meter achalasia balloon to 5 psi was attempted. A mucosal tear was noted during the procedure proximal to the stricture site, and the procedure was aborted. Despite clinical stability post-operatively, the patient had notable crepitus on exam. An esophagram obtained shortly after revealed a contained leak at the mid-to-distal posterior esophagus. On the CT chest it followed, a large defect in the posterior lateral wall of the mid-esophagus with contrast pooling and extensive pneumomediastinum was noted. Given these findings, the patient was taken for endoscopic closure approximately 4 hours from the time of perforation. Please refer to the following diagrams. On the right is an illustration of the perforation site and distal esophageal stenosis. The table on the left contains icons that will be placed on the esophagus diagram to highlight the location and intervention type used. These include endoscopic clipping, over-the-scope suturing, through-the-scope suturing and stenting. Upon scope insertion, a large transmural defect was noted 22 cm to 30 cm from the incisors. While a series of endoscopic closure techniques would be employed, over-the-scope suturing was decided upon as the best initial modality. In our experience, the perforation was too large for endoscopic sponge therapy, too wide to oppose ends of the perforation with clips, and too long to employ a helical tag system. An over-the-scope suturing device was loaded with a 2.0 polypropylene suture and attached to the endoscope. Taking full thickness bites lateral to the perforation, sutures were placed in running fashion along the defect. This was accomplished by positioning the top jaw of the device in the mediastinum and the bottom in the esophageal lumen. Limited mobility in the proximal esophagus led to the entanglement of lead and trailing sutures. In this instance, biopsy forceps were used to prevent suture crossover. While a mattress suture approach was considered, limited mobility, particularly in the proximal esophagus, made this not feasible. Throughout endoscopic suturing, careful attention had to be paid to the close proximity of mediastinal structures as to not cause atrogenic injury. Upon removing the over-the-scope suturing device, re-inspection of the esophageal perforation was conducted. Suboptimal tissue approximation was noted at the mid-esophageal perforation and at the proximal perforation site. Due to closely approximated tissue, additional repair of the proximal defect was attempted using through-the-scope endoscopic clip placement. After the first clip failed to grasp both sides of the perforation, placement of a second clip was attempted. Despite several attempts at closing the proximal perforation with clips, re-inspection revealed suboptimal tissue approximation and ultimately, both clips were removed. Because the residual perforation was not long, closure with a through-the-scope helical tach system was employed. To reinforce the most proximal end of the perforation, four tachs that were previously connected to a 3.0 polypropylene suture were placed in running Z fashion along the edge of the perforation. After suture completion, a cinch was placed to anchor the tachs in place. To reinforce tissue approximation at the most proximal end of the defect, an endoscopic clip was placed. Compared to the previous attempt at perforation closure using clips, improved tissue approximation was noted. Inspection of the distal perforation site revealed dehiscence and suboptimal approximation of the perforation edges. Given the opposing ends of the perforation were still relatively close, a total of three endoscopic clips were able to successfully reinforce the previously placed sutures and close the persisting defect. Exam of the distal esophagus revealed the intrinsic and severe stenosis at the site of the prior esophageal fundoplication. In an effort to reduce upstream pressure and foster healing of the perforation, an 18 millimeter by 97 millimeter covered metal stent was placed across the stricture. To prevent stent migration, the stent was sutured into place using an over-the-scope suturing device. The stent was not placed over the entire perforation as stent expansion could possibly disrupt the previously conducted endoscopic repair. While the patient would be kept NPO and started on TPN, a nasogastric tube was placed and put to suction to prevent reflux of gastric contents into the esophagus. Repeat esophogram obtained two days after the procedure was negative for any contrast extravasation. A CT scan obtained on the same day revealed persistent but decreased air fluid collections within the mediastinum posterior to the region of the repaired esophageal perforation. Repeat EGD to assess healing occurred approximately one week after the index procedure. Exam of the perforation site revealed persistence of a small defect at the mid esophagus where the previously placed clips had fallen off. Over-the-scope suturing was decided upon to close the persistent defects and repeat suturing of the distal esophagus was conducted. Initial attempts to close the persisting mid esophageal perforation was complicated by an inability to anchor sutures into the friable edges of the perforation. This obstacle was remedied by taking wide bites into the less inflamed and irritated tissue lateral to the longitudinal perforation site. Occasionally, suturing was complicated by deep openings at the site where sutures were placed. To fortify suboptimal tissue approximation of the mid esophagus, a total of three sutures in running fashion were placed. Ultimately, suturing would extend from the most distal to proximal end of the perforation with cinches placed on both ends. At two weeks post-perforation, laparoscopic G2 placement allowed for the initiation of enterofeeding. A CT scan that followed was negative for contrast extravasation and with the aid of previously placed chest tubes, the pneumomediastinum and periesophageal air fluid levels had resolved. Approximately one month after the perforation, the patient underwent repeat EGD for stent removal and post-procedure evaluation. To avoid driving the stent across the prior perforation site, the stent was carefully removed by pulling it through a large endoscope with a 6mm working channel. Subsequently, a thorough exam of the esophagus revealed healthy-appearing mucosa and no evidence of persistent perforation. In over six months since this intervention, the patient has not required additional procedures. This case demonstrates the feasibility of using endoscopic therapy to manage massive defects in the esophagus and highlights the importance of early intervention and enterofeeding when doing so.

esophageal tear

endoscopic closure

covered metal stent

suturing device

early intervention