Thirty-five studies (n=2,240) were included: 14 RCTs (n=789), 13 prospective CCTs (n=797) and 8 retrospective studies (n=654).
The studies scored between three and nine stars for quality. Sixteen studies (14 RCTs and 2 CCTs) were classified as the highest quality (six or more stars).
Failure to construct a J-pouch reservoir was found in 4.8% of the patients (12 out of 249 patients in 7 studies).
J-pouch versus transverse colostomy (5 studies, n=280).
There was no significant difference between J-pouch and transverse colostomy for any of the adverse post-operative events or functional outcomes.
Hospital stay and operating time were significantly shorter with J-pouch, but only one study (n=88) reported these outcomes: the WMD was -1.40 days (95% CI: -1.86, -0.94, p<0.001) for hospital stay and 33.60 minutes (95% CI: 24.55, 42.65, p<0.001) for operating time.
J-pouch versus CAA (31 studies, n=1,976).
J-pouch was associated with a reduction in anastomotic leak compared with CAA ,but the reduction was not statistically significant; the OR was 0.71 (95% CI: 0.48, 1.0, p=0.069; based on 17 studies, n=1,225). No evidence of statistical heterogeneity was found. There were no statistically significant differences between J-pouch and CAA for any of the other adverse post-operative events.
J-pouch was associated with a significant reduction in the frequency of defaecation at 6 months compared with CAA (WMD -1.88, 95% CI: -3.33, -0.42, p=0.011). The reduction was less but remained statistically significant at 1 year (WMD -1.35, 95% CI: -1.92, -0.78, p<0.001) and 2 or more years (WMD -0.74, 95% CI: -1.31, -0.18, p=0.010). Significant statistical heterogeneity was found for all three meta-analyses (p<0.001 for all).
J-pouch was associated with a significant reduction in faecal urgency at 6 months compared with CAA (OR 0.27, 95% CI: 0.12, 0.60, p=0.001) and at 1 year (OR 0.21, 95% CI: 0.12, 0.37, p<0.001), but there was no significant difference between treatments at 2 or more years (based on 3 studies, n=134). Significant statistical heterogeneity was found for the meta-analysis at 6 months (p=0.022) but not for the other two meta-analyses.
A funnel plot based on post-operative anastomotic leak showed no evidence of publication bias.
The sensitivity analysis showed similar results. For faecal frequency, heterogeneity remained statistically significant for high-quality studies and for studies with more than 50 patients.