CT scan-based modelling of anastomotic leak risk after colorectal surgery.

ABSTRACT

AIM: Prolonged ileus, low-grade fever and abdominal discomfort are common during the first week after colonic resection. Undiagnosed anastomotic leak carries a poor outcome and computed tomography (CT) scan is the best imaging tool for assessing postoperative abdominal complications. We used a CT scan-based model to quantify the risk of anastomotic leak after colorectal surgery. METHOD: A case-control analysis of 74 patients who underwent clinico-radiological evaluation after colorectal surgery for suspicion of anastomotic leak was undertaken and a multivariable analysis of risk factors for leak was performed. A logistic regression model was used to identify determinant variables and construct a predictive score. RESULTS: Out of 74 patients with a clinical suspicion of anastomotic leak, 17 (23%) had this complication confirmed following repeat laparotomy. In multivariate analysis, three variables were associated with anastomotic leak (1) white blood cells count > 9 × 10(9) /l (OR = 14.8); (2) presence of ≥ 500 cm(3) of intra- abdominal fluid (OR = 13.4); and (3) pneumoperitoneum at the site of anastomosis (OR = 9.9). Each of these three parameters contributed one point to the risk score. The observed risk of leak was 0, 6, 31 and 100%, respectively, for patients with scores of 0, 1, 2 and 3. The area under the receiver operating characteristic curve for the score was 0.83 (0.72-0.94). CONCLUSION: This CT scan-based model seems clinically promising for objective quantification of the risk of a leak after colorectal surgery.