Twelve studies (1,002 patients) were included in the review. One study with 38 participants was prospective, while the others were retrospective.
Ten studies were considered to be high quality.
Early post-operative adverse events.
A pooled analysis of 6 studies showed a significantly lower need for reoperation within 30 days in the IRA group compared with the IPAA group (OR 2.11, 95% confidence interval, CI: 1.21, 3.70), with no significant heterogeneity between the studies. There was no significant difference between groups in bowel obstruction (reported in 10 studies), post-operative haemorrhage (3 studies), intra-abdominal sepsis (8 studies), anastomotic separation (5 studies) or wound infection (6 studies).
Long-term adverse events.
The meta-analysis showed a statistically significant difference in favour of IRA over IPAA in perianal irritation in 7 studies (OR 2.48, 95% CI: 1.36, 4.55) and in anastomotic stricture in 5 studies (OR 3.84, 95% CI: 1.46, 10.11). Conversely, the benefit was significantly in favour of IPAA over IRA for reduction in cancer in the pouch or rectum in 5 studies (OR 0.13, 95% CI: 0.03, 0.61) and the need for further operation on the pouch or rectum in 6 studies (OR 0.10, 95% CI: 0.04, 0.23). No difference was shown in desmoid formation (reported in 6 studies).
Functional outcomes.
IRA was shown to have a significant benefit over IPAA in a pooled analysis of bowel frequency in 24 hours as reported in 8 studies (WMD 1.62, 95% CI: 1.05, 2.20), night defaecation in 4 studies (OR 6.64, 95% CI: 2.99, 14.74), incontinence day or night in6 studies (OR 2.71, 95% CI: 1.81, 4.07) and the need for incontinence pads day or night in 4 studies (OR 2.72, 95% CI: 1.02, 7.23). There was significant heterogeneity between studies in the meta-analysis of bowel frequency, which was no longer significant when only the 4 high-quality studies were included whilst the difference in effect remained statistically significant. In 5 studies significantly fewer patients who underwent IPAA were affected by faecal urgency compared with the IRA group (OR 0.43, 95% CI: 0.23, 0.80). No difference was shown between IRA and IPAA in the need for antidiarrhoeal medication (reported in 8 studies).
Quality of life.
Social restriction was significantly lower after IRA compared with IPAA in 2 studies (OR 6.04, 95% CI: 1.53, 23.78). No difference was shown in dietary restriction in 6 studies or in male (6 studies) or female (5 studies) sexual dysfunction.
The meta-analysis results were unchanged when only studies with mean or median follow-up of 3 years or more were included. The sensitivity analysis based on study quality, size and publication date reduced heterogeneity for some outcomes. There was no evidence of publication bias in a funnel plot of antidiarrhoeal medication use.