Three randomized trials demonstrated that male circumcision (MC) reduces HIV acquisition by 51% to 60%.1–3 The World Health Organization (WHO) and the Joint United Nations Programme on HIV/AIDS recommend encouraging voluntary MC as part of a comprehensive strategy for reducing heterosexual transmission of HIV in regions with HIV epidemics and low MC prevalence.4 Furthermore, ministries of health in several sub-Saharan African countries have proposed reaching 80% MC coverage among men 15 to 49 years old by 2015.5
The MC trials demonstrated that MC is associated with reduced HIV as well as reduced low-risk and high-risk human papillomavirus (HR-HPV), herpes simplex virus type 2 (HSV-2), and genital ulcer disease (GUD).6 In addition, female partners of circumcised men were found to have reduced rates of low-risk HPV, HR-HPV, trichomoniasis, bacterial vaginosis (BV), and GUD.6–8
Although mathematical models have shown that MC scale-up may avert substantial costs by reducing HIV,9,10 no study has analyzed the comprehensive impact of MC in sub-Saharan Africa, incorporating MC’s effects on other sexually transmitted infections (STIs). With recently proposed funding cuts from the United States President’s Emergency Plan for AIDS Relief and increasing obligations of donor agencies, rigorous cost and effectiveness evaluations of programs to reduce HIV/AIDS and other STIs are needed to guide global health policy.
This study analyzes the health care costs and savings associated with MC, accounting for all MC-reduced infections and evaluating outcomes under alternative potential MC scale-up plans. It focuses on the Rakai District, Uganda, a prototypical rural sub-Saharan African community, to assess short- and long-term financial implications of MC scale-up.
A dynamic Markov model was constructed using TreeAge Pro (Wiliamstown, MA) to assess changes in STI-related health care costs associated with MC scale-up, accounting for impact on men and women in the Rakai District, Uganda. Monte Carlo microsimulations were used to compare 4 strategies, varying by the scope and rate of MC scale-up, to a baseline scenario, defined by a continuation of current prevalence and trends of MC (23.6% of 15- to 49-year-old men circumcised).11
MC Procedures and MC-Reduced Infections
This model simulated MC scale-up incorporating infant (age, 0–1 year) and adolescent/adult (age, 15–49 years) MC procedures and accounted for any potential procedure-related adverse events. In addition, it tracked all STIs among men and women demonstrated through randomized trials to be reduced by MC.3,6–8,12 Risk of acquiring each STI was assumed to be independent of infection history (individuals who had previously acquired particular STIs did not face an elevated risk from other infections). The model did not incorporate potential secondary transmission events after initial infections, and MC protection among females paralleled MC-status among males; the proportion of females experiencing protective benefits was similar to the proportion of circumcised males. Direct medical expenses associated with MC procedures, procedure-related adverse events, and STIs were accumulated through the simulation.
Male circumcision–reduced infections modeled among men included heterosexually transmitted HIV (excluding mother-to-child transmission), GUD, and HR-HPV. Although the Ugandan MC trial also found reduced HSV-2 among circumcised men, most etiologically defined GUD was attributed to HSV-2.13 Thus, HSV-2 was not incorporated as a separate STI. Male circumcision–reduced infections modeled among women included GUD, HR-HPV, trichomoniasis, and BV. Low-risk HPV was not incorporated because its primary symptomatic presentation, genital warts, is not often treated in sub-Saharan Africa.
Individual microsimulations followed individuals through the model in a series of 1-year cycles, tracking MC procedures, and MC-reduced infections. The model (illustrated in Fig. 1A) simulated an “initial population,” defined by current age and sex profiles, supplemented annually by a birth cohort, defined by birth rate14 and sex distribution at birth.15 These background parameters were assumed to remain constant. A dynamic population was explicitly generated by assigning individuals to either the initial population or to a birth cohort, based on the ratio of the current population size to the total number of individuals expected to be part of the population during the period of simulation and assuming that the size of each annual birth cohort remained constant. Further specification on how the dynamic population was explicitly generated is provided in the supplemental text (http://links.lww.com/OLQ/A65). National age- and sex-specific mortality16 was also incorporated through a terminal Markov state to account for death because of background or STI causes.
Individuals assigned to a birth cohort were not initiated into the model until an individual-specific birth year. Individuals of the initial population were assessed at the beginning of the simulation for existing infections and MC protection status, using STI prevalence estimates11,17 and current MC rates in central Uganda.11
Each potential health condition (≥1 STIs) was modeled as a sex-specific Markov state, with associated costs incorporated as incremental rewards (accrued each time an individual experienced a particular state) or as transition rewards (accrued each time an individual passed between particular states; Fig. 1B). Transition probabilities between states were defined by incidence and MC efficacy, such that individuals protected by MC faced a reduced risk of transitioning to particular STI Markov states. Individuals with initial infections populated the corresponding Markov states at the start of the simulation.
The simulation was analyzed separately for 5 and 25 years to assess short- and long-term impact of potential MC scale-up. Only costs incurred during these periods were incorporated. Furthermore, input parameters (described later) were assumed to remain constant. Costs were discounted to the beginning of the analysis period, corresponding to the start of scale-up, and adjusted to US$2012 using the medical care component of the Consumer Price Index.18
MC Protection Status
Each MC scale-up strategy was defined by 2 components: one targeting uncircumcised males of the initial population (“catch-up MC”) and another targeting males of future birth cohorts.9,19 Male circumcisions under each program were performed among infants (age, 0–1 year) or among adolescents/adults (age, 15–49 years). Evaluated strategies (Table 1) varied by the ages of males becoming circumcised, as well as by the rate of scale-up (rapid or gradual). Although strategy 1 reflected a continuation of baseline MC rates, without catch-up MCs or any change in MC trends among future birth cohorts, strategies 2 to 5 defined alternative MC scale-up plans.
Under strategies 2 to 5, catch-up MCs were conducted until 80% coverage was achieved among men aged 15 to 49 years, but the rate of scale-up (rapid or gradual) varied. Under strategies 2 and 4, 10 years of scale-up was needed to reach 80% coverage, whereas under strategies 3 and 5, this level of coverage was achieved after only 5 years. Further details on identifying individuals who would gain MC protection through this “catch-up” program component are provided in the supplemental text (http://links.lww.com/OLQ/A65).
Among annual birth cohorts, MC rates differed by scale-up strategy and birth year, with men born later experiencing greater MC rates than those born earlier. For simplicity, the simulation period was divided into 5 stages, where each stage consisted of a 5-year period of scale-up. Among the men born during these stages, MC procedures were assumed to be conducted as infants (age, 0–1 year), adolescents (age, 15 years), or adults (age, 25 years). Under strategy 1 (baseline), 23.6% of males were circumcised, with 13.6% obtaining infant MCs, 5% obtaining adolescent MCs, and 5% obtaining adult MCs. These rates persisted through stage 5 of the simulation. However, under strategies 2 to 5, the MC rate increased through the course of the simulation, such that 80% of males born during stage 5 would ultimately become circumcised. Under strategies 2 and 3, infant MCs occurred at the same rate as under the baseline strategy (13.6%), but adolescent and adult MCs were scaled up, with the adolescent MC rate ultimately reaching 40% and the adult MC rate ultimately reaching 26.4%. Strategies 4 and 5 focused on the scale-up of infant MCs (in addition to scaling up adolescent and adult MCs), such that 80% of men born during stage 5 were circumcised as infants. Strategy 2 characterized a more gradual scale-up than strategy 3, and strategy 4 characterized a more gradual scale-up than strategy 5.4,5
Input parameters (Table 2) describe the costs, prevalence, and incidence of MC procedures and MC-reduced infections and were drawn from existing literature, demographic surveys, and global databases. Efficacy estimates were drawn from the incidence rate ratios (IRRs) and prevalence risk ratios (PRR) reported in randomized trials in Rakai, Uganda, which evaluated the impact of MC on male participants and their female partners. These efficacy parameters described the effect of MC on an individual’s acquisition, but did not explicitly describe the extent to which transmission itself was reduced. Trials in South Africa and Kenya found similar efficacy results.1,2,20,21
Infection prevalence and incidence were age-specific, whenever possible. Declines in these rates were modeled for the ages below or beyond the age range of the group studied in the original data source (ie, HIV rates were 0 at birth and declined to 0 by age 75 years). Where age-specific parameters were unavailable, rates were assumed to be constant over the age range reported. Incidence and prevalence rates specific to MC protection status were derived from published incidence and prevalence rates to account for an existing portion of the current population who is already experiencing protective effects from MC. Details are provided in the supplemental text (http://links.lww.com/OLQ/A65).
Genital ulcer disease incidence, for men and women, was assumed to be equivalent to annual GUD prevalence, as reported in the Ugandan 2006 Demographic and Health Survey.22
Because BV is a transient state23 and incidence estimates were unavailable, BV was modeled as a potentially prevalent infection and not as a Markov health state. A parameter defining age at infection was drawn from a triangular distribution (mode, 24.7; minimum, 15; maximum, 59). A female with BV infection accumulated associated costs if she became infected during the period of analysis.
Only HR-HPV cases that led to cervical cancer (among females) or penile cancer (among males) were modeled because these cases drive costs associated with HR-HPV. Because both of these cancers develop at least 5 to 10 years after HR-HPV infection,24 when evaluating individuals for incidence, reduced risk was only incorporated if MC protection had been acquired before HR-HPV infection.
Because data on costs associated with HPV-induced penile cancer in Uganda were unavailable, the cost parameter associated with penile cancer was extrapolated from a cost estimate from the United States,25 using a ratio of intercountry costs for cervical cancer.26,27 This extrapolation assumed that the ratio of costs for cervical cancer between Uganda and the United States is roughly equivalent to the intercountry ratio of costs for penile cancer.
Costs associated with GUD, BV, and trichomoniasis were incorporated in a piecewise method, gathering individual components from the WHO’s “CHOICE” database,28 published literature, and the International Drug Price Indicator Guide.29 Each infection was associated with costs from care conducted according to the Ugandan Clinical Treatment Guidelines for STIs,30 including expenses for a visit to a primary health center, as well as costs for the appropriate medical treatment (acyclovir, metronizadole and clotrimazole, or metronidazole, respectively). Additional model parameters accounted for the portion of infections that would not cause symptoms or cause treatment seeking. Potential recurrence of BV (but none of the other STIs) was incorporated. For HIV, potential patient loss to follow-up was also incorporated. Furthermore, HIV-associated expenses incurred before initiating antiretroviral treatment (ART) were distinguished from post-ART expenses. Similarly, costs associated with HPV-induced cervical cancer incorporated separate values for stages I and II and for advanced stages. Nononcogenic or precancerous effects of HPV were not incorporated.
Under the base-case scenario, strategies were analyzed over each period running 500,000 individual microsimulations (simulating 500,000 individuals). Cost and health outcomes for the district were extrapolated by calculating the total population present during the analysis period (from the initial population and each annual birth cohort). Cost outcomes were also separated by attributable infection to identify primary components of financial impact. Probabilistic sensitivity analysis, using 1000 samples of 10,000 trials each, was conducted to evaluate the impact of uncertainty in sample-level input parameters. Table 2 shows the ranges and/or distributions characterizing each varied input parameter. Costs were varied by 25% in either direction using an adjustment factor sampled from a triangular distribution (mode, 1; minimum, 0.75; maximum, 1.25). Efficacy estimates and other rates were drawn from beta distributions, using the 95% confidence interval (CI) reported by the original data source.
Under the base-case scenario, cost savings from averted MC-reduced infections vary from US$0.20 million (US$197,531) after 5 years of scale-up focusing on adolescent and adult procedures to US$13.71 million after 25 years, under a strategy incorporating increased infant MCs (Table 3). Under the strategies evaluated, net savings, accounting for scale-up costs and reduced STIs, is not expected to occur until after 10 years.
For all periods evaluated, changes in infection costs are more substantial under strategies incorporating infant MC (strategies 4 and 5) than those focusing exclusively on adolescent and adult procedures (strategies 2 and 3), and the differences in cost savings is expected to increase over time. Furthermore, although under the 5-year analysis period, the net change in costs is positive and strategies incorporating infant MCs are more costly than those focusing only on adolescent and adult procedures, MC scale-up becomes cost-saving over a longer period of analysis, and savings under strategies including infant MCs are ultimately more substantial.
Comparing outcomes between strategies of gradual and accelerated scale-up strategies demonstrates that accelerated strategies are associated with greater cost savings from averted MC-reduced infections. In the short term, accelerated scale-up is associated with a greater (positive) net cost but would result in greater net cost savings in the long term.
Further analysis of cost savings associated with averted MC-reduced infections, shown in Table 4, shows the portion of the total infection cost attributable to each MC-reduced infection, as well as the portion of the total change in infection cost attributable to reductions in each infection. The portion of total costs attributable to each infection is reported as an average across all strategies (strategies 1–5), and the portion of the change in costs is reported as an average across all scale-up strategies (strategies 2–5). Attributable fractions specified for each strategy, under the base-case scenario, are provided in Supplemental Table 1 (http://links.lww.com/OLQ/A65). In the short term, after 5 years of scale-up, reduced HIV contributes 46.24% of MC-reduced infection costs and 50.04% of cost savings. By 25 years, HIV contributes 78.77% of costs and 89.77% of cost savings. In addition to HIV, GUD among males and females, HPV-associated cervical cancer, and trichomoniasis contribute significantly to MC-reduced infection costs.
The cost to avert a single HIV infection was calculated for each period evaluated, using the difference in costs of MC scale-up and prevalence of HIV infections between scale-up strategies and the baseline strategy. Over a 5 year period, the cost to avert a single HIV infection is US$5500 to US$9200, but over 25 years, this cost is less than US$500.
Financial outcomes from probabilistic sensitivity analysis are described in Table 5. Variation of input parameters does not substantially alter results. Cost savings from averted MC-reduced infections varies from US$0.30 million (US$295,514) after 5 years under strategy 2, a gradual scale-up focusing on adolescent and adult MCs, to US$14.73 million after 25 years under strategy 5, an accelerated scale-up incorporating infant MCs.
Male circumcision has been demonstrated to be an effective preventative intervention for reducing HIV incidence.1–3,31 As a result, the WHO, Joint United Nations Programme on HIV/AIDS, and other health agencies have promoted plans for MC scale-up.4 However, additional medical benefits of MC exist—reduced HPV, HSV-2, and GUD among men6,12,20,21 and reduced HPV, BV, trichomoniasis, and GUD among female partners.6–8 This model suggests that previous cost analyses of MC considering only averted HIV have underestimated financial savings by 10% to 50%. Although these findings are unlikely to dramatically alter policy, in a context of tightened budgets, financial assessment of MC-related programs that incorporates a comprehensive set of effects is essential.
This analysis is the first to evaluate the health and financial impact of MC in sub-Saharan Africa, incorporating all STIs among males and females shown through randomized controlled trials to be reduced by MC. Results suggest that accelerated scale-up may lead to cost savings from averted infections earlier than gradual scale-up and that programs increasing infant MC procedures are associated with greater savings and fewer infections in the long-term than programs incorporating only adolescent or adult MC procedures.
The model used in this analysis allowed for substantial flexibility in the scenarios evaluated, particularly because strategy-related parameters varied both by the age of an individual at the time of MC and by the stage of the scale-up program. Strategies were defined in consideration of recent MC scale-up proposals and trends among existing MC programs in sub-Saharan Africa. Although rapid scale-up incorporating increased infant MC procedures over time has been advocated, gradual scale-up focusing on adolescent/adult MC procedures may be more practical in some settings.4,31 Although infant MC is less costly and circumvents the possibility of sexual activity before the procedure, it may be difficult to implement because many births in sub-Saharan Africa occur outside formal health facilities. Furthermore, although adolescent or adult men may obtain the procedure on their own, infants rely on parents or caretakers. The additional costs or burden associated with implementing infant MC procedures has not been incorporated in this analysis.
A prior analysis suggested that a rapid scale-up of MC in 14 target African countries that reached 80% coverage among newborns and males aged 15 to 49 years within 6 years of scale-up would ultimately result in 19% fewer cases of HIV and a comprehensive net savings of US$20.2 billion for 26 years.32 It is difficult to effectively compare these results with those reported here because of slight variations in context, scale, and evaluated scale-up strategies. However, these data support our findings that investment in infant MCs as a component of scale-up proposals can have substantial health and financial impact in the long-run.
This study has several limitations. Because input parameters used in the model characterized Rakai, Uganda, it may not be appropriate to extrapolate results to other regions in sub-Saharan Africa. Furthermore, the model likely underestimates the impact of MC scale-up. First, although all STIs associated with MC were incorporated, the transmission dynamics between men and women leading to infection were not explicitly modeled. Incidence rates did not vary by infection prevalence and were held constant throughout the period analyzed. If incidence increased from initial rates because of increased prevalence, health and financial outcomes may be underestimated.
The presented outcomes assume a continuation of the cost levels, incidence rates, and MC efficacy estimates used as input parameters. Because these input parameters may vary over time, the long-term health and financial outcomes estimated by this model may be less certain than short-term estimates. Furthermore, because MC preventative efficacy increases over time,6 long-term averted infections and associated costs may be underestimated. Potential behavioral changes associated with MC were not incorporated in this model because several studies have suggested that MC is not associated with any adverse postprocedural behavioral modifications.2,3,33
The only MC-reduced infections incorporated were those shown by the Ugandan randomized trials to be associated with MC. However, other studies have suggested that MC is also associated with reduced infant urinary tract infections,34 syphilis, chlamydia, chancroid, and Mycoplasma genitalium.34–36 Potential averted costs from these infections or from additional transmission events were not included. Although the Ugandan trial did not demonstrate reduced HIV among female partners of male MC participants,37 this finding is likely caused by resumption of sexual activity before full healing from the MC procedure. Other studies suggest that with proper postoperative care, reduced HIV among men will indirectly affect women (and, subsequently, other men) through reduced transmission.4,38 Incorporating potential transmission events would likely increase the portion of costs attributable to HIV.
Efficacy estimates for STIs other than HIV were based on a study of HIV-negative men and their female partners. Because it is not clear how preexisting male HIV infection affects MC-protective efficacy for all other STIs evaluated and because transmission dynamics were not explicitly modeled, this analysis did not incorporate the potential effects of male HIV status on MC efficacy estimates.
Efficacy input parameters were drawn exclusively from randomized trials in Rakai, Uganda, although trials in Kenya and South Africa demonstrated similar results.1,2,21 Although the Kenyan trial did not show reduced HSV-2 among male MC participants, it did find a similar reduction in GUD,39 which was assumed to incorporate symptomatic HSV-2 in this model.
Finally, this analysis included direct medical costs only, excluding indirect and nonmedical costs from productivity loss, transportation, infrastructure, and caregiver or patient time. These other costs may be substantially higher than the direct medical costs considered.
This analysis suggests that short- and long-term health and financial impact of MC may be more substantial than previously reported. Furthermore, strategies achieving target MC rates at an accelerated rate may become cost-saving earlier than more gradual scale-up strategies, and scale-up programs which incorporate infant MC procedures may have a greater long-term impact than programs focusing exclusively on adolescent and adults. In an environment of increasingly tightened health care budgets, this assessment of the comprehensive impact of MC further supports the health and financial benefit of MC scale-up as a component of STI-related programs.
1. 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.
2. 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: 643–656.
3. Gray RH, Kigozi G, Serwadda D, et al. Male circumcision for HIV prevention in men in Rakai, Uganda: A randomised trial. Lancet 2007; 369: 657–666.
4. UNAIDS. Joint Strategic Action Framework to Accelerate the Scale-Up of Voluntary Medical Male Circumcision for HIV Prevention in Eastern and Southern Africa: 2012-2016. Geneva: UNAIDS: Joint United Nations Programme on HIV/AIDS, 2011.
5. Ministry of Public Health and Sanitation. Progress report on Kenya’s Voluntary Medical Male Circumcision Programme, 2008–2010. NASC , ed. Nairobi, Kenya: Government of Kenya, 2012.
6. Tobian AA, Gray RH. The medical benefits of male circumcision. JAMA 2011; 306: 1479–1480.
7. Wawer MJ, Tobian AA, Kigozi G, et al. Effect of circumcision of HIV-negative men on transmission of human papillomavirus to HIV-negative women: A randomised trial in Rakai, Uganda. Lancet 2011; 277: 209–218.
8. Gray RH, Kigozi G, Serwadda D, et al. The effects of male circumcision on female partners’ genital tract symptoms and vaginal infections in a randomized trial in Rakai, Uganda. Am J Obstet Gynecol 2009; 200: 42 e41–e47.
9. Hankins C, Forsythe S, Njeuhmeli E. Voluntary medical male circumcision: an introduction to the cost, impact, and challenges of accelerated scaling up. PLoS Med 2011; 8: e1001127.
10. UNAIDS/WHO/SACEMA Expert Group on Modelling the Impact and Cost of Male Circumcision for HIV Prevention. Male circumcision for HIV prevention in high HIV prevalence settings: What can mathematical modelling contribute to informed decision making? PLoS Med 2009; 6: e1000109.
11. Ministry of Health (MOH) Uganda and ORC Macro. Uganda HIV/AIDS Sero-behavioral Survey 2004–2005. Calverton, MA, USA: Ministry of Health and ORC Macro, 2006.
12. Tobian AA, Serwadda D, Quinn TC, et al. Male circumcision for the prevention of HSV-2 and HPV infections and syphilis. New Engl J Med 2009; 360: 1298–1309.
13. Suntoke TR, Hardick A, Tobian AA, et al. Evaluation of multiplex real-time PCR for detection of H. ducreyi
, T. pallidum
, HSV-1, and HSV-2 in the diagnosis of genital ulcer disease in Rakai District, Uganda. Sex Transm Infect 2009; 85: 97–101.
14. The World Bank. World DataBank: World Development Indicators & Global Development Finance, 2010.
15. Uganda Bureau of Statistics. Projections of Demographic Trends in Uganda: 2007–2017, 2007.
17. Hladik W, Musinguzi J, Kirungi W, et al. The estimated burden of HIV/AIDS in Uganda, 2005–2010. AIDS 2008; 22: 503–510.
18. United States Department of Labor. Bureau of Labor Statistics. Consumer Price Index—All Urban Consumers. 2013. Available at: http://data.bls.gov/cgi-bin/surveymost
. Accessed January 22, 2013.
19. UNAIDS. Scaling-up male circumcision programmes in the Eastern and Southern Africa Region: Country update meeting to share lessons, explore opportunities and overcome challenges to scale-up. Paper presented at: Joint United Nations Programme on HIV/AIDS; 2010; Arusha, Tanzania.
20. Tobian AA, Gray RH, Quinn TC. Male circumcision for the prevention of acquisition and transmission of sexually transmitted infections: The case for neonatal circumcision. Arch Pediatr Adolesc Med 2010; 164: 78–84.
21. Auvert B, Sobngwi-Tambekou J, Cutler E, et al. Effect of male circumcision on the prevalence of high-risk human papillomavirus in young men: Results of a randomized controlled trial conducted in orange farm, South Africa. J Infect Dis 2009; 199: 14–19.
22. Uganda Bureau of Statistics (UBOS) and Macro International Inc. Uganda Demographic and Health Survey 2006. Calverton, MD, USA: UBOS and Macro International Inc, 2007.
23. Thoma ME, Gray RH, Kiwanuka N, et al. The short-term variability of bacterial vaginosis diagnosed by Nugent Gram stain criteria among sexually active women in Rakai, Uganda. Sex Transm Dis 2011; 38: 111–116.
24. Schiffman M, Kjaer SK. Chapter 2: Natural history of anogenital human papillomavirus infection and neoplasia. J Natl Cancer Inst Monogr 2003; (31): 14–19.
25. Hu D, Goldie S. The economic burden of noncervical human papillomavirus disease in the United States. Am J Obstet Gynecol 2008; 198: 500 e501–e507.
26. Goldie SJ, O’Shea M, Campos NG, et al. Health and economic outcomes of HPV 16,18 vaccination in 72 GAVI-eligible countries. Vaccine 2008; 26: 4080–4093.
27. Goldhaber-Fiebert JD, Stout NK, Salomon JA, et al. Cost-effectiveness of cervical cancer screening with human papillomavirus DNA testing and HPV-16,18 vaccination. J Natl Cancer Inst 2008; 100: 308–320.
29. Frye JE. International Drug Price Indicator Guide. Management Sciences for Health, 2011.
30. Sexually Transmitted Infections Treatment Guidelines for Use by Operational Level Health Workers in Uganda. STD Control Unit. STD/AIDS Control Programme. Ministry of Health U, 2003.
31. UNAIDS. World Health Organization Technical Consultation on Male Circumcision and HIV Prevention: Policy and Programme Implications. Conclusions and Recommendations. World Health Organization/Joint United Nations Programme on HIV/AIDS: Geneva: World Health Organization, 2007.
33. Gray R, Kigozi G, Kong X, et al. The effectiveness of male circumcision for HIV prevention and effects on risk behaviors in a posttrial follow-up study. AIDS 2012; 26: 609–615.
34. Schoen EJ, Colby CJ, Ray GT. Newborn circumcision decreases incidence and costs of urinary tract infections during the first year of life. Pediatrics 2000; 105 (4 pt 1): 789–793.
35. Weiss HA, Thomas SL, Munabi SK, et al. Male circumcision and risk of syphilis, chancroid, and genital herpes: A systematic review and meta-analysis. Sex Transm Infect 2006; 82: 101–109; discussion 110.
36. Mehta SD, Gaydos C, Maclean I, et al. The effect of medical male circumcision on urogenital Mycoplasma genitalium
among men in Kisumu, Kenya. Sex Transm Dis 2012; 39: 276–280.
37. Wawer MJ, Makumbi F, Kigozi G, et al. Circumcision in HIV-infected men and its effect on HIV transmission to female partners in Rakai, Uganda: A randomised controlled trial. Lancet 2009; 374: 229–237.
38. Weiss HA, Hankins CA, Dickson K. Male circumcision and risk of HIV infection in women: A systematic review and meta-analysis. Lancet Infect Dis 2009; 9: 669–677.
39. Mehta SD, Moses S, Parker CB, et al. Circumcision status and incident herpes simplex virus type 2 infection, genital ulcer disease, and HIV infection. AIDS 2012; 26( 9): 1141–1149.