Computed tomography versus water-soluble contrast swallow in the detection of intrathoracic anastomotic leak complicating esophagogastrectomy (Ivor Lewis): a prospective study in 97 patients.

Abstract

Water-soluble contrast swallow (CS) is usually performed before refeeding for anastomosis assessment after esophagectomy with intrathoracic anastomosis but the sensitivity of CS is low. Another diagnostic approach is based on analysis of computed tomography (CT) scan with oral contrast and of CT mediastinal air images. We undertook to compare them prospectively.
Ninety-seven patients with an esophageal carcinoma operated by intrathoracic anastomosis were included prospectively in a study based on a CT scan at postoperative day 3 (without oral and intravenous contrast) and CT scan and CS at day 7. CT scan analysis consisted of assessing contrast and air leakage. In case of doubt, an endoscopy was done.
A diagnosis of anastomotic leak was made in 13 patients (13.4%), in 2 cases before day 7 and in 3 beyond day 7. At day 3, 94 CT scans were performed, but the diagnostic value was poor. In 95 patients with both CS and CT scan at day 7, CS disclosed a leak in 5 of 11, and CT scan was abnormal in 8 of 11. Leakage of contrast and/or presence of mediastinal gas had the best negative predictive value (95.8%). Endoscopy was done in 16 patients with only mediastinal gas at day 7 CT scan. It disclosed a normal anastomosis in 11, fibrin deposits in 4, and a leak in 1.
In comparison with CS only, CT at day 7 improves the sensitivity and negative predictive value for diagnosing an anastomotic leak. In case of doubt endoscopy is advisable. This approach provides an accurate assessment of the anastomosis before refeeding.