Risk of Stent-Related Aortic Erosion After Endoscopic Stent Insertion for Intrathoracic Anastomotic Leaks After Esophagectomy

Michael Schweigert, MD, Attila Dubecz, MD, Rudolf J. Stadlhuber, MD, Herbert Muschweck, MD, and Hubert J. Stein, MD Departments of General and Thoracic Surgery, and Gastroenterology, Klinikum Nürnberg Nord, Nürnberg, Germany

Background. Intrathoracic anastomotic leakage after esophagectomy is associated with high morbidity and mortality. Because of disappointing results after surgical reexploration endoscopic stent implantation was introduced as primary treatment option with improved outcome. Aortoesophageal fistula is a very rare complication and has thus far only anecdotally been reported after esophagectomy. The aim of this retrospective study was to investigate if endoscopic stent implantation increases the incidence of postoperative aortoesophageal fistula by reason of stent-related erosion of the thoracic aorta. Methods. Between January 2004 and October 2010, 213 patients underwent esophageal resection mainly for esophageal cancer. An intrathoracic esophageal anastomotic leak was endoscopically verified in 25 patients. Seventeen patients received endoscopic implantation of a self-expanding stent as primary treatment. In 8 patients a rethoracotomy was mandatory. Results. After successfully accomplished endoscopic stent placement, complete closure of the anastomotic leak

was radiologically proven in all 17 patients. In 13 cases, definitive closure and healing of the leak was achieved and the stent could subsequently be removed. In 1 patient, because of early recurrence of very malignant small cell cancer, the stent remained in situ. Three patients developed an erosion of the thoracic aorta with subsequent massive hemorrhage. The mean time between stent insertion and occurrence of aortoesophageal fistula was 26 days. All 3 patients died of exsanguination with severe hypovolemic shock. Postmortem examination confirmed an aortoesophageal fistula in each case. Conclusions. While endoscopic stent implantation seems to be effective in the control of intrathoracic anastomotic leakage, nevertheless the incidence of aortoesophageal fistula caused by stent-related aortic erosion exceeds the thus far reported numbers. Awareness of this life-threatening complication after stent insertion is therefore mandatory. (Ann Thorac Surg 2011;92:513–9) © 2011 by The Society of Thoracic Surgeons

A

tervention. Upon left rethoracotomy, closure of a 2-mm fistula between the anastomosis and the aorta was accomplished. The patient, who also received cervical esophagostomy and gastrostomy, survived and reconstruction was done a few months later [4]. Ever since 1967 only anecdotally reports of single cases have been published [1, 5]. In contrast to aortoesophageal fistula intrathoracic anastomotic leakage is a quite frequent complication after esophagectomy. Due to advances in surgical technique and perioperative management, including intensive care and anesthesia, the outcome in esophageal surgery has considerably improved during the last decades with in-hospital mortality decreasing from 29% to 7.5% between1950 and 2000 [6 – 8]. Notwithstanding this favorable trend, anastomotic leakage remains the most severe postoperative complication and intrathoracic anastomotic leakage accounts for approximately 40% of postoperative death after esophagectomy [9 –11]. Because of disappointing results after surgical reexploration, conservative treatment [8 –11] and recently endoscopic stent insertion [12–17] had been introduced as treatment options to deal with this devastating complication.

ortoesophageal fistula is a very rare life-threatening complication after esophagectomy [1]. The first 2 cases of a fistula between the intrathoracic esophagogastric anastomosis and the thoracic aorta were reported by Merendino and Emerson in Seattle (1946 and 1947) [2]. Both patients died from massive hemorrhage and autopsy confirmed the diagnosis of an aortoesophageal fistula with involvement of the anastomosis [2]. Since that time postoperative aortoesophageal fistula after esophagectomy has remained an infrequent but nevertheless fatal complication. In 1961, Le Roux (from the University of Edinburgh) reported about 8 fatal cases of aortoesophageal fistula after esophagectomy [3]. In 1967 Maillard and colleagues (from France) described 5 cases of aortoesophageal-gastric fistula they had experienced between 1957 and 1967 after left thoracoabdominal esophagogastrectomy [4]. While 4 patients died, they reported the first successful rein-

Accepted for publication Feb 24, 2011. Address correspondence to Dr Schweigert, Klinik für Allgemein-, Viszeral- und Thoraxchirurgie, Klinikum Nürnberg Nord, Prof-ErnstNathan-Str. 1, 90419 Nürnberg, Germany; e-mail: michael.schweigert@ klinikum-nuernberg.de.

© 2011 by The Society of Thoracic Surgeons Published by Elsevier Inc

0003-4975/$36.00 doi:10.1016/j.athoracsur.2011.02.083

GENERAL THORACIC

Risk of Stent-Related Aortic Erosion After Endoscopic Stent Insertion for Intrathoracic Anastomotic Leaks After Esophagectomy

GENERAL THORACIC

514

SCHWEIGERT ET AL AORTOESOPHAGEAL FISTULA AFTER ESOPHAGECTOMY

Ann Thorac Surg 2011;92:513–9

While on the one hand endoscopic stent implantation has shown promising results in several series [12–17], on the other hand there have been reports on stent-related vascular erosion in other collectives [18 –23]. These reports are mostly about patients with benign or malignant stenosis of the esophagus who received endoscopic stent insertion in palliative intention to reopen passage for oral intake [18 –23]. Thus far there has been only one brief case report [24] regarding the risk of aortoesophageal fistula in patients with stent insertion for treatment of intrathoracic anastomotic leaks after esophagectomy. Therefore, the aim of this retrospective study was to investigate if endoscopic stent implantation increases the incidence of postoperative aortoesophageal fistula by reason of stent-related erosion of the thoracic aorta.

swallow for diagnosis of anastomotic leaks because of the possibility of direct visual examination of the anastomosis, quantification of the leak, and determination of whether the pulled up gastric tube is ischemic or not. A straight decision, if an endoscopic stent insertion is feasible or not, can be made out of the endoscopic aspect and the direct visualization of the leak. In addition to the endoscopy a computed tomography of the chest and abdomen is mandatory to rule out advanced pleural empyema or mediastinal abscess, which would require either percutaneous interventional or surgical drainage.

Patients and Methods Patients The study includes 213 consecutive patients who underwent an esophageal resection between January 2004 and October 2010 at the Department of General and Thoracic Surgery of the Klinikum Nürnberg Nord. The median age was 62.5 years and there were 177 male (83.1%) and 36 female (16.9%) patients. The ethics committee approved this retrospective study and waived the need for individual consent. The main reason for surgery was esophageal cancer. There were 84 squamous cell carcinomas (39.4%), 85 adenocarcinomas of the esophagogastric junction (AEG) type I (39.9%), 28 AEG type II (13.1%), which is a cancer of the gastric cardia with involvement of the distal esophagus [25–28], 3 AEG type III (1.4%), which is a subcardial gastric carcinoma infiltrating the esophagogastric junction [25–28], 5 other malignant tumors (2,4%), and 8 benign lesions (3.8%). A total of 52 patients (24,4%) received neoadjuvant therapy. In 154 patients an abdominothoracic esophagectomy, with interposition of a pulled up gastric tube and intrathoracic stapled anastomosis, was performed above the level of the tracheal carina (72.3%) [29]. Twenty-two patients received extended total gastrectomy with transhiatal resection of the distal esophagus and intrathoracic stapled esophagojejunostomy below the level of the tracheal carina (10.3%) [29]. Furthermore, 7 transmediastinal esophagectomies (3.3%), 13 transthoracic esophagectomies with delayed reconstruction (6.1%) [30], 5 limited resections of the distal esophagus with jejunum interposition and stapled intrathoracic esophagojejunostomy (Merendino procedure) [31], and 12 other procedures were carried out.

Stent Insertion For closure of the anastomotic leak either a selfexpanding, covered metal stent (Ultraflex; Boston Scientific, Natick, MA) or a self-expanding, covered silicone stent (Polyflex, esophageal stent; Boston Scientific) were endoscopically inserted. The stent placement was performed by a gastroenterologist well trained in interventional endoscopy. The exact position of the leakage was marked on the patient’s skin, and afterward the stent was inserted under radioscopic guidance. After implantation the correct placement of the stent and the successful closure of the leak were always endoscopically and radioscopically controlled. In 17 endoscopically verified intrathoracic anastomotic leaks endoscopic implantation of either a self-expanding metal or silicone stent was successfully accomplished [Table 1]. Rethoracotomy was mandatory in 8 other cases because of either necrosis or ischemia of the pulled-up gastric tube, fistulization between the airways and the esophagus, or advanced pleural empyema that required surgical debridement.

Patient Care After Stent Insertion After stent placement, contrast swallow esophagography and endoscopy showed a complete sealing of the anastomotic leak. All patients were treated at the intensive care unit and received sepsis therapy including antibiotherapy, hemodynamic monitoring and management, hemofiltration if necessary, and adjunctive therapies. Physiotherapy and inhalation with positive airway pressure were administered as soon as possible. Tube thoracostomy was performed in all patients and some patients received irrigation of the pleural cavity with Ringer solution through the inserted chest tubes.

Results Out of the 17 patients with endoscopic stent insertion three developed an aortoesophageal fistula with subsequent severe hemorrhage.

Diagnosis of Anastomotic Leakage

First Case

If leakage of the intrathoracic anastomosis was clinically suspected immediate endoscopy and computed tomography of the chest and abdomen were performed. An intrathoracic leakage of the esophageal-intestinal anastomosis was endoscopically verified in 25 patients. Endoscopy was preferred to radiographic contrast-medium

The first patient was a 62-year-old male with adenocarcinoma of the distal esophagus (AEG type I). A primary abdominothoracic esophagectomy with intrathoracic stapled anastomosis between the pulled up gastric tube and the esophagus was performed. Initially the patient recovered well but at postoperative day 5 his condition dete-

SCHWEIGERT ET AL AORTOESOPHAGEAL FISTULA AFTER ESOPHAGECTOMY

515

Table 1. Characteristics of Patients With Stent Insertion Mean age Histology: Squamous cell carcinoma AEG type I, II, III Neuroendocrine cancer Procedures: Neoadjuvant therapy Subtotal en-bloc esophagectomy (interposition of a pulled up gastric tube and intrathoracic stapled anastomosis above the level of the tracheal carina) Transhiatal extended gastrectomy with resection of the distal esophagus (intrathoracic stapled esophagojejunostomy below the level of the tracheal carina) Results: Stent explantation (definitive closure of the leak) Aortoesophageal fistula Mean time between stent insertion and fatal bleeding In-hospital mortality In hospital mortality of patients with surgical reexploration (rethoracotomy)

60.4 years 2 13 2 2 14

3

13 (77%) 3 (17%) 26 days (16-41 days) 4 (23%) 4/8 (50%)

AEG ⫽ adenocarcinoma of the esophagogastric junction.

riorated and endoscopy showed an anastomotic leak. A self-expanding, covered silicone stent was inserted. Radiographically the leakage ceased after stent implantation and endoscopy showed correct stent placement. The patient recuperated quickly and could be transferred from the intensive care unit to the surgical ward at the 10th day after stent insertion. Oral intake was resumed and transfer into a rehabilitation facility was planned. At the 16th day after stent insertion the patient suddenly started bleeding with hematemesis and signs of hypovolemic shock. Upon immediate endoscopy bleeding intensity increased at once and the patient rapidly died from exsanguination. Postmortem examination showed a stent-related erosion of the thoracic aorta with an aortoesophageal fistula.

Second Case The second patient was a 72-year-old male suffering from subcardial gastric carcinoma infiltrating the esophagogastric junction (AEG type III), diffuse type (Lauren classification) with signet-ring cells. He received a transhiatal extended gastrectomy with resection of the distal esophagus and intrathoracic esophagojejunostomy. At postoperative day 5, because of clinical deterioration a computed tomography was arranged which showed an intrathoracic anastomotic leakage. Placement of a selfexpanding, covered silicone stent was accomplished but repeated stent migration occurred. Because of several stent dislocations a larger self-expanding, covered silicon stent was implanted 21 days after initial stent insertion. Fourteen days after the exchange stent dislocation happened again but endoscopic reintervention was feasible and correct placement of the stent could be reestablished. The patient recovered well and was finally discharged from hospital and transferred to a rehabilitation facility at the 41st day after initial stent insertion. Upon admission at the rehabilitation clinic the patient col-

lapsed with hematemesis. He died a few hours later in a nearby emergency department from hypovolemic shock. Autopsy confirmed an erosion of the aorta by the stent with aortoesophageal fistula.

Third Case The third patient was a 68-year-old male with adenocarcinoma of the gastric cardia infiltrating the distal esophagus (AEG type II), who had an extended gastrectomy with transhiatal resection of the distal esophagus and reconstruction by means of intrathoracic esophagojejunostomy. At postoperative day 4 the chest tubes contained dirty secretion and reintubation because of respiratory failure was necessary. Immediate endoscopy showed no anastomotic leak, and thus only a nasogastric tube was placed. Afterward the patient developed a right-sided pleural empyema and received videoassisted thoracoscopy with pleural debridement and lavage. Repeated endoscopy failed again to show a leak but because of clinically suspected leakage a self-expanding, covered metal stent was endoscopically inserted at postoperative day 13. Henceforth the patient recovered well, healing of the pleural empyema was achieved, and oral food intake was successfully started. There were no signs of persisting anastomotic leakage and the chest tubes producing just clear fluid were subsequently removed. At postoperative day 35 and the 22nd day after stent implantation the patient suddenly collapsed with hematemesis and died rapidly from peracute shock. Postmortem examination revealed stent-related ulceration of the esophageal and intestinal mucosa with aortoesophageal fistula.

Cases Without Aortoesophageal Fistula In the 14 remaining patients who had received endoscopic stent implantation, complete closure of the leak was accomplished without exception. All patients finally

GENERAL THORACIC

Ann Thorac Surg 2011;92:513–9

GENERAL THORACIC

516

SCHWEIGERT ET AL AORTOESOPHAGEAL FISTULA AFTER ESOPHAGECTOMY

Ann Thorac Surg 2011;92:513–9

recuperated well. Routine stent removal was performed in 13 patients. After stent explantation the underlying mucosa showed neither necrosis nor severe morphologic changes. However, mucosal hyperproliferation was visible at the stent margins. After stent explantation a tiny fistula with a diameter of around 3 mm was detected in one case and successfully sealed with cyanoacrylate glue. Notwithstanding complete healing of the anastomotic leak 1 patient died 24 days after successful stent explantation from massive pulmonary aspiration with subsequent pneumonia and severe sepsis. In a 47-year-old patient with exceptional malignant G4 neuroendocrine and small cell cancer of the esophagus, the stent remained in situ by reason of early tumor recurrence in the mediastinum 3 months after initial abdominothoracic esophagectomy. Only 3 months later the patient died from rapid progressing cancer with local recurrence and distant metastasis.

toesophageal fistula had simultaneously with the generally improved outcome after esophageal surgery become a very rare complication that had, even in specialized esophageal surgery units with high volume, only sporadically been seen. Nearly all reported fistulas originated from an insufficient intrathoracic anastomosis [1-5]. Therefore intrathoracic anastomotic leak remains the main reason for postoperative fistulization between the esophageal anastomosis and the aorta. Since 1950 the outcome in esophageal surgery has substantially improved. In 1961 Le Roux reported an operative mortality of 30% in his series of 418 resections for esophageal carcinoma [3]. The in-hospital-mortality has thenceforth decreased from 30% to 7.5% in 2000 [6 – 8, 32]. Intrathoracic anastomotic leak, however, remains a frequent and potential fatal complication that still accounts for 40% of postoperative death after esophagectomy [9 –11, 32]. If proper drainage of the leakage is not achievable the lethality reaches even 80% [10, 33]. Surgical reexploration in cases of intrathoracic anastomotic leak has shown disappointing results and is associated with a high mortality; between 60% [34] and even 100% [33]. Therefore, conservative treatment with drainage, total parenteral nutrition, nasogastric decompression, and antibiotherapy has been suggested [8, 9, 11]. The disadvantage of this conservative treatment is the ongoing pollution of the mediastinum and the pleural cavity through the anastomotic leak. The continuing contamination results in mediastinitis, pleural empyema, and sepsis. This accounts for a mortality rate similar to that of surgical reexploration and up to 40% [16]. Therefore, adequate drainage of the leak as well as prevention of further mediastinal contamination are the primary goals in the management of intrathoracic anastomotic leakage [10]. While sufficient drainage can be achieved by tube thoracostomy or interventional drainage, endoscopic stent implantation is able to prevent further pollution of the mediastinum [14]. Endoscopic stent insertion has shown good and reliable results in the treatment of intrathoracic anastomotic leaks in several series [12–17]. Self-expanding covered silicone or metal stents can close leaks regardless of size as long as the pulled-up gastric tube is not ischemic and there is no complete dehiscence of the anastomosis. These preconditions have to be verified by endoscopy with direct visualization of the anastomosis before stent insertion can be considered [10 –12]. At the same time, first reports of stent-related vascular erosions were published [18 –23]. These reports are mostly on patients with benign or malignant stenosis of the esophagus who received endoscopic stent insertion in palliative intention to reopen passage for oral intake. Erosion of the stent through the esophageal wall into the aortic arch were reported after prolonged stent placement of several months as well as after quite short placement of just a few days [19 –23]. Also, stent-related erosions of other major blood vessels were reported [18]. The aim of this retrospective study was to investigate if stent insertion, while on the one hand providing im-

Comment Aortoesophageal fistula is an infrequent complication after esophagectomy [1, 3, 5]. Merendino and Emerson, from Seattle [2], were the first to publish 2 cases of postoperative aortoesophageal fistula they had observed (in 1946 and 1947) after left transthoracic esophagectomy. In both cases the fistula had been clearly related to defects of the intrathoracic esophagogastric anastomosis with a technical defect being responsible in the first and an anastomotic leak in the second case [2]. Lookmann (from the Thoracic Surgery of the Royal Infirmary in Bradford) reported in 1959 [32] 3 fatal cases of aortogastric fistula after left thoracoabdominal partial esophagectomy. In all 3 cases the fistula was related to the intrathoracic esophagogastric anastomosis [32]. In 1961 Le Roux (from the University of Edinburgh) published 8 cases [3] of aortic erosion in a series of 418 consecutive patients with abdominothoracic esophagectomy and intrathoracic end-to-end esophagogastric anastomosis. In 7 cases the fistula was related to the anastomosis and in 1 case to the closure line of the pulled up stomach. In his first case, which he had encountered in 1953, postmortem examination showed, for example, an anastomotic defect with an abscess cavity the size of a golf-ball in the mediastinum “and perforation into the aortic lumen.” All 8 patients died from massive hemorrhage. Le Roux concluded that the “aortic fistula usually communicates with the alimentary tract through a defect of the esophagogastric anastomosis.” He expected to encounter one such fistulization between the insufficient anastomosis and the thoracic aorta in every 52 operated patients [3]. The first successful surgical reintervention in a case of postoperative aortoesophageal fistula was described by Maillard and colleagues (from France) [4]. In 1967 they reported about 4 patients with aortoesophageal fistula after left thoracoabdominal partial esophagectomy. All fistulas were related to anastomotic leaks. Since 1967 no larger series, but only reports of single cases, have been published [1, 5, 24]. Apparently aor-

SCHWEIGERT ET AL AORTOESOPHAGEAL FISTULA AFTER ESOPHAGECTOMY

proved results in closure of intrathoracic anastomotic leaks, on the other hand leads to increased incidence of aortoesophageal fistula through the backdoor by reason of stent-related vascular erosion. In this study endoscopic stent placement was feasible in 17 patients with endoscopically verified intrathoracic anastomotic leak without ischemia of the pulled-up gastric tube or the pulled-up jejunum. Immediate closure of the leak was achieved without exception and, invariably, all 17 patients first recovered well. Mediastinitis and pleural empyema healed, respiratory failure and sepsis were overcome, and all patients could again start oral intake. These are very promising short-term results. In 13 patients stent removal was routinely performed after complete healing of the anastomotic leak, and in 1 patient the stent remained in situ only because of early mediastinal tumor recurrence with rapid progress of a high malignant small cell carcinoma. Thus, in 14 out of 17 patients (82%) endoscopic stent insertion led to recuperation from intrathoracic anastomotic leakage, even though 1 patient died 24 days after stent explantation from severe aspiration, which is a frequent problem in all patients after esophagectomy and not primarily related to the anastomotic leak. But notwithstanding these favorable results, stentrelated erosion of the thoracic aorta occurred in 3 cases. The diagnosis was always confirmed by postmortem examination. Aortoesophageal fistula established itself between the 16th and the 41st day after stent implantation, always becoming recognizable from a herald bleeding with hematemesis. Afterward, invariably hypovolemic shock with death from exsanguination following rapidly. In 1 case endoscopic reintervention was tried but, upon endoscopy, bleeding intensity became immense and the patient died immediately in spite of mass transfusion and resuscitation efforts. In none of the 196 cases without stent implantation was an aortoesophageal fistula seen. Thus, the development of fistulization between the anastomosis and the thoracic aorta in this series seems to be related to the presence of anastomotic leakage and stent insertion. At first glance 3 aortoesophageal fistulas in a series of 213 cases are nearly as many as in the series of 418 cases with 8 aortic erosions reported by Le Roux in 1961 [3]. Does this mean that there had been no improvement in the last 50 years regarding this devastating complication? Or has, rather, the incidence newly risen again by reason of stent-related vascular erosion? While Le Roux reports an overall mortality of 30% [3], mortality in our series was only 8%. Aortoesophageal fistula itself has in almost all reports and series been related to anastomotic leakage and anastomotic leakage was, ever since the days of Merendino and Emerson, a main reason for postoperative mortality [34]. The vastly reduced mortality is among others attributable to the improved management of postoperative anastomotic defects. Endoscopic stents have shown themselves to be very effective tools in dealing with intrathoracic anastomotic leakage [12, 14]. But it seems that simultaneous to increasing numbers of implanted

stents the previously described risk of stent-related vascular erosion becomes more imminent. None of the 3 patients, who finally developed an aortoesophageal fistula, had received neoadjuvant therapy [35]. Therefore neoadjuvant chemotherapy or chemoradiotherapy does not provide an explanation for the origin of aortoesophageal fistula in this series. In 2 cases a silicone stent and in 1 case a metal stent had been implanted. Thus fistulization is unlikely attributable to the material of the stent. In the second described case stent migration and stent dislocation happened several times. After initial successful stent placement within 5 weeks multiple endoscopic interventions were mandatory, including exchange of the stent with an increase in stent size. During those 5 weeks the closure of the leak was probably less than ideal and ongoing contamination of the mediastinum can easily be assumed. Prolonged leakage in combination with mechanical irritation by an inappropriate stent might be explanation enough for inflammatory fistulization between the deficient anastomosis and the aorta. In the third case an anastomotic leak had been clinically suspected for a long time. At postoperative day 4 the chest tubes contained dirty fluid suspicious for an anastomotic defect. Afterward the patient developed a pleural empyema that finally required surgical debridement by video-assisted thoracoscopy. But not until postoperative day 13 was a stent finally inserted. Between the postoperative days 4 and 13 the mediastinum was probably continually polluted through an endoscopic invisible but clinically apparent anastomotic leakage. Hence, in the second and third cases ongoing mediastinal contamination either by inappropriate stent or delayed stent placement can easily be presumed. Le Roux described in his 1953 encountered case [3] an “abscess cavity the size of a golf-ball in the mediastinum” and “where the aorta formed the wall of the abscess cavity its adventitia was necrotic and there was a perforation into the aortic lumen.” Thus, mediastinal inflammation with involvement of the aortic wall caused by persisting anastomotic leakage favors the development of postoperative aortoesophageal fistula [3, 34]. In such a situation additional mechanical irritation by the stent may be all that is needed to finally cause fistulization. Therefore, in cases when correct stent placement is not readily achievable, or in cases when the diagnosis of anastomotic leakage has been considerably delayed, treatment options other than endoscopic stent implantation should be considered to prevent a situation where prolonged mediastinal inflammation and mechanical irritation by the stent may contribute to an increased risk for developing an aortoesophageal fistula. In conclusion, increase in the treatment of intrathoracic anastomotic leaks by means of endoscopic stent insertion puts the already well-known risk of stent-related vascular erosion on the spot. Awareness for this lifethreatening and often fatal complication is therefore mandatory. Better prevention seems achievable by improved patient selection. Notwithstanding the risk of aortoesophageal fistula endoscopic stent implantation

GENERAL THORACIC

517

Ann Thorac Surg 2011;92:513–9

GENERAL THORACIC

518

SCHWEIGERT ET AL AORTOESOPHAGEAL FISTULA AFTER ESOPHAGECTOMY

Ann Thorac Surg 2011;92:513–9

remains an effective and reliable approach to deal with intrathoracic anastomotic leakage.

temporary stenting with self-expanding plastic stents. Br J Surg 2009;96:887–91. Zisis C, Guillin A, Heyries L, et al. Stent placement in the management of oesophageal leaks. Eur J Cardiothoc Surg 2008;33:451– 6. Vanden Eynden F, Devière J, Laureys M, de Cannière D. Erosion of a retroesophageal subclavian artery by an esophageal prosthesis. J Thorac Cardiovasc Surg 2006;131:1183– 4. Siersema PD, Tan TG, Sutorius FF, Dees J, van Blankenstein M. Massive hemorrhage caused by a perforating Gianturco-Z stent resulting in an aortoesophageal fistula. Endoscopy 1997;29:416 –20. Unosawa S, Hata M, Sezai A, et al. Surgical treatment of an aortoesophageal fistula caused by stent implantation for esophageal stenosis: report of a case. Surg Today 2008;38: 62– 4. Um SJ, Park BH, Son. An aortoesophageal fistula in patient with lung cancer after chemo-irradiation and subsequent esophageal stent implantation. J Thorac Oncol 2009;4:263–5. Rogart J, Greenwald A, Rossi F, Barrett P, Aslanian H. Aortoesophageal fistula following polyflex stent placement for refractory benign esophageal stricture. Endoscopy 2007; 39(Suppl 1):E321–2. Ahn M, Shin BS, Park MH. Aortoesophageal fistula secondary to placement of an esophageal stent: emergent treatment with cyanoacrylate and endovascular stent graft. Ann Vasc Surg 2010;24:555.e1-5. Whitelocke D, Maddaus M, Andrade R, D=Cunha J. Gastroaortic fistula: a rare and lethal complication of esophaeal stenting after esophagectomy. J Thorac Cardiovasc Surg 2010;140:e49 –50. Stein HJ, Feith M, Siewert JR. Cancer of the esophagogastric junction. Surg Oncol 2000;9:35– 41. Stein HJ, Feith M, Siewert JR. Individualized surgical strategies for cancer of the esophagogastric junction. Ann Chir Gynaecol 2000;89:191– 8. Rusch VW. Are cancers of the esophagus, gastroesophageal junction, and cardia one disease, two, or several? Semin Oncol 2004;31:444 –9. Siewert JR, Stein HJ, Sendler A, Fink U. Surgical resection for cancer of the cardia. Semin Surg Oncol 1999;17:125–31. von Rahden BH, Stein HJ, Siewert JR. Surgical management of esophagogastric junction tumors. World J Gastroenterol 2006;12:6608 –13. Siewert JR, Stein HJ. Esophagectomy as therapeutic principle for squamous cell esophageal cancer. [Article in German] Chirurg 2005;76:1033– 43. Stein HJ, Hutter J, Feith M, von Rahden BH. Limited surgical resection and jejunal interposition for early adenocarcinoma of the distal esophagus. Semin Thorac Cardiovasc Surg 2007;19:72– 8. Lookman AA. Aortogastric fistula. Br J Surg 1959;46:652–5. Urschel JD. Esophagogastric anastomotic leaks complicating esophagectomy: a review. Am J Surg 1995;169:634 – 40. Alanezi K, Urschel JD. Mortality secondary to esophageal anastomotic leak. Ann Thorac Cardiovasc Surg 2004;10:71–5. Lordick F. Neoadjuvant therapy for squamous cell carcinoma of the esophagus. [Article in German] Chirurg 2005;76:1025–32.

References 1. Molina-Navarro C, Hosking SW, Hayward SJ, Flowerdew AD. Gastroaortic fistula as an early complication of esophagectomy. Ann Thorac Surg 2001;72:1783– 8. 2. Merendino KA, Emerson EC. Aorto-esophagogastric fistula, an unusual complication of esophagogastrostomy performed under the aortic arch following esophageal resection for carcinoma; a report of two cases. J Thorac Surg 1950;19: 405–11. 3. Le Roux BT. Aortic erosion complicating oesophagogastrectomy. Br J Surg 1961;49:271–7. 4. Maillard JN, Blanc L, Brucher C. Aorto-esophago-gastric fistula following resection of an esophageal cancer. Description of an operative cure. [Article in French] Ann Clin Thorac Cardiovasc 1967;6:103–9. 5. Okita R, Mukaida H, Takiyama W, Oheda M, Uchida N, Ishihara H. Successful surgical treatment of aortoesophageal fistula after esophagectomy. Ann Thorac Surg 2005; 79;1059 – 61. 6. Bonavina L, van Lanschot JJB. Complications in oesophageal and gastric surgery. Dig Surg 2002;19:86 –7. 7. Abunasra H, Lewis S, Beggs L, Duffy J, Beggs D, Morgan E. Predictors of operative death after oesphagectomy for carcinoma. Br J Surg 2005;92:1029 –33. 8. Lerut T, Coosemans W, Decker G, De Leyn P, Nafteux P, Van Raemdonck D. Anastomotic complications after esophagectomy. Dig Surg 2002;19:92– 8. 9. Sauvanet A, Baltar J, Le Mee J, Belghiti J. Diagnosis and conservative management of intrathoracic leakage after oesophagectomy. Br J Surg 1998;85:1446 –9. 10. Siewert JR, Stein HJ, Bartels H. Anastomotic leaks in the upper gastrointestinal tract. [Article in German] Chirurg 2004;75:1063–70. 11. Griffin SM, Lamb PJ, Dresner SM, Richardson DL, Hayes N. Diagnosis and management of mediastinal leak following radical oesophagectomy. Br J Surg 2001;88:1346 –51. 12. 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. 13. Langer FB, Wenzel E, Prager G, et al. Management of postoperative esophageal leaks with the polyflex selfexpanding covered plastic stent. Ann Thorac Surg 2005;79: 398 – 404. 14. Schweigert M, Dubecz A, Stadlhuber RJ, Muschweck H, Stein HJ. Treatment of intrathoracic esophageal anastomotic leaks by means of endoscopic stent implantation. Interact Cardiovasc Thorac Surg 2011;12:147–51. 15. Peters JH, Craanen ME, van der Peet DL, Cuesta MA, Mulder CJ. Self-expanding metal stents for the treatment of intrathoracic esophageal anastomotic leaks following esophagectomy. Am J Gastroenterol 2006;101:1393–5. 16. Dai YY, Gretschel S, Dudeck O, Rau B, Schlag PM, Hünerbein M. Treatment of oesophageal anastomotic leaks by

17. 18. 19.

20.

21. 22.

23.

24.

25. 26. 27. 28. 29. 30. 31.

32. 33. 34. 35.

INVITED COMMENTARY This report by Schweigert and colleagues [1] reviewing their experience treating anastomotic leaks after esophagogastrectomy with expandable stents, and the subsequent rate of fatal aortoesophageal fistula (AEF), is commendable. Esophagogastrectomy is a technically demanding operation associated with significant morbidity and mortality. Intrathoracic anastomotic leaks are © 2011 by The Society of Thoracic Surgeons Published by Elsevier Inc

perhaps the most dreaded complication, justifiably, when one considers the increased rate of adverse events, tripling of mortality, and decreased long-term survival associated with this complication [2]. The profound effect on patient outcomes has led some surgeons to routinely perform a cervical esophagogastric anastomosis. Major leaks, associated with anastomotic 0003-4975/$36.00 doi:10.1016/j.athoracsur.2011.03.090