Risk factors, short and long term outcome of anastomotic leaks in rectal cancer

Background: An anastomotic leak (AL) after colorectal surgery is one major reason for postoperative morbidity and mortality. There is growing evidence that AL affects short and long term outcome. This prospective German multicentre study aims to identify risk factors for AL and quantify effects on short and long term course after rectal cancer surgery. Methods: From 1 January 2000 to 31 December 2010 381 hospitals attributed patients to the prospective multicentre study Quality Assurance in Colorectal Cancer managed by the Otto-von-Guericke-University Magdeburg (Germany). Included were 17 867 patients with histopathologically confirmed rectal carcinoma and primary anastomosis. Risk factor analysis included 13 items of demographic patient data, surgical course, hospital volume und tumour stage. Results: In 2 134 (11.9%) patients an AL was diagnosed. Overall hospital mortality was 2.1% (with AL 7.5%, without AL 1.4%; p < 0.0001). In multivariate analysis male gender, ASA-classification >= III, smoking history, alcohol history, intraoperative blood transfusion, no protective ileostomy, UICC-stage and height of tumour were independent risk factors. Overall survival (OS) was significantly shorter for patients with AL (UICC I-III; UICC I, II or III - each p < 0.0001). Disease free survival (DFS) was significantly shorter for patients with AL in UICC I-III; UICC II or UICC III (each p < 0.001). Rate of local relapse was not significantly affected by occurrence of AL. Conclusion: In this study patients with AL had a significantly worse OS. This was mainly due to an increased in hospital mortality. DFS was also negatively affected by AL whereas local relapse was not. This emphasizes the importance of successful treatment of AL related problems during the initial hospital stay.