Eligible for inclusion were randomised controlled trials (RCTs) of patients receiving abdominal surgery, for conditions other than trauma, with entry into the gastrointestinal tract, biliary tract or both. Trials had to use an occlusive barrier device to protect the laparotomy wound, and compare this with no use of an occlusive barrier device. The primary outcome of interest was surgical site infection (defined by the Centers for Disease Control and Prevention, as any infection of the superficial or deep tissues or the organ or space affected by surgery, which occurred within 30 days of surgery, when no prosthesis was implanted). Further inclusion criteria for the primary outcome were reported. Trials without extractable data, and those for which the authors could not be contacted, were excluded.
The included trials were conducted in Sweden, Ireland, Japan, the USA, or Australia, and published between 1984 and 2010. The patients received upper or lower gastrointestinal, abdominal, or open appendix surgery. Where reported, most surgery was elective; the mean age of patients ranged from 33 to 67 years; and there was a fairly equal distribution of men and women. Equal numbers of trials used single ring versus dual ring wound protectors. There was variation in the use of preoperative antibiotics, mechanical bowel preparation, and brand of wound protector.
Three reviewers independently selected the studies for inclusion. Discrepancies were resolved by consensus.