Comparison of pleural drain amylase and serum C-reactive protein for early detection of intrathoracic esophago-gastric anastomotic leaks

INTRODUCTION: Early detection of anastomotic leaks following esophagectomy has the potential to reduce hospital length of stay and mortality. The aim of this study was to compare the predictive value of pleural drain amylase and serum C-reactive protein for the early diagnosis of leak. METHODS: A retrospective observational cohort study was conducted on 121 patients who underwent Ivor Lewis esophagectomy and intrathoracic gastric conduit reconstruction. Pleural drain amylase levels were measured daily until postoperative day (POD) 5 and compared with CRP values measured on POD 3, 5, and 7. Specificity and sensitivity for both tests, and the respective ROC curves, were calculated. RESULTS: Anastomotic leak occurred in 12 patients. There was a significant statistical association between pleural drain amylase and serum CRP levels and the presence of anastomotic leakage. Pleural drain amylase cutoff of 209 IU/L on POD 2 yielded a sensitivity of 75% and a specificity of 94% (AUC = 0.813), whereas CRP cutoff value of 22.5 mg/dL on POD 3 yielded a sensitivity of 56% and a specificity of 92% (AUC = 0.772). The negative likelihood ratio of pleural drain amylase was 0.27 and 0.12 on POD 2 and 5, respectively. There was no statistically significant difference between ROC curves of amylase and CRP on POD 3 and 5 (p = 0.79 and p = 0.14, respectively). CONCLUSIONS: Pleural drain amylase seems more efficient than serum CRP for early detection of esophago-gastric anastomotic leak. The practice of monitoring drain amylase and CRP may allow safer implementation of enhanced postoperative recovery pathway.