Five non-randomised studies were included in the review (5,012 patients, range 40 to 4,600): three prospective studies (271 patients) and two retrospective studies (4,741 patients). Follow-up was reported in the text as ranging from 17 to 60 months (median 33). One trial was assessed as being poor quality (score 4). Four trials were assessed as high quality (two scored 5 and two scored 7). All of the pooled results reported were calculated using the random-effects model.
Rectal washout significantly reduced the risk of local recurrence compared with no rectal washout (OR 0.57, 95% CI 0.43 to 0.74; five trials; Ι²=1%). Similar results were found when the analysis was restricted to patients treated with a radical resection (OR 0.54, 95% CI 0.39 to 0.76; five trials; Ι²=0%) and patients treated by curative resection (OR 0.55, 95% CI 0.42 to 0.72; four trials; Ι²=0%). One trial demonstrated a similar result among patients treated by pre-operative radiotherapy but this result was of borderline statistical significance (OR 0.62, 95% CI 0.39 to 0.98).
Sensitivity analyses that included only studies with a quality score of 5 or more revealed similar findings for risks of local recurrence overall (four studies; Ι²=11%) and among patients treated with local radical resection (four trials; Ι²=0%) and curative resection (three trials; Ι²=0%). Results were reported fully in the paper.
No evidence of publication bias was found.