784. NO SUPPORT FOR IMMEDIATE POSTOPERATIVE REMOVAL OF NASOGASTRIC TUBE AFTER ESOPHAGECTOMY, RESULTS FROM A NORDIC MULTICENTER RANDOMIZED CONTROLLED TRIAL

Abstract Background Esophagectomy is central in curative treatment of esophageal and gastroesophageal junctional cancer. Postoperative nasogastric (NG) tube for drainage of the gastric conduit is routine in most centers. The NG tube is typically associated with significant discomfort for the patients and may have risks of its own. Immediate postoperative removal of the NG tube has been suggested and deemed safe in several smaller trials. We hypothesized that immediate postoperative removal of the NG tube is non-inferior to keeping the NG-tube in place for 5 days postoperatively, with regard to anastomotic leak and other early outcomes. Methods Esophagectomy patients with a gastric conduit reconstruction were included in a multicenter, randomized controlled trial between February 2022 and March 2024. A 1:1 randomization stratified for center, age and anastomotic site (thorax/neck) was performed between immediate postoperative removal of NG tube (intervention) and 5 days of NG tube use (control). All patients underwent chest CT postoperative day 7. Anastomotic leak was the primary endpoint and a non-inferiority threshold of -9% difference in proportion of anastomotic leak, including 95% confidence interval, was set. Secondary endpoints included overall complications, pneumonia, length of stay as well as days in high dependency ward. Results Intention to treat analyses were performed on 444 patients whereof 215 were randomized to immediate NG-tube removal and 229 to control. Mean age was 67 years and neck anastomosis was performed in 78 patients (18%). 47 patients (22 %) had leak in the no NG tube group compared to 35 (15%) in the control group. Non-inferiority could not be established with a difference leak proportion to the advantage of NG tube use of 6.6% (95% CI: 13.8%-(-0.7%)). Overall complications (Clavien-Dindo >2) occurred in 94 patients (44%) in the experimental group and 91 patients (40%) in the control group. Overall 30d mortality was 1.1%. Conclusion In this randomized controlled multicenter trial, the by far largest to be performed so far, we could not establish non-inferiority for abstaining from postoperative NG tube use after esophagectomy with regards to anastomotic leak, and therefore support NG tube use. The somewhat high proportion of leak could be partly due to meticulous complication registration, including CT with peroral contrast on day seven in all patients. Use of postoperative decompression of the gastric conduit with an NG tube should be encouraged.