A Prospective Analysis of Risk Factors, Presentation and Outcome of Leakage After Gastrointestinal Anastomoses

Abstract

Background: Despite advancements in modern surgery and postoperative care, disruption of gastrointestinal anastomosis remains the most dreaded complication, even in experienced surgical hands. The cause of leakage is multifactorial consisting of a complete spectrum of pre, intra and postoperative factors. Search for an ideal gastrointestinal anastomosis still remains an unquenched thirst. Study Design: Prospective, hospital based, time bound observational study. Methods: After ethical clearance, 288 consenting adult patients who underwent gastrointestinal anastomosis were observed for risk factors, presentation and outcome of leakage and evaluated using appropriate statistical tools. Results: An overall gastrointestinal anastomotic leak rate of 15.28% with peak incidence at 41-50 years (19.51%) was seen. Peritonitis (p=0.0009, OR=2.9611), COPD (p=0.0181, OR=2.7306), low serum albumin concentration (p=0.0028, OR=3.1442), ASA status of ≥III (p=0.0001, OR=4.0281) and a perioperative blood transfusion requirement of ≥2 units (p=0.0028, OR=3.1442) were the most signicant risk factors associated with leakage. Obstruction (p=0.0160, OR=2.2310), malignancy (p=0.0149, OR=2.6961), steroid therapy (p=0.0176, OR=2.2741), chemoradiation (p=0.0400, OR=2.4889), diabetes (p=0.0427, OR=2.2689), undernutrition (p= 0.0308, OR= 2.1099), anaemia (p=0.0325, OR=2.0183) and sepsis (p=0.0187, OR=2.2702) also showed clear risk augmentation. Risk of leakage was increased with a surgical duration of >4 hours (p=0.0078, OR=2.5610), when anastomosis was done as an emergency procedure (p=0.0427, OR=2.6571) or by a surgeon with expertise of ≤5 years (p=0.0338, OR=2.7733). Neither the level, type, technique of anastomosis; nor the usage of surgical staplers had an impact on leakage. Preoperative bowel preparation and creation of a proximal stoma also had minimal effect on leakage rates; though, the infectious complications that follow were greatly reduced. The most common presentation of anastomotic leak was a suspicious drain output with a mean time of 7.59± 2(2.48) postoperative days; resulting in a prolongation of hospitalization by more than ten days (p<0.0001), along with an increased mortality rate (p<0.0001). Conclusions: Accurately predicting anastomotic leakage still requires more evidence-based information. Even with good risk stratication, many causative factors may not be amenable to immediate correction in the pre-operative period. In such cases, the patient must be considered as a candidate for an enterostomy to help tide the crisis over.