Peroral transthoracic rendezvous for endoscopic closure of persistent gastroploidal fistula that has failed surgical management. These are our disclosures. Gastroploidal fistula is a rare but serious complication that can occur after sleeve gastrectomy. Management strategies include conservative management with NPO and TPN as well as percutaneous drainage, endoscopic drainage and or closure, and surgical management. As with all fistulas, factors such as inflammation, infection, epithelialization and distal obstruction can reduce the chance of spontaneous closure of the fistula. In patients with sleeve gastrectomy anatomy, chronic fistulas are particularly challenging as epithelialization and the high-pressure system of the gastric sleeve further complicate fistula management. This is the case of a 38-year-old female who underwent laparoscopic sleeve gastrectomy complicated by a sleeve leak near the GE junction. This was MAJ-V exploratory laparotomy. However, she developed a gastroploidal fistula that failed further surgical management with a repeat exploratory laparotomy as well as an open left lower pulmonary lobectomy. She presented to our group having been NPO and TPN with a chronic indwelling chest tube for four years. She had frequent ED visits with fever and purulent drainage from the chest tube. Due to failure of prior surgical management, she was offered a staged endoscopic attempt at fistula closure. The planned endoscopic intervention was based on the ligation of intrasphyncteric fistula tract or LIFT procedure used to manage perianal fistula surgically. This involves an incision between the inner and outer anal sphincter muscles with identification of the fistula tract and subsequent ligation and separation of the tract. Our endoscopic approach borrows principles of LIFT. In the first stage, we perform submucosal endoscopic tunnel dissection from the esophagus to the fistula tract at the EGJ to fenestrate the fistula tract, then perform an ESD to resect the epithelialized opening of the fistula tract on the gastric side. In the second stage, we perform thoracic fistuloscopy and rendezvous a cholangioscope from the chest to the stomach to APC ablate the entire fistula tract prior to closure with an over-the-scope clip. Here we see the fistula just distal to the Z line. Closed inspection of the fistula reveals a well epithelialized tract consistent with chronic fistula as seen here. We begin the first stage by making a mucosotomy six centimeters proximal to the Z line and creating a submucosal tunnel to the fistula. At the fistula site, there is significant inflammation and obliteration of the submucosal plane. We used a digital cholangioscope as seen here in the lower left corner, introduced periorally into the fistula to identify the fistula tract within the inflamed, scarred submucosa. We followed the light to identify the area where fenestration of the fistula tract could be performed. We confirmed fenestration by injecting methylene blue solution from the fistula and observing it enter the submucosal plane. We repeated this fenestration every few millimeters along a two centimeter length of the tract. We then performed an ESD around the opening of the fistula tract on the gastric side, dissecting circumferentially around the tract to isolate the opening. Once isolated, we then dissected distally along the tract in an attempt to resect as much of the fistula tract as we were able to safely. This included dissecting into the muscular wall. We then used hemostatic graspers to seal the fistula tract within the muscle. The resultant defect was then carefully examined for hemostasis. An over-the-scope suturing device was used to place a figure of eight suture along the distal end of the resection bed. The more proximal end presented a difficult angle for over-the-scope suture closure, so a through-the-scope suturing device was used to close the defect proximally. Post-closure inspection confirmed complete closure of the ESD site. We then used through-the-scope clips to close the esophageal mucosotomy site. The post-op day 1 upper GI study showed no leak. The patient was discharged home with the plan to remain NPO and on TPN until subsequent evaluation. One week later, the patient reported a cough and left upper quadrant pain. An upper GI study confirmed a recurrent fistula. As we had counseled the patient on this possibility preoperatively, we moved forward with the second stage of the procedure, the transthoracic rendezvous to ablate the fistula tract. Unfortunately, a technical issue prevented capture of video from the endoscopic view of the second stage of the procedure. However, the fluoroscopic images clearly demonstrate the technique. We began with an EGD injection of the small residual opening of the fistula tract at the area where we had used the through-the-scope suturing device. The fluoroimages show the fistula tract extending into the thorax. We thus advanced a 450 cm long flexible-tip wire through the endoscope and into the fistula tract. Simultaneously, we fed a flexible-tip wire through the patient's chest tube into the thoracic cavity. Under fluoroscopy, the wire from the gastric side was advanced through the fistula tract until it was level with the wire from the chest tube site. The chest tube was removed over the wire and a digital cholangioscope was introduced through the thoracic opening of the fistula tract into the pleural space. The fistula tract was explored using the cholangioscope until the gastric side wire was identified. A cholangioscope biopsy forcep was used to grasp the gastric wire and pull it out of the thoracic opening of the fistula, establishing wire access from the patient's mouth through the fistula to the chest wall. The cholangioscope was then passed over the wire down the fistula from the chest to the gastric lumen to rendezvous with our gastroscope. We then passed an APC catheter through the gastroscope into the fistula tract and used the cholangioscope to monitor ablation of the fistula tract from the stomach out to the pleural space to ensure complete deepithelialization of the fistula tract. Once completed, we advanced the percutaneous chest tube to the approximate level of the diaphragm to promote drainage of any residual fluid. We then used an over-the-scope clip to close the gastric opening of the fistula tract. An influoroscopy showed no leak and a post-update to upper GI series also confirmed no leak. The patient's chest tube was slowly backed out over the course of several months to promote collapse of the remaining fistula tract. An 18-month upper GI swallow study shows no continuing leak. EGD shows a small indentation at the site of the prior fistula tract. The patient is tolerating a regular diet and gaining weight and overall doing well. Deepithelialization of a fistula tract can make management of chronic fistulas difficult. It is thus important to achieve complete deepithelialization of a chronic fistula tract to help promote the fistula closure. In this case, we demonstrate that principles of fistula management can be borrowed from other procedures such as the lift procedure in colorectal surgery. In conclusion, multidisciplinary collaboration can facilitate unique approaches to the management of difficult fistulas. Persistent gastropleural fistulas may be managed endoscopically even after failure of surgical management. However, patients should be counseled on the likely need for multiple procedures.

gastropleural fistula

endoscopic intervention

LIFT procedure

sleeve gastrectomy

multidisciplinary approach