Diagnostic performance of postoperative endoscopy versus water-soluble esophagram in detecting anastomotic leak after...

Water-soluble esophagram and endoscopy are commonly used to detect anastomotic leak after esophagectomy. However, the comparative performance of these two modalities in detecting leaks has not been well established. We retrospectively reviewed all patients who underwent esophagectomy for cancer between January 2021 and December 2022 from a prospectively maintained database. Occurrence, management and consequences of leaks were examined. The primary outcome was to compare the diagnostic performance of esophagram and endoscopy in detecting anastomotic leak. Therefore, a total of 846 patients were identified. The overall incidence of anastomotic leak was 10.4%. Patients with leaks had older age and more retrosternal routes for reconstruction. Of them, 466 patients underwent endoscopy and 217 underwent esophagram. After matching, 434 patients, 217 per group were analyzed. The sensitivity and negative predictive value of endoscopy were 79.2% and 97.5%, whereas those of esophagram were only 37.9% and 91.3%, respectively (P = 0.003; P = 0.007). In the subset of patients with cervical anastomosis, endoscopy also demonstrated superior sensitivity (78.3% vs 36.0%, P = 0.004) and negative predictive value (97.4% vs 90.8%, P = 0.007). Esophagram showed false-negative results in 18 patients, while only 5 leaks were missed by initial endoscopy. Regarding the management of false negatives, esophagram group showed non-significant trends toward longer hospital stay and more interventions, including stenting (11.1% vs 0%) and clipping (5.6% vs 0%) (both P > 0.05). In conclusion, endoscopy showed higher sensitivity and negative predictive value than esophagram in detecting anastomotic leak after esophagectomy in this study. However, a large-scale, randomized study is warranted to validate these results.