Safety and Efficacy of Nontherapeutic Male Circumcision: A Systematic Review

Inclusion Criteria

Eligible studies were those reporting on circumcision in an operative setting in male patients of any age, with no contraindications to or medical indications for circumcision. Generally, the main comparator was intact genitalia. To assess the effect of circumcision on subsequent pain response, however, it was necessary to include studies where all patients were circumcised. The clinical efficacy outcomes assessed were the incidence of sexually transmitted infection, urinary tract infection, HIV/AIDS, and penile cancer. Safety and patient-reported outcomes were also assessed.

Search Strategy and Selection Criteria

Only randomized controlled trials (RCTs) were included for analysis in this level-1 systematic review. Research articles were excluded if they reported solely on cost-effectiveness analyses or methodological issues. The reviewers believed that 1997 was an appropriate date limit, as a comprehensive review was performed in 1996 by the Canadian Paediatric Society.¹¹ It was deemed not useful to duplicate this summary of the literature; hence, studies were restricted to those published in English from January 1997 to August 2008 (Table 1⇓).

RCTs were retrieved when they were judged to possibly meet the inclusion criteria. These criteria were then applied fully to the retrieved studies to identify those to be appraised and included in the review. RCTs that did not meet the inclusion criteria were excluded.

Data Extraction and Synthesis

Data from all included RCTs were extracted by 1 reviewer and checked by a second reviewer using standardized data extraction tables that were developed a priori, with any differences resolved through discussion. The included RCTs were classified according to the National Health and Medical Research Council hierarchy of evidence.¹² An analysis of the methodological quality was undertaken with regard to such factors as randomization, blinding, and sample size.

Formal statistical pooling (meta-analysis) was only performed if 2 or more RCTs addressed the same comparison, with data available for comparable outcomes. A test for statistical heterogeneity was then performed, with P <.10 chosen to indicate statistical heterogeneity. Using RevMan (the Cochrane Information Management System Review Manager software, http://www.ims.cochrane.org/revman), relative risks (random effects model) with 95% confidence intervals (CIs) were calculated for dichotomous outcomes, and the data for outcomes that were not suitable for statistical pooling were reported narratively.

Neonatal Nontherapeutic Male Circumcision RCTs

Two studies reported upon neonatal circumcision. Taddio et al⁸ reported upon the effect neonatal circumcision had upon subsequent pain response at vaccination, and Macke¹⁹ reported upon the relationship between neonatal circumcision and pain distress/maternal-infant bonding.

Description of Adult Nontherapeutic Male Circumcision RCTs

Three studies reported upon circumcision in adult men, specifically the efficacy of circumcision for preventing HIV/AIDS acquisition in sub-Saharan African men.^13,14,16 Three studies^15,17,18 reported further unique outcomes for the Bailey et al¹⁴ and Gray et al¹⁶ RCT cohorts. Mattson et al¹⁵ reported sexual risk behavior outcomes, whereas Kigozi et al^17,18 reported safety and sexual satisfaction and function outcomes.

Another study²¹ also reported on the Bailey et al¹⁴ cohort and included data for a further 141 men originally randomized to the control group who received circumcision at the 2-year follow-up. As outcomes were not reported separately for these 141 patients, it was not possible to discern outcomes for those originally randomized to the circumcision group, so this study was methodologically ineligible for inclusion in this systematic review.

Clinical Efficacy

Two RCTs^15,16 reported on the efficacy of adult nonther-apeutic male circumcision for preventing sexually transmitted infections. One RCT¹⁶ found that the prevalence of self-reported genital ulcers was significantly lower in the circumcised group than in the uncircumcised group (3.1% vs 5.8%; prevalence risk ratio 0.53, 95% CI, 0.4–0.64; P <.001). The other study¹⁵ found that circumcised men were more likely than uncircumcised men to have a prevalent infection diagnosed at baseline (P=.02) and an incident infection diagnosed at the 6-month follow-up (P = .05). At the 12-month follow-up there were no significant differences between the proportions of circumcised and uncircumcised men with incident sexually transmitted infections.¹⁵

No RCTs that reported on the association between nontherapeutic male circumcision and penile cancer or urinary tract infections could be identified.

Three RCTs^13,14,16 reported on HIV/AIDS acquisition after adult nontherapeutic male circumcision in sub-Saharan African men. As studies were homogeneous, a meta-analysis was performed controlling for those who were found to have been HIV positive at baseline (Figure 2⇓). Circumcised men in this population were at significantly lower risk of acquiring HIV/AIDS (random effects OR = 0.44, 95% CI, 0.32–0.59; P <.001). All 3 RCTs were stopped by their data- and safety-monitoring boards before their designed completion because of significant reductions in HIV/AIDS incidence in the circumcision groups. It was deemed unethical to continue without offering circumcision to control group participants.²²

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Efficacy of circumcision in preventing HIV infection in men who were HIV-negative at randomization.

CI = confidence interval; HIV = human immunodeficiency virus; M-H = Mantel-Haenszel odds ratio.

Study Limitations

There was a paucity of RCT evidence upon which to draw conclusions. Further, not all outcomes were reported in all RCTs, and several outcomes were not reported uniformly. It is internationally recommended that circumcisions should be performed in a surgical setting,^1,23 and this systematic review considered circumcision only in such a setting. It is likely that circumcisions performed in unhygienic conditions and with little or no postoperative care would have higher adverse events.²⁴ Mass circumcisions performed using a common instrument may also have associated morbidity.

Another limitation may have been any introduced performance bias in that some outcome assessors were not blinded to circumcision status. Importantly, only 5 of the 8 included randomized cohorts reported that their outcome assessors were blinded.

Neonatal Circumcision

No reports of mortality or adverse events resulting from the procedure itself were found in any of the included studies. Pain control is advocated, as neonates circumcised with analgesia can experience significantly less postoperative pain distress than those circumcised with placebo.¹⁹ Further, neonatal circumcision performed without pain relief may affect neonatal behavior, interfere with maternal-infant interaction,¹⁹ and cause increased responses to subsequent pain.⁸

Adult Circumcision

No reports of mortality due to the procedure itself were found in any of the included studies. The reported adverse effects in adult circumcisions were infrequent, suggesting that circumcision in this population can be performed safely. Studies differed in the manner in which adverse events were defined, detected, and reported, yet severity of complications were generally reported as mild to moderate, and most resolved within hours or days. Serious, permanent adverse effects, such as erectile dysfunction, occurred rarely.

Adult circumcision may potentially have more associated adverse events than childhood or neonatal circumcision, as adult circumcisions usually require surgical wound suturing. Complications can occur if patients resume intercourse before wound healing.^24,18 Various age-dependent factors, such as the type of anesthesia or surgical technique used, and skills of the circumciser (clinical vs nonclinical traditional circumciser) may also affect outcomes.

As safety outcomes were reported only for sub-Saharan African adults with strict inclusion criteria, the adverse events reported may not reflect the broader population with a wider range of comorbidities. Some patient groups have contraindications to circumcision, including patients with congenital anomalies of the penis, chordee, buried penis, or patients with a family history of a bleeding disorder.¹⁰

One RCT¹⁶ found that circumcision can protect men against genital ulcers and postulated that it could be protective against cutaneously acquired infections but may not protect against urethral infections. It is difficult to assess this statement, as no other sexually transmitted infections were reported upon in the RCT literature.

In developed countries penile cancer is a relatively rare disease, with an incidence of approximately 1 in 100,000.¹⁰ The absence of RCT evidence, combined with the rarity of penile cancer, suggests that circumcision is not justified for the sole purpose of protecting against penile cancer.¹⁰

The presence of uropathogens is thought to be a risk factor for urinary tract infection, as uropathogens may attach to the foreskin and reach the renal tract via the ascending route.¹⁰ The absence of RCT evidence, in addition to the incidence of urinary tract infection in normal boys of approximately 1% to 2%, makes it unlikely that preventive circumcision of normal boys would outweigh the adverse events associated with the procedure.^10,11

A meta-analysis found that circumcised sub-Saharan African men had a significantly lower risk of acquiring HIV/AIDS after nontherapeutic male circumcision. It is not clear whether the findings from these African trials can be extrapolated to other countries with ready access to condoms and a lower prevalence of HIV/AIDS.

Several biologic mechanisms for a protective effect of circumcision for HIV/AIDS infection have been proposed. For example, the prepuce contains a high density of Langerhans cells, which express HIV-1 receptors. By removing this source of Langerhans cells, circumcision could potentially reduce the risk of HIV/AIDS acquisition and transmission.^25–27 The prepuce may be susceptible to mild trauma during intercourse, facilitating transmission of infection,²⁸ whereas the warm and moist environment under the prepuce is also thought to be conducive to prolonged viral survival.^27,28 Finally, the layer of keratin that develops in a circumcised penis may provide additional defense against infection.²⁶ In contrast, a protective effect of the prepuce for HIV/AIDS infection has been suggested. The apocrine glands in the inner prepuce secrete a lysozyme that reportedly kills HIV-1 in vitro.²⁹ The immunological effects of sub-preputial flora, secretory immunoglobulins, and lytic secretions from the prostate, urethra, and seminal residues are unclear.²⁸ Finally, the prepuce may reduce the likelihood of abrasions occurring during intercourse, thereby removing a potential site for viruses to enter the body.²⁸

The evidence did not show a decrease in sexual satisfaction and function after circumcision. Interestingly, uncircumcised men showed significant increases from baseline to 2-year follow-up for most measures of satisfaction and function, which may be due to the health education and medical care provided during the study.¹⁷ Overall the frequency of most problems either did not change or diminished during the follow-up period, and problems were rare (less than 2%) at all time points. The difficulties with penetration and pain experienced after circumcision may have been transient problems caused by incomplete keratinization of the scar.¹⁷

Research Recommendation

Although approximately 30% of the global male population is circumcised, there is a paucity of high-quality evidence.²⁴ Depending on the indication under investigation, prospective RCTs and case-control studies should be conducted to strengthen the evidence base and allow more informed conclusions on nontherapeutic male circumcision to be drawn.

Although the evidence for the efficacy of adult nontherapeutic male circumcision in preventing HIV/AIDS acquisition in sub-Saharan African men is strong, it is unclear whether these findings can be extrapolated to male populations in other countries. The role of adult nontherapeutic male circumcision in preventing sexually transmitted infection, urinary tract infection, and penile cancer is less clear, whereas the role of neonatal circumcision in preventing HIV/AIDS, sexually transmitted infection, urinary tract infection, and penile cancer is not presently supported by RCT evidence.

Patients who request circumcision in the belief that it bestows clinical benefits must be made aware of the lack of consensus and robust evidence, as well as the potential medical and psychosocial harms of the procedure. As the efficacy of prophylactic nontherapeutic male circumcision has not been comprehensively studied in neonates, it would be inappropriate to recommend widespread neonatal circumcision for this purpose.