Thirty-three RCTs (5,381 patients) were included in the review and 28 RCTs (5,191 patients) were included in the meta-analysis. Four RCTs had a low risk of bias and 16 RCTs (57%) were industry sponsored. The risk of random error was small for adhesion score in eight RCTs and overall incidence of small bowel obstruction by any cause in one trial.
Oxidised regenerated cellulose: There were 11 RCTs in gynaecological surgery (two at low risk of bias). There was no evidence on the primary outcome, reoperations for adhesive small bowel obstruction. For secondary outcomes, oxidised regenerated cellulose statistically significantly reduced the incidence of adhesions (RR 0.51, 95% CI 0.31 to 0.86; NNT=6; three RCTs).
Hyaluronate carboxymethyl cellulose: Nine RCTs included six in colorectal surgery and one each in gynaecological, hepatic and gastric surgery (one at low risk of bias). Hyaluronate carboxymethylcellulose statistically significantly reduced incidence of reoperations for adhesive small bowel obstruction (RR 0.49, 95% CI 0.28 to 0.88; five RCTs); separate analyses showed that the differences were not statistically significant for the single trials in hepatic and gastric surgery.
Icodextrin: Four RCTs showed no statistically significant differences between treatment groups for reoperation for adhesive small bowel obstruction. There was insufficient evidence to assess the effects of icodextrin on operation time. Icodextrin statistically significantly reduced the incidence of small bowel obstruction (RR 0.20, 95% CI 0.04 to 0.88; one RCT).
Polyethylene glycol: Four RCTs (one at low risk of bias) showed no evidence on the primary outcome, reoperations for adhesive small bowel obstruction. There were no statistically significant differences between treatment groups for the incidence of adhesions. Polyethylene glycol showed a beneficial effect on operation time in one trial at low risk of bias in colorectal surgery (SMD -0.84, 95% CI -1.49 to -0.19).
There were no statistically significant differences between any adhesion barrier and controls in the incidence of serious adverse events. Other results were reported in the review, including results from subgroup analyses and findings from trials not included in the meta-analyses.
There was no evidence of publication bias using funnel plots.