Fourteen RCTs (n=4,859 patients) were included. Three RCTs had a quality score of 2 points and were classified as poor quality; the remaining trials scored 3 or, most commonly, 4 points. Excluding the two large trials of more than 1,300 participants each, sample size ranged from 47 to 380. Where reported, length of follow-up ranged from seven to 90 days.
Following mechanical bowel preparation, 4.02% of participants experienced anastomotic leakage compared with 3.44% of participants without mechanical preparation. There was no statistically significant difference in anastomotic leakage between the two groups (OR 1.12, 95% CI: 0.82 to 1.53, 14 trials). There was no statistically significant difference between groups for the other outcomes, except for overall surgical site infections, where there was a higher risk in the mechanical bowel preparation group (OR 1.40, 95% CI: 1.05 to 1.87, 14 trials). There was statistically significant heterogeneity in this analysis. The subgroup analyses did not substantially alter the results. It was not possible to undertake subgroup analysis of rectal surgery as, for the majority of trials, this was an exclusion criterion in the trial or data were not reported separately for this group.
The authors stated that visual inspection of the funnel plot did not suggest publication bias, but interpretation was limited by the small number of trials.