Risk Factors for Anastomotic Leak After Colon Resection for Cancer Multivariate Analysis and Nomogram From a Multicentric, Prospective, National Study With 3193 Patients

Objective: To determine pre-/intraoperative risk factors for anastomotic leak after colon resection for cancer and to create a practical instrument for predicting anastomotic leak risk. Background: Anastomotic leak is still the most dreaded complication in colorectal surgery. Many risk factors have been identified to date, but multicentric prospective studies on anastomotic leak after colon resection are lacking. Methods: Fifty-two hospitals participated in this prospective, observational study. Data of 3193 patients, operated for colon cancer with primary anastomosis without stoma, were included in a prospective online database (September 2011-September 2012). Forty-two pre-/intraoperative variables, related to patient, tumor, surgical procedure, and hospital, were analyzed as potential independent risk factors for anastomotic leak (60-day follow-up). A nomogram was created to easily predict the risk of anastomotic leak for a given patient. Results: The anastomotic leak rate was 8.7%, and widely varied between hospitals (variance of 0.24 on the logit scale). Anastomotic leak significantly increased mortality (15.2% vs 1.9% in patients without anastomotic leak, P < 0.0001) and length of hospitalization (median 23 vs 7 days in uncomplicated patients, P < 0.0001). In the multivariate analysis, the following variables were independent risk factors for anastomotic leak: obesity [P = 0.003, odds ratio (OR) = 2.7], preoperative serum total proteins (P = 0.03, OR = 0.7 per g/dL), male sex (P = 0.03, OR = 1.6), ongoing anticoagulant treatment (P = 0.05, OR = 1.8), intraoperative complication (P = 0.03, OR = 2.2), and number of hospital beds (P = 0.04, OR = 0.95 per 100 beds). Conclusions: Anastomotic leak after colon resection for cancer is a frequent, relevant complication. Patients, surgical technique, and hospital are all important determining factors of anastomotic leak risk.