Li 1994.

Methods	Unit of randomisation: pelvic sidewall Method of randomisation: computer‐generated random numbers Time of randomisation: at completion of surgery Blinding: assessor unaware of which sidewall was treated 28 participants randomly assigned 1 exclusion (declined second‐look laparoscopy) 27 participants (54 pelvic sidewalls) analysed Trial supported by manufacturers of Interceed (Johnson & Johnson Medical, Ltd., UK)
Participants	Patients undergoing adhesiolysis by laparotomy (N = 28; 54 pelvic sidewalls, 1 withdrawal) Indication for surgery: infertility 21; chronic pelvic pain 6 Pre‐existing adhesions: yes Cause of adhesions: endometriosis 9; PID 10; previous surgery 5; uncertain 3 Microsurgery: yes Age, years, mean: 30.4 (SD 4.5) No details on participants with infertility Location: UK Timing and duration: 1989 to 1992
Interventions	Oxidised regenerated cellulose on pelvic sidewall vs uncovered opposite sidewall (oxidised regenerated cellulose sutured in 3 cases) Other adjuvants used: intraperitoneal hydrocortisone in Intralipid Second‐look laparoscopy Timing: 3 to 14 weeks after initial surgery Surgeon unaware: yes Adhesiolysis: not stated
Outcomes	Adhesions at second‐look laparoscopy: 3 to 14 weeks after initial surgery Incidence (per pelvic sidewall) Extent: area, area differential, % improvement Severity Area: measured (not specified) Adverse effects Pregnancy outcomes: not stated
Notes	Documented with photographs and drawings at initial surgery and with video recording at second‐look laparoscopy Sponsored by Johnson & Johnson Video at second‐look laparoscopy No power calculations No intention‐to‐treat analysis
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomisation stated but no details of method provided Quote: "the assignment of the test side for each woman was carried out before the study and was randomised blocks of two to provide balance"
Allocation concealment (selection bias)	Low risk	Allocation concealment using envelopes opened intraoperatively, reducing risk of bias Quote: "a sealed envelope was then opened revealing which of the side walls was to receive Interceed"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Unit of randomisation was either right or left side of pelvic wall, so every woman also acted as her own control Surgeon was not blinded, but treatment allocation was revealed only intraoperatively
Blinding of outcome assessment (detection bias) All outcomes	Low risk	An independent blinded surgeon performed assessments at second‐look laparoscopy
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 exclusion (declined second‐look laparoscopy)
Selective reporting (reporting bias)	High risk	Data presented as numbers of cases as well as graphically. P values stated for significant outcomes. No omission of outcomes and no subsets of data. Unclear whether power calculation performed Furthermore, despite 21 participants undergoing adhesiolysis for infertility, no pregnancy outcomes stated It should also be noted that centres other than Sheffield ‐ Li 1994 ‐ were randomly assigning participants around the UK for a study of oxidised regenerated cellulose and no treatment in the early 1990s (personal communication, A Watson). However, data have never been published
Other bias	Unclear risk	During surgery, "constant irrigation of the tissue with a physiological solution that contained heparin" was performed. Also, all patients received instillation of hydrocortisone in Intralipid