Early and late symptomatic anastomotic leakage following low anterior resection of the rectum for cancer: are they different entities?

Aim The aim of the study was to compare patients with symptomatic anastomotic leakage following low anterior resection of the rectum (LAR) for cancer diagnosed during the initial hospital stay with those in whom leakage was diagnosed after hospital discharge. Method Forty-five patients undergoing LAR (n=234) entered into a randomized multicentre trial (NCT 00636948), who developed symptomatic anastomotic leakage, were identified. A comparison was made between patients diagnosed during the initial hospital stay on median postoperative day 8 (early leakage, EL; n=27) and patients diagnosed after hospital discharge at median postoperative day 22 (late leakage, LL; n=18). Patient characteristics, operative details, postoperative course and anatomical localization of the leakage were analysed. Results Leakage from the circular stapler line of an end-to-end anastomosis was more common in EL, while leakage from the stapler line of the efferent limb of the J-pouch or side-to-end anastomosis tended to be more frequent in LL (P=0.057). Intra-operative blood loss (P=0.006) and operation time (P=0.071) were increased in EL compared with LL. On postoperative day 5, EL performed worse than LL with regard to temperature (P=0.021), oral intake (P=0.006) and recovery of bowel activity (P=0.054). Anastomotic leakage was diagnosed most often by a rectal contrast study in EL and by CT scan in LL. The median initial hospital stay was 28days for EL and 10days for LL (P<0.001). Conclusion The present study has demonstrated that symptomatic anastomotic leakage can present before and after hospital discharge and raises the question of whether early and late leakage after LAR may be different entities.