Utility of Removable Esophageal Covered Self-Expanding Metal Stents for Leak and Fistula Management

Transcription

1 Utility of Removable Esophageal Covered Self-Expanding Metal Stents for Leak and Fistula Management Shanda H. Blackmon, MD, MPH, Rachel Santora, MD, Peter Schwarz, MD, Alberto Barroso, MD, and Brian J. Dunkin, MD Departments of Surgery and Medicine, The Methodist Hospital, Houston, Texas Background. Esophageal or gastric leakage from anastomotic wound dehiscence, perforation, staple line dehiscence, or trauma can be a devastating event. Traditional therapy has often consisted of either surgical repair for rapidly diagnosed leaks or diversion for more complicated cases, commonly associated with a delayed diagnosis. This study summarizes our experience treating leaks or fistulas with novel, covered self-expanding metal stents (csems). The primary objective of this study was to determine the efficacy and safety of covered selfexpanding metal stents when used to treat complicated leaks and fistulas. Methods. Over 15 months, 25 patients with esophageal or gastric leaks were evaluated for stenting as primary treatment. A prospective database was used to collect data. Stents were placed endoscopically, with contrast evaluation used for leak evaluation. Patients who did not improve clinically after stenting or whose leak could not be sealed underwent operative management. Results. During a mean follow-up of 15 months, 23 of the 25 patients with esophageal or gastric leaks during a 15-month period were managed with endoscopic stenting as primary treatment. Healing occurred in patients who were stented for anastomotic leakage after gastric bypass or sleeve gastrectomy (n 10). One patient with three esophageal iatrogenic perforations healed with stenting. Eight patients successfully avoided esophageal diversion and healed with stenting and adjunctive therapy. Two of the 4 patients with tracheoesophageal fistulas sealed with the assistance of a new pexy technique to prevent stent migration; 1 additional patient had this same technique used to successfully heal an upper esophageal perforation. Conclusions. Esophageal leaks and fistulas can be effectively managed with csems as a primary modality. The potential benefits of esophageal stenting are healing without diversion or reconstruction and early return to an oral diet. (Ann Thorac Surg 2010;89:931 7) 2010 by The Society of Thoracic Surgeons Accepted for publication Oct 28, Presented at the Fifty-fifth Annual Meeting of the Southern Thoracic Surgical Association, Austin, TX, Nov 5 8, Address correspondence to Dr Blackmon, Department of Surgery, Weill Cornell Medical College, The Methodist Hospital, 6550 Fannin St, Houston, TX 77030; shblackmon@tmhs.org. Esophageal or gastric leakage can be a devastating event. Traditional therapy includes either surgical repair for rapidly diagnosed leaks or diversion for more complicated cases commonly associated with a delayed diagnosis [1, 2]. Esophageal stents are now being used with increasing frequency for disorders such as tracheoesophageal fistula [3, 4], corrosive burn injury [5], spontaneous perforation [6 10], obstructing tumor [11], iatrogenic injuries [8, 12], anastomotic leaks [13 20], and to reinforce traditional repairs [21]. Although the stents may offer protection of esophageal and gastric mucosa while healing takes place, adverse events include stent migration, patient discomfort, difficult removal [22], or even direct injury to adjacent structures [22 25]. Comparison of stent efficacy is difficult due to the variety of etiologies requiring stent placement and lack of standardization with such procedures. There are two commercially available covered selfexpanding metal esophageal stents (csems) in the United States, and neither is approved by the Food and Drug Administration (FDA) for late removal or temporary use for leaks and fistulas. The Alimaxx-e stent (Alveolus, Charlotte, NC) is made in diameters of 18 mm or 22 mm and the length ranges from 12 cm to 7 cm. The Ultraflex (Boston-Scientific, Natick, MA) stent is made in proximal/distal diameters of 23/18 mm and 28/23 mm, respectively, and the length ranges from 10 cm to 15.5 cm, with approximately 1.5 cm of the stent uncovered at each end to prevent migration. Retrieval sutures are attached at the proximal end of the stents for removal or repositioning. This study summarizes our experience using esophageal csems for definitive management of complicated leaks and fistulas. The primary objective of this study was to determine the efficacy and safety of csems when used to treat complicated leaks and fistulas by The Society of Thoracic Surgeons /10/$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 932 BLACKMON ET AL Ann Thorac Surg CSEMS FOR ESOPHAGEAL LEAKAGE 2010;89:931 7 Fig 1. Esophageal pexy procedure to prevent stent migration from upper esophagus. Patients and Methods From August 24, 2007, until October 30, 2008, 25 patients with esophageal or gastric leaks were treated within a single institution. A multidisciplinary team of surgeons and gastroenterologists evaluated patients for endoluminal stenting. A prospective database was used to collect data. Institutional Review Board approval was granted for evaluation of patients with esophageal and gastric leakage in a retrospective manner from a prospective database and consent was waived. The Alveolus Alimaxx-e esophageal stent and Boston Scientific Ultraflex stents were used in patients. Patients were informed these stents are not FDAapproved for temporary use. Traditional surgical consent was obtained before each procedure. Covered SEMS were placed endoscopically using fluroscopy as a guide with a guidewire-assisted deployment. Leak isolation was evaluated using esophagram with limited small bowel follow-through within 24 hours of placement. Patients who did not improve clinically after stenting or whose leak could not be sealed with stent change, reposition, or addition underwent traditional operative treatment or diversion or both. Stents were removed endoscopically before 30 days of treatment, and the esophagus was reevaluated for leak with contrast. Patients with upper esophageal leakage requiring stenting in the area of the upper esophageal sphincter also underwent a novel neck pexy technique to prevent migration. Once the upper esophageal stent was deployed in the desired position under fluoroscopy, an ultrasound probe was used to guide the deployment of a suture-passing needle (W. L. Gore, Phoenix, AZ) lateral to the trachea and medial to the carotid artery through the stent and into the esophagus. A pledgeted suture was passed into the esophagus with a scope and held into position with a grasper as the suture-passing needle was opened and used to capture one side of the suture. This technique was repeated as the second end of the suture was captured, and the two ends of the suture, now externalized, were passed through a second pledget and tied on the outside of the neck (Fig 1). Statistical assessment including duration of follow-up (defined as time from evaluation to last contact), duration of stent therapy, and delay in treatment was calculated using SPSS software version 15 (SPSS, Chicago, IL). Results Patient Characteristics Twenty-three of the 25 patients evaluated for stenting had stents placed. Nineteen women and 6 men were Fig 2. Etiology, chronicity, and outcome of patients undergoing esophageal stenting. One patient had two episodes of stenting for different reasons, thus, n 24 for etiology. (GBP gastric bypass; S successfully treated with esophageal stenting with complete healing; S successful seal with stenting but no healing noted; TE tracheo-esophageal; U unsuccessfully treated with esophageal stenting.)

3 Ann Thorac Surg BLACKMON ET AL 2010;89:931 7 CSEMS FOR ESOPHAGEAL LEAKAGE 933 Table 1. Patient Characteristics Characteristic Number Number of patients 25 Sex, male/female 6/19 Mean age, years Mean follow-up, months Mean delay in treatment, days Mean duration of stent, days evaluated with leakage (1 patient had two episodes of stenting). The mean age was years. Mean follow-up for patients, including the 2 who did not have stents placed, was months. Mean duration of stenting was days. Mean delay to treatment was days. For a tabulation of these results, please see Table 1. Presentation Two patients presented with septic spontaneous (Boerhaave s) esophageal perforation. Five patients had iatrogenic perforations. Four patients presented with tracheoesophageal fistulas (one as a complication from laryngectomy, one as a complication from radiation and chemotherapy for lung cancer, and two from prolonged intubation) and 3 patients presented with esophageal anastomotic leaks. Ten patients had leaks associated with previous gastric bypass; 1 with a fistula to the left chest and empyema. The remaining gastric bypass patients had leakage confined to the abdomen. Etiology of leak, chronicity, and outcomes are detailed in Figure 2. Postoperative Events The most common postoperative event from stent placement was migration (10 of 23; 43%). Six patients required two or more procedures to correct stent position. Other complications included kinking (n 1), small bowel obstruction (n 2), cardiac compression (n 1), pain (n 2), hiccups (n 1), erosion (n 1), and reflux (n 4). For a full report of adverse events, please see Table 2. Stent folding was corrected with repeat endoscopic exchange to a smaller diameter stent. Of those patients having small bowel obstruction, one was unrelated to stenting and caused by a transabdominal jejunal feeding tube and another one was caused by distal migration of the stent; both patients improved with laparoscopic and endoscopic intervention, respectively. The patient who had ventricular tachycardia with echocardiographic appearance of stent compression of the left atrium improved clinically when the stent was removed and immediately underwent surgical diversion. Only 1 of the 2 patients complaining of severe pain from the stent required removal to relieve the pain; the second patient was successfully treated with pain medication. One patient had erosion of the stent into the aorta, had a right-sided aortic arch and a complete vascular ring, a previous laryngectomy with two previous failed muscle flaps to close a tracheoesophageal fistula, and left against medical advice once the stent was deployed. The patient would not return for follow-up or immediate stent removal. He later had a fistula to his aorta develop and returned to our hospital for stabilization, esophagectomy, muscle flap reinforcement of the aorta, division of the Kommerell s diverticulum, endografting, and esophageal reconstruction. The same patient had a leak from an esophageal anastomosis to a small bowel conduit, and this leak healed with stenting. Four patients had severe reflux during the time they had stents (defined as burning in the upper esophagus or reflux of gastric or enteric material into the back of the mouth). No patients required stent removal for reflux, and all were medically treated successfully with acid suppression therapy. Outcomes The majority of patients (61%, 14 of 23) were able to begin a stent diet after a contrast study was performed and confirmed appropriate seal and no leakage. All patients were referred to a nutritionist and given a stent diet. One patient had clinical signs of migration and new leakage after a stent diet was begun (Fig 3), and required stent repositioning before reinstitution of the diet. There were no stent impactions from eating. Of the 23 patients presenting with leakage and fistulas, 11 patients were offered immediate surgical diversion or stenting as primary treatment owing to a long duration of Table 2. Adverse Events After Stenting Complication Total (n) Requiring Reintervention Technical Migration 10 b 10 Stent extraction/new stent 2 Additional stent placed 8 Stent repositioning only 4 Removal of stent 2 c Continued leak requiring 1 1 diversion Continued leak requiring 1 1 exclusion Increasing size of 1 0 perforation or leak during deployment Difficult removal 2 1 Bleeding a 1 1 Erosion 1 1 Symptomatic Pain 2 1 Left atrial 1 1 compression/hypotension (with ventricular tachycardia) Hiccups 2 0 Reflux 4 0 a Clinically significant bleeding requiring transfusion. b One patient had a stent removed for migration into the airway (TEF) and an additional patient had a stent removed for bowel obstruction. c Some patients had more than one intervention for migration.

4 934 BLACKMON ET AL Ann Thorac Surg CSEMS FOR ESOPHAGEAL LEAKAGE 2010;89:931 7 Fig 3. Continued leakage of contrast into left side of chest after initial deployment of stent. leak. Nine patients healed a complicated leak or fistula with the use of stenting instead of traditional diversion. One patient refused initial diversion, failed stenting, and subsequently consented to have diversion performed. One patient successfully healed two of three separate perforations over a 30-day period with stents and eventually healed the third lesion with the aid of esophageal exclusion. One patient presented with a Boerhaave s perforation and had at least a 3-day delay to treatment. Both patients with spontaneous perforation underwent decortication and debridement of the pleural space, muscle flap closure, had stents used to buttress their repair, and healed. Another patient presented 90 days after surgery with a fistulous track from her abdomen to the left side of the chest and healed with fibrin glue injection of the track, decortication, and stenting (Fig 4). Stenting was successful in healing leakage in every gastric bypass patient (n 6) and sleeve gastrectomy patient (n 4). Thirty-day mortality for every patient presenting with an esophageal leak or fistula of the foregut was 16% (4 of 25 patients). Three more patients died after discharge to the hospital. Every patient in whom the stent could not seal a leak or in whom immediate surgical intervention was performed (because stenting was not an option) eventually died. Of the 2 patients who died of sepsis and anastomotic dehiscence after esophagectomy, both presented with empyema and respiratory failure more than a month from the date of their leak with a gap between the esophagus and the stomach conduit. Tracheoesophageal Fistula Outcomes Regarding the efficacy of csems in palliating tracheoesophageal fistulas, only 2 patients successfully sealed. One late-stage lung cancer patient did not want to be intubated and therefore could not have a tracheal stent. Placement of an esophageal stent resulted in intratracheal migration, which was corrected by removal of the stent. This patient died of progressive lung cancer in hospice care 7 days later. A second patient with tracheoesophageal fistula presented in septic shock and died of pneumonia complications after placing the esophageal stent, and a seal could not be confirmed before death. Two of the 4 patients with tracheoesophageal fistulas sealed with the assistance of a new pexy technique to prevent stent migration when placed adjacent to the upper esophageal sphincter; 1 additional patient had this same technique used to successfully heal an upper esophageal perforation. Comment Although not currently FDA-approved for such procedures, there appears to be a real advantage to using esophageal csems. This study summarizes a single institution s experience treating leaks or fistulas of the Fig 4. Esophageal fistula (more than 90 days old) that healed after decortication of chest and stenting. (A) Computed tomography reconstructed image. (B) Contrast leakage into abdomen tracking toward left pleural space.

5 Ann Thorac Surg BLACKMON ET AL 2010;89:931 7 CSEMS FOR ESOPHAGEAL LEAKAGE 935 foregut with two different esophageal stents. With the exception of tracheoesophageal fistulas, every patient in whom a stent was able to seal a leak healed without diversion. Esophageal stents cannot bridge a completely separated anastomosis because these patients lack adequate tissue to seat the stent, which results in migration and poor positioning. Neither of the intrathoracic anastomotic leaks in this series healed, likely because of the more than 30-day delay in treatment, but other studies have shown promise in this area, especially with immediate treatment [13 20]. The two patient groups who received the most benefit in our study were the patients with leaks associated with gastric bypass procedures and patients who would have undergone esophageal diversion and instead had hybrid procedures performed, with stenting used as the primary mode of sealing the leak. Adjunctive procedures, such as decortication and muscle flap reinforcement, still appear to play an integral role when patients present with delayed esophageal or gastric leakage and fistulas. Stenting appears to allow these patient to be treated without traditional diversion and offers early oral nutrition once the leak is sealed. Migration is the predominant problem with esophageal stents both in our own study and in others [18, 26 32]. We utilized several new techniques to limit problems with migration, temporarily suturing stents into position when using them as a buttress to repair [21], bridging stents into distal small bowel, and suturing them in place from the outside. When faced with placing a stent across the upper esophageal sphincter, we developed a new way to secure the stent in the upper neck. Our neck pexy technique is not previously reported, and our team found such a technique to be critical for those patients requiring upper esophageal stenting (Fig 1). A T-fastener was less helpful in this circumstance. None of the patients with upper esophageal stent pexy had tracheal compression or respiratory difficulties, as was expected. In morbidly obese patients with an anastomotic leak after a Roux-en-Y gastric bypass or sleeve gastrectomy, we found the deployment of two stents with a greater than 3 cm overlap prevented further migration. We hypothesize migration often occurs within the larger gastric pouch or sleeve, and such migration was prevented by extending the stent into the distal small bowel. None of the patients in this study have had late strictures or leakage. Other complications such as severe bleeding, erosion into adjacent structures, and obstruction have been reported for a variety of esophageal stents [16, 23 25, 33 36]. Continued leakage should not be accepted if the patient is not clinically improving, as progressive sepsis and death will likely result [26]. Our overall 28% mortality was quite high when compared with other stenting studies [14, 21, 30, 33], and this high rate is likely due to lengthy delays in referral for therapy. There was no immediate procedural mortality. The patients who died were either diverted, had a tracheoesophageal fistula, or died as a complication of their cancer progression. Stent extraction was not difficult if the stents were removed within a 30-day window, except with occasional fracturing encountered removing the Alveolus stents. Animal studies suggest that 30 days should be adequate time for healing [37]. Based on that, we recommend extracting esophageal stents within 30 days and reevaluating the area for leakage. If leakage is discovered after the stents are removed, new stents should be placed and removed before the end of another 30-day period. Only 1 of our patients required two courses of stenting to heal the esophageal injury. Although no direct comparison can be made when comparing this small series to that for plastic stents, it appears the metal stents have a potential to fracture more but may migrate less. Based on this current study, the potential benefits of esophageal stenting with csems in complicated esophageal fistulas and leakage remain healing without diversion or reconstruction and early return to an oral diet. References 1. Bufkin BL, Miller JI, Mansour KA. Esophageal perforation: emphasis on management. Ann Thorac Surg 1996;61: Zwishchenberger JB, Savage C, Bidari A. Surgical aspects of esophageal disease: perforation and caustin injury. Am J Respir Crit Care Med 2002;165: Adler DG, Pleskow DK. Closure of a benign tracheoesophageal fistula by using a coated, self-expanding plastic stent in a patient with a history of esophageal atresia. Gastrointest Endosc 2005;61: Pennathur A, Chang AC, McGrath KM, et al. Polyflex expandable stents in the treatment of esophageal disease: initial experience. Ann Thorac Surg 2008;85: Zhou JH, Jiang YG, Wang RW, et al. Management of corrosive esophageal burns in 149 cases. J Thorac Cardiovasc Surg 2005;130: Eubanks PJ, Hu E, Nguyen D, Procaccino VE, Eysselein VE, Klein SR. Case of Borehaave s syndrome successfully treated with a self-expandable metallic stent. Gastrointest Endosc 1999;49: Fischer A, Thomusch O, Benz S, von Dobschuetz E, Baier P, Hopt UT. Nonoperative treatment of 15 benign esophageal perforations with self-expandable covered metal stents. Ann Thorac Surg 2006;81: Raijman I, Siddique I, Ajnani J, Lynch P. Palliation of malignant dysphagia and fistulae with coated expandable metal stents: experience with 101 patients. Gastrointest Endosc 1998;48; Tsunoda S, Shimada Y, Watanabe G, Nakau M, Inamura M. Covered metallic stent treatment of a patient with spontaneous rupture of the esophagus. Dis Esophagus 2001;14: Petruzziello L, Tringali A, Riccioni, et al. Successful early treatment of Borehaave s syndrome by endoscopic placement of a temporary self-expandable plastic stent without fluroscopy. Gastrointest Endosc 2003;58: Davies N, Thoms HG, Eyre-Brook IA. Palliation of dysphagia from inoperable oesophageal carcinoma using Atkinson tubes or self-expanding metal stents. Ann R Coll Surg Engl 1998;80: Freeman RK, Van Woerkom JM, Ascioti AJ. Esophageal stent placement for the treatment of iatrogenic intrathoracic esophageal perforation. Ann Thorac Surg 2007;83: Langer FB, Wenzl E, Prager G, et al. Management of postoperative esophageal leaks with the Polyflex self-expanding covered plastic stent. Ann Thorac Surg 2005;79: Hunernbein M, Stroszynski C, Moesta KT, Schlag PM. Treatment of thoracic anastomtoic leaks after esophagec-

6 936 BLACKMON ET AL Ann Thorac Surg CSEMS FOR ESOPHAGEAL LEAKAGE 2010;89:931 7 tomy with self-expanding plastic stents. Ann Surg 2004;240: Kauer WKH, Stein HJ, Dittler HJ, Siewert JR. Stent implantation as a treatment option in patients with thoracic anastomotic leaks after esophagectomy. Surg Endosc 2008;22: Roy-Choudry SH, Nicholson AA, Wedgwood KR, et al. Symptomatic malignant gastroesophageal anastomotic leak: management with covered metallic esophageal stents. Am J Roentgenol 2001;176: Freeman RK, Ascioti AJ, Wozniak TC. Postoperative esophageal leak management with the Polyflex esophageal stent. J Thorac Cardiovasc Surg 2007;133: Eubanks S, Edwards CA, Fearing NM, et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg 2008;206: Nguyen NT, Mailey B, Hinojosa M, Chang K. Natural orifice management of anastomotic leaks after minimally invasive esophagogastrectomy. Surg Innov 2008;15: Schubert D, Scheidbach H, Kuhn R, et al. Endoscopic treatment of thoracic esophageal anastomotic leaks by using silicone-covered, self-expanding polyester stents. Gastrointest Endosc 2005;61: Wright C, Mathisen DM, Wain JC, Moncure AC, Hilgenberg AD, Grillo HC. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60: Yoon CJ, Shin JH, Song HY, Lim JO, Yoon HK, Sung KB. Removal of retrievable esophageal and gastrointestinal stents: experience in 113 patients. AJR Am J Roentgenol 2004;183: Kennedy C, Steger A. Fatal hemorrhage in stented esophageal carcinoma: tumor necrosis of the aorta. Cardiovasc Intervent Radiol 2001;24: Muto M, Ohtsu A, Miyata Y, Shioyama Y, Boku N, Yoshida S. Self-espandable metallic stents for patients with recurrent esophageal carcinoma after failure of primary chemotherapy. J Clin Oncol 2001;31: Tomaselli F, Maier A, Pinter H, Smolle-Juttner F. Management of iatrogenous esophageal perforation. Thorac Cardiovasc Surg 2002;50: Ott C, Ratiu N, Endlicher E, et al. Self-expanding Polyflex plastic stents in esophageal disease: various indications, complications, and outcomes. Surg Endosc 2007;21: Sabharwal T, Hamady MS, Chui S, Atkinson S, Mason R, Adam A. A randomised prospective comparison of the Flamingo Wallstent and Ultraflex stent for palliation of dysphagia associated with lower third oesophageal carcinoma. Gut 2003;52: Shin JH, Song HY, Ko GY, Lim JO, Yoon HK, Sung KB. Esophagorespiratory fistula: long-term results of palliative treatment with covered expandable metallic stents in 61 patients. Radiology 2004;232: Adam A, Ellul J, Watkinson A, et al. Palliation of inoperable esophageal carcinoma: a prospective randomized trial of laser therapy and stent placement. Radiology 1997;202: Wang MQ, Sze DY, Wang ZP, Wang ZQ, GaoYA, Dake MD. Delayed complications after esophageal stent placement for treatment of malignant esophageal obstruction and esophagorespiratory fistulas. J Vasc Interv Radiol 2001;12: Fiorni A, Fleischer D, Valero J, Israeli E, Wengrower D, Goldin E. Self-expandable metal coil stents in the treatment of benign esophageal strictures refractory to conventional therapy: a case series. Gastrointest Endosc 2000;52: Radecke K, Gerken G, Treichel U. Impact of self-expanding plastic esophageal stent on various esophageal stenosis, fistulas and leakages: a single-center experience in 39 patients. Gastrointest Endosc 2005;61: Kim AW, Liptay MJ, Snow N, Donahue P, Warren WH. Utility of silicone esophageal bypass stents in the management of delayed complex esophageal disruptions. Ann Thorac Surg 2008;85: Low DE, Kozarek, RA. Removal of esophageal expandable metal stents. Surg Endosc 2003;17: Hramiec JE, O Shea MA, Quinlan RM. Expandable metallic esophageal stents in benign disease: a cause for concern. Surg Laparosc Endosc 1998;8: Farivar AS, Vallieres E, Kowdley KV, Wood DE, Mulligan MS. Airway obstruction complicating esophageal stent placement in two post-pneumonectomy patients. Ann Thorac Surg 2004;78:e Takimoto Y, Nakamura T, Yamamoto Y, Kiyotani T, Teramachi M, Shimizu Y. The experimental replacement of a cervical esophageal segment with an artificial prosthesis with the use of collagen matrix and a silicone stent. J Thorac Cardiovasc Surg 1998;116: DISCUSSION DR RICHARD K. FREEMAN (Indianapolis, IN): I would like to congratulate Dr Blackmon and her colleagues on their presentation and for the intellectual curiosity required to approach these difficult problems in an unconventional manner. The series of 25 patients with esophageal or gastric fistula or anastomotic leak presented today is built on work that has evaluated esophageal stenting for perforations, chronic fistula, and anastomotic leak. However, there are several unique elements contained in this investigation that deserve mention. First, although there are diverse subgroups of patients in this review, one of the accomplishments of this study is the relatively large group of patients treated for leaks after bariatric surgery. In what can be a challenging patient population, endoluminal stent therapy prevented the need for further surgery in all of these patients. Second, although stents have been anchored with endoluminal clips or by means of a suture placed through a counterincision, the technique Dr Blackmon has developed and described in her manuscripts of accomplishing this percutaneously is intriguing. Lastly, the majority of our work as well as others has focused on a silicone-coated polyester or plastic stent. Doctor Blackmon has chosen to use a covered self-expanding metal stent with comparable results. Based on these observations, I would like to offer the following two questions. First, for those of us whose practice utilizes esophageal stent placement and for those who may be considering it, what factors do you consider when deciding which stent to use? Do you find a product you feel comfortable with and adapt it to different situations or do you use different stents for different indications? Second, the patients in this series comprised a diverse group with multiple etiologies for esophageal or gastric leak. With the experience you have accumulated, which patients will you continue to offer endoluminal stent placement and which, if any, will you treat in a different manner? I would like to thank the Association for allowing me to initiate the discussion of this paper. DR BLACKMON: Thank you, Dr Freeman, for your questions and for the contributions you and your partners have made on this topic. Regarding your first question, the migrating stents were a problem. We specifically chose the covered selfexpanding metal stents because I felt like they might migrate less than the Polyflex. And then regarding the ones that we

7 Ann Thorac Surg BLACKMON ET AL 2010;89:931 7 CSEMS FOR ESOPHAGEAL LEAKAGE 937 chose, I prefer the Ultraflex in some circumstances because it has an uncovered beginning and ending and it appears to migrate less. However, the Alveolus stent also seems to have its own advantages. It has a new silicone lining that is thicker and it also has more options with regard to size. I have not seen any clear evidence one covered self-expanding metallic stent is better than the other at this point. Based on your second question, I would not offer stenting for patients with a 360-degree separation or a gap. I would not offer it for tracheoesophageal fistula patients who don t want to be intubated for their tracheal stenting. I would not offer it as a prophylaxis to stent anastomoses. I would possibly use it to treat an anastomotic leak. I would also not use it in a patient with a right-sided aortic arch, a Kommerell s diverticulum, a tracheoesophageal fistula, who had been radiated and left the hospital immediately after stent placement against medical advice, as one of ours did. I would continue to offer stenting for everyone else where it seems appropriate, and we have written a prospective protocol I am happy to share with others. I think it is very important that these new data are continuously scrutinized and that any new patients who are given stents are very carefully followed, as this is an off-label use and strict discussions with patients regarding the fact that this is off-label use should take place before stenting. DR MICHAEL J. DIMAIO (Dallas, TX): Have you tried using a T-bar fastener like you saw in Dallas when you came for our course to fixate either tracheal or esophageal stents as are used for laparoscopic feeding jejunostomy tubes? DR BLACKMON: That was the part of the presentation that I gleaned very quickly over. We use a Gore-Tex suture with a suture passer, and we use ultrasonography to guide the suture passer medial to the carotid and lateral to the trachea with endoscopic guidance. We did try the T-bar fastener, and it put a hole in the Alveolus stent and made it leak, and so we patched it with a Gore-Tex pledget. The pledget and suture technique seemed to work much better and was easier to remove at the time of stent extraction.