This is a presentation on endoscopic management of an esophago-jejunal-anastomotic disruption with endoluminal vacuum therapy after total gastrectomy. The disclosures are as follows. Esophago-jejunal-anastomotic leak can be a devastating complication after total gastrectomy. Depending on the severity of the leak, therapeutic options can be endoscopic or surgical with the same goal or source control. Here we present a patient with an esophago-jejunal-anastomotic leak that was successfully managed with multimodal endoscopic therapy. Patient is a 76-year-old male with history of peptic ulcer disease who underwent an enterectomy and biliroth 2 reconstruction over 50 years ago who developed high-grade dysplasia in his remnant stomach concerning for invasive malignancy. He was taken to the operating room by surgical oncology and underwent completion total gastrectomy with Roux-en-Y esophago-jejunal reconstruction. His initial postoperative course was complicated by midline fascial dehiscence and esophago-jejunal-anastomotic leak found on imaging. He was taken back to the operating room for washout, drainage, nasal jejunal tube placement, and fascial closure with negative pressure wound therapy dressing. However, the esophago-jejunal leak persisted and minimally invasive surgery was consulted. Patient was taken for endoscopic intervention on postoperative day 14 from his gastrectomy. A GIF HQ190 gastroscope was advanced through the oropharynx into the esophagus. The EJ anastomosis was located 40 cm from the incisors. There was a 30% posterior disruption of the anastomosis with subfibrinous material. There was purulent fluid draining from the right posterior aspect. The anterior 70% of the anastomosis was intact without any concern. We removed the diagnostic scope and advanced a GIF N180 ultraslim scope to explore the abscess cavity outside the wall of the anastomosis. This communicated with the patient's right most strain and it was seen to fill with air. We also saw air and fluid bubbling out of the midline wound suggesting communication with the midline in the upper abdomen. After washing out this cavity, we proceeded with endoluminal vacuum therapy. We re-advanced the diagnostic scope to clear out any remaining fluid and then secured a 3 cm long by 2 cm wide negative pressure therapy foam to the end of a nasogastric tube circumferentially using permanent sutures. This was guided into position endoscopically, placed outside the lumen. Once in place, suction at negative 125 mmHg was applied through the nasogastric tube. The tube was then secured at the nares and the patient was awakened from anesthesia. The patient continued with serial endoscopic exchanges of the endoluminal vacuum therapy system using the same steps as the initial procedure. On postoperative day 8 from his initial endoluminal vacuum therapy, patient was noted to have persistent TAM fluid from his midline negative pressure wound therapy dressing and from his two intra-abdominal drains. These were both assessed to have elevated amylase and lipase levels consistent with a pancreatic leak, which likely caused his midline wound dehiscence and EJ anastomotic disruption. On postoperative day 14, the cavity was notably smaller and had good evidence of granulation tissue circumferentially. There was no contamination on the right side of the abscess cavity and the more lateral drain was present and functional. However, the left side of this cavity, which we explored with an FN190 scope through a small midline defect, had a large amount of pancreatic secretions. We could also see the more medial drain, but this did not appear functional even after flushing it. We endoscopically used a grasper to reposition this left drain deep into the abscess cavity near the location of the suspected pancreatic leak and ensured that it was draining. After repositioning the medial drain, we replaced the endoluminal vacuum system and the patient was awakened from anesthesia. With the confirmed pancreatic leak, our maneuver to reposition the medial drain became the turning point in his care. Shown here is a graph of the patient's daily wound and drain output throughout his admission. Highlighted is the day we repositioned the medial drain and seen after is a consistently lower output volume. Once we effectively controlled the pancreatic leak, the patient's anastomotic disruption and wound began to heal appropriately. Shown is the progression of the anastomotic disruption with each exchange. On post-operative day 24, endoscopic evaluation demonstrated continued healing at the EJ anastomosis and heavy granulation tissue within the abscess cavity. The anastomosis still had two areas of defect on the left side. Definitive closure was attempted endoscopically in this area with five through the scope clips. On post-operative day 31, the right-sided cavity continued to collapse. The lateral drain on the right side, which had previously been seen, continued to scar down and was no longer visible. We insufflated the right-sided drain and the cavity did not fill with air. Therefore, we elected not to replace the indoluminal vacuum device understanding that this area will likely continue to granulate inward without any further endoscopic therapy. We turned our attention to the left side of the anastomosis. When we flushed the medial drain, a large amount of purulent material came through the defect in the anastomosis, localizing a 2-3mm defect in the staple line. We advanced a GIF 1T scope with a 12-6T type over-the-scope clip affixed to the end. The wall of the defect was pulled in with an anchor grasper and centered in the cap. Aggressive suction was applied to draw the walls of the defect into the cap and the clip was fired. At the conclusion of the case, contrast was injected to fill the left-sided abscess cavity and no extravasation was seen fluoroscopically or endoscopically entering the lumen. This allowed the pancreatic leak to be drained exclusively by the left-sided medial drain. No additional endoscopic therapy was indicated. The following day, patient continued to tolerate tube feeds without issues and the drain output continued to be low. On postoperative day 35, the lateral drain was removed. On postoperative day 36, patient was deemed stable for discharge with visiting nurse service for wound and drain care. At 3 weeks, patient underwent a CT scan that did not show extravasation of contrast and was allowed to slowly start oral intake again. His left-sided drain was removed at 4 weeks. After 5 months, patient is tolerating a regular diet, no longer needing his J-tube, and has been without issues. In conclusion, the esophago-jejuno-anastomotic disruption was successfully managed with multimodal endoscopic therapy following the same principles as surgical revision. Patient was able to recover effectively while avoiding additional major operations.

endoscopic management

esophago-jejunal-anastomotic disruption

total gastrectomy

endoluminal vacuum therapy

patient recovery