| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Hull and East Yorkshire Hospitals NHS Trust, Anlaby
Rd., Kingston Upon Hull, East Yorkshire, HU3 2JZ, United Kingdom.
2
Department of Surgery, Hull and East Yorkshire Hospitals NHS Trust, Kingston
Upon Hull, East Yorkshire, HU3 2JZ, United Kingdom.
3
Department of Radiology, York District Hospital NHS Trust, Wigginton Rd.,
York, YO31 8HE, United Kingdom.
Received March 21, 2000;
accepted after revision June 8, 2000.
Address correspondence to D. J. Breen.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. During a 52-month period, 14 patients had placement of stents for clinically significant postoperative leaks: 10 patients had an esophagogastrectomy and four patients had a total gastrectomy with esophagojejunal anastomosis. Thirteen of 14 patients had tumors that were histologically staged as T3 N1 M0 or worse. Significant anastomotic leaks were revealed by a contrast-enhanced study at 3-28 days after surgery. Stents were inserted in patients in whom the leakage was debilitating or initial conservative treatment had failed. Stenting outcome in terms of clinical and radiologic healing, hospital stay, survival, and complications was assessed.
RESULTS. No procedural morbidity or 30-day mortality occurred. Immediate postprocedural leak occlusion was obtained in all patients. Clinical healing of the leak occurred in 13 (92.8%) of 14 patients, with a median healing time of 6 days. Of the 13 patients, healing occurred within 10 days in 10 patients (76.9%). Eight of these 10 early closures received a knitted nitinol stent (p = 0.02). One patient (7%) died as a consequence of leakage at 135 days. Median survival for all 14 patients was 11 months (Kaplan-Meier method). Complications included five episodes of food blockages in three patients, which required endoscopic clearance, and one case of stent-related upper gastrointestinal hemorrhage. No patients developed anastomotic stricture or occlusive epithelial hyperplasia.
CONCLUSION. Covered esophageal stenting appears to reduce the mortality and morbidity of symptomatic anastomotic leakage after surgery for gastroesophageal cancer. Knitted nitinol stents may be best suited to this purpose.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A key issue is that there are no uniform methods for treating patients with symptomatic intrathoracic leakage. Some authors suggest surgical reexploration and repair [2, 6, 7], whereas others recommend a more conservative approach using enteral or parenteral nutrition, perianastomotic drainage in the form of intercostal drains, or CT-guided percutaneous drainage [5], and broad-spectrum antibiotics [6]. In spite of current advances, the mortality from clinically apparent thoracic leakage remains high [2, 3, 5], often approaching 60% [6]. Clinically asymptomatic small-volume leaks revealed on fluoroscopic contrast studies are now usually treated conservatively.
Covered esophageal stents are emerging as the treatment of choice in the palliative treatment of patients with advanced esophageal cancer [8, 9]. Various types of covered stents have also been used in the treatment of iatrogenic esophageal perforation or malignant esophago-respiratory fistulas with excellent results and good short-term quality of life [8, 10,11,12].
The aim of our study was to analyze the effectiveness of covered stents in the treatment of symptomatic postoperative gastroesophageal anastomotic leak (EAL) in comparison with other established methods of treatment in terms of reduction of leak-related mortality and time to EAL healing.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Ten of the 14 patients had an esophagogastrectomy either through a left thoracotomy (3 patients) or a right thoracoabdominal approach (Ivor-Lewis approach) (7 patients) in which a laparotomy to mobilize the stomach and a right thoracotomy are performed to complete the anastomosis. Four patients had total gastrectomies and end-to-side esophagojejunal anastomosis. The anastomosis was fashioned using a circular stapling device in 11 of 14 patients and was handsewn in three. The gastric suture line was stapled in all but one patient. Surgery was performed with curative intent in seven of the 14 patients and was deemed palliative at exploratory surgery in the remaining patients. Resection margins were clear in 12 patients. The tumors of 13 of 14 patients were histologically staged as at least T3 N1 M0 using the TNM classification; 13 patients had adenocarcinoma and one patient had a squamous cell carcinoma.
All patients had a major or moderate anastomotic leak [2] before recourse to stenting was undertaken. These leaks were defined by uncontained passage of contrast material into the pleural space or the extraluminal passage of most of the swallowed contrast material, with clinical sepsis and active chest-drain efflux.
If the leakage was of small volume, then conservative treatment [5] was initiated until the aforementioned criteria were met. Patients with contained leaks who remained asymptomatic were excluded from the study.
Anastomotic leakage was detected using a water-soluble contrast agent (meglumine amidotrizoate [Gastrografin]; Schering, West Sussex, United Kingdom) between 2 and 9 days after surgery. The median time to confirmation of a significant leak was 13.5 days after surgery (range, 3-28 days). Two patients had leakage from the gastric staple line where the gastric tube had been fashioned. In one of these patients, an additional leak occurred at the stapled gastroesophageal anastomosis.
Informed consent was obtained from all patients. Stents were inserted between 9 and 34 days (median, 19.5 days) after surgery. A total of 19 stents was placed in 14 patients (Table 1). Ten were covered Ultraflex stents (Boston Scientific, Watertown, MA) in its wide-diameter configuration, three were covered Gianturco Rosch Z stents (Cook, Bjaeverskov, Denmark), and six were covered Telestep Wallstents (Schneider, Minneapolis, MN). The Ultraflex stent has a 16-French delivery sheath, knitted nitinol structure, and the covered segment is coated with a single layer of polyurethane. The Gianturco Z stent is a polyethylene-covered stent with a 24-French delivery system. The Schneider Telestep Wallstents were coated with polyurethane and were delivered using a 18-French system. Three patients had multiple coaxial stents, and in one of these patients, four coaxial stents were placed in an attempt to cover a persistent symptomatic leak.
Stents were inserted under conscious sedation with IV midazolam and fentanyl as necessary. The insertion technique has been previously described [8, 9, 11]. Using fluoroscopic guidance, we clearly identified the anastomosis and crossed using standard guidewire-and-catheter technique. The exact site of the leak and the efferent postanastomotic bowel was identified, particularly in the case of end-to-side esophagojejunal anastomoses (Fig. 1A,1B). The covered segment of the stent was deployed to occlude the leakage site. A contrast-enhanced examination was performed after the procedure to assess immediate leak occlusion.
|
|
The patients continued to be treated with broad-spectrum antibiotics (12/14 patients) and enteral (2/14) or parenteral (12/14) nutrition. A contrast swallow was performed 1-7 days after stenting and was repeated later as necessary to assess sustained radiologic closure. Patients were examined in the erect and supine positions to provoke any persistent leak (Fig. 2). This form of provocative contrast assessment revealed significant ongoing leaks as well as smaller clinically asymptomatic leaks, which were seen in two patients. Three patients had adjunctive surgical procedures (two intercostal drain placements and one empyema drainage) for persistent leakage. In addition, fibrin tissue glue was tried in one of these patients with continuing leakage without any benefit.
|
Clinical healing of the EAL was deemed to have occurred when the patient became asymptomatic and the consulting surgeon recommenced oral feeding. Radiologic evidence of EAL healing occurred when no anastomotic leakage could be shown with provocative contrast assessment. We defined expedited healing to have occurred in comparison with other established methods of EAL treatment when healing was seen in less than 10 days.
All patients were followed up clinically, endoscopically, or radiologically. Accrued patient follow-up to date is 128 patient-months (median, 8 patient-months; range, 3-18 patient-months).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Immediate leak occlusion was obtained on erect contrast assessment after the procedure in all patients. Healing of the EAL occurred in 13 (92.8%) of 14 patients. In one of 14 patients, the original leak closed after 117 days, but a recurrent leak developed and the patient died of empyema and septicemia. Clinical healing occurred within 10 days (and was considered expedited) in 10 (71.4%) of 14 patients. Of these 10 patients, two had small contained asymptomatic leaks, one of which completely closed at 24 days. The other patient was not referred for radiologic confirmation of leak closure. Both patients were receiving oral intake, and the leak visible on radiography was considered clinically insignificant. Therefore, radiologic healing occurred within 10 days in eight of 14 patients.
In three of 14 patients, delayed fistula closure occurred between 42 and 85 days. The median period to clinical healing was 6 days (range, 3-85 days; mean, 17.5 days), and the median period to radiologic healing was 8 days (range, 2-135 days). The median hospital stay after the stent procedure in the 13 surviving patients was 17 days (range, 4-117 days).
Of the 10 patients with early healing, eight were treated with a covered Ultraflex stent. In this small series there was a significant difference in early closure when the 10 Ultraflex stents were compared with the other two stent types (p = 0.02, Fisher's exact test). No statistically significant correlation was found between early healing and type of resection (palliative or curative) or type of anastomosis. None of the patients developed symptomatic anastomotic stricture or occlusive epithelial hyperplasia. The median survival of the whole patient group was 11.0 months (standard error, 2.7; Kaplan-Meier survival analysis).
In three patients, five episodes of food blockages occurred that required endoscopic clearance. In one patient, significant upper gastrointestinal hemorrhage that prompted removal of the Z stent using a rigid esophagoscope occurred from an erosion associated with the inferior margin of the stent 7 months after stent placement.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Surgical reexploration or conservative treatment of EAL are well described in the surgery literature [5,6,7]. Sauvanet et al. [5] describe conservative treatment in a series of 38 patients, of which 27 were symptomatic. The treatment involved assisted nutrition and perianastomotic drainage. In nine (33%) of these 27 patients, death was attributable to leakage. Viste et al. [3] analyzed 22 clinically symptomatic leakages, of which 10 required a second surgery. In this series nine (40.9%) of the 22 patients died. In another series of 13 patients with both symptomatic and smaller asymptomatic leaks [13], six patients (46.2%) died.
In our series of 14 patients with clinically significant EAL, stenting incurred no procedural morbidity or mortality. Thirty-day mortality was nil. If outcome of this stented group is analyzed as defined by the outcome of the previously mentioned studies [3, 5, 13], mortality due to leakage has been reduced to 7%. This outcome, in our view, was the principal advantage of stenting.
The major limitation of this study is the lack of a suitable control group from the same setting with which to compare the stenting results. Formal statistical analysis of outcome against other treatment protocols is problematic because of the small and heterogeneous nature of the patient groups involved. Although outcome studies are difficult to compare, it is worth looking into published figures, particularly those on conservative management of EAL. Mortality from leakage seems to be markedly reduced after stenting. The main measures of patient morbidity remain the time to leak closure, the beginning of oral feeding, and the period of inpatient stay. Surprisingly little has been written in the recent literature on these issues and on the natural history of EAL with conservative treatment. Fan et al. [14] reported the conservative treatment of 26 patients with clinical anastomotic leaks, of which only nine patients had intrathoracic leakage. Five of these 26 patients died of uncontrolled sepsis, and two patients died as a result of pulmonary infection after leakage. Nine of 26 patients had a persistent fistula. Closure occurred in 15 patients, with a mean time to closure of 24.6 days (range, 10-56 days). In an additional series with 17 patients [15], 10 patients healed (mean healing time, 33 days) and seven died. EAL healing occurred in our series in 13 of 14 patients at a median time of 6 days (mean, 17.5 days) after stenting. Compared with published figures, our data seems to show that stent insertion expedited clinical healing.
Patients in this study usually underwent a period of trial or failed conservative treatment (median, 19.5 days) before referral for stenting. This policy has invariably prolonged the total hospital stay. Median hospital stay after the stenting procedure in the 13 surviving patients in this study was 17 days (range, 4-117 days). In a comparable series of 18 surviving patients with a leaked gastroesophageal anastomosis [5], the median postoperative hospital stay was 44 days (range, 21-102 days).
In two of the patients who had delayed or nonhealing of EAL, a large leak was discovered within 4 days. The usual cause of this early leakage is perianastomotic tissue necrosis. Surgical reexploration, which has a high mortality rate [6], is usually advised for this group of patients [2, 6, 16, 17]. In retrospect, these patients were perhaps not suitable candidates for stenting, although the results are likely to be poor whatever treatment is undertaken.
The healing of EAL is determined by a constellation of clinical and radiologic criteria. These include the absence of sepsis, a decreased chest-drain efflux, and a contrast-enhanced study that confirms healing. Oral feeding is initiated at the discretion of the supervising surgeon. In addition to the eight patients who had expedited radiologic healing of the EAL within 10 days, two additional patients became clinically asymptomatic after stenting. In these two patients, although a small leak could still be defined radiologically, the chest-drain efflux and other clinical indicators had improved. As a result, the patients were considered to have early EAL closure, and oral feeding was initiated by the consulting surgeon on the basis of clinical findings.
Covered metallic stents have been used in the treatment of esophageal cancer, malignant fistulas, and iatrogenic esophageal injury with significant success in poor-risk patients [8, 9, 11]. In one series [11], 19 perforations after endoscopic dilatation were stented successfully. Such a series would inevitably include cases tending to have a better outcome because the injury is diagnosed shortly after endoscopic perforation [18]. In this situation, a number of cases are likely to be self-limited contained mediastinal leaks, and the covered stent appears to configure better to the intact tubular esophagus than to a postoperative anastomosis in which there is a propensity to "back leak" around the stent along the refashioned gastric tube (Fig. 2).
Comparison with the above group of patients would explain why sustained closure of all EALs after stenting was not attained immediately. However, the diversion of the fluid stream away from the fistula appears to be a clinical benefit. In this small series, no particular type of anastomosis healed more effectively than any other after stenting.
The idea of an intraluminal prosthetic device to protect from the adversities of leakage is not new. Ravo and Ger [19] described a surgically placed intraluminal bypass tube in an experimental trial for the treatment of esophageal dehiscence. Isolated cases of stenting across leaking symptomatic gastroesophageal anastomoses have also been described. Segalin et al. [20] described temporary Wilson-Cook esophageal prosthesis for recurrent postoperative leaks in two patients. Radiologic healing occurred in 2-3 weeks and oral feeding was resumed on days 25-26 after prosthesis insertion. Occasional case reports of persistent leaks successfully closed using covered stents are found in the literature [12, 21]. In one patient reported [21], the stent was subsequently endoscopically removed because of a hyperplastic stricture causing dysphagia.
Doubts have been raised about the long-term effectiveness of covered stents because of their high rate of migration, chest pain after insertion, and upper gastrointestinal hemorrhage [22]. Lower rates of complications have been reported with Gianturco Rosch Z stents [8, 23]. In our series, one significant stent-related complication occurred during the accrued follow-up of 128 patient-months. This patient experienced a significant upper gastrointestinal hemorrhage that prompted endoscopic stent removal. Otherwise, the only complications were five cases of food blockages in three patients during the same period.
The issue of epithelial hyperplasia or granulation tissue in response to a covered stent in the potentially benign (i.e., after a curative resection) situation is also of concern. Epithelial hyperplasia has been noted in animal models [24] and in relation to stenting of benign strictures, in up to two of five patients at 4 and 7.5 months of follow-up [25]. Although we did not experience any clinically significant epithelial hyperplasia, the place of long-term metallic stents in the benign situation remains to be seen. In this series, however, stenting was reserved for the treatment of patients with tumors staged as at least T3 N1 M0 (13 of 14 patients) in whom the long-term clinical outlook was inherently poor. Although they are likely to be of overall clinical benefit, the use of currently available stents to treat EAL in young patients with early-stage disease cannot be advocated.
Three patients survived 14, 16, and 18 months after stenting without significant complications. A total of three patients remain alive at the time of this report. The median survival of the entire patient group was 11 months (range, 4-18 months), which compares favorably with the expected 3- to 4-month survival reported in other series of palliative stenting of the esophagus [11, 23].
There is an increased incidence of stricture formation caused by anastomotic leaks [6, 17]. Two of the 18 symptomatic patients in one study developed anastomotic stricture requiring endoscopic dilatation [5]. In our follow-up, none of the patients have experienced this complication. The metallic endoprosthesis may protect against stricturing after mediastinal leakage.
In summary, on the basis of this outcome analysis, covered esophageal stenting across a surgical anastomosis is a safe and effective technique. We believe that the procedure is more effectively carried out by the radiologist using careful fluoroscopic guidance (Fig. 3A,3B). Covered stenting appears to reduce the morbidity and mortality of symptomatic mediastinal leakage and expedited healing in 10 (71.4%) of 14 patients, thereby reducing cost, and possibly length of hospital stay, from that incurred with current treatment. Our experience suggests that the covered knitted nitinol stent is best suited to this procedure by virtue of its closer configuration to the postoperative anatomy. Consideration should be given to covered esophageal stenting in the highly morbid postoperative complication of gastroesophageal resection for carcinoma.
|
|
Acknowledgments
We thank Eric D. Gardiner for doing the statistical analysis of the data;
Bridget Freer and Karon Turner for preparing the manuscript; and M. E. Cowen,
G. Miller, and M. P. Tilston for permission to report on their patients.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
|
|
F. B. Langer, E. Wenzl, G. Prager, A. Salat, J. Miholic, T. Mang, and J. Zacherl Management of Postoperative Esophageal Leaks With the Polyflex Self-Expanding Covered Plastic Stent Ann. Thorac. Surg., February 1, 2005; 79(2): 398 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S H Lee The role of oesophageal stenting in the non-surgical management of oesophageal strictures Br. J. Radiol., October 1, 2001; 74(886): 891 - 900. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
