Table 3.

Critical Appraisal of Studies of Nontherapeutic Male Circumcision

StudyRandomization TechniqueOutcome Assessors BlindedInclusion/Exclusion Criteria ProvidedOperatorBaseline CharacteristicsPower CalculationsFollow-UpLost to Follow-Up No. (%)
HIV=human immunodeficiency virus; NR = not reported; STI = sexually transmitted infection.
a Study identifier where more than 1 report for a single study exists.
b Note: this cohort was originally randomly assigned to receive lidocaine-prilocaine or placebo cream prior to circumcision.20
Adult circumcision
Auvert et al,13 2005Opaque, sealed envelopesYesYesGeneral practitioners experienced in male circumcisionPatients in both groups well matched for age, religion, ethnic group, and educational level; authors did not provide information on statistical significance3,035 patients required to obtain power of 80% to detect 50% reduction in HIV21 mo251 (8)
 100 (6.5%) in circumcision group
 151 (9.5) in control group (P=.0016).
Bailey et al,14 2007aOpaque, sealed envelopesYesYesCliniciansPatients in both groups well matched for demographic and physical characteristics, prevalence of STI, and reported sexual history with women; authors did not provide information on statistical significance2,776 patients required to detect 50% difference in 2-y HIV seroincidence24 mo240 (8.6)
 126 (9.1) in circumcision group
 114 (8.2) in control group
Mattson et al,15 2008
Gray et al,16 2007aOpaque, sealed envelopesNRYesTrained and certified physiciansPatients in both groups well matched for sociodemographic characteristics, sexual risk behaviors, rates of self-reported symptoms of STI; authors did not provide information on statistical significanceStudy had 80% power to detect rate ratio of 0.5 for incident HIV in intervention group relative to control group24 mo229 (10.4)
 114 (10.4) in circumcision group
 115 (10.4) in control group
Kigozi et al,17 2008
Kigozi et al,18 2008aPrevious receipt of voluntary counselling and testing slightly higher in intervention than control groups (no figures provided)
Neonatal circumcision
Macke 19 2001Groups determined by pharmacy staffYesInclusion criteria onlyPrivate and resident physiciansNo significant differences between the groups for maternal or newborn variablesNR1-h post circumcisionNR
Taddio et al,8 1997NRbYesNoPediatriciansNo significant differences between groups for demographic characteristics30 infants per group needed20 sec before, 20 sec during, and 20 sec after vaccinationNR