Efficacy of pelvic floor peritoneal reconstruction after laparoscopic resection of middle and low rectal cancer: A meta-analysis

Figure 1 Results and process of the records screening. PubMed, Cochrane Library, EMBASE, and MEDLINE databases were searched. In total, 20 relevant English records were retrieved after screening based on the inclusion and exclusion criteria; 11 English articles[1-11] met the inclusion criteria.

Figure 2 Forest plot. A: Comparing the risk of postoperative intestinal obstruction between the reconstruction and control groups. The pooled association was relative risk (RR) = 0.30 (95%CI: 0.18-0.49, P_effect < 0.001), with I² = 0% and P_{heterogeneity} = 0.52. A fixed-effect model was used; B: Comparing the risk of postoperative intestinal obstruction following laparoscopic abdominoperineal resection (LAPR) between the reconstruction and control groups. For LAPR/extralevator abdominoperineal excision (ELAPE), the pooled association was RR = 0.25 (95%CI: 0.12-0.52; P_effect < 0.001), with I² = 0% and P_{heterogeneity} = 0.85; a fixed-effect model was used; C: Comparing the risk of postoperative intestinal obstruction following laparoscopic anterior rectal resection (LAR) between the reconstruction and control groups. For LAR, the pooled association was RR = 0.34 (95%CI: 0.17-0.70; P_effect = 0.003) with I² = 32% and P_{heterogeneity} = 0.21; a fixed-effect model was used; D: Comparing the risk of postoperative internal hernia between the reconstruction and control groups. The pooled association was RR = 0.36 (95%CI: 0.17-0.79; P_effect = 0.010), with I² = 0% and P_{heterogeneity} = 0.52; a fixed-effect model was used; E: Comparing the risk of postoperative anastomotic leak between the reconstruction and control groups. The pooled association was RR = 0.69 (95%CI: 0.51-0.93; P_effect = 0.020), with I² = 5% and P_{heterogeneity} = 0.64; a fixed-effect model was used; F: Comparing the risk of postoperative unplanned reoperations between the reconstruction and control groups. The pooled association was RR = 0.33 (95%CI: 0.19-0.57; P_effect < 0.001), with I² = 0% and P_{heterogeneity} = 0.52; a fixed-effect model was used; G: Comparing the risk of unplanned reoperations following LAPR between the reconstruction and control groups. For LAPR/ELAPE, the pooled association was RR = 0.40 (95%CI: 0.16-0.98; P_effect = 0.046), with I² = 38% and P_{heterogeneity} = 0.20; a fixed-effect model was used; H: Comparing the risk of unplanned reoperations following LAR between the reconstruction and control groups. For LAR, the pooled association was RR = 0.30 (95%CI: 0.15-0.60; P_effect < 0.001), with I² = 0% and P_{heterogeneity} = 0.78; a fixed-effect model was used; I: Comparing postoperative length of hospital stay between the reconstruction and control groups. The random-effects estimate was MD = -2.40 days (95%CI: -4.33 to -0.47; P_effect = 0.015), with I² = 87% and P_{heterogeneity} < 0.001. The results were unstable in leave-one-out sensitivity analyses. PC: Pelvic floor peritoneal reconstruction.

Figure 3 Funnel plot of the risk of intestinal obstruction. Visual inspection did not show marked asymmetry. However, with only nine studies, the funnel plot had limited power, and publication bias could not be excluded. RR: Relative risk.