Loop transverse colostomy versus loop ileostomy for defunctioning of colorectal anastomosis: a systematic review, updated conventional meta-analysis, and cumulative meta-analysis.

Defunctioning of colorectal anastomosis either with loop transverse colostomy or ileostomy was evaluated using updated and cumulative meta-analyses. Studies were identified by a systematic search of Embase, PubMed, Cochrane Library, and Google Scholar databases and were selected as per the PRISMA checklist. Both randomised control trials (RCTs) and retrospective studies were included. A sensitivity analysis was performed, and a cumulative meta-analysis was performed to monitor evidence over time. Significantly more male patients underwent loop ileostomy than transverse colostomy [odds ratio (OR) = 0.59 (95% confidence interval (CI) 0.39, 0.90), p < 0.001, I2 = 48%]. Significantly more colostomies were complicated by stoma prolapse than by ileostomies [OR = 6.32 (95% CI 2.78, 14.35), p < 0.001, I2 = 0%). Patients with ileostomy demonstrated a significantly higher complication rate of high-output stoma than patients with colostomies [Peto OR = 0.16 (95% CI 0.04, 0.55), p = 0.004, I2 = 0%]. Patients with colostomies demonstrated significantly more complications related to stoma reversal, such as wound infections and incisional hernias, than patients with ileostomies [OR = 3.45 (95% CI 2.00, 5.95), p < 0.001, I2 = 0%; OR = 4.80 (95% CI 1.85, 12.44), p < 0.001, I2 = 0%, respectively]. Overall complications related to stoma formation and closure did not demonstrate significant differences; however, their I2 values were 82% and 76%, respectively, suggesting high heterogeneity, which may have influenced the results. A subgroup analysis of RCTs showed no discrepancies when compared to the whole sample. In the cumulative meta-analysis, the effect size of each study was non-significant for the entire period. The demonstrated significant differences did not translate in favour of ileostomy when the overall complications of stoma formation and reversal were evaluated. Confounding factors and underpowered samples may have influenced the results. Future multicentre RCTs with homogeneous populations and adequate power may demonstrate more conclusive evidence regarding the superiority of one procedure over the other.