Fuchs 2014.

Methods	Parallel‐group randomised controlled trial. Single centre, gynaecologic endoscopy unit of a tertiary care medical centre in Zerifin, Israel. Protocol approved by IRB: yes. Post hoc statistical power calculation; non‐inferiority design. No external funding and no conflicts of interest (query clarified by Dr Moty Pansky).
Participants	Number recruited: 110 women. Number excluded before randomisation: 58 women (14 did not meet inclusion criteria; 37 declined to participate; 7 excluded for other reasons). Number randomly assigned: 52 women. Number lost to follow‐up: 11 women. Number analysed: 41 women. Women who underwent hysteroscopic surgery because of suspected RPOC between September 2009 and June 2012 invited to participate in study, and enrollees gave signed informed consent. Inclusion criteria: aged 18‐50 years; suspicion of RPOC on transvaginal ultrasound, diagnostic clinic hysteroscopy, or both. Study duration: 34 months; September 2009 to June 2012. Mean age (± SD) in intervention group: 29.5 (± 5.1) years. Mean age (± SD) in control group: 31.4 (± 6.5) years. Quote: "The study didn't include women with primary subfertility" (query clarified by Dr Moty Pansky). Comment: only women with confirmed fertility included in study.
Interventions	Oxiplex gel (intervention: n = 21) vs no gel (control: n = 20) All hysteroscopic procedures performed under general anaesthesia. Pelvic bimanual examination performed under anaesthesia, and findings recorded in the medical records. Uterus considered enlarged when uterine fundus was palpated above pelvic brim. Sodium chloride 0.9% solution used as distension medium. Suspected RPOC removed via blunt dissection, with 4‐mm loop resectoscope (Stryker Corp., Kalamazoo, MI, USA) as a curette and under direct hysteroscopic view. All specimens sent for pathological analysis. Intervention group: after completion of hysteroscopic dissection, Oxiplex gel inserted into uterine cavity, up to complete filling of the cavity or up to 10 mL gel, whichever occurred first. All women discharged from the hospital several hours after procedure. Control group: no gel. Both intervention and control groups received sequential hormonal treatment (oestradiol valerate 2 mg/day for 11 days, followed by oestradiol valerate 2 mg/day + norgestrel 0.5 mg/day for 10 days) and antibiotic therapy (amoxicillin‐clavulanic acid, 875 mg, twice daily for 7 days). All women underwent diagnostic clinic hysteroscopy at 6‐8 weeks after operative procedure, performed by a surgeon blinded to treatment group.
Outcomes	Intraoperative and postoperative complication rates, incidence of moderate or severe adhesions and pregnancy defined as a positive heartbeat (query clarified by Dr Moty Pansky). Comment: primary and secondary outcomes not determined.
Notes	Quote: "Because this was a pilot study using a non‐inferiority design, post hoc power analysis was performed. This calculation showed that the power for detection of a statistically significant difference in rates of intrauterine adhesions between the 2 groups was 24%." Comment: study was substantially underpowered for the outcome of incidence of moderate or severe IUAs. * Correspondence with authors on 19 January 2015: 1. The first citation is an interim analysis that included 30 women, and was presented at AAGL [American Association of Gynecologic Laparoscopists] on 2011. The second citation is the final analysis that was published in JMIG [Journal of Minimally Invasive Gynecology] 2014 and included 52 women. The study population of the second citation includes all 30 women from the first one and 22 additional women. 2. Allocation was based on a computer‐generated randomisation scheme that was prepared in advance by the study coordinator. Sealed envelopes containing allocation were opened only following consent by the treating physician. The study coordinator documented the allocation on a password protected computer. 3. The control group received NS [normal saline] at the end of the procedure. The participants didn't know which group they were allocated to, nor did the outcome assessors. Naturally, the treating physician at time of procedure was aware of the treatment. Treating physicians' identity was documented and the study coordinator made sure that different physicians performed the treatment and the assessment per patient. 4. The gel was provided by J&J [Johnson & Johnson]. There was no funding for the study. There was no conflict of interest. 5. The study didn't include women with primary subfertility. 6. This was a pilot study designed to assess safety, hence there was no distinction between primary and secondary outcomes. 7. Pregnancy was defined as a positive heartbeat. 8. We are not aware of any ongoing research on anti‐adhesion therapy following operative hysteroscopy."
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The study entrants, in blocks of 12, were randomly allocated via a computer‐generated randomisation schedule, using institutional computer software, to treatment with (study group) or without (control group) Oxiplex gel." Comment: probably done.
Allocation concealment (selection bias)	Low risk	Quote: "Allocation was based on a computer generated randomisation scheme that was prepared in advance by the study coordinator. Sealed envelopes containing allocation were opened only following consent by the treating physician. The study coordinator documented the allocation on a password protected computer" (query clarified by Dr Moty Pansky). Comment: probably done.
Blinding of participants and personnel (performance bias) Live birth, pregnancy or miscarriage	Low risk	Comment: unequivocal outcome.
Blinding of participants and personnel (performance bias) Adhesions	High risk	Quote: "Different surgeons performed the operative hysteroscopy and the follow‐up diagnostic hysteroscopy. Both the patients and the surgeons who performed the follow‐up studies were unaware of patient group assignment." Quote: "The participants didn't know which group they were allocated to, nor did the outcome assessors. Naturally, the treating physician at time of procedure was aware of the treatment. Treating physicians' identity was documented and the study coordinator made sure that different physicians performed the treatment and the assessment per patient" (query clarified by Dr Moty Pansky). Comment: participants probably blinded, as they were under general anaesthesia, but treating physicians not blinded.
Blinding of outcome assessment (detection bias) Live birth, pregnancy or miscarriage	Low risk	Comment: unequivocal outcome.
Blinding of outcome assessment (detection bias) Adhesions	Low risk	Quote: "Different surgeons performed the operative hysteroscopy and the follow‐up diagnostic hysteroscopy." Quote: "All patients underwent diagnostic office hysteroscopy at 6 to 8 weeks after the operative procedure, performed by a surgeon who was blinded to the treatment group." Quote: "The participants didn't know which group they were allocated to, nor did the outcome assessors. Naturally, the treating physician at time of procedure was aware of the treatment. Treating physicians' identity was documented and the study coordinator made sure that different physicians performed the treatment and the assessment per patient" (query clarified by Dr Moty Pansky). Comment: probably done.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote from the figure 1 CONSORT flow diagram of the study report: "In the intervention group five women were excluded from analysis after randomisation: the intervention was discontinued but no further clarification was given." Quote from the figure 1 CONSORT flow diagram of the study report: "In the control group six women were excluded from analysis after randomisation: lost to follow‐up (3) and discontinuation of the intervention (3) without further clarification." Comment: likely to cause attrition bias.
Selective reporting (reporting bias)	High risk	Comment: at high risk of selective outcome reporting, as live birth rates not reported for a study from September 2009 to June 2012, and publication of the final study report in 2014.
Other bias	High risk	Quote: "Patients with a diagnosis of adhesions (AFS grade 1) were offered an additional procedure for adhesiolysis." Quote: "At follow‐up hysteroscopy, 3 patients in the control group (14%) had AFS stage 2 or 3 (moderate to severe) intrauterine adhesions, compared with 1 woman in the study group (4%), who had AFS stage 3 intrauterine adhesions (P = 0.30)." Comment: imbalance between groups for a cointervention.