Systematic review and meta-analysis of incisional hernia post-reversal of ileostomy

Abstract

Purpose: Incisional hernia following closure of loop ileostomy is a common problem. Assessment of the proportion of this complication is limited by small sample size and inconsistent reporting. The aim of this review was to provide an estimate of the proportion of incisional hernia following closure of loop ileostomy according to clinical and radiological diagnostic criteria and to investigate the association of bibliometric and study quality parameters with reported proportion. Methods: A systematic review of PubMed, Embase, CENTRAL, ISRCTN Registry and Open Grey from 2000 onwards was performed according to PRISMA standards. Reporting on the type of stoma and mesh reinforcement after closure was mandatory for inclusion, whereas studies on paediatric populations were excluded. Fixed effect or random effects models were used to calculate pooled proportion estimates. Meta-regression models were formed to explore potential heterogeneity. Results: 42 studies with 7166 patients were included. The pooled estimate of the proportion of incisional hernia after ileostomy closure was 6.1% (95% confidence interval, CI 4.4–8.3%). Proportion estimates for higher quality studies and studies reporting on incisional hernia as primary outcome were 9.0% (95% CI 6.3–12.7%) and 13.1% (95% CI 8.8–19.1%). Significant between-study heterogeneity was identified (P < 0.001, I2 = 87%) and the likelihood of publication bias was high (P = 0.028). Mixed effects regression showed that both year of publication (P = 0.034, Q = 4.484, df = 1.000) and defining hernia as a primary outcome (Q = 20.298, P < 0.001) were related to effect size. Method of follow-up and quality of the studies affected the proportion. Conclusion: The proportion of incisional hernia at ileostomy closure site is estimated at 6.1%. Reporting incisional hernia as primary or secondary outcome, the method of diagnosis, the year of publication and methodological quality are associated with reported proportion.