with respect to other predictors of HIV-1 infection, and adjustments for these predictors had little impact on the magnitude of the association. Also, we verified circumcision status by physical examination and thus avoided potential misclassification of exposure status.

In theory, male circumcision could either increase or decrease the prevalence of HIV-1. There is a possibility that if performed under unhygienic conditions, circumcision could actually be responsible for transmitting the infection.

On the other hand, several physiological and biologic processes have been suggested as ways in which an intact foreskin could raise the risk of HIV-1 acquisition. It has been reported that the HIV-1 virus targets the Langerhans cells and macrophages in the inner mucosal surface of the foreskin, both of which are coated with CD4 receptors.6,20,24,28 Most cases of primary HIV-1 infection are thought to involve HIV-1 binding initially to CD4-positive cells, making them the primary point of viral entry into the penis of an uncircumcised man.20,28 This is particularly important because during heterosexual intercourse, the foreskin is pulled back down the shaft of the penis, and the inner surface of the foreskin is exposed to vaginal secretions, providing a relatively large area over which HIV-1 transmission could take place.20,29 Male circumcision therefore would provide protection by removing most of the vulnerable mucosal lining.6,20

It has also been suggested that the mucosal lining of the foreskin may be susceptible to minor injuries during intercourse, given that in uncircumcised men, the epithelium lining the glans penis and preputial sac is thinner and less cornified than that of circumcised men.6,16,20 Finally, the warm and relatively moist environment under the foreskin may act as an incubator and thus facilitate replication of the virus, in addition to allowing longer viral persistence and greater exposure time and opportunity for viral penetration.6,12,20,29

In conclusion, we believe that the results of this and most prior studies support the notion that male circumcision should be seriously considered as an intervention to slow the spread of HIV-1 in uncircumcised populations. Circumcision intervention research is needed to address issues of cultural acceptability, medical and economic costs and benefits, and the potential for increased risk-taking by circumcised men who may falsely assume complete protection from HIV-1. The high prevalence of HIV-1 among Luo men who are circumcised (20%) demonstrates that circumcision alone is at most a partial means of blocking female-to-male transmission of HIV-1 infection. Thus, circumcision intervention programs would need to be accompanied by other interventions such as programs that promote condom use and reduction in risky sexual behaviors.


Our gratitude to the research team: D. O. Oguso, E. M. Aroko, J. M. Osodo, T. O. Liwindi, V. O. Onyango, E. K. Odeny, N. O. Yogo, G. A. Odhiambo, R. Achieng, the late G. Maitha, J. O. Kogan, A. Vander Stoep, and E. A. Onyango.


1.UNAIDS. 2003. AIDS Epidemic Update: December 2002. Available at: Accessed in May 2003.
2.UNAIDS. 2002. Global HIV/AIDS Surveillance by Country. Available at: Accessed in November 2003.
3.Cohen MS. Preventing sexually transmitted infection of HIV: new ideas from sub-Saharan Africa. N Engl J Med. 2000;342:970–972.
4.Newell J, Senkoro K, Mosha F, et al. A population-based study of syphilis and sexually transmitted disease syndromes in north-western Tanzania. 2. Risk factors and health seeking behaviour. Genitourin Med. 1993;69:421–426.
5.Pepin 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;6:489–493.
6.Lavreys L, Rakwar JP, Thompson ML, et al. 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. J Infect Dis. 1999;180:330–336.
7.Aral SO. Sexual network patterns as determinates of STD rates: paradigm shift in the behavioral epidemiology of STDs made visible. Sex Transm Dis. 1999;26:262–264.
8.Morris M, Kretzschmar M. Concurrent partnerships and the spread of HIV. AIDS. 1997;11:641–648.
9.Hudson CP. AIDS in rural Africa: a paradigm for HIV-1 prevention. Int J STD AIDS. 1996;7:236–243.
10.Bwayo JJ, Omari AM, Mutere AN, et al. Long distance truck drivers: prevalence of sexually transmitted diseases (STDs). East Afr Med J. 1991;68:425–429.
11.Nasio JM, Nagelkerke NJD, Mwatha A, et al. Genital ulcer disease among STD clinic attenders in Nairobi: association with HIV-1 and circumcision. Int J STD AIDS. 1996;7:410–414.
12.Tyndall MW, Ronald MW, Agoki E, et al. Increased risk of infection with human immunodeficiency virus type 1 among uncircumcised men presenting with genital ulcers in Kenya. Clin Infect Dis. 1996;23:449–453.
13.O'Farell N, Egger M. Circumcision in men and the prevention of HIV infection: a ‘meta-analysis’ revisited. Int J STD AIDS. 2000;11:137–142.
14.Bailey R, Neerma S, Othieno R. Sexual behaviors and other risk factors in circumcised and uncircumcised men in Uganda. J Acquir Immune Defic Syndr. 1999;22:294–301.
15.Moses S, Bailey RC, Ronald AR. Male circumcision: an assessment of health benefits and risks. Sex Transm Infect. 1998;74:368–373.
16.Kelly R, Kiwanuka N, Wawer MJ, et al. Age of male circumcision and risk of prevalent HIV infection in rural Uganda. AIDS. 1999;13:399–405.
17.Cameron DW, Simonsen JN, D'Costa LJ, et al. Female-to-male circumcision of HIV-1 risk factors for seroconversion in men. Lancet. 1989;2:403–407.
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.Auvert B, Buvè A, Lagarde E, et al. Male circumcision and HIV infection in four cities in sub-Saharan Africa. AIDS. 2001;15(suppl 4):S31–S40.
20.Szabo R, Short RV. How does male circumcision protect against HIV infection? BMJ. 2000;320:1592–1594.
21.De Vincenzi I, Mertens T. Male circumcision: a role in HIV prevention? AIDS. 1994;8:153–160.
22.Ntozi JPM. Using circumcision to prevent HIV infection in sub-Saharan Africa: the view of an African. Health Transit Rev. 1995;5:97–100.
23.Van Howe RS. Circumcision and HIV infection: review of the literature and meta-analysis. Int J STD AIDS. 1999;10:8–16.
24.Jessamine PG, Plummer FA, Ndinya Achola JO, et al. Human immunodeficiency virus, genital ulcers, and the male foreskin: synergism in HIV–1 transmission. Scand J Infect Dis. 1990;69:181–186.
25.Mbugua GG, Muthiami LN, Mutura CW, et al. Epidemiology of HIV infection among long distance truck drivers in Kenya. East Afr Med J. 1995;72:515–518.
26.Rakwar J, Lavreys L, Thompson ML, et al. Cofactors for the acquisition of HIV-1 among heterosexual men: prospective cohort study of trucking company workers in Kenya. AIDS. 1999;13:607–614.
27.Government of Kenya. 1999 Population and Housing Census: Counting Our People for Development, vol 1. Nairobi: Central Bureau of Statistics and Ministry of Finance and Planning; 2001.
28.Bailey RC, Plummer FA, Moses S. Male circumcision and HIV prevention: current knowledge and future research directions. Lancet. 2001;1:223–231.
29.Moses S, Nagelkerke NJD, Blanchard J. Analysis of the scientific literature on male circumcision and risk factor for HIV infection. Int J STD AIDS. 1999;10:626–628.

Supplemental Digital Content

© 2004 Lippincott Williams & Wilkins, Inc.