Risk Factors for Anastomotic Leakage after Rectal Cancer Resection and Reconstruction with Colorectostomy. A Retrospective Study with Bootstrap Analysis

Abstract

This study was designed to apply modern statistical methods to evaluate risk factors for anastomotic leakage after rectal cancer resection in a retrospective cohort of patients who received a colorectostomy. Whereas a diverting stoma and tumor height are considered proven risk factors for anastomotic leakage, a lack of evidence about additional risk factors persists. In a single-center study, 527 consecutive patients who received a colorectostomy after rectal cancer resection between 1991 and 2008 were retrospectively assessed. In addition to traditional uni- and multivariate regression, locally weighted scatterplot smoothing (LOWESS) regression and bootstrap analysis were applied to increase internal validity. Anastomotic leakage occurred in 70 patients (13.3%; 95% confidence interval (CI), 10.5-16.5%) and mortality was 2.5% (95% CI, 1.4-4.2%). Diverting stoma (odds ratio (OR), 0.4; 95% CI, 0.17-0.61) and tumor height (OR, 0.88; 95% CI, 0.8-0.94) were proven to be protective. Neoadjuvant radiotherapy (OR, 2.15; 95% CI, 1.58-4.24) and intraoperative blood loss (OR, 1.05; 95% CI, 1.02-1.09) had a derogatory effect. Bootstrap analysis identified pre-existing vascular disease (95.5%), more advanced UICC stage III or IV tumors (95.7% or 91.5%, respectively), and intraoperative (96.1%) and postoperative (99.4%) blood substitution as harmful. Both intraoperative and postoperative blood substitution caused a dose-dependent increase in risk. Applying statistical resampling methods identified intraoperative blood loss, blood substitution, vascular disease, and advanced UICC stage as risk factors for anastomotic leakage. Greater distances between the tumor and the anal verge and performance of a diverting stoma were associated with a decreased risk of anastomotic leakage.