Classification of Esophageal Stent Leaks: Leak Presentation, Complications, and Management Elizabeth H. Stephens, MD, PhD, Arlene M. Correa, PhD, Min P. Kim, MD, Puja Gaur, MD, and Shanda H. Blackmon, MD, MPH The Methodist Hospital, and The University of Texas MD Anderson Cancer Center, Houston, Texas; and Weill Cornell College of Medicine, Columbia University, New York, New York
Background. Esophageal stent leaks can have catastrophic consequences if not promptly recognized and managed appropriately. However, there are different mechanisms for esophageal stent leaks that may demonstrate unique features in presentation and response to management strategy. The objective of this study was to develop a classification system for esophageal leaks and assess distinctions between leak types. Methods. Patients with esophageal stent leaks from 2007 to 2010 managed at The Methodist Hospital were classified into the following 5 leak types: type 1, proximal; 2, distal retrograde; 3, stent lining; 4, between stents; and 5, migrated stent. Patients’ baseline characteristics, procedural data, and outcomes were analyzed. Results. Of the 89 patients who underwent esophageal stenting, 23 stent leaks were identified after the first procedure. Mean age was 57 ± 14 years, 61% were male, 43% had esophageal cancer, and 52% were status
postesophagectomy. Seven of the leaks were type 1, 6 were type 2, 2 were type 3, 4 were type 4, and were type 5. The vast majority (70%) of leaks were detected within the first 48 hours. The management of leaks varied significantly depending on the leak type (p < 0.001) and included additional stenting, placement of a larger stent, bridle, percutaneous gastrostomy, stent exchange, observation, and surgery. The majority of leaks (65%) ultimately resolved. Survival according to leak type was not different (p [ 0.072). Conclusions. Esophageal leaks tend to be managed differently depending on leak type. The majority of leaks ultimately resolve with stenting. Our proposed leak classification may enhance esophageal stent management strategy.
E
varies, but can be seen in approximately 10% to 40% of patients [14–17]. Esophageal stent leaks can have catastrophic consequences if not recognized and managed appropriately; furthermore, they may demonstrate unique clinical presentations and response to treatment strategy depending on the patients’ underlying pathophysiology and mechanism of leak. For example, esophageal stent leak secondary to stent migration can be managed or prevented by bridling, pexy, anti-migration stent design, or oversizing. However, no classification exists for defining the type of stent leak. Furthermore, the optimal treatment strategy for certain types of esophageal stent leaks, such as those secondary to stent migration, in specific patient populations, remains to be determined. The objective of this study was to develop a classification system for esophageal leaks, similar to that used in classification of endovascular stent leaks [18], and assess distinctions between leak types in terms of presentation and management strategies to aid in the prompt recognition and optimal treatment of stent leaks in specific patient populations.
sophageal stents are used for a large variety of patients and clinical scenarios including esophageal perforation [1, 2], tracheoesophageal fistulas [3, 4], anastomotic leaks encountered after esophagectomy or gastric bypass [5–8], iatrogenic injuries [9, 10], corrosive burn injuries [11], obstructive lesions such as tumors [12], and to reinforce traditional repairs [13]. While the traditional treatment for esophageal and gastric leaks includes surgical diversion or surgical repair, esophageal stenting provides a less invasive treatment strategy that can be beneficial in patients who may not tolerate surgery [1, 2]. Stenting can also allow the salvage of an esophageal leak that otherwise would not heal because of delayed presentation. While esophageal stenting has become an increasingly utilized technique in esophageal disease, the optimal management of these patients remains to be determined. Esophageal stent leaks are a dreaded complication of stent placement. The reported incidence of stent leak
(Ann Thorac Surg 2014;-:-–-) Ó 2014 by The Society of Thoracic Surgeons
Accepted for publication Jan 28, 2014. Presented at the Sixtieth Annual Meeting of the Southern Thoracic Surgical Association, Scottsdale, AZ, Oct 30–Nov 2, 2013. Address correspondence to Dr Blackmon, Division of Thoracic Surgery, Houston Methodist Hospital, 6550 Fannin St, Smith Tower, Ste 1661, Houston, TX 77030; e-mail: [email protected].
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
Dr Blackmon discloses financial relationships with Covidien, Ethicon, and Maquet; and Dr Kim with Covidien and Ethicon.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.01.063
2
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
Patients and Methods Patients A prospectively maintained database of esophageal stents placed from 2007 to January 2013 at The Methodist Hospital was searched for esophageal stent leaks, which were then classified into the following 5 leak types on the basis of radiographic and intraoperative assessment (Fig 1): type 1 ¼ proximal; type 2 ¼ distal retrograde; type 3 ¼ stent lining; type 4 ¼ between stents; and type 5 ¼ migrated stent. Patients’ baseline characteristics, procedural data, and outcomes were analyzed. Time to leak detection was defined as the time from clinical signs and symptoms suggestive of leak to diagnosis of leak. A waiver of informed consent was obtained along with Institutional Review Board approval. An investigational device exemption from the US Food and Drug Administration was obtained to investigate the use of off-label esophageal stenting, and some of the patients in this database elected to participate in a prospective study to evaluate the safety and efficacy of esophageal stenting in the setting of leaks or fistulae. All stents were covered stents and included Wallflex (Boston Scientific, Natick, MA), Allimax (Alveolus, Charlotte, NC), and Ultraflex (Boston Scientific).
Statistical Analysis Statistical comparisons between groups were made using c2 testing and log-rank analysis of Kaplan-Meier curves were used to compare survival. Statistical significance was defined as a p value less than 0.05. All statistical analyses were performed using SPSS version 19.0 (SPSS, Chicago, IL).
Results Patient Characteristics There were 108 patients who presented to Houston Methodist Hospital for evaluation of an esophageal leak
Ann Thorac Surg 2014;-:-–-
from 2007 to January 2013; of these, 89 (82%) patients had a stent placed to initially treat their foregut leak or fistula. Of these 89 patients, 23 (26%) patients presented with a leak after their initial stent was placed and required further intervention to seal the leak (Table 1). Fourteen patients (61%) were males and the mean age was 57 14 years old. A large portion (43%) of patients had esophageal cancer and 52% were status postesophagectomy. Seven of the leaks were type 1, 6 were type 2, 2 were type 3, 4 were type 4, and 5 were type 5. The vast majority (70%) of leaks were detected within the first 48 hours and confirmed radiographically (see examples of each type of leak in Figure 1). The management of leaks varied significantly dependent on leak type (p < 0.001) and included additional stenting, placement of a larger stent, bridling the stent to the nasal septum, gastrostomy feeding tube placement, stent exchange, observation, and surgery. The majority of these leaks (65%) ultimately resolved and the majority of those that did not resolve with stent placement were patients with a significant delay between diagnosis and appropriate intervention. There was a wide range in the delay in diagnosis, with some patients transferring into our institution leaking for 1 to 2 weeks before diagnosis or intervention. However, once the diagnosis was made intervention was made usually that day. The delay between diagnosis and seal or diversion also widely varied from 1 to 2 days to 1 to 3 months. Delay in diagnosis and delay in obtaining seal or diverting was not found to be different between leak types or associated with adverse overall outcome. Similarly, presentation at outside hospital (7 of the patients) was not associated with decreased survival or likelihood of seal. Overall survival was not significantly different between leak types (p ¼ 0.072). Type 1 (proximal) leaks were predominantly seen after esophagectomy, most commonly performed for esophageal cancer. Out of the 7 patients with type 1 leaks, 4 were successfully managed with additional stent placement,
Fig 1. Fluoroscopic imaging illustrating the different types of esophageal stent leaks (type 1 to type 5).
Ann Thorac Surg 2014;-:-–-
3
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
Table 1. Pathology, Management and Clinical Presentation of Stent Leaks Classified by Leak Type (1 Through 5) Total Cohort (n¼23) 1 (n ¼ 7, 30%) 2 (n ¼ 6, 26%) 3 (n ¼ 2, 9%) 4 (n ¼ 4, 17%) 5 (n ¼ 4, 17%) p Value
Variable Average age Gender (% male) Previous surgery/interventions: Gastric bypass or sleeve Gastrectomy Esophagectomy Eso dilation Eso hernia repair Cardiac Underlying pathology: Eso cancer Stricture Iatrogenic Prolonged vent Boerhaaves Management of leak: Additional stent only Larger stent Bridle PEG Stent exchange Observation Surgery Timing of leak detection: 14 days Resolution of leak: Mean follow-up (months) % survival last follow-up
57 14 61%
56 9 71%
59 11 66%
66 5 50%
50 16 50%
59 27 50%
5 1 12 1 1 1
(22%) (4%) (52%) (4%) (4%) (4%)
1 (14%) 1 (14%) 3 (43%) 0 1 (14%) 0
1 (17%) 0 3 (50%) 1 (17%) 0 0
0 0 2 (100%) 0 0 0
2 (50%) 0 1 (25%) 0 0 1 (25%)
1 (25%) 0 2 (50%) 0 0 0
10 2 2 1 1
(43$) (9%) (9%) (4%) (4%)
4 (57%) 0 2 (29%) 0 0
2 (33%) 1 (17%) 0 0 1 (17%)
1 (50%) 1 (50%) 0 0 0
1 (25%) 0 0 0 0
2 (50%) 0 0 1 (25%) 0
4 3 3 3 2 2 6
(17%) (13%) (13%) (13%) (9%) (9%) (26%)
4 (57%) 2 (29%) 0 0 0 1 (14%) 0
0 0 0 3 (50%) 0 0 3 (50%)
0 0 0 0 2 (100%) 0 0
0 0 0 0 0 1 (25%) 3 (75%)
0 1 (25%) 3 (75%) 0 0 0 0
7 (100%) 0 0 5 (71%) 10 13 86%
2 (33%) 3 (50%) 1 (17%) 4 (66%) 11 10 100%
1 (50%) 1 (50%) 0 1 (50%) 20 100%
2 (50%) 0 2 (50%) 2 (50%) 13 10 75%
4 (100%) 0 0 3 (75%) 99 100%
NS NS NS
NS
6 cm transthoracic repair Stent PEG for gastric decompression
Type 2 leak, stent migration, incomplete healing of leak
G-tube ¼ gastrostomy tube;
j-tube ¼ jejunostomy tube;
repair. For intraabdominal esophageal or gastric perforations they recommended transabdominal repair and jejunostomy. While specific recommendations vary between authors, there is a consensus that the use of stents must be selective and individualized, taking into account the individual patient’s underlying etiology and anatomy, as well as clinical status and comorbidities. Continued reevaluation of the patient is also required to determine whether the leak has sealed and when additional interventions are necessary. At our institution all patients undergo esophagram within 24 hours to confirm that the stent leak has sealed. Of note, in this 2012 study Freeman and colleagues also developed a classification system for esophageal leaks prior to stent placement based on the location, mechanism, and size of the leak, as well as if the leak was associated with malignancy or distal stricture. As our classification system is validated with more patients, refinement in the classification system may include parameters such as location and quantification of leak. For instance, if the leak was large but well contained with chest tube drainage, this would be managed differently than a freely extravasating leak without chest tube drainage or a contained leak; these aspects could be incorporated into the classification system to further aid management decisions. Similarly, a type 5 leak in the neck is distinct in terms of leak mechanism and management compared with a type 5 leak in the midthoracic region. Equally important in the determination of when and which patients may benefit from stenting is recognition of which patients would not benefit from stenting. Patients with conduit ischemia and large, uncontrolled leak are patients who will not benefit from stenting. Repeat stenting is used if the patient is stable, is tolerating diet, or is a particularly high surgical risk. If the leak is not sealed within 24 to 48 hours they are not likely to seal; depending on the clinical scenario and patient status either another stent is placed, with an adjunctive muscle flap if appropriate, or surgery is warranted. Of note, stents are not left in for over a month in order to avoid ingrowth that can cause problems on removal and erosion into adjacent vascular structures, but rather a
Type 2 leak, aspiration if stenting across esophagojejunostomy anastomosis Stent migration
Stent migration
Type 2 leak, aspiration (open GEJ) PEG ¼ percutaneous endoscopic gastrostomy.
stent exchange is performed if the patient continues to need a stent to seal the leak. Given a recent report demonstrating that most esophageal-aortic fistulas occur at 26 to 36 days [21], our current recommendation is to not leave stents in for longer than 3 weeks. Overall, esophageal stents are an excellent tool in the appropriate clinical scenario with certain pathophysiology and anatomy. Not all patients with esophageal leaks require operative management and in many cases esophageal stenting is an effective management strategy with lower risk of morbidity and mortality than open operative intervention. However, in patients who do not seal after 2 attempts or become unstable, prompt diversion should be considered as further delay with a persistent esophageal leak carries a poor prognosis. As we have gained experience using this stent leak classification system, certain problems in specific populations and stent leak types were identified, and we have made several changes in stent placement and management that have led to improved care of this challenging population. For instance, after recognizing the relatively high incidence of stent leaks secondary to migration in cervical stents, we now we routinely use pexy or bridle in cervical stent leaks and endoluminal suturing in other stents; this has eliminated stent leaks secondary to stent migration over the last several years. Given the problems of distal retrograde type 2 stent leaks seen in stents placed across esophagojejunostomy or esophagogastrostomy anastomoses we now routinely place a PEG. Furthermore, after experience in poor outcomes in patients who do not seal after 24 to 48 hours, we now recommend prompt intervention with either another attempt at stent placement with adjunctive procedures to enhance the likelihood of seal or surgical diversion.
Limitations While randomized trials in larger cohorts of patients would provide the most robust data on distinguishing the optimal management of different leak types in distinct patient populations, such studies have not yet been performed and may not be ethical given certain treatment
Ann Thorac Surg 2014;-:-–-
strategies are more appropriate for certain leak types and clinical scenarios. This study, however, provides information on different management strategies utilized by experienced thoracic surgeons in distinct clinical situations with excellent results. The strategies utilized and classification system reported herein can form a basis for further refinement of the optimal management of these patients. Furthermore, as additional data are collected validation or enhancement of this classification system can be made; such refinement may include factors such as location and quantification of leak, as well as pertinent clinical parameters that may be revealed to be important in management and prognosis.
Conclusions Esophageal leaks tend to be managed differently depending on leak type. The majority of leaks ultimately resolve with stenting, and do so within 24 to 48 hours; if the leak is not sealed within 24 to 48 hours surgery is warranted. Our proposed leak classification may aid in the refinement of optimal esophageal stent management strategy for distinct patient populations and clinical scenarios. Dr Blackmon is a speaker for Covidien, Ethicon, and Maquet but all proceeds are donated to the Houston Methodist Hospital Fund. Special thanks go to Elaine Jordan for her administrative assistance.
References 1. D’Cunha J, Rueth NM, Groth SS, Maddaus MA, Andrade RS. Esophageal stents for anastomotic leaks and perforations. J Thorac Cardiovasc Surg 2011;142:39–46.e31. 2. Eroglu A, Turkyilmaz A, Aydin Y, Yekeler E, Karaoglanoglu N. Current management of esophageal perforation: 20 years experience. Dis Esophagus 2009;22: 374–80. 3. 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;61:765–8. 4. Pennathur A, Chang AC, McGrath KM, et al. Polyflex expandable stents in the treatment of esophageal disease: Initial experience. Ann Thorac Surg 2008;85:1968–73. 5. Kauer WK, Stein HJ, Dittler HJ, Siewert JR. Stent implantation as a treatment option in patients with thoracic anastomotic leaks after esophagectomy. Surg Endosc 2008;22:50–3.
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
7
6. Freeman RK, Ascioti AJ, Wozniak TC. Postoperative esophageal leak management with the Polyflex esophageal stent. J Thorac Cardiovasc Surg 2007;133:333–8. 7. Eubanks S, Edwards CA, Fearing NM, et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg 2008;206:935–9. 8. 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;61:891–6. 9. Raijman I, Siddique I, Ajani J, Lynch P. Palliation of malignant dysphagia and fistulae with coated expandable metal stents: experience with 101 patients. Gastrointest Endosc 1998;48:172–9. 10. Freeman RK, Van Woerkom JM, Ascioti AJ. Esophageal stent placement for the treatment of iatrogenic intrathoracic esophageal perforation. Ann Thorac Surg 2007;83:2003–8. 11. Zhou JH, Jiang YG, Wang RW, et al. Management of corrosive esophageal burns in 149 cases. J Thorac Cardiovasc Surg 2005;130:449–55. 12. Davies N, Thomas 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;80:394–7. 13. Wright CD, Mathisen DJ, Wain JC, Moncure AC, Hilgenberg AD, Grillo HC. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60: 245–9. 14. Salminen P, Gullichsen R, Laine S. Use of self-expandable metal stents for the treatment of esophageal perforations and anastomotic leaks. Surg Endosc 2009;23:1526–30. 15. Schoppmann SF, Langer FB, Prager G, Zacherl J. Outcome and complications of long-term self-expanding esophageal stenting. Dis Esophagus 2013;26:154–8. 16. Freeman RK, Ascioti AJ, Giannini T, Mahidhara RJ. Analysis of unsuccessful esophageal stent placements for esophageal perforation, fistula, or anastomotic leak. Ann Thorac Surg 2012;94:959–65. 17. Eisendrath P, Cremer M, Himpens J, Cadi ere GB, Le Moine O, Deviere J. Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery. Endoscopy 2007;39:625–30. 18. Baum RA, Stavropoulos SW, Fairman RM, Carpenter JP. Endoleaks after endovascular repair of abdominal aortic aneurysms. J vasc intervent radiol 2003;14(9 Pt 1):1111–7. 19. Dai YY, Gretschel S, Dudeck O, Rau B, Schlag PM, H€ unerbein M. Treatment of oesophageal anastomotic leaks by temporary stenting with self-expanding plastic stents. Br J Surg 2009;96:887–91. 20. 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;79:398–404. 21. Schweigert M, Dubecz A, Stadlhuber RJ, Muschweck H, Stein HJ. Risk of stent-related aortic erosion after endoscopic stent insertion for intrathoracic anastomotic leaks after esophagectomy. Ann Thorac Surg 2011;92:513–8.
DISCUSSION DR M. BLAIR MARSHALL (Washington, DC): That was an excellent presentation, and I commend you on trying [to] categorize how these leaks are managed in the modern era. I found with cervical leaks, the patients don’t tolerate stents in that area. They frequently migrate proximally. What has been your experience? My second question is, do you place a second stent versus a G tube? The patter may require an operation in many of these patients who have had previous surgery. I find for persistent leaks, that it is easy to put a stent within a stent, stenting all the way to the pylorus. This can solve the problem endoscopically. I
was wondering if you could highlight which time you use a G tube and which time you would consider multiple stents? DR BLACKMON: Blair, that’s a great question. We typically have not had great success stenting all the way through the pylorus in some of the patients that have had a pyloromyotomy. We had one patient where we attempted that and opened the pylorus causing a new leak. So we did have a complication from stenting over the pylorus. Because of that, we found in patients who have an intact stomach, we just put a G tube in. We do a laparoscopic G tube or a PEG [percutaneous endoscopic gastrostomy] before.
8
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
There’s a lot of concern of dragging a PEG through a tear in the esophagus, and so we have converted almost all of these to just laparoscopic G tube now. If they have had an esophagectomy, you can’t put a G tube in. And typically in those people you have a good distal landing zone. If you don’t have a good distal landing zone, a retrograde drainage tube is another option. DR RICHARD K. FREEMAN (Indianapolis, IN): Great paper. I would congratulate you on devising a classification for stent failures such as other specialties have done, such as vascular surgery for stent graft leaks . We’ll talk a little bit tomorrow about stent failures in a population that we are going to report, and I also applaud you for saying, follow these patients closely. That’s why we think a thoracic surgeon should be involved in placing the stent and following these patients allowing failures to be recognized early. You are going to have failures with any form of treatment. Recognize it early and fix it and prevent a disaster. DR STEPHENS: Thank you for your comments and for pointing out the need for thoracic surgeons to be intimately involved in the evaluation and treatment of these patients. On that note, I should highlight that a number of patients in this series were interinstitutional transfers, in which their stent was placed by other specialists at Methodist, later a leak was discovered and only at that point was the thoracic surgery team consulted. Secondly, we do a lot of hybrid procedures. These are 2 reasons that thoracic surgeons need to remain involved in the management and treatment of these patients. DR BRIAN E. LOUIE (Seattle, WA): A nice presentation and I think a great attempt at trying to classify them. I have 2 questions. In your classification, did you try to account for how big the leak is, because some of those leaks we have seen are free flowing out in the mediastinum and some of them are a little trickle, and I would manage those endoscopically quite differently. Number 2, we have also gone to a hybrid endoscopic strategy where we have been using the over-the-scope (Ovesco) clip, where we have closed those and then stented across them. Sometimes we just use the Ovesco clip. And the third point I guess is, at 24 to 48 hours when you see a leak, what are your driving factors to go back to the OR [operating room]? Is it that you just simply have a leak or do they have mediastinitis that drives you back to the OR? Because even if they’re leaking at 48 hours and it’s a little bit and they are otherwise stable with a normal white count, I don’t take them back to the OR. I will keep them NPO [nothing by outh] and use the J tube to feed them, and almost all those people resolve their leaks without anything else that you need to do, even if it’s an additional stent or anything else. DR STEPHENS: Thank you for your insightful comments. One of the main points of this paper is not so much our results per se but developing a system that can be used moving forward as we continue to determine the best management strategies for leaks in certain clinical scenarios. We certainly have learned from our own experience with stents to change our management strategy in certain situations; for instance, we now always pexy/bridle, which is particularly important in the cervical stents and has decreased our leak rate in that setting to virtually zero, and we use a PEG whenever we are going across the EG [esophagogastric ] junction. We have not looked more closely at some
Ann Thorac Surg 2014;-:-–-
sort of quantification of leaks, thank you for pointing that out, and that clearly would be useful in terms of guiding management. And in terms of what clinical situations drive us back to the OR, that is a patient by patient decision dependent on many factors. There are certainly people that we watch and they resolved and everything was fine. DR MARSHALL: Just 1 comment. I was asked how many patients we had put a stent within a stent past the pylorus, and I have done it in 2 and saw it done with 3 stents, I’ve never used 3, in 1 of my previous partner’s patients. But I have also seen and I wonder your experience, that with stent migration you can actually have the stent make the leak worse, and I wonder if you had seen that at all? DR BLACKMON: Yes, we have made several leaks worse. We had a sleeve patient that we stented and turned the situation into a gastrectomy when we unzipped the entire sleeve and ultimately reconstructed that patient and she is doing well today, but she certainly got sicker when her entire sleeve was disrupted. We also see a lot of patients who have had an indwelling stent for a long period of time managed at an outside institution and then they get sent to us, and they have had a stent in their conduit with an anastomotic leak. The stent has been hanging out inside the esophagus like a foreign body doing absolutely nothing with the leak ongoing, and we’re really behind the eight ball on those. And then with regards to the migration, we started endoluminally suturing those, and we have no migration anymore and they don’t have the pain from the bridle. That’s off label also, but that has really helped us a lot, and it has also helped with our bigger leaks. DR MARSHALL: And what is your algorithm for when you take the stent out? DR STEPHENS: Based on The Annals of Thoracic Surgery 2011 paper [21] showing that aortic esophageal fistulas tend to occur between days 26 and 36, we now only leave stents in for 21 days, at most. If the patient still needs a stent at that point, we perform a stent exchange. As far as surveillance for leaks, for all patients who have had stents placed we perform a Gastrogafin upper GI [gastrointestinal] swallow within 24 hours after stent placement to assess for leak. DR ERIC L. GROGAN (Nashville, TN): There is another technique for PEG and PEJ [percutaneous endoscopic jejunostomy] placement that many of you probably know about, which is the push technique. Here you can place 4 T fasteners to tack the stomach to the abdominal wall before you do the push technique. So I’ll put the scope in, identify the leak, put the PEG J tube in endoscopically, and as I’m coming out, stent the patient. With the leaks across the EG junction this technique is nice because you can drain the stomach, feed the jejunum, and then obviously study the patient to make sure you don’t have an ongoing leak. The only thing that I would add is in the Boerhaave patients, those are ones that I’m pretty quick to do a thoracoscopy and open up the mediastinum to make sure you have really adequate drainage so that you are not having ongoing sepsis.
Background. Esophageal stent leaks can have catastrophic consequences if not promptly recognized and managed appropriately. However, there are different mechanisms for esophageal stent leaks that may demonstrate unique features in presentation and response to management strategy. The objective of this study was to develop a classification system for esophageal leaks and assess distinctions between leak types. Methods. Patients with esophageal stent leaks from 2007 to 2010 managed at The Methodist Hospital were classified into the following 5 leak types: type 1, proximal; 2, distal retrograde; 3, stent lining; 4, between stents; and 5, migrated stent. Patients’ baseline characteristics, procedural data, and outcomes were analyzed. Results. Of the 89 patients who underwent esophageal stenting, 23 stent leaks were identified after the first procedure. Mean age was 57 ± 14 years, 61% were male, 43% had esophageal cancer, and 52% were status
postesophagectomy. Seven of the leaks were type 1, 6 were type 2, 2 were type 3, 4 were type 4, and were type 5. The vast majority (70%) of leaks were detected within the first 48 hours. The management of leaks varied significantly depending on the leak type (p < 0.001) and included additional stenting, placement of a larger stent, bridle, percutaneous gastrostomy, stent exchange, observation, and surgery. The majority of leaks (65%) ultimately resolved. Survival according to leak type was not different (p [ 0.072). Conclusions. Esophageal leaks tend to be managed differently depending on leak type. The majority of leaks ultimately resolve with stenting. Our proposed leak classification may enhance esophageal stent management strategy.
E
varies, but can be seen in approximately 10% to 40% of patients [14–17]. Esophageal stent leaks can have catastrophic consequences if not recognized and managed appropriately; furthermore, they may demonstrate unique clinical presentations and response to treatment strategy depending on the patients’ underlying pathophysiology and mechanism of leak. For example, esophageal stent leak secondary to stent migration can be managed or prevented by bridling, pexy, anti-migration stent design, or oversizing. However, no classification exists for defining the type of stent leak. Furthermore, the optimal treatment strategy for certain types of esophageal stent leaks, such as those secondary to stent migration, in specific patient populations, remains to be determined. The objective of this study was to develop a classification system for esophageal leaks, similar to that used in classification of endovascular stent leaks [18], and assess distinctions between leak types in terms of presentation and management strategies to aid in the prompt recognition and optimal treatment of stent leaks in specific patient populations.
sophageal stents are used for a large variety of patients and clinical scenarios including esophageal perforation [1, 2], tracheoesophageal fistulas [3, 4], anastomotic leaks encountered after esophagectomy or gastric bypass [5–8], iatrogenic injuries [9, 10], corrosive burn injuries [11], obstructive lesions such as tumors [12], and to reinforce traditional repairs [13]. While the traditional treatment for esophageal and gastric leaks includes surgical diversion or surgical repair, esophageal stenting provides a less invasive treatment strategy that can be beneficial in patients who may not tolerate surgery [1, 2]. Stenting can also allow the salvage of an esophageal leak that otherwise would not heal because of delayed presentation. While esophageal stenting has become an increasingly utilized technique in esophageal disease, the optimal management of these patients remains to be determined. Esophageal stent leaks are a dreaded complication of stent placement. The reported incidence of stent leak
(Ann Thorac Surg 2014;-:-–-) Ó 2014 by The Society of Thoracic Surgeons
Accepted for publication Jan 28, 2014. Presented at the Sixtieth Annual Meeting of the Southern Thoracic Surgical Association, Scottsdale, AZ, Oct 30–Nov 2, 2013. Address correspondence to Dr Blackmon, Division of Thoracic Surgery, Houston Methodist Hospital, 6550 Fannin St, Smith Tower, Ste 1661, Houston, TX 77030; e-mail: [email protected].
Ó 2014 by The Society of Thoracic Surgeons Published by Elsevier Inc
Dr Blackmon discloses financial relationships with Covidien, Ethicon, and Maquet; and Dr Kim with Covidien and Ethicon.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.01.063
2
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
Patients and Methods Patients A prospectively maintained database of esophageal stents placed from 2007 to January 2013 at The Methodist Hospital was searched for esophageal stent leaks, which were then classified into the following 5 leak types on the basis of radiographic and intraoperative assessment (Fig 1): type 1 ¼ proximal; type 2 ¼ distal retrograde; type 3 ¼ stent lining; type 4 ¼ between stents; and type 5 ¼ migrated stent. Patients’ baseline characteristics, procedural data, and outcomes were analyzed. Time to leak detection was defined as the time from clinical signs and symptoms suggestive of leak to diagnosis of leak. A waiver of informed consent was obtained along with Institutional Review Board approval. An investigational device exemption from the US Food and Drug Administration was obtained to investigate the use of off-label esophageal stenting, and some of the patients in this database elected to participate in a prospective study to evaluate the safety and efficacy of esophageal stenting in the setting of leaks or fistulae. All stents were covered stents and included Wallflex (Boston Scientific, Natick, MA), Allimax (Alveolus, Charlotte, NC), and Ultraflex (Boston Scientific).
Statistical Analysis Statistical comparisons between groups were made using c2 testing and log-rank analysis of Kaplan-Meier curves were used to compare survival. Statistical significance was defined as a p value less than 0.05. All statistical analyses were performed using SPSS version 19.0 (SPSS, Chicago, IL).
Results Patient Characteristics There were 108 patients who presented to Houston Methodist Hospital for evaluation of an esophageal leak
Ann Thorac Surg 2014;-:-–-
from 2007 to January 2013; of these, 89 (82%) patients had a stent placed to initially treat their foregut leak or fistula. Of these 89 patients, 23 (26%) patients presented with a leak after their initial stent was placed and required further intervention to seal the leak (Table 1). Fourteen patients (61%) were males and the mean age was 57 14 years old. A large portion (43%) of patients had esophageal cancer and 52% were status postesophagectomy. Seven of the leaks were type 1, 6 were type 2, 2 were type 3, 4 were type 4, and 5 were type 5. The vast majority (70%) of leaks were detected within the first 48 hours and confirmed radiographically (see examples of each type of leak in Figure 1). The management of leaks varied significantly dependent on leak type (p < 0.001) and included additional stenting, placement of a larger stent, bridling the stent to the nasal septum, gastrostomy feeding tube placement, stent exchange, observation, and surgery. The majority of these leaks (65%) ultimately resolved and the majority of those that did not resolve with stent placement were patients with a significant delay between diagnosis and appropriate intervention. There was a wide range in the delay in diagnosis, with some patients transferring into our institution leaking for 1 to 2 weeks before diagnosis or intervention. However, once the diagnosis was made intervention was made usually that day. The delay between diagnosis and seal or diversion also widely varied from 1 to 2 days to 1 to 3 months. Delay in diagnosis and delay in obtaining seal or diverting was not found to be different between leak types or associated with adverse overall outcome. Similarly, presentation at outside hospital (7 of the patients) was not associated with decreased survival or likelihood of seal. Overall survival was not significantly different between leak types (p ¼ 0.072). Type 1 (proximal) leaks were predominantly seen after esophagectomy, most commonly performed for esophageal cancer. Out of the 7 patients with type 1 leaks, 4 were successfully managed with additional stent placement,
Fig 1. Fluoroscopic imaging illustrating the different types of esophageal stent leaks (type 1 to type 5).
Ann Thorac Surg 2014;-:-–-
3
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
Table 1. Pathology, Management and Clinical Presentation of Stent Leaks Classified by Leak Type (1 Through 5) Total Cohort (n¼23) 1 (n ¼ 7, 30%) 2 (n ¼ 6, 26%) 3 (n ¼ 2, 9%) 4 (n ¼ 4, 17%) 5 (n ¼ 4, 17%) p Value
Variable Average age Gender (% male) Previous surgery/interventions: Gastric bypass or sleeve Gastrectomy Esophagectomy Eso dilation Eso hernia repair Cardiac Underlying pathology: Eso cancer Stricture Iatrogenic Prolonged vent Boerhaaves Management of leak: Additional stent only Larger stent Bridle PEG Stent exchange Observation Surgery Timing of leak detection: 14 days Resolution of leak: Mean follow-up (months) % survival last follow-up
57 14 61%
56 9 71%
59 11 66%
66 5 50%
50 16 50%
59 27 50%
5 1 12 1 1 1
(22%) (4%) (52%) (4%) (4%) (4%)
1 (14%) 1 (14%) 3 (43%) 0 1 (14%) 0
1 (17%) 0 3 (50%) 1 (17%) 0 0
0 0 2 (100%) 0 0 0
2 (50%) 0 1 (25%) 0 0 1 (25%)
1 (25%) 0 2 (50%) 0 0 0
10 2 2 1 1
(43$) (9%) (9%) (4%) (4%)
4 (57%) 0 2 (29%) 0 0
2 (33%) 1 (17%) 0 0 1 (17%)
1 (50%) 1 (50%) 0 0 0
1 (25%) 0 0 0 0
2 (50%) 0 0 1 (25%) 0
4 3 3 3 2 2 6
(17%) (13%) (13%) (13%) (9%) (9%) (26%)
4 (57%) 2 (29%) 0 0 0 1 (14%) 0
0 0 0 3 (50%) 0 0 3 (50%)
0 0 0 0 2 (100%) 0 0
0 0 0 0 0 1 (25%) 3 (75%)
0 1 (25%) 3 (75%) 0 0 0 0
7 (100%) 0 0 5 (71%) 10 13 86%
2 (33%) 3 (50%) 1 (17%) 4 (66%) 11 10 100%
1 (50%) 1 (50%) 0 1 (50%) 20 100%
2 (50%) 0 2 (50%) 2 (50%) 13 10 75%
4 (100%) 0 0 3 (75%) 99 100%
NS NS NS
NS
6 cm transthoracic repair Stent PEG for gastric decompression
Type 2 leak, stent migration, incomplete healing of leak
G-tube ¼ gastrostomy tube;
j-tube ¼ jejunostomy tube;
repair. For intraabdominal esophageal or gastric perforations they recommended transabdominal repair and jejunostomy. While specific recommendations vary between authors, there is a consensus that the use of stents must be selective and individualized, taking into account the individual patient’s underlying etiology and anatomy, as well as clinical status and comorbidities. Continued reevaluation of the patient is also required to determine whether the leak has sealed and when additional interventions are necessary. At our institution all patients undergo esophagram within 24 hours to confirm that the stent leak has sealed. Of note, in this 2012 study Freeman and colleagues also developed a classification system for esophageal leaks prior to stent placement based on the location, mechanism, and size of the leak, as well as if the leak was associated with malignancy or distal stricture. As our classification system is validated with more patients, refinement in the classification system may include parameters such as location and quantification of leak. For instance, if the leak was large but well contained with chest tube drainage, this would be managed differently than a freely extravasating leak without chest tube drainage or a contained leak; these aspects could be incorporated into the classification system to further aid management decisions. Similarly, a type 5 leak in the neck is distinct in terms of leak mechanism and management compared with a type 5 leak in the midthoracic region. Equally important in the determination of when and which patients may benefit from stenting is recognition of which patients would not benefit from stenting. Patients with conduit ischemia and large, uncontrolled leak are patients who will not benefit from stenting. Repeat stenting is used if the patient is stable, is tolerating diet, or is a particularly high surgical risk. If the leak is not sealed within 24 to 48 hours they are not likely to seal; depending on the clinical scenario and patient status either another stent is placed, with an adjunctive muscle flap if appropriate, or surgery is warranted. Of note, stents are not left in for over a month in order to avoid ingrowth that can cause problems on removal and erosion into adjacent vascular structures, but rather a
Type 2 leak, aspiration if stenting across esophagojejunostomy anastomosis Stent migration
Stent migration
Type 2 leak, aspiration (open GEJ) PEG ¼ percutaneous endoscopic gastrostomy.
stent exchange is performed if the patient continues to need a stent to seal the leak. Given a recent report demonstrating that most esophageal-aortic fistulas occur at 26 to 36 days [21], our current recommendation is to not leave stents in for longer than 3 weeks. Overall, esophageal stents are an excellent tool in the appropriate clinical scenario with certain pathophysiology and anatomy. Not all patients with esophageal leaks require operative management and in many cases esophageal stenting is an effective management strategy with lower risk of morbidity and mortality than open operative intervention. However, in patients who do not seal after 2 attempts or become unstable, prompt diversion should be considered as further delay with a persistent esophageal leak carries a poor prognosis. As we have gained experience using this stent leak classification system, certain problems in specific populations and stent leak types were identified, and we have made several changes in stent placement and management that have led to improved care of this challenging population. For instance, after recognizing the relatively high incidence of stent leaks secondary to migration in cervical stents, we now we routinely use pexy or bridle in cervical stent leaks and endoluminal suturing in other stents; this has eliminated stent leaks secondary to stent migration over the last several years. Given the problems of distal retrograde type 2 stent leaks seen in stents placed across esophagojejunostomy or esophagogastrostomy anastomoses we now routinely place a PEG. Furthermore, after experience in poor outcomes in patients who do not seal after 24 to 48 hours, we now recommend prompt intervention with either another attempt at stent placement with adjunctive procedures to enhance the likelihood of seal or surgical diversion.
Limitations While randomized trials in larger cohorts of patients would provide the most robust data on distinguishing the optimal management of different leak types in distinct patient populations, such studies have not yet been performed and may not be ethical given certain treatment
Ann Thorac Surg 2014;-:-–-
strategies are more appropriate for certain leak types and clinical scenarios. This study, however, provides information on different management strategies utilized by experienced thoracic surgeons in distinct clinical situations with excellent results. The strategies utilized and classification system reported herein can form a basis for further refinement of the optimal management of these patients. Furthermore, as additional data are collected validation or enhancement of this classification system can be made; such refinement may include factors such as location and quantification of leak, as well as pertinent clinical parameters that may be revealed to be important in management and prognosis.
Conclusions Esophageal leaks tend to be managed differently depending on leak type. The majority of leaks ultimately resolve with stenting, and do so within 24 to 48 hours; if the leak is not sealed within 24 to 48 hours surgery is warranted. Our proposed leak classification may aid in the refinement of optimal esophageal stent management strategy for distinct patient populations and clinical scenarios. Dr Blackmon is a speaker for Covidien, Ethicon, and Maquet but all proceeds are donated to the Houston Methodist Hospital Fund. Special thanks go to Elaine Jordan for her administrative assistance.
References 1. D’Cunha J, Rueth NM, Groth SS, Maddaus MA, Andrade RS. Esophageal stents for anastomotic leaks and perforations. J Thorac Cardiovasc Surg 2011;142:39–46.e31. 2. Eroglu A, Turkyilmaz A, Aydin Y, Yekeler E, Karaoglanoglu N. Current management of esophageal perforation: 20 years experience. Dis Esophagus 2009;22: 374–80. 3. 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;61:765–8. 4. Pennathur A, Chang AC, McGrath KM, et al. Polyflex expandable stents in the treatment of esophageal disease: Initial experience. Ann Thorac Surg 2008;85:1968–73. 5. Kauer WK, Stein HJ, Dittler HJ, Siewert JR. Stent implantation as a treatment option in patients with thoracic anastomotic leaks after esophagectomy. Surg Endosc 2008;22:50–3.
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
7
6. Freeman RK, Ascioti AJ, Wozniak TC. Postoperative esophageal leak management with the Polyflex esophageal stent. J Thorac Cardiovasc Surg 2007;133:333–8. 7. Eubanks S, Edwards CA, Fearing NM, et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg 2008;206:935–9. 8. 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;61:891–6. 9. Raijman I, Siddique I, Ajani J, Lynch P. Palliation of malignant dysphagia and fistulae with coated expandable metal stents: experience with 101 patients. Gastrointest Endosc 1998;48:172–9. 10. Freeman RK, Van Woerkom JM, Ascioti AJ. Esophageal stent placement for the treatment of iatrogenic intrathoracic esophageal perforation. Ann Thorac Surg 2007;83:2003–8. 11. Zhou JH, Jiang YG, Wang RW, et al. Management of corrosive esophageal burns in 149 cases. J Thorac Cardiovasc Surg 2005;130:449–55. 12. Davies N, Thomas 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;80:394–7. 13. Wright CD, Mathisen DJ, Wain JC, Moncure AC, Hilgenberg AD, Grillo HC. Reinforced primary repair of thoracic esophageal perforation. Ann Thorac Surg 1995;60: 245–9. 14. Salminen P, Gullichsen R, Laine S. Use of self-expandable metal stents for the treatment of esophageal perforations and anastomotic leaks. Surg Endosc 2009;23:1526–30. 15. Schoppmann SF, Langer FB, Prager G, Zacherl J. Outcome and complications of long-term self-expanding esophageal stenting. Dis Esophagus 2013;26:154–8. 16. Freeman RK, Ascioti AJ, Giannini T, Mahidhara RJ. Analysis of unsuccessful esophageal stent placements for esophageal perforation, fistula, or anastomotic leak. Ann Thorac Surg 2012;94:959–65. 17. Eisendrath P, Cremer M, Himpens J, Cadi ere GB, Le Moine O, Deviere J. Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery. Endoscopy 2007;39:625–30. 18. Baum RA, Stavropoulos SW, Fairman RM, Carpenter JP. Endoleaks after endovascular repair of abdominal aortic aneurysms. J vasc intervent radiol 2003;14(9 Pt 1):1111–7. 19. Dai YY, Gretschel S, Dudeck O, Rau B, Schlag PM, H€ unerbein M. Treatment of oesophageal anastomotic leaks by temporary stenting with self-expanding plastic stents. Br J Surg 2009;96:887–91. 20. 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;79:398–404. 21. Schweigert M, Dubecz A, Stadlhuber RJ, Muschweck H, Stein HJ. Risk of stent-related aortic erosion after endoscopic stent insertion for intrathoracic anastomotic leaks after esophagectomy. Ann Thorac Surg 2011;92:513–8.
DISCUSSION DR M. BLAIR MARSHALL (Washington, DC): That was an excellent presentation, and I commend you on trying [to] categorize how these leaks are managed in the modern era. I found with cervical leaks, the patients don’t tolerate stents in that area. They frequently migrate proximally. What has been your experience? My second question is, do you place a second stent versus a G tube? The patter may require an operation in many of these patients who have had previous surgery. I find for persistent leaks, that it is easy to put a stent within a stent, stenting all the way to the pylorus. This can solve the problem endoscopically. I
was wondering if you could highlight which time you use a G tube and which time you would consider multiple stents? DR BLACKMON: Blair, that’s a great question. We typically have not had great success stenting all the way through the pylorus in some of the patients that have had a pyloromyotomy. We had one patient where we attempted that and opened the pylorus causing a new leak. So we did have a complication from stenting over the pylorus. Because of that, we found in patients who have an intact stomach, we just put a G tube in. We do a laparoscopic G tube or a PEG [percutaneous endoscopic gastrostomy] before.
8
STEPHENS ET AL ESOPHAGEAL LEAK CLASSIFICATION SYSTEM
There’s a lot of concern of dragging a PEG through a tear in the esophagus, and so we have converted almost all of these to just laparoscopic G tube now. If they have had an esophagectomy, you can’t put a G tube in. And typically in those people you have a good distal landing zone. If you don’t have a good distal landing zone, a retrograde drainage tube is another option. DR RICHARD K. FREEMAN (Indianapolis, IN): Great paper. I would congratulate you on devising a classification for stent failures such as other specialties have done, such as vascular surgery for stent graft leaks . We’ll talk a little bit tomorrow about stent failures in a population that we are going to report, and I also applaud you for saying, follow these patients closely. That’s why we think a thoracic surgeon should be involved in placing the stent and following these patients allowing failures to be recognized early. You are going to have failures with any form of treatment. Recognize it early and fix it and prevent a disaster. DR STEPHENS: Thank you for your comments and for pointing out the need for thoracic surgeons to be intimately involved in the evaluation and treatment of these patients. On that note, I should highlight that a number of patients in this series were interinstitutional transfers, in which their stent was placed by other specialists at Methodist, later a leak was discovered and only at that point was the thoracic surgery team consulted. Secondly, we do a lot of hybrid procedures. These are 2 reasons that thoracic surgeons need to remain involved in the management and treatment of these patients. DR BRIAN E. LOUIE (Seattle, WA): A nice presentation and I think a great attempt at trying to classify them. I have 2 questions. In your classification, did you try to account for how big the leak is, because some of those leaks we have seen are free flowing out in the mediastinum and some of them are a little trickle, and I would manage those endoscopically quite differently. Number 2, we have also gone to a hybrid endoscopic strategy where we have been using the over-the-scope (Ovesco) clip, where we have closed those and then stented across them. Sometimes we just use the Ovesco clip. And the third point I guess is, at 24 to 48 hours when you see a leak, what are your driving factors to go back to the OR [operating room]? Is it that you just simply have a leak or do they have mediastinitis that drives you back to the OR? Because even if they’re leaking at 48 hours and it’s a little bit and they are otherwise stable with a normal white count, I don’t take them back to the OR. I will keep them NPO [nothing by outh] and use the J tube to feed them, and almost all those people resolve their leaks without anything else that you need to do, even if it’s an additional stent or anything else. DR STEPHENS: Thank you for your insightful comments. One of the main points of this paper is not so much our results per se but developing a system that can be used moving forward as we continue to determine the best management strategies for leaks in certain clinical scenarios. We certainly have learned from our own experience with stents to change our management strategy in certain situations; for instance, we now always pexy/bridle, which is particularly important in the cervical stents and has decreased our leak rate in that setting to virtually zero, and we use a PEG whenever we are going across the EG [esophagogastric ] junction. We have not looked more closely at some
Ann Thorac Surg 2014;-:-–-
sort of quantification of leaks, thank you for pointing that out, and that clearly would be useful in terms of guiding management. And in terms of what clinical situations drive us back to the OR, that is a patient by patient decision dependent on many factors. There are certainly people that we watch and they resolved and everything was fine. DR MARSHALL: Just 1 comment. I was asked how many patients we had put a stent within a stent past the pylorus, and I have done it in 2 and saw it done with 3 stents, I’ve never used 3, in 1 of my previous partner’s patients. But I have also seen and I wonder your experience, that with stent migration you can actually have the stent make the leak worse, and I wonder if you had seen that at all? DR BLACKMON: Yes, we have made several leaks worse. We had a sleeve patient that we stented and turned the situation into a gastrectomy when we unzipped the entire sleeve and ultimately reconstructed that patient and she is doing well today, but she certainly got sicker when her entire sleeve was disrupted. We also see a lot of patients who have had an indwelling stent for a long period of time managed at an outside institution and then they get sent to us, and they have had a stent in their conduit with an anastomotic leak. The stent has been hanging out inside the esophagus like a foreign body doing absolutely nothing with the leak ongoing, and we’re really behind the eight ball on those. And then with regards to the migration, we started endoluminally suturing those, and we have no migration anymore and they don’t have the pain from the bridle. That’s off label also, but that has really helped us a lot, and it has also helped with our bigger leaks. DR MARSHALL: And what is your algorithm for when you take the stent out? DR STEPHENS: Based on The Annals of Thoracic Surgery 2011 paper [21] showing that aortic esophageal fistulas tend to occur between days 26 and 36, we now only leave stents in for 21 days, at most. If the patient still needs a stent at that point, we perform a stent exchange. As far as surveillance for leaks, for all patients who have had stents placed we perform a Gastrogafin upper GI [gastrointestinal] swallow within 24 hours after stent placement to assess for leak. DR ERIC L. GROGAN (Nashville, TN): There is another technique for PEG and PEJ [percutaneous endoscopic jejunostomy] placement that many of you probably know about, which is the push technique. Here you can place 4 T fasteners to tack the stomach to the abdominal wall before you do the push technique. So I’ll put the scope in, identify the leak, put the PEG J tube in endoscopically, and as I’m coming out, stent the patient. With the leaks across the EG junction this technique is nice because you can drain the stomach, feed the jejunum, and then obviously study the patient to make sure you don’t have an ongoing leak. The only thing that I would add is in the Boerhaave patients, those are ones that I’m pretty quick to do a thoracoscopy and open up the mediastinum to make sure you have really adequate drainage so that you are not having ongoing sepsis.
