Oncological outcome following anastomotic leak in rectal surgery

Background: The influence of anastomotic leak on local recurrence and survival remains debated in rectal cancer. Methods: This was a multicentre observational study using data from the Spanish Rectal Cancer Project database. Demographics, American Society of Anesthesiologists classification, tumour location, stage, use of defunctioning stoma, administration of neoadjuvant and adjuvant treatment, invasion of circumferential resection margin, quality of mesorectal excision and anastomotic leakage were recorded. Anastomotic leak was defined as an anastomotic event requiring surgical intervention or interventional radiology, including pelvic abscesses without radiological evidence of leakage and early rectovaginal fistulas. Variables associated with oncological outcome were assessed by multivariable Cox regression analysis. Results: A total of 1181 consecutive patients were included. Rates of anastomotic leak and 30-day postoperative mortality were 9.4 and 2.4 per cent respectively. Data from 1153 patients were analysed after amedian follow-up of 5 years. Cumulative rates of local recurrence, overall recurrence, overall survival and cancer-specific survival were 4.9, 19.4, 77.5 and 84.7 per cent respectively. In the multivariable regression analysis, anastomotic leakage was not associated with local recurrence (hazard ratio (HR) 0.80, 95 per cent c.i. 0.28 to 2.26; P = 0.669), overall recurrence (HR 1.14, 0.70 to 1.85; P = 0.606), overall survival (HR 1.10, 0.73 to 1.65; P = 0.648) or cancer-specific survival (HR 1.23, 0.75 to 2.02; P = 0.421). Conclusion: Anastomotic leak after low anterior resection did not affect oncological outcomes in these patients.