Lagarde 2009.

Study characteristics
Methods	Design: RCT conducted in South Africa. Compared the TK device with FG male circumcision. 69 participants of the control group of a male circumcision RCT conducted in South Africa were recruited. Ethical approval: University of the Witwatersrand Human Research Ethics Committee (Medical). Adverse events forms were transmitted to a Data Safety Monitoring Board. Consent: participants provided informed consent. Duration of study: September–November 2004
Participants	Inclusion criteria: uncircumcised males from the control group of a male circumcision RCT with no contraindication to circumcision, of good general health with normal physical and genital condition and consenting to participate in trial. They had to agree to avoid sexual contacts (except with condom protection) during the 6 weeks following circumcision, consent to a medical visit 6 weeks after circumcision and consent to reporting of any adverse events. Exclusion criteria: contraindications to circumcision, abnormal physical and genital conditions and not consenting to participate in the trial. Baseline characteristics: groups comparable with no statistically significant differences between them for the listed characteristics. Circumcision device group mean age was 21.5 , interquartile range (IQR) 20 ‐ 23.25 years , traditional method mean age was 22, IQR (21 ‐ 24) years.
Interventions	Providers: 3 GPs who had extensive experience with FG circumcision method conducted both procedures. Intervention group (TK): 35 participants. Appropriate size TK ascertained. Penis disinfected using povidone‐iodine solution. Dorsal penile nerve block and circumferential block performed with 1% lidocaine. Foreskin pulled slightly forward over rim of an inner tube and positioned inside an outer ring. 2 plastic arms locked into place to force 2 surfaces (inner tube and outer clamping ring) into tight contact with foreskin trapped between them. If foreskin was tight, a dorsal slit was required to gain access for ring to be applied over glans. Then foreskin trapped forward of clamping device was cut away. Device was intended to remain on the penis for 7–10 days until removal or it fell off with the necrotized foreskin. Control group (FG): 34 participants. Foreskin was pulled outwards in front of the glans, and forceps clamped across it, parallel to the corona of the glans and immediately in front of the glans. The scalpel was run across the face of the foreskin, and absorbable sutures used to close the cut edges. Excess bleeding controlled with ligature, direct pressure or cautery. Sutured area was covered with sterile paraffin tulle gras, sterile gauze and paper tape; dressing was removed 24–48 hours after surgery by the GP who performed it. Follow‐up: For all groups, 3 days after surgery and participants advised to return to center 6 weeks after surgery for a genital exam and completion of a short questionnaire. Adverse events recorded using a standardized form.
Outcomes	Primary outcomes: number and nature of adverse events penile infection delay in wound healing excessive or insufficient skin removed erectile dysfunction pain score bleeding within 2 weeks following the procedure lesions to the penis, swelling or hematoma within 2 weeks following the procedure, any problem when urinating satisfaction with penis appearance
Funding sources	Agence Nationale de Recherche sur le Sida et le Hépatites, Institut Nationale de la Santé et de la Recherche Médicale (Paris) and National Institute for Communicable Diseases (NICD) (Johannesburg). Quote: "INSERM and the NICD signed collaboration and licensing agreements with a private South African company for providing all sterilised instruments and consumables for performing a single male circumcision according to the FG method."
Declarations of interest	Funders had no role in study design, data collection and analysis, decision to publish or manuscript preparation.
Notes	All participants who attended the center for their last visit of the male circumcision RCT follow‐up received a ZAR 150 (EUR 20) payment whether or not they participated in the study. Participants were asked to return to the center 6 weeks after surgery for a genital exam and completion of a short questionnaire. Participants who attended this postcircumcision visit were paid ZAR 40.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Process for development of the random sequence not described; unclear risk of bias.
Allocation concealment (selection bias)	Unclear risk	Some information about allocation concealment was provided. Quote: "For randomisation, each participant chose an envelope containing the group name from a basket of 10 envelopes. After each choice, a new envelope taken sequentially from a set of envelopes prepared in such a way that each set of 10 envelopes contained the same number of Usual and TK tokens, was added to the basket." Comment: certain details that would inform whether the allocation could be tampered with were not provided. For example it was not reported whether envelopes were sealed or opaque; unclear risk of bias.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding of participants and personnel was not possible with this surgical intervention; unclear whether this may have impacted the performance of either group. Quote: "Neither GPs, participants nor investigators were blinded to the randomisation group. At interview, the nurse was not aware of the method used but on examination could conclude which technique was used."
Blinding of outcome assessment (detection bias) Subjective outcomes (severe adverse events, moderate adverse events, minor adverse events, pain, patient satisfaction)	Unclear risk	Outcome assessment was not blinded. Quote: "…nurse who performed the interview and the clinical examination was blinded to the intervention group, but was obviously able to identify the group from seeing the scar." As many of the outcomes were subjective, this may have introduced a high risk of detection bias.
Incomplete outcome data (attrition bias) All outcomes	High risk	There was greater attrition in the TK group (20% with TK vs 10% with surgery). 4 participants crossed over to the standard surgical group. 19/35 participants randomized to the TK group were analyzed for all the outcomes; 7 participants did not visit the GP after circumcision and were excluded from the analyses. 29/34 participants randomized to the FG group were analyzed for all outcomes; 6 did not visit the GP in the FG group and were excluded from further analyses. High risk of attrition bias.
Selective reporting (reporting bias)	Unclear risk	No study protocol and, therefore, it was unclear whether there may have been selective outcome reporting. No mention of clinical trials database registration.
Other bias	High risk	There was potential for bias due to early stopping of the trial and the trial not meeting its planned sample size for the primary outcome (power). The planned number of participants of 400 in each group was not achieved because of trial interruption owing to a high number of adverse events in the TK group. Quote: "The protocol was designed to randomize participants to three groups, including men circumcised with a single‐use sterile surgery kit, but because of delay in its availability, the study started with only two (the TK vs. FG method with re‐usable instruments). The postcircumcision visit was originally planned exactly 6 weeks after surgery, but only 3 participants made the visit 1–3 days before 6 weeks. Median and mean intervals between circumcision and visit were more than 6 weeks. The trial investigators informed the DSMB and stopped the trial." Comment: the TK supplier attributed the high number of adverse events to the inexperience of GPs using the device and an experienced practitioner was sent to provide additional training to the GPs.