Diagnosing anastomotic leak post-esophagectomy: a systematic review

Esophagectomy is the gold-standard treatment for esophageal cancer; however, postoperative anastomotic leakage remains the primary concern for surgeons. No consensus exists on the optimal investigations to predict an anastomotic leak. This systematic review aims to identify a single test or combination of tests with acceptable sensitivity and specificity to identify anastomotic leak after esophagectomy and to formulate a diagnostic algorithm to facilitate surgical decision-making. A systematic review of PubMed and EMBASE databases was undertaken to evaluate diagnostic investigations for anastomotic leak post-esophagectomy. Each study was reviewed and where possible, the sensitivity, specificity, positive predictive value, and negative predictive value were extracted. The review identified 3,204 articles, of which 49 met the inclusion criteria. Investigations most commonly used for diagnosis of anastomotic leak were: C-reactive protein (CRP), oral contrast imaging, computed tomography (CT), pleural drain amylase concentration, and the 'NUn score'. The sensitivity of CRP for detecting anastomotic leak varied from 69.2% to 100%. Oral contrast studies sensitivities varied between 16% and 87.5% and specificity varied from 20% to 100%. Pleural drain amylase sensitivities ranged between 75% and 100% and specificity ranged from 52% to 95.5%. The NUn score sensitivities ranged from 0% to 95% and specificity from 49% to 94.4%. No single investigation was identified to rule out anastomotic leak in asymptomatic patients. However, the authors propose a diagnostic algorithm incorporating CRP, pleural drain amylase concentration, and CT with oral contrast to aid clinicians in predicting anastomotic leak to facilitate safe, timely discharge post-esophagectomy.