Results of observational studies suggest that male circumcision provides some degree of protection against acquisition of HIV,1-3 ranging from 1.5-fold4 to 8.2-fold5 among circumcised men compared to uncircumcised men. More recently, the first randomized controlled trial (RCT) conducted in Orange Farm, South Africa, showed a 60% reduction of HIV infection among circumcised men.6 Two other RCTs are ongoing, one in Kisumu, Kenya, and another in Rakai, Uganda. If results from the two ongoing RCTs confirm those from South Africa, then demand for circumcision is likely to increase in ethnic communities in Sub-Saharan Africa, where male circumcision is not a traditional practice. Although the RCTs should be able to assess the overall impact of circumcision on HIV acquisition, incorporating both physical and social consequences, the standard of care found in clinical trial settings, especially thoroughly trained service providers, a careful selection of healthy participants, repeated HIV counseling and testing, and counseling on sexual risk reduction, may influence subsequent sexual behaviors of trial participants in ways that are different from those receiving circumcision in other settings. We sought to investigate the effect of circumcision on sexual behaviors of men in a typical public health care setting.
We conducted a prospective cohort study in Siaya and Bondo District Hospitals, western Kenya, to assess if men from the Luo ethnic community who underwent circumcision engaged afterward in more risky sexual behaviors than other Luo men. The study was approved by the Kenyatta National Hospital Ethical Review Committee, Nairobi, Kenya, and the institutional review board at University of Washington, Seattle.
Between January 2002 and August 2004, we used fliers, posters, word of mouth, and referrals from study participants to invite men to enroll in the study. To be eligible, men had to be uncircumcised, at least 18 years or older, and residents of Siaya or Bondo districts who were not planning to move out of the district for one year. Men were asked to choose between getting circumcised and remaining uncircumcised. Those who chose to be circumcised were matched with those who chose to remain uncircumcised on the basis of age (±2 years), marital status (married/single), and residential location (urban/rural).
At enrollment, men were asked questions relating to demographic characteristics, current and past medical history, and sexual behavior in the previous 3 months. Participants also received a medical examination. All participants were asked to return for follow-up visits at 1, 3, 6, 9, and 12 months after enrollment. During these visits, participants were asked questions relating to medical history and sexual behavior for the period since the previous visit, given a medical exam, and counseled on risk reduction. Questions focused on sexual behaviors with other girlfriends, casual partners, or sex workers. We asked about the number of sex partners, number of sex acts, condom use, and frequency of condom use. With regard to sexual transmitted infection (STI) history, participants were asked at baseline if they were having, or had experienced, various symptoms of sexually transmitted infections (urethral discharge, genital ulcers, genital warts) in the previous 6 months and, during subsequent follow-up visits, whether they were having STI symptoms. In addition, all participants were physically examined to ascertain the presence of these STI symptoms and to verify their circumcision status. Men who were circumcised were asked to return 3 and 8 days post-surgery to have the bandage removed and healing assessed. All participants were encouraged to return to the clinic for free outpatient treatment of complications arising from the surgery or for illnesses such as malaria, STIs, and respiratory and abdominal complaints. They were given USD 1.25 as compensation for time spent during a study visit.
For this study, we were primarily interested in sexual behaviors known to place men at increased risk of contracting HIV (“risky sex”), specifically having sexual intercourse with partners other than their legal wife/wives or, for unmarried men, with partners other than their “regular” girlfriends. Participants were asked the duration of each relationship, and to identify the partners they considered regular. The main two variables of interest were (1) simply having had sex acts with partners other than legal wife/wives or regular girlfriends (“risky sex”) and (2) having had unprotected sex acts with partners other than legal wife/wives or regular girlfriends (“risky unprotected sex”). The average rate of sex acts per week was determined by taking the total number of reported sex acts divided by total weeks of follow-up. Participants were then categorized according to three levels of risky sex: low risk (no risky sex acts), medium risk (0-0.5 risky sex acts per week), and high risk (>0.5 risky sex acts per week). Other sexual risk behaviors assessed included average number of risky sex partners per month, which was also categorized into three risk levels [low risk (no risky sex partners), medium risk (0-1 risky sex partner per month), and high risk (>1 risky sex partners per month)], and condom use (no use, inconsistent use, and consistent use).
Risk ratios (RRs) and 95% confidence intervals (CIs) were used to assess the association between risk of having each risky sexual behavior of interest and circumcision status at every follow-up visit. In addition, a summary RR was calculated to describe the same association over the entire follow-up period, excluding the first month. RRs were adjusted for risky unprotected sex behavior during the 3 months before baseline. Differences between STI symptoms and baseline variables were assessed using Pearson χ2 tests or Fisher exact tests. All analyses were performed using SAS statistical software, version 9.1.
Characteristics of the 648 men who participated in the study are shown in Table 1. At study entry, there were no differences between men who were circumcised and men who remained uncircumcised with respect to age, income, education, marital status, history of major medical illness or surgical procedures, presence of STIs, condom use, and average number of risky sex partners. Men who chose to be circumcised were 1.31 times more likely to have had risky sex acts and 1.79 times more likely to have had risky unprotected sex acts in the three months before study entry than men who chose to remain uncircumcised (95% CI: 1.03, 1.67 and 1.20, 2.66, respectively). Men who were circumcised were more likely to have current problems with the foreskin or preputial complaints and to report deviation during erection (8.4% vs. 1.6%), difficulty inserting the penis during intercourse (7.7% vs. 2.5%), and difficulty achieving erection because skin of the penis was too tight (5.0% vs. 0.6%).
One month after circumcision, men who chose to be circumcised were 63% less likely to report having 0 to 0.5 risky sex acts per week and 61% less likely to report having >0.5 risky sex acts per week than men who remained uncircumcised (95% CI: 21 to 83% and 27% to 79%, respectively; Table 2). Similarly, men who chose to be circumcised were 87% less likely to report having 0 to 0.5 risky unprotected sex acts per week than men who remained uncircumcised (95% CI: 44% to 97%; Table 3). Later in the 12-month follow-up period, this difference largely disappeared, but at no point during the year was there any appreciable reported excess of risky sex or risky unprotected sex among circumcised men. These results applied equally to all men and to the subgroup who reported no history of risky unprotected sex at baseline. Correspondingly, there were no differences in average number of risky sex partners per month or condom use among circumcised and uncircumcised men after the first month (Table 4). In addition, the incidence of STI symptoms during follow-up among circumcised men was 0.25 per 100 person months (95% CI: −0.29, 0.80), lower, but not significantly different than the incidence among uncircumcised men (0.59 per 100 person months, 95% CI: −0.25, 1.42).
At 12 months, 86% of study participants remained in follow-up (89% among circumcised men, 83% among uncircumcised men). Those who were lost to follow-up were more likely to be single (22.3% vs. 11.0%), have a lower income (median Kenya Shillings 200 vs. 1000), and to have finished secondary school (53.4% vs. 34.7%) but were not different from those who had completed follow-up with regard to other baseline characteristics. Of the 324 men who were circumcised, the most common reason for coming for circumcision was for protection from STIs/HIV (151 men, 47%). However, there was no difference in risky sexual behavior among those who reported protection from STIs/HIV as reason for circumcision compared to those who reported such other reasons as improving genital hygiene (24%), avoiding injuries during sex (14%), or being influenced by friends (10%).
Unlike the highly controlled research settings where circumcision studies have been conducted, we based our study in a public health care setting where circumcision services are typically offered in Kenya. We found that, except for the first month when uncircumcised men engaged in more risky sex, sexual behaviors of circumcised men associated with increased risk of HIV acquisition did not differ from those of uncircumcised men during the first year post-enrollment/circumcision. Circumcised men likely engaged in less risky sex during the first month because they were healing and also because they were given verbal and written postoperative instructions not to engage in sex for at least 1 month after surgery, even if the penis looked healed or felt painless. The finding of no difference in risky sex behaviors between men who were circumcised and those who remained uncircumcised differs somewhat from the results of a recent RCT,6 which found sexual behavioral factors that might predispose to HIV transmission to have increased slightly in the intervention (circumcised) group than in the control (uncircumcised) group during the first year after study enrollment. However, only the mean number of sexual contacts was significantly different at the month-12 follow-up visit.
In our study, almost half of the men (47%) who underwent circumcision cited protection from STIs/HIV infections as their main reason for doing so. However, after the procedure, there was no difference in risky sexual behavior among these men compared to those circumcised for such reasons as being influenced by friends, improving genital hygiene, and avoiding injuries during sex. In this population, it seems that feeling protected from STIs and HIV through circumcision did not translate into actual sexual risk behaviors. Also, because men who were circumcised were likely to report more risky sex and problems with foreskin, if circumcision interventions are introduced, these groups of men may be the first to take advantage of the service.
Our study had a number of limitations. First, men were not tested for HIV or asked about HIV status before or during the study, and thus we were unable to assess whether HIV status may have confounded our results. Second, the duration of follow-up of 12 months may have been too short to adequately observe long-term sexual behavior after circumcision. There is concern that the protective effect of male circumcision could be diluted over months or years if the circumcised men believe they are protected from infection and engage in high-risk behaviors.7 The ongoing male circumcision trial in Kisumu, Kenya, has been following up participants for 24 to 66 months and will be able to address this concern. Also, frequent visits entailed in this study could have had a dampening effect on risk behavior, which would have reduced differences between the two groups. Finally, we relied on self-reported sexual activity; conceivably, the accuracy of the information provided could differ between circumcised and uncircumcised men.
Studies have been conducted to examine if the introduction of HIV preventive and therapeutic interventions, such as highly active antiretroviral therapy, vaccines, needle-exchange programs, and microbicides, have elicited changes toward more high-risk sexual behaviors for HIV acquisition.8-11 Although concerns have been expressed over potential behavioral disinhibition with regard to male circumcision as a HIV prevention strategy,1,2,4,12-15 to our knowledge, our study is the first to address this concern outside the context of a clinical trial. Our results suggest that, within the context of adequate counseling on risk reduction, any physical benefits arising from circumcision are not likely to be appreciably offset by an adverse behavioral impact of the procedure. Similar studies need to be replicated in other settings to assess whether potential promotion of male circumcision could lead to increased risk behaviors offsetting the potential benefit of the intervention.
This project was supported by NIH Research Grant D43 TW000007, funded by the Fogarty International Center, and Population Services International through support provided by the Global Bureau of Health/HIV-AIDS, U.S. Agency for International Development, under the terms of Award No. HRN-A-00-97-00021-00. We appreciate the role played by Billy Agot, Festus Tiema, John Osodo, and theater staff and administrators of Siaya and Bondo district hospitals in conducting this study. We are also indebted to the participants who made this study possible.
1. Weiss HA, Quigley MA, Hayes RJ. Male circumcision
and risk of HIV infection in sub-Saharan Africa: a systematic review and meta-analysis. AIDS
2. Bailey RC, Plummer FA, Moses S. Male circumcision
and HIV prevention
: current knowledge and future research directions. Lancet Infect Dis
3. Siegfried N, Muller M, Deeks J, et al. HIV and male circumcision
-a systematic review with assessment of the quality of studies. Lancet Infect Dis
4. Agot KE, Ndinya-Achola JO, Kreiss JK, et al. Risk of HIV-1 in rural Kenya
: a comparison of circumcised and uncircumcised men. Epidemiology
5. Cameron DW, Simonsen JN, D'Costa LJ, et al. Female to male transmission of human immunodeficiency virus type 1: risk factors for seroconversion in men. Lancet
6. Auvert B, Taljaard D, Lagarde E, et al. Randomized, controlled intervention trial of male circumcision
for reduction of HIV infection risk: The ANRS 1265 Trial. PLoS Med
. 2005;2:e298 (Epub 2005 Oct 25).
7. Mills E, Menard J, Devereaux PJ, et al. Trial of circumcision for HIV prevention
should be interpreted with caution. E-Letter responding to article: Auvert B, Taljaard D, Lagarde E, Sobngwi-Tambekou J, Sitta R, et al. Randomized, controlled intervention trial of male circumcision
for reduction of HIV infection risk: The ANRS 1265 Trial. PLoS Med
8. Crepaz N, Hart TA, Marks G. Highly active antiretroviral therapy and sexual risk behavior: a meta-analytic review. JAMA
9. Tun W, Gange SJ, Vlahov D, et al. Increase in sexual risk behavior associated with immunologic response to highly active antiretroviral therapy among HIV-infected injection drug users. Clin Infect Dis
10. Donoghoe MC, Stimson GV, Dolan K, et al. Changes in HIV risk behaviour in clients of syringe-exchange schemes in England and Scotland. AIDS
11. Chesney MA, Chambers DB, Kahn JO. Risk behavior for HIV infection in participants in preventive HIV vaccine trials: a cautionary note. J Acquir Immune Defic Syndr Hum Retrovirol
12. Mattson CL, Bailey RC, Muga R, et al. Acceptability of male circumcision
and predictors of circumcision preference among men and women in Nyanza Province, Kenya
. AIDS Care
13. Bailey RC, Muga R, Poulussen R, et al. The acceptability of male circumcision
to reduce HIV infections in Nyanza Province, Kenya
. AIDS Care
14. Quigley MA, Weiss HA, Hayes RJ. Male circumcision
as a measure to control HIV infection and other sexually transmitted diseases. Curr Opin Infect Dis
15. Halperin DT, Fritz K, McFarland W, et al. Acceptability of adult male circumcision
for sexually transmitted disease and HIV prevention
in Zimbabwe. Sex Transm Dis