RESUMEN
BACKGROUND: Anastomotic leak (AL) in upper gastrointestinal (UGI) surgery continues to be a diagnostic challenge. We seek to identify clinical parameters that predict AL and examine the effectiveness of investigations in evaluating AL following UGI surgeries. METHODS: 592 patients underwent UGI surgeries with an anastomosis between January 2011 and January 2021. Data on patient characteristics, surgery, postoperative investigations and outcomes were prospectively collected and analysed. RESULTS: The overall occurrence of AL was 6.4 %. Tachycardia >120 BPM (OR 6.959, 95 % CI 1.856-26.100, p = 0.004) and leukocyte count >19 × 109/L (OR 3.327, 95 % CI 1.009-10.967, p = 0.048) were independent predictors of AL. On multivariate analysis, patients whose anastomosis was deemed high risk and had pre-emptive investigation done postoperatively to exclude a leak were less likely to require intervention and were more likely to be managed conservatively (66.7 % vs 14.3 %, p = 0.025). Methylene blue test, oral contrast study and Computed Tomography scan with intravenous and oral contrast had 50.0 %, 20.0 % and 9.1 % false negative results, while esophagogastroduodenoscopy had none. There was no misdiagnosed AL when more than 1 investigation (n = 15, 39.5 %) were performed. CONCLUSION: Our study demonstrates that the presence of a triad including desaturation, tachycardia and leucocytosis predicts for AL following UGI surgery and for confirmation of a leak, evaluation with 2 or more investigation is needed. A practice of evaluating high risk anastomosis prior to commencement of feeding decreased the need for surgical intervention and improves success of conservative treatment.