|
|
||||||||
Ann Thorac Surg 2005;79:398-403
© 2005 The Society of Thoracic Surgeons
a Department of Surgery, Division of General Surgery, Medical University of Vienna, Vienna, Austria
b Department of Radiology, Medical University of Vienna, Vienna, Austria
Accepted for publication July 6, 2004.
* Address reprint requests to Dr Zacherl, University Clinic of Surgery, General Hospital Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria (E-mail: johannes.zacherl{at}meduniwien.ac.at).
| Abstract |
|---|
|
|
|---|
METHODS: Between April 2000 and November 2003, 24 patients were included into this prospective study and underwent Polyflex stent placement for postoperative esophageal anastomotic leaks. The primary operation was esophagectomy in 13 patients, gastrectomy in 7, cardia resection in 2, and other procedures in 2 patients. The median interval between operation and stent placement was 19 days (range, 4 to 65). The effectiveness of leak occlusion was evaluated by water-soluble contrast swallow and the clinical course.
RESULTS: In 2 patients stent misplacement produced an enlarged anastomotic dehiscence that necessitated reoperation. Radiologic evaluation was impossible in 4 patients because of their generally restricted condition. Among 18 evaluable patients, leak occlusion was successful with a single stent in 16 patients (89%) based on radiologic evaluation. Immediate oral feeding was well tolerated by these patients. After a median follow-up of 220 days (range, 7 to 1221), 9 cases of late stent dislocation were observed. Stent removal in patients after esophagectomy with gastric pull-up led to dysphagia from anastomotic strictures in 2 patients. Symptomatic strictures did not develop in the 5 evaluable postgastrectomy patients after stent removal.
CONCLUSIONS: The placement of self-expanding plastic stents is a highly effective treatment for esophageal anastomotic leaks. Because clinically-relevant anastomotic strictures can be expected, we do not recommend stent removal after esophagectomy with gastric pull-up reconstruction.
| Introduction |
|---|
|
|
|---|
Various types of covered stents are successfully used in the treatment of malignant esophageal fistulas and esophageal tumor perforation after dilatation [1012]. There have been a few recent reports on the implantation of self-expanding, covered metal stents (SEMS) as an effective treatment for symptomatic esophageal anastomotic leaks after cancer resection [13]. However, potential restrictions in the use of SEMS are high cost, difficulties in stent removal or repositioning, mucosal hyperproliferation, permigration, and bleeding [1417].
These disadvantages may be avoided by the placement of the Polyflex (Willy Ruesch GMBH, Kernen, Germany) self-expanding plastic stent. A number of reports have indicated that inoperable esophageal obstructions can be managed successfully with Polyflex stents [1821].
This study analyzes the effectiveness of Polyflex plastic stent placement for closure of postsurgical esophageal anastomotic leaks in a series of 24 patients.
| Material and Methods |
|---|
|
|
|---|
|
Table 2 depicts the size distribution of the implanted endoprostheses. All interventions were done under general anesthesia. After endoscopic evaluation of the local condition and positioning of the guide-wire, stent positioning and deployment was performed under fluoroscopic guidance as described in the literature [18, 20].
|
The effectiveness of leak occlusion was evaluated radiographically by Gastromiro (iopamidol), a water-soluble contrast-medium swallow (Gerot Pharma, Vienna, Austria). If the occlusion of the fistula was successful, oral feeding was started within 18 hours after stent implantation.
Whenever occlusion of the fistula could not be achieved after the first intervention, either a different Polyflex stent with a wider diameter was used or the stent position was changed.
In patients of the gastrectomy group, the stent was used for bridging until leak healing and its removal was scheduled for at least 3 weeks after placement. To prevent dysphagia from anastomotic stricture, the stent was not routinely removed from patients who underwent esophagectomy.
The follow-up was done by telephone and outpatient clinic visits. In cases of suspected stent dislocation or dysphagia, a contrast-swallow esophagography was performed.
| Results |
|---|
|
|
|---|
Among the remaining 22 patients, 18 underwent a postinterventional radiographic investigation. Contrast-swallow esophagography was impossible in 4 patients because of their poor general condition. One of these patients subsequently underwent esophagectomy after initial cardia resection because of clinical deterioration. Of the remaining 3 patients, 1 died of bilateral pneumonia and hepatic cirrhosis on day 7, 1 of thrombosis of the superior mesenteric artery on day 8, and 1 of multiorgan failure from an unknown source of sepsis on day 67 after successful stent placement with an unremarkable anastomosis at autopsy.
After placement of a single self-expanding Polyflex plastic stent, a radiographic investigation revealed successful initial closure of the anastomotic leakage in 16 of the 18 patients (89%). In these patients, semisolid oral feeding initiated within 18 hours was well tolerated and drainage efflux ceased. Persistent leakage was revealed in 2 of the 18 patients (11%), both of whom had undergone a prior gastrectomy. The persisting fistula was treated by the placement of a different Polyflex stent with a wider diameter. In 1 patient, the stent was exchanged twice for a larger size until sufficient leak occlusion could be achieved (Table 3).
|
No early dislocations (within 72 hours) were observed in the 24 patients. Over a median follow-up of 220 days, 9 instances (37.5%) of late stent dislocation after placement were observed on day 8, 28, 30, 42, 57, 59, 66, 70, and 95. Of those patients experiencing dislocation, the primary operation was an esophagectomy with gastric pull-up in 4 patients, jejunal interposition after cervical esophageal resection in 1, gastrectomy in 1, and laser transsection for Zenker's diverticulum in 1 patient.
Late stent dislocation occurred in both patients who underwent colon interposition. Therapy consisted of stent removal after the fistula was healed in 3 patients, stent repositioning in 4, and stent exchange in 2 patients (Table 4). The stent was repositioned at day 8, 28, 57, and 95, respectively. One patient underwent a stent switch to an Ultraflex stent (Boston Scientific, Natick, MA) 119 days later owing to re-dislocation and stricture formation. Another patient underwent surgical abscess drainage and stent repositioning because of a late leak at day 57 presenting with a cervical abscess. The stent was removed after leak healing 179 days later because of extensive heartburn. During 83 days of observation, the patient tolerated oral feeding well without reflux. In the third patient, another dislocation was observed 21 days after stent repositioning and the stent was changed. In the fourth patient, repositioning was successful without further intervention over a follow-up period of 56 days. Thus, repositioning was successful in 1 of 4 patients (25%).
|
The stents were routinely removed from 5 postgastrectomy patients on day 22, 23, 29, 58 and 70 after successful leak occlusion. In every case, neither radiographic nor endoscopic examination revealed any sign of leak persistence. After a median follow-up of 240 days (range, 88 to 1199) after stent removal, none of these 5 patients developed any clinically relevant stricture.
In all the cases of stent removal, the underlying mucosa showed an almost normal endoscopic morphology. Rarely, mild mucosal hyperproliferation was observed at the stent margins.
Eight of the 24 patients (33%) were in poor clinical condition at the time of stent placement and required intensive care treatment and ventilation owing to sepsis syndrome and organ failure. One of these patients underwent surgical revision and anastomotic diversion because of ischemia of the interposed jejunal segment. In retrospect, stenting was obviously an inadequate treatment in this case.
The 30-day mortality was 12.5% (3/24), overall in-hospital mortality was 25% (6/24). These patients were critically ill before stent placement. In one patient, the lethal outcome might have been associated with stent misplacement since reoperation did not reveal any other surgical complication. The patient subsequently developed pneumonia and never recovered from reoperation. A further patient was reoperated for anastomotic leak enlargement that was due to stent malpositioning. But a large subhepatic abscess caused by bile duct necrosis after extended lymphadenectomy was revealed during surgery. The anastomotic treatment was locally successful, but the patient died of complications from the bile duct necrosis. As confirmed by autopsy, the remaining 5 patients died as a result of complications associated with the primary surgery: 1 of pneumonia, 1 of bilateral pneumonia and decompensated liver cirrhosis, 1 of thrombosis of the superior mesenteric artery, 1 of cardiac arrest, and 1 patient of sepsis from an unknown source (Table 1). In none of these cases the anastomotic region showed any sign of persisting mediastinitis or peritonitis at autopsy.
The median hospital stay of the surviving patients (n = 18) from stent placement until discharge was 8 days (range, 3 to 34). Five of these patients stayed for more than 10 days, mostly because of reasons not associated with the leak.
Overall, 14 of 24 (58%) patients are still alive after a median follow-up of 220 days, as an additional 4 patients have died from their tumor disease. At follow-up, 7 patients had their stent removed and were without complaints, and another 7 are doing well with a stent (6 Polyflex, 1 Ultraflex) in situ.
| Comment |
|---|
|
|
|---|
Radiologic and endoscopic placement of stents in the esophagus is widely accepted as effective palliation of nonresectable malignant esophageal strictures with instant relief of dysphagia [11, 1821, 27, 28]. Nowadays, the most frequently used stents are SEMS. Rigid plastic endoprostheses have been abandoned because of the high incidence of procedural complications [29, 30].
A few reports have been published on the use of esophageal stents to manage postoperative leaks or tumor perforation [10, 31, 32]; however, this report deals with the application of the Polyflex self-expanding plastic stent in a large series of patients with postoperative esophageal anastomotic leaks. We were able to evaluate leak occlusion after Polyflex stent placement in 18 patients. In almost 90% (16 of 18), we observed instant complete sealing of the fistula with the Polyflex stent. Leaks of cervical esophagogastrostomies were initially occluded in all evaluated patients. Immediate oral feeding was well tolerated, and drainage was suspended. This corresponds with the observations of Roy-Choudhury and colleagues [13] who used various stent types, with exception of the Polyflex stent. The median 8-day hospital stay after stent placement in the surviving 18 patients in our study was comparably shorter than the 17 days in the series of Roy-Choudhury and colleagues. However, a possible explanation for this is that they treated intrathoracic anastomotic fistulas, which are associated with higher morbidity and mortality than cervical anastomoses.
Postinterventional hospital stay was prolonged by the limited general condition in some of our patients. It is important to mention that we also included patients who had a persistent leak after protracted conservative management. Some of them were able to leave the hospital within few days after stenting, despite a preceding long time of failed conservative treatment. Others experienced long hospitalization even after stent placement because postoperative complications other than leakage resulted in a protracted reconvalescence.
The reintervention rate is remarkably high in our study, mainly because of stent dislocation. Migration rates of 6% to 18% have been reported for the use of Polyflex stents in patients with inoperable esophageal strictures [1820]. In contrast, early and late dislocations of different types of SEMS occur in 6% or less [3335]. In this regard, it is relevant to notice that luminal stenoses, which act as anchors in palliative stent treatment, are initially absent in patients with anastomotic fistula. Thus, direct contact with the mucosa is the only brace for the stent in patients with postoperative leaks. This might explain why migration occurs more frequently in the postoperative situation compared with stenting for incurable tumors. However, reports on stents for postoperative leak occlusion are rare, and the migration rate is not specifically mentioned in previous literature [13]. Additionally, the Polyflex stent is characterized by a very smooth surface that tends toward migration, especially in the absence of a stricture or stenosis. On the other hand, this might be advantageous for patients with a benign disorder such as anastomotic leakage because it allows easy endoscopic removal or adjustment.
Interestingly, stent dislocation was the predominating problem after esophagectomy, although no leak persistence was observed. In contrast, leakage persisted solely after stenting of dehiscent esophagojejunostomies, but no dislocation was encountered in these cases.
Most of the patients with stent dislocation or leak persistence experienced successful endoscopic reintervention. The success of the second intervention seemed to be more likely when the stent was not only repositioned but exchanged with a wider-diameter stent (Table 3).
The primary stent size depends on the location of the leak and the patient size. For patients with cervical anastomotic leaks after esophagectomy, we primarily recommend Polyflex stents with a diameter of 18 to 23 mm and a length of 90 mm. Stents with a diameter of 21 to 25 mm should be reserved for large patients or those with leaks after gastrectomy. These stents might cause tracheal compression or a foreign body sensation if placed in the cervical esophagus; however, in the patients who presented with esophageal leaks, we never observed tracheobronchial obstruction after esophageal stenting, even with large diameter stents. Cervical discomfort usually vanishes within a few days.
Different stent removal strategies were adopted for patients after gastrectomy and esophagectomy. No clinically relevant anastomotic strictures were revealed during an average follow-up period of 240 days that followed stent removal after leak healing in postgastrectomy patients. In contrast, 5 of 6 patients with cervical anastomotic leaks after esophagectomy, and 1 patient whose failed leak closure came after resection of a Zenker's diverticulum, presented with dysphagia because of suture strictures after removal or dislocation of the stent. Based on our experience, we do not recommend stent removal after even minor cervical leaks that follow esophagectomy, generally because of the frequent development of anastomotic stenosis. We share the opinion that in these patients, the stent should be removed only when it causes complaints.
Sheding light on the economic aspect, it should be mentioned that the Polyflex stent is substantially cheaper than the SEMS. With growing experience, it should be possible to reduce the reintervention rate and increase the economic advantage of plastic stents in contrast to SEMS and repeated dilatations.
Overall, the hospital mortality of 25% was comparably high (6/24). The 6 patients who died in the hospital were severely ill at the time of stent placement. Death was caused by complications of primary surgery in almost all cases. In one patient, the lethal outcome might have been associated with complications of stent misplacement. In the second case of misplacement, death occurred as a sequela of common bile duct necrosis and local peritonitis. Both cases of leak enlargement owing to misplacement happened in the initial study period in patients with Roux-en-Y reconstructions and were associated with difficult placement of the guidewire. Therefore, stent placement should not be performed without radiologic confirmation of the proper guide-wire positioning, with the tip at least 15 cm distal to the anastomosis to certify the intraluminal location.
The indication for stent placement in our series was closure of the postoperative leak at the esophagus that impaired regular food intake. Endoscopic placement of covered, self-expanding stents provides instant occlusion of the fistula that allows immediate oral feeding. Inner sealing of the leak shortens the healing time of anastomotic and esophagocutaneous fistula to an average 6 days after stent implantation, as previously reported [13]. Stenting enables earlier discharge from the hospital in the absence of serious comorbidities and apparently improves the quality of life of these patients.
We believe that it is appropriate to confine leaks smaller than one third of the circumference for stenting. In our opinion, larger dehiscences reflect a major problem with the reconstruction (ie, ischemia, tension) which is assumed to be insufficiently treated by stenting.
In summary, this outcome analysis indicates that the self-expanding covered Polyflex plastic stent for occlusion of esophageal anastomotic leaks is an effective treatment modality to immediately seal leaks and to enable oral nutrition. However, the main drawback of the Polyflex stent for fistula occlusion is its tendency to dislocate. This can be effectively managed by placing a large-diameter prosthesis. Mortality remains high in critically ill patients, and sepsis source control is crucial in addition to (or instead of) stent placement.
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
N. T. Nguyen, B. A. Mailey, M. W. Hinojosa, and Ken Chang Natural Orifice Management of Anastomotic Leaks After Minimally Invasive Esophagogastrectomy Surgical Innovation, December 1, 2008; 15(4): 249 - 252. [Abstract] [PDF] |
||||
![]() |
A. W. Kim, M. J. Liptay, N. Snow, P. Donahue, and W. H. Warren Utility of Silicone Esophageal Bypass Stents in the Management of Delayed Complex Esophageal Disruptions Ann. Thorac. Surg., June 1, 2008; 85(6): 1962 - 1967. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Pennathur, A. C. Chang, K. M. McGrath, G. Steiner, M. Alvelo-Rivera, O. Awais, W. E. Gooding, N. A. Christie, S. Gilbert, R. J. Landreneau, et al. Polyflex Expandable Stents in the Treatment of Esophageal Disease: Initial Experience Ann. Thorac. Surg., June 1, 2008; 85(6): 1968 - 1973. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Zisis, A. Guillin, L. Heyries, P. Lienne, X.-B. D'Journo, C. Doddoli, R. Giudicelli, and P.-A. Thomas Stent placement in the management of oesophageal leaks Eur. J. Cardiothorac. Surg., March 1, 2008; 33(3): 451 - 456. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Lindenmann, V. Matzi, C. Porubsky, U. Anegg, O. Sankin, S. Gabor, N. Neuboeck, A. Maier, and F. M. Smolle-Juettner Self-Expandable Covered Metal Tracheal Type Stent for Sealing Cervical Anastomotic Leak After Esophagectomy and Gastric Pull-Up: Pitfalls and Possibilities Ann. Thorac. Surg., January 1, 2008; 85(1): 354 - 356. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Bona, D. Sarli, G. Saino, M. Quarenghi, and L. Bonavina Successful Conservative Management of Benign Gastro-Bronchial Fistula After Intrathoracic Esophagogastrostomy Ann. Thorac. Surg., September 1, 2007; 84(3): 1036 - 1038. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. T. Nguyen, M. W. Hinojosa, C. Fayad, and S. E. Wilson Minimally Invasive Management of Intrathoracic Leaks After Esophagogastrectomy Surgical Innovation, June 1, 2007; 14(2): 96 - 101. [Abstract] [PDF] |
||||
![]() |
R. K. Freeman, A. J. Ascioti, and T. C. Wozniak Postoperative esophageal leak management with the Polyflex esophageal stent J. Thorac. Cardiovasc. Surg., February 1, 2007; 133(2): 333 - 338. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Tuncozgur, M. C. Savas, A. F. Isik, A. Sarimehmetoglu, M. Sanli, and L. Elbeyli Removal of Metallic Stent by Using Polyflex Stent in Esophago-Colic Anastomotic Stricture Ann. Thorac. Surg., November 1, 2006; 82(5): 1913 - 1914. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |