Adult male circumcision: results of a standardized procedure in Kisumu District, Kenya


John N. Krieger, Department of Urology, University of Washington School of Medicine, Box 356510, Seattle, WA 98195, USA.


In a combined study from the USA and Kenya, the safety of adult male circumcision in the latter country was reviewed, particularly with the purported association between this procedure and a lower incidence of HIV and other sexually-transmitted infections. It was found that safe and acceptable adult male circumcision services could be delivered in developing countries should this be advocated as a public health measure.


To develop a standard procedure for male circumcision in a resource-poor medical setting and prospectively evaluate the outcome in a randomized, controlled trial with the incidence of human immunodeficiency virus (HIV) as the main outcome, as studies suggest that circumcision is associated with a lower incidence of HIV and other sexually transmitted infections in high-risk populations.


Healthy, uncircumcised, HIV-seronegative men aged 18–24 years from Kisumu District, Kenya, were offered participation in a clinical trial using a standard circumcision procedure based on ‘usual’ medical procedures in Western Kenya. The follow-up included visits at 3, 8 and 30 days after circumcision, with additional visits if necessary. Healing, satisfaction and resumption of activities were assessed at these visits and 3 months from randomization.


Overall, 17 (3.5%) of the 479 circumcisions were associated with adverse events judged definitely, probably or possibly related to the procedure. The most common adverse events were wound infections (1.3%), bleeding (0.8%), and delayed wound healing or suture line disruption (0.8%). After 30 days, 99% of participants reported being very satisfied with the procedure; ≈ 23% reported having had sex and 15% reported that their partners had expressed an opinion, all of whom were very satisfied with the outcome. About 96% of the men resumed normal general activities within the first week after the procedure.


Safe and acceptable adult male circumcision services can be delivered in developing countries should male circumcision ultimately be advocated as a public-health measure.


sexually transmitted infection


Kenya shilling.


Of African men, ≈ 62% are circumcised [1]; more African men are seeking clinical circumcision to improve hygiene, to reduce the risk of sexually transmitted infections (STIs), to appear more ‘urban’, or to be more like others [2–6]. Epidemiological studies suggest that male circumcision is associated with a ≥ 50% reduction in the risk of HIV infection, and with a lower risk of ulcerative STI [4,7–10]. Male circumcision eliminates the genital problems of balanoposthitis, phimosis and paraphimosis, substantially reduces penile carcinoma rates and, in one study, reduced the cervical cancer risk in female partners [3].

Despite increased interest among medical authorities and greater acceptance among African men, there are substantial concerns about recommending male circumcision as a public-health measure. These reservations include the lack of data from randomized, controlled clinical trials showing that circumcision decreases the risk of STIs and HIV. Considerable anecdotal evidence, including reports based on medical record reviews, suggests that circumcision in traditional settings exposes young men to an excessive risk of complications [11–15]. Very few data on circumcision outcomes are available from African clinical settings. One study from Tanzania found that infant circumcision using the Plastibell technique resulted in a 2% complication rate [16], comparable to reported rates in developed countries [17].

We are conducting a prospective, randomized controlled clinical trial of adult male circumcision to evaluate its impact on the incidence of HIV among 18–24-year-old men in Kisumu District, Kenya. One important objective was to develop a standard circumcision procedure that would be applicable to resource-poor medical settings. A second objective was to determine the adverse event rates associated with this procedure. Thus the purposes of this report are to: (i) describe the technical aspects of the standard procedure; (ii) outline the procedures for monitoring and classifying adverse events; and (iii) to describe outcomes and adverse events associated with the first 479 circumcisions.


Uncircumcised, HIV-seronegative men aged 18–24 years residing within Kisumu District, Kenya, were offered admission to the trial. After providing written, informed consent, participants were randomized equally to either circumcision or delayed circumcision after a 2-year follow-up (the control group). Exclusion criteria included: any medical indication for circumcision (e.g. significant phimosis, recurrent balanitis, paraphimosis, or extensive genital warts), history of a bleeding disorder, keloid formation, or other condition that might unduly increase the risks of elective surgery. Men in both arms of the trial were given extensive counselling about the reduction in risk of STI/HIV, including the provision of unlimited supplies of free condoms.

This study was approved by institutional review boards at the Universities of Nairobi, Illinois at Chicago, Manitoba, Washington, and RTI International. A Data Safety and Monitoring Board reports to the USA National Institute of Allergy and Infectious Diseases.

For circumcision, a standard procedure was based on ‘usual’ medical circumcision procedures in Western Kenya [18]. One Medical Officer (medical doctor) and two Clinical Officers (Kenyan paramedicals who conduct medical consultations and selected surgical procedures) were trained in techniques and details of the procedure. All instruments and supplies were purchased from suppliers in Western Kenya.

The genitalia were prepared with a povidone-iodine solution and draped in sterile fashion. A dorsal nerve block was applied using 2% xylocaine, with additional anaesthetic administered circumferentially about the penile base (maximum 10 mL). The prepuce was grasped at the 3 and 9 o’clock positions using two mosquito clamps, then pulled over the glans. The incision was outlined with a marking pen parallel to and 1 cm proximal to the coronal sulcus. The mosquito clamps were then used to pull the prepuce over the glans while a Kocher clamp was applied below the planned incision, taking care to avoid injuring the glans. The prepuce was excised by cutting above the Kocher clamp, which was then removed. Bleeding vessels were ligated using 3/0 plain gut. The wound was irrigated with physiological saline.

The skin and mucosal incisions were aligned using 3/0 chromic vertical mattress sutures at the 12, 3 and 9 o’clock positions, and a horizontal mattress suture at the 6 o’clock (frenulum) position. The four initial sutures were tagged with mosquito clamps to provide traction during the remainder of the closure. Four additional vertical mattress sutures were placed midway between the initial sutures, with simple sutures then placed to alternate between the mattress sutures. The wound was covered using a Vaseline gauze dressing.

Each patient was given verbal and written instructions, including clothing and bathing practices, and instructed to urinate frequently to minimize erections. They were given acetaminophen, 1 g three times daily as needed for 5 days, and tramadol hydrochloride, 100 mg at bedtime as needed for two nights. Participants were strongly counselled to abstain from sex for at least a month.

Follow-up visits were scheduled for 3, 8 and 30 days after surgery. Patients were instructed to return to the clinic if there was bleeding, infection or excessive pain. At each visit the wound was checked, and standard questions assessed activities of daily living, potential adverse events, and the participants’ satisfaction. Additional detailed evaluations were conducted at 1, 3, 6, 12, 18 and 24 months from randomization.

Regular case conferences reviewed all procedures and outcomes. Adverse events were categorized by severity and relatedness to the study. Severity categories were mild, moderate and severe. Relatedness to the procedure was classified as: (1) definitely unrelated, clearly explained by another cause; (2) possibly related, likely explained by other causes; (3) probably related, most likely explained by the circumcision; and (4) definitely related, a clear complication.

Patient satisfaction was analysed at ≈ 30 and 90 days after surgery as (1) very satisfied, (2) somewhat satisfied, (3) somewhat dissatisfied, or (4) very dissatisfied. We evaluated whether patients had resumed work and sexual activities and, if so, whether their sexual partners had expressed an opinion about their circumcision status.

The critical endpoints and definitions were determined before starting the study. Data were collected prospectively and entered on site, and evaluated for quality and consistency continuously.


Between 4 February 2002 and 11 March 2004, 2160 men were screened, including 47% who agreed to participate. Of these, 505 were randomized to circumcision, including 479 (95%) who had the procedure. The remainder either declined after randomization or deferred circumcision until later. The median age of circumcised participants was 21 years and 96% were of Luo ethnic origin; 89% had completed primary school, 53% had completed secondary school and >90% were unmarried. Almost half classified themselves as unskilled workers and about a quarter were students. Only 37% reported regular employment.

The median (range) duration of circumcision was 30 (18–63) min; Fig. 1 shows the duration for each successive group with time (Cochran-Mantel-Haenszel rank correlation statistic, chi-square 149.4, 1 d.f., P < 0.001).

Figure 1.

Box-and-whisker plot comparing procedure duration with the number of circumcisions. For each group of circumcisions, the line in the box indicates the median (50th percentile), and the upper and lower bounds of the box the interquartile range (25–75th percentiles). The error bars indicate the 10th and 90th percentiles. Values above the 90th or below the 10th percentiles are plotted as points. For all 479 circumcisions, the median (range) duration was 30 (18–63) min; the median duration was 38 min for the first 100, 30 min for the second 100, 30 min for the third, 28 min for the fourth and 26 min for procedures 401–479 (Cochran-Mantel-Haenszel rank correlation P < 0.001).

Of the 479 circumcisions, 17 (3.5%) were associated with 18 adverse events considered definitely, probably or possibly related, including six wound infections (1.3%), four bleeds (0.8%), four delayed healing or disrupted wounds (0.8%), two with excessive swelling (0.4%), one anaesthetic reaction, and one participant who reported erectile dysfunction.

Of the 18 adverse events, none was classified as severe, 11 were moderate (requiring additional treatment, e.g. drainage of an infection or revising a suture to control untoward bleeding), and seven were mild (requiring minimal additional treatment). Fourteen adverse events were considered definitely related (including most infections and bleeding problems, plus the anaesthetic reaction), one was considered probably related (an infection away from the wound that represented a reaction to the dressing tape), and three possibly related (including diminished erectile function reported, folliculitis at the pubic base, and a pubic abscess in one participant each). Overall, nine of 18 adverse events were considered to be both moderate and definitely/probably procedure-related, occurring after eight of 479 procedures (1.7%).

The cosmetic results were excellent, with no need for revisions or secondary procedures. There were no cases of excess preputial skin removal, inadequate preputial skin removal, penile torsion, keloid, other poor cosmetic or functional outcomes.

At this analysis, 411 of 479 participants were evaluated at ≈ 30 days after surgery; 99% reported being very satisfied, 23% reported having had sex since the circumcision, and 15% of these men reported that their female partner had expressed an opinion about the procedure, all of whom were very satisfied with the outcome. In all, 321 men had assessments at ≈ 90 days from surgery, when 65% reported having resumed sexual intercourse, and 54% of their sex partners had expressed an opinion about the procedure. Of these partners, 95% were very satisfied with the outcome, 4% were somewhat satisfied and 2% were somewhat dissatisfied. None was very dissatisfied.

At 30 days after surgery, 98% of men who were employed reported having returned to work, 91% within a week of the procedure and 97% within 2 weeks. At this visit, all men reported having resumed normal general activities, 96% within the first week and another 3% during the second week after surgery.


We developed a circumcision procedure that was implemented effectively in an African medical setting. All instruments and supplies were purchased locally and local clinicians performed all procedures. The standard method was based on medical procedures used in eastern Africa [18]. To avoid reliance on a technology that is not widely available in developing countries, we did not use electrocautery. Nevertheless, the overall adverse event rate was comparable to reported rates from the developed world. Postoperative wound infection (1.3%) and bleeding (0.8%) were the most common adverse events, as would be predicted. The vast majority of participants and their partners who expressed opinions were very satisfied at 30 and 90 days after surgery.

This is one of the largest series of adult male circumcisions and it is unique in that outcomes were monitored prospectively. Participants were followed using a rigorous protocol. We used simple definitions that may prove useful for other surgical trials, especially studies evaluating circumcision as a public-health measure.

The median time to complete a circumcision was 30 min; operating duration decreased from a median of 38 min for the first 100 procedures to 26 min for procedures 401–479 (P < 0.001). This is important, because if circumcision were recommended as a public-health measure for preventing HIV, then it would be necessary to complete procedures within an acceptable period in resource-poor countries. The instruments and supplies were modest and all were obtained locally. In Kisumu the cost of supplies is 1500 Kenya shillings (KES) per procedure at our facility (≈ US $20). In the Kisumu community, charges for circumcision range from 1000 KES (US $13) at the government hospital to 5800 KES (US $77) at the leading private hospital. Thus, this study suggests that our standard circumcision method is consistent with existing practice in Kenya and is likely to prove practical within the developing world. We also showed that it is possible to obtain results that are equivalent to series of adult male circumcisions from the developed world.

The present study has some advantages; it represents a large series of adult male circumcisions and one of very few reports of studies designed to provide an optimum prospective assessment of outcomes. We used rigorous definitions for events. The circumcision procedure and resources were based on local medical practice and resource availability. All instruments and supplies were obtained locally, and clinicians trained and working in their own community completed all procedures. Thus, our results should be directly applicable to similar clinical settings in other areas. Two of the three clinicians were Clinical Officers, the cadre of professionals likely to perform most circumcisions in sub-Saharan Africa.

An important lesson was that clear patient instructions are critical for wound care after surgery. Young men in many developing country settings may lack easy access to water and have difficulty in maintaining hygiene. Their opportunistic employment patterns often require high levels of physical activity that can disrupt wound healing. Instructions, both verbal and written in indigenous languages, can assist patients to limit the potential consequences of such lifestyle challenges.

A finding that requires further attention is that 23% of men reportedly resumed sexual activities before their 30-day follow-up visit, despite intensive counselling to the contrary. Fortunately, only one man reported having had sexual intercourse at one of the early visits, and that was 19 days after the procedure. If men resume sexual contact before their incisions are fully healed, there is the danger of increased risk of infection with HIV or other STIs. Once the trial is complete, we will test whether circumcised men have a higher incidence of HIV/STI than uncircumcised controls during this 1-month period after randomization. Effective counselling of young men is challenging, but developing effective ways to counsel against sexual contact while the surgical wound is healing will be crucial if male circumcision is to be introduced as an intervention to prevent HIV.

The present study has important limitations; the men were circumcised in a carefully designed, controlled research project, with thorough training of clinicians, careful selection of healthy patients and continuous monitoring of the outcome. It will be necessary to show similar outcomes in less rigorous settings. The follow-up was short, but we will obtain more data on sexual function. Most importantly, we need to determine, through the trial, whether adult male circumcision is effective in reducing the incidence of HIV.

In conclusion, we developed a standard procedure for adult male circumcision that proved suitable for resource-poor clinical settings. The adverse event rate compares favourably with rates reported from developed countries. Participants had excellent clinical outcomes, acceptable morbidity and excellent patient satisfaction. They returned to work and to normal general activities very quickly. These findings suggest that it should be possible to deliver adult male circumcision services with acceptable complication rates in developing countries.


We especially thank the young men in Kisumu who have so enthusiastically participated in the study. We acknowledge the assistance and unflagging support provided by Dr J. Otieno. We greatly appreciate the tireless efforts of the UNIM project staff. This study was funded by grants from the National Institute of Allergy and Infections Diseases, National Institutes of Health, Bethesda, Maryland USA, Grant No. AI150440, and the Canadian Institutes of Health Research (CIHR), Grant No. HCT 44180. S. Moses is the recipient of a CIHR investigator award.


None declared.