CLINICAL STUDY
Treatment of Esophagopleural Fistulas Using Covered Retrievable Expandable Metallic Stents Tae-Hyung Kim, PhD, Ji Hoon Shin, MD, Kyung Rae Kim, MD, Jung-Hoon Park, RT, Jin Hyoung Kim, MD, and Ho-Young Song, MD
ABSTRACT Purpose: To evaluate the clinical efficacy of placement of covered retrievable expandable metallic stents for esophagopleural fistulas (EPFs). Materials and Methods: During the period 1997–2013, nine patients with EPF were treated using covered retrievable expandable metallic stents. The underlying causes of EPF were esophageal carcinoma (n ¼ 6), lung cancer (n ¼ 2), and postoperative empyema for Boerhaave syndrome (n ¼ 1). Results: Technical success was achieved in eight patients (88.9%). In one patient, incomplete EPF closure was due to incomplete stent expansion. Clinical success, defined as complete EPF closure within 7 days, was achieved in five patients (55.6%). Overall fistula persistence (n ¼ 1) or reopening (n ¼ 4) occurred in five patients (55.6%) 0–15 days after stent placement. The causes of reopening were due to the gap between the stent and the esophagus (n ¼ 3) or stent migration (n ¼ 1). For fistula persistence or reopening, additional interventional management, such as gastrostomy, stent removal, or stent reinsertion, was performed. Stent migration occurred as a complication in one patient with EPF from a benign cause secondary to postoperative empyema. In the eight patients who died during the follow-up period, the mean and median survival times were 78.8 days and 46 days, respectively. Conclusions: Placement of a covered expandable metallic esophageal stent for the palliative treatment of EPF is technically feasible, although the rate of clinical success was poor secondary to fistula persistence or reopening. Fistula reopening was caused by the gap between the stent and the esophagus or by stent migration, and additional interventional treatment was useful to ensure enteral nutritional support.
ABBREVIATIONS EPF = esophagopleural fistula, PTFE = polytetrafluoroethylene
Esophagopleural fistula (EPF) is very rare and is associated with high morbidity and mortality because of ensuing empyema and nutritional debilitation (1,2). EPF is an uncommon complication of pneumonectomy and is associated with advanced esophageal carcinoma, tuberculosis, diverticulum, Boerhaave syndrome, surgical procedures,
From the Department of Radiological Science (T.-H.K.), College of Health Science, Kangwon National University, Gangwon Province, South Korea; Department of Radiology and Research Institute of Radiology (J.H.S., J.-H.P., J.H.K., H.-Y.S.), University of Ulsan College of Medicine, Asan Medical Center, 388-1, Pungnap-2dong, Songpa-gu, Seoul 138-736, Korea; and Department of Radiology (K.R.K.), University of North Carolina at Chapel Hill, Chapel Hill, North Carolina. Received October 7, 2013; final revision received December 13, 2013; accepted December 15, 2013. Address correspondence to J.H.S.; E-mail: [email protected] & SIR, 2014 J Vasc Interv Radiol 2014; 25:623–629 http://dx.doi.org/10.1016/j.jvir.2013.12.015
endoscopic examinations, chemical injury after ingestion of corrosive substances, and radiation therapy (1–8). Treatment of EPF begins with control of the empyema, repair of the esophagus, and nutritional support (9,10). Eradication of empyema by direct repair combined with a cover of viable tissue and thoracoplasty should be considered as the treatment of choice; however, because of its invasiveness, it is associated with high mortality rates (2,10). Endoscopic obliteration of the fistulous tract using electrocoagulation, a clip, or a suturing device has been reported to be successful (9,11– 13). To our knowledge, there have been only a few reports describing the placement of covered retrievable expandable metallic stents for the treatment of EPF (14). The purpose of this study is to investigate the clinical efficacy of treating EPF with covered retrievable expandable metallic stents.
M/69
’
JVIR
ca. ¼ cancer; EPF ¼ esophagopleural fistula; Eso. ¼ esophageal; Explo. ¼ exploratory; LE ¼ lower esophagus; Lt. ¼ left; M ¼ male; ME ¼ midesophagus; NA ¼ not available; Postop. ¼ postoperative; Rt. ¼ right; RT ¼ radiation therapy; SP ¼ stent placement; UE ¼ upper esophagus; ↓ ¼ decreased; ↑ ¼ increased. *Empyema developed after primary esophageal repair for Boerhaave syndrome.
NA Empyema, EPF tract Thoracoplasty for empyema LE—Lt. lung
NA
9
Postop. empyema*
NA Hydropneumothorax, EPF tract
NA Chemotherapy and RT; explo. thoracotomy UE—Rt. lung
M/53 M/55 7 8
Eso. ca. M/55 6
(left lung)
NA Pneumonia ↑ Hydropneumothorax, esophagopulmonary fistula Aspiration pneumonia Chemotherapy and RT Chemotherapy and RT; explo. thoracotomy ME—Rt. lung ME—Rt. lung M/58 M/55 4 5
Eso. ca. Eso. ca.
Pleural effusion, EPF tract Chemotherapy and RT LE—Rt. lung M/79 3
Eso. ca.
Hydropneumothorax, EPF tract Unremarkable
Before SP Treatment before SP
Chemotherapy and RT Chemotherapy and RT LE—Lt. lung UE—Rt. lung
Fistula Site Diagnosis or Cause
Eso. ca. Eso. ca. M/76 M/44
Sex/Age (y)
1 2
No.
Table 1 . Baseline Clinical and Radiologic Information
Baseline characteristics of the patients are detailed in Table 1. From September 1997 to June 2013, 91 patients underwent esophageal or airway stent placement for esophagorespiratory fistula; 9 of these patients were treated using covered retrievable expandable metallic stents for EPF. All nine patients were men with a mean age of 60 years (range, 44–79 y). None of the patients was considered a surgical candidate at the time of stent placement. The diagnosis of EPF was established with esophagography or computed tomography (CT) examination in all patients. The diagnostic esophagogram finding was a direct communication between the esophagus and pleura without visualization of the trachea or bronchial trees. The diagnostic CT scan finding was a direct communication between the esophagus and pleural space. The underlying causes of EPF were esophageal carcinoma with or without exploratory thoracotomy (n ¼ 6), non–small cell lung cancer with right pneumonectomy or right lower lobectomy (n ¼ 2), and open thoracoplasty for chronic empyema from Boerhaave syndrome (n ¼ 1). Chemotherapy with (n ¼ 7) or without (n ¼ 1) radiation therapy was performed in all eight patients with a diagnosis of malignancy. The diagnosis of malignancy was established using endoscopic biopsy in all eight patients. The location of the EPF was evaluated on esophagography before stent placement; fistulas were located in the upper (n ¼ 2), middle (n ¼ 3), or lower thoracic (n ¼ 4) esophagus. The EPF was to the right and left pleura in seven and two patients, respectively. Patients were evaluated for symptoms of dysphagia before and after stent placement. The grade of dysphagia at the time of clinical presentation was defined using a 0–4 grading system and with a previously published grading system (15): grade 0, normal swallowing; grade 1, ability to swallow semisolids; grade 2, ability to swallow soft foods; grade 3, ability to swallow liquids only; and grade 4, absolute dysphagia. All patients reported coughing while swallowing food or saliva. Dysphagia was grade 4 in six patients, grade 3 in one patient, and grade 2 in two patients. The average dysphagia score was 3.44. Seven patients underwent contrast-enhanced CT before stent placement, and CT scan abnormalities were seen in six patients (Table 1). In one (patient no. 4) of the three patients with hydropneumothorax, there was another esophagopulmonary fistula with a resulting lung
CT Findings
Patient Population
(chest tube), Pneumonia
Effusion ↓
After SP
This retrospective study was performed with the approval of our institutional review board. Written informed consent was obtained from all patients at the time of their enrollment.
NA Unremarkable
MATERIALS AND METHODS
NA Hydropneumothorax ↓
Kim et al
Chemotherapy and RT; Rt. lower lobectomy Chemotherapy; Rt. pneumonectomy
Covered Retrievable Expandable Metallic Stent to Treat EPF
ME—Rt. lung LE—Rt. lung
’
Lung ca. Lung ca.
624
Volume 25
’
Number 4
’
April
’
2014
abscess. The EPF tract was visible in four patients (patient nos. 1, 3, 8, 9).
Techniques of Stent Placement We used polyurethane-covered or polytetrafluoroethylene (PTFE)-covered self-expandable esophageal stents (Niti-S esophageal covered stent; Taewoong Medical, Ilsan, Korea). PTFE was used as a covered material beginning in 2001; we have used a polyurethane-covered stent in five patients and a PTFE-covered stent in four patients. A stent at least 4 cm longer than the stricture was placed so that its proximal and distal parts rested on the upper and lower margins of the stricture as well as on the fistula opening (15). The body of each stent was 18 mm in diameter and 60–120 mm long when fully expanded. Both ends of the stents were 4–6 mm wider in diameter than the body of the stents to prevent stent migration. Before the procedure, a topical anesthetic, lidocaine hydrochloride (Dai Han Scientific Co, Seoul, Korea), was administered to the pharynx via aerosol spray. Neither sedatives nor general anesthesia was used. Under fluoroscopic guidance, a 0.035-inch, angled exchange guide wire (Radifocus Guide wire M; Terumo, Tokyo, Japan) was inserted through the patient’s mouth, across the stricture and fistula, and into the stomach. After measuring the stricture length, a stent was placed so that its proximal and distal parts rested on the upper and lower margins of the stricture and on the fistula. Stents were removed if stent-related complications occurred. The nylon drawstrings attached to the stent were grasped with the hook wire, the hook wire was withdrawn into the sheath to collapse the proximal stent, and the entire assembly was removed together. An esophagogram was obtained immediately after stent placement. If the stent completely sealed the fistula, the patient was allowed to consume a liquid diet 1 hour after the procedure and was encouraged to resume a tolerable diet gradually. If the esophagogram showed persistent leakage through the fistula and that resulted from incomplete stent expansion, a follow-up esophagogram was obtained 1–3 days after stent placement to verify the stent expansion before food intake was allowed.
Follow-up An esophagogram or a CT scan was obtained 3–7 days after stent placement to evaluate the closure of the fistula and again every 1 or 2 months until the patient died or whenever stent-related problems were suspected. Followup information was obtained from each patient, either in the outpatient clinic or by telephone interviews, concerning possible coughing while swallowing, dysphagia, or complications such stent migration or pain. Technical success was defined as accurate stent placement in the target area so as to cover the fistula or
625
stricture or both immediately after stent placement. Clinical success was defined as complete closure of the fistula and with dysphagia improvement within 7 days after stent placement. Clinical failure was defined as incomplete closure of the fistula or reopening of a completely closed fistula and without dysphagia improvement within 7 days after stent placement. Dysphagia was evaluated 1 week after stent placement in patients with clinical success because only the patients with complete fistula closure were encouraged to swallow. Placement of an additional stent or percutaneous feeding gastrostomy was performed in cases of clinical failure or fistula reopening during follow-up. Final information regarding patient survival and the cause of a patient’s death was obtained.
RESULTS Stent placement outcomes are summarized in Table 2. Technical success was achieved in eight of nine patients (88.9%) using a single stent. In one patient with technical failure (patient no. 1), the esophagogram obtained immediately after stent placement demonstrated fistula persistence (ie, incomplete closure of the EPF), probably owing to incomplete expansion of the stent. The stent was removed 2 days later, and a second stent of the same size was reinserted at the same location to solve the problem. Esophagograms obtained within 7 days after stent placement showed complete occlusion of the EPF and with no further symptoms of aspiration, achieving clinical success in five patients (55.6%) (Fig 1a–f). The remaining four patients (44.4%) had persistent aspiration secondary to fistula persistence (n ¼ 1) or fistula reopening (n ¼ 3) within 7 days after successful stent placement (ie, clinical failure). In three patients with fistula reopening, the fistula was caused by contrast material passage through the gap between the apparently well-expanded stent and the esophagus in two patients (patient nos. 6, 8) (Fig 2a–f) and by stent migration in one patient (patient no. 9). In one patient (patient no. 5), fistula reopening developed 15 days after stent placement and was caused by contrast material passage through the gap between the apparently wellexpanded stent and the esophagus. The overall fistula persistence or reopening occurred in five patients (55.6%) 0–15 days after stent placement. Four of five patients with clinical success were evaluated for dysphagia 1 week after stent placement. The average dysphagia score in these four patients improved from 3.3 to 1.5 after stent placement. Follow-up CT scans were available in three of the six patients in whom an abnormality was demonstrated on CT scans before stent placement (Table 1). A complication occurred in one (patient no. 9) of the nine patients (ie, stent migration occurred 1 day after
eso. ¼ esophageal; NA ¼ not available; PTFE ¼ polytetrafluoroethylene; PU ¼ polyurethane; SP ¼ stent placement; SR ¼ stent removal.
PTFE 9
gastrostomy
1 d, stent migration/spontaneous stent passage 567 d: alive with thoracocutaneous fistula NA 4 Failure Success
3 d/myocardial infarction 372 d/ Disease progression None 2 d/SR and eso. diversion with surgical 4 4 Success Failure Success Success 18 mm, 8 cm 18 mm, 12 cm PU PTFE 7 8
18 mm, 6 cm
63 d/pneumonia 30 d/pneumonia 15 d/gastrostomy 5 d/gastrostomy
NA NA
21 d/disease progression None 2
2 NA 2 4
3 Success
Success Failure Success Success
Success
PU PU 5 6
18 mm, 12 cm 18 mm, 10 cm
PTFE 4
18 mm, 10 cm
64 d/disease progression 15 d/pneumonia initial SP
None None 1 1 4 4 Success Success Success Success PU PTFE
18 mm, 10 cm 18 mm, 12 cm
62 d/cachexia
2 3
Fistula Reopening/Management
Persistent fistula/SR and 2nd SP 2 d after NA 2 Failure Failure 18 mm, 10 cm
Dysphagia score Stent
No. Covering Diameter, Length Technical Success Clinical Result Before SP After SP
PU
Survival Days/Cause of Death
Covered Retrievable Expandable Metallic Stent to Treat EPF
1
’
Table 2 . Clinical Outcome after Stent Placement
626
Kim et al
’
JVIR
stent placement and caused fistula reopening). The migrated stent, which was 18 mm in diameter and 6 cm long, spontaneously passed through this patient’s rectum 5 days after stent placement. This patient had a thoracocutaneous fistula, which had developed before stent placement. Although exclusion of the EPF failed, he has remained alive as a result of percutaneous jejunostomy, which had been performed before stent placement. Four of the five patients with fistula persistence or reopening underwent percutaneous gastrostomy (n ¼ 2, patient nos. 5, 6), stent removal and reinsertion of the same size 18-mm-diameter stent at the same site (n ¼ 1, patient no. 1), or stent removal and esophageal diversion with surgical gastrostomy (n ¼ 1, patient no. 8) to ensure enteral nutritional support. Eight patients died during the follow-up period (mean survival, 78.8 d ⫾ 120.9; median survival, 46 d; range, 3–372 d). Causes of death included pneumonia (n ¼ 3), disease progression (n ¼ 3), myocardial infarction (n ¼ 1), and cachexia (n ¼ 1).
DISCUSSION Because of the close anatomic relationship of the esophagus and pleura, abnormal communications or fistulas may develop between these structures secondary to various benign and malignant processes. Early diagnosis is important because of the life-threatening consequences of these fistulas (4). EPF is an unusual but potentially adverse complication observed in 0.5%–0.65% of patients undergoing pneumonectomy (2,10,12). Anatomically, EPF occurs more frequently after a right pneumonectomy because the thoracic esophagus is directly covered by the parietal pleura on the right, whereas on the left the thoracic aorta is interposed between the mediastinal pleura and the esophagus (7). We treated five patients who developed EPF after chest surgery (ie, two patients with right pneumonectomy or right lower lobectomy, two with exploratory thoracotomy, and one with open thoracoplasty performed for empyema of the left lung). In this study, the predisposition of the fistula between the esophagus and right pleura seems comparable to the predilection for EPF after right pneumonectomy. Occasional reports of EPF after radiation are described in the literature (7,16). In this study, all eight patients with malignancy underwent chemoradiation (n = 7) or chemotherapy (n = 1), which contributed to the development of EPF because malacia and ulceration induced by radiation therapy and chemotherapy can produce EPF. Because many patients with EPF are not ideal surgical candidates owing to advanced disease and poor performance status, less invasive methods to achieve fistula closure are preferable. Several clinical investigators have reported success after electrocoagulation; clips; suturing
Volume 25
’
Number 4
’
April
’
2014
627
Figure 1. A 79-year-old man (patient no. 3) with EPF caused by esophageal cancer. (a) Chest CT image shows esophageal thickening (arrows) with a fistulous tract (arrowheads) between the esophagus and the right pleural space and a large amount of bilateral pleural effusion. (b) Esophagogram obtained at the time of stent placement shows fistulous tract (arrows) from the esophagus to the right pleural space (arrowhead). (c) An 18-mm-diameter, 12-cm PTFE-covered esophageal stent was used. (d, e) Esophagograms obtained immediately (c) and 2 days (d) after esophageal stent placement show gradual stent expansion (arrows) and occluded fistula. (f) Chest CT image obtained 10 days after stent placement shows the tight apposition of the stent (arrows) and the esophagus. Right pleural effusion decreased after stent placement, although the right minor fissure effusion increased.
device; or combination of vascular plug, coils, and n-butyl cyanoacrylate (Histoacryl; B. Braun, Sempach, Switzerland) (11,13,17). However, these interventional procedures are not always possible in patients with a large fistula or short fistulous tract because a large fistula could not be covered with clips or suturing device and a short fistulous tract is difficult to close with embolic materials. Covered expandable metallic stents are considered a safe and effective treatment for various types of malignant esophagorespiratory fistulas, including esophagotracheal, esophagobronchial, and esophagopulmonary fistulas (14,15,18). There are only a few published
reports describing the use of covered metallic stents for the treatment of EPF. Kang et al (14) reported a successful case of temporary esophageal stent placement for a 12-mm-long fistula. They implanted an 18-mm-diameter, covered retrievable metallic stent to treat a benign spontaneous EPF and removed the stent endoscopically 3 months later. The fistula seemed to be well remodeled and obliterated based on the 6-month follow-up endoscopy. However, in this study, clinical success was achieved in only five of the nine patients (55.6%) owing to frequent fistula persistence or reopening. The fistula
628
’
Covered Retrievable Expandable Metallic Stent to Treat EPF
Kim et al
’
JVIR
Figure 2. A 55-year-old man (patient no. 8) with EPF caused by right pneumonectomy performed to treat lung cancer. (a) Chest CT image shows a fistulous tract (arrow) between the esophagus and the right pneumonectomy space. (b, c) Anteroposterior (b) and lateral (c) views of esophagograms obtained immediately after esophageal stent placement (18 mm diameter, 12 cm in length) show complete closure of the fistula. (d) Esophagogram obtained 2 days after esophageal stent placement shows reopening of the fistula (arrows). Contrast material passed through the gap between the esophagus and the apparently well-expanded stent. (e) Chest CT image obtained immediately after the esophagogram (ie, 2 days after esophageal stent placement) shows that contrast material remained in both the right pneumonectomy space and the stent lumen. (f) After the stent was removed using a retrievable hook and under fluoroscopic guidance, esophageal diversion with surgical gastrostomy was performed. The collapsed proximal portion (arrows) of the stent was seen during stent removal.
reopening was caused by the gap between the stent and the esophagus or by stent migration. Incomplete covering of a fistula secondary to the gap between the stent and the esophagus or secondary to stent migration is also a common cause of fistula reopening in esophagotracheal, esophagobronchial, or esophagopulmonary fistulas (15,18). Looseness of the esophagus surrounding
the stent, possibly attributed to past history such as exploratory thoracostomy, pneumonectomy, or thoracoplasty, may have resulted in failure of approximation between the stent and esophageal wall and, ultimately, in sealing of the fistula. Chemotherapy and radiation therapy might contribute to the weakness of the esophagus and the soft tissue around the fistula and the stent.
Volume 25
’
Number 4
’
April
’
2014
Although the clinical success rate in the present study was not high, the dysphagia scores decreased in patients with clinical success, and secondary complications of respiratory diseases could be prevented. For fistula persistence or reopening, interventional management using percutaneous feeding gastrostomy or stent removal with secondary stent placement or surgical gastrostomy was useful to allow oral intake. The deaths of eight patients during this study were caused by malignancies, and mean survival time (ie, 78.8 d), was shorter than the 93.8 days seen in patients with esophagotracheal or esophagobronchial fistulas (15) or the 100.9 days seen in patients with esophagopulmonary fistula (18). The high clinical failure rate in our study might explain the worse patient prognosis. Failure to control pulmonary contamination can decrease the ability to achieve palliation and reduce patient survival and quality of life (15,19). This worse prognosis can also be attributed to the higher proportion of study patients who underwent chemoradiation or surgery or both. For patients with simple esophageal strictures, a 16-mmdiameter stent is generally used (20), but an 18-mmdiameter stent could enhance the sealing effect in patients with various types of esophagorespiratory fistula (14,15,18). Nevertheless, in our study, fistula persistence or the reopening rate was high when using 18-mm-diameter stents, and further research using a stent 4 18 mm in diameter is necessary for determining the successful treatment of EPF. The covered retrievable expandable metallic stents used in this study had the advantage of allowing stent removal when a stent-related complication occurred or if elective removal was necessary. In a report by Kang et al (14), temporary esophageal stent placement was possible secondary to being able to pull the removal snare with an endoscopic forceps in a patient with benign EPF. In our clinical experience, two stents were removed when the fistula persisted or reopened, and secondary stent placement was achieved for one of the patients. The present study has some limitations. First, the retrospective nature presented some inherent limitations, including limited availability of data and lack of a common protocol with respect to management and follow-up. Second, the number of patients in this series is small. Third, the use of two types of esophageal stents could be a potential confounder of the study results. In conclusion, placement of a covered expandable metallic esophageal stent for the palliative treatment of EPF is technically feasible, although our clinical success rate was poor secondary to fistula persistence or reopening. Fistula reopening was caused by the gap between the
629
stent and the esophagus or by stent migration, and additional interventional treatment was useful to ensure enteral nutritional support.
REFERENCES 1. Liu PS, Levine MS, Torigian DA. Esophagopleural fistula secondary to esophageal wall ballooning and thinning after pneumonectomy: findings on chest CT and esophagography. AJR Am J Roentgenol 2006; 186:1627–1629. 2. Sethi GK, Takaro T. Esophagopleural fistula following pulmonary resection. Ann Thorac Surg 1978; 25:74–81. 3. Agarwal V, Singh SK, Siddiqi MS, Joshi LM, Tandon S. Esophagopleural fistula following spontaneous rupture of traction diverticulum. Asian Cardiovasc Thorac Ann 2003; 11:344–345. 4. Gimenez A, Franquet T, Erasmus JJ, Martinez S, Estrada P. Thoracic complications of esophageal disorders. Radiographics 2002; 22:S247–S258. 5. Guerra JM, Zuil M, Garcia I, Moreno E. Epiphrenic diverticula, esophageal carcinoma and esophagopleural fistula. Hepatogastroenterology 2001; 48:718–719. 6. Inoue M, Nakanishi R, Osaki T, Yoshimatsu T, Yasumoto K. Esophagopleural fistula originating from diverticulum after pneumonectomy: a case report and review of the literature. J Cardiovasc Surg (Torino) 1999; 40:761–763. 7. Oliaro A, Filosso PL, Casadio C, et al. Right post pneumonectomy pleural empyema caused by an esophagopleural fistula due to an esophageal carcinoma. J Cardiovasc Surg (Torino) 1995; 36:607–609. 8. Prasad H, Poddar U, Thapa BR, Narashiman KL, Singh K. Esophagopleural fistula after endoscopic sclerotherapy in a child. Gastrointest Endosc 2000; 52:804–806. 9. Dosios T, Karavokyros I, Felekouras E, et al. Presternal gastric bypass for late postpneumonectomy esophagopleural fistula. Dis Esophagus 2005; 18:202–203. 10. Massard G, Ducrocq X, Hentz JG, et al. Esophagopleural fistula: an early and long-term complication after pneumonectomy. Ann Thorac Surg 1994; 58:1437–1440, discussion 1441. 11. Adler DG, McAfee M, Gostout CJ. Closure of an esophagopleural fistula by using fistula tract coagulation and an endoscopic suturing device. Gastrointest Endosc 2001; 54:652–653. 12. Massard G, Wihlm JM. Early complications. Esophagopleural fistula. Chest Surg Clin N Am 1999; 9:617–631. 13. van Bodegraven AA, Kuipers EJ, Bonenkamp HJ, Meuwissen SG. Esophagopleural fistula treated endoscopically with argon beam electrocoagulation and clips. Gastrointest Endosc 1999; 50:407–409. 14. Kang GH, Yoon BY, Kim BH, et al. A case of spontaneous esophagopleural fistula successfully treated by endoscopic stent insertion. Clin Endosc 2013; 46:91–94. 15. Shin JH, Song HY, Ko GY, Lim JO, Yoon HK, Sung KB. Esophagorespiratory fistula: long-term results of palliative treatment with covered expandable metallic stents in 61 patients. Radiology 2004; 232:252–259. 16. Seaman WB, Ackerman LV. The effect of radiation on the esophagus: a clinical and histologic study of the effects produced by the betatron. Radiology 1957; 68:534–541. 17. Koo JH, Park KB, Choo SW, Kim K, Do YS. Embolization of postsurgical esophagopleural fistula with AMPLATZER vascular plug, coils, and Histoacryl glue. J Vasc Interv Radiol 2010; 21:1905–1910. 18. Kim KR, Shin JH, Song HY, et al. Palliative treatment of malignant esophagopulmonary fistulas with covered expandable metallic stents. AJR Am J Roentgenol 2009; 193:W278–W282. 19. Burt M, Diehl W, Martini N, et al. Malignant esophagorespiratory fistula: management options and survival. Ann Thorac Surg 1991; 52: 1222–1228, discussion 1228–1229. 20. Park JH, Song HY, Kim JH, Jung HY, Kim SB, Lee H. Polytetrafluoroethylene-covered retrievable expandable nitinol stents for malignant esophageal obstructions: factors influencing the outcome of 270 patients. AJR Am J Roentgenol 2012; 199:1380–1386.
Treatment of Esophagopleural Fistulas Using Covered Retrievable Expandable Metallic Stents Tae-Hyung Kim, PhD, Ji Hoon Shin, MD, Kyung Rae Kim, MD, Jung-Hoon Park, RT, Jin Hyoung Kim, MD, and Ho-Young Song, MD
ABSTRACT Purpose: To evaluate the clinical efficacy of placement of covered retrievable expandable metallic stents for esophagopleural fistulas (EPFs). Materials and Methods: During the period 1997–2013, nine patients with EPF were treated using covered retrievable expandable metallic stents. The underlying causes of EPF were esophageal carcinoma (n ¼ 6), lung cancer (n ¼ 2), and postoperative empyema for Boerhaave syndrome (n ¼ 1). Results: Technical success was achieved in eight patients (88.9%). In one patient, incomplete EPF closure was due to incomplete stent expansion. Clinical success, defined as complete EPF closure within 7 days, was achieved in five patients (55.6%). Overall fistula persistence (n ¼ 1) or reopening (n ¼ 4) occurred in five patients (55.6%) 0–15 days after stent placement. The causes of reopening were due to the gap between the stent and the esophagus (n ¼ 3) or stent migration (n ¼ 1). For fistula persistence or reopening, additional interventional management, such as gastrostomy, stent removal, or stent reinsertion, was performed. Stent migration occurred as a complication in one patient with EPF from a benign cause secondary to postoperative empyema. In the eight patients who died during the follow-up period, the mean and median survival times were 78.8 days and 46 days, respectively. Conclusions: Placement of a covered expandable metallic esophageal stent for the palliative treatment of EPF is technically feasible, although the rate of clinical success was poor secondary to fistula persistence or reopening. Fistula reopening was caused by the gap between the stent and the esophagus or by stent migration, and additional interventional treatment was useful to ensure enteral nutritional support.
ABBREVIATIONS EPF = esophagopleural fistula, PTFE = polytetrafluoroethylene
Esophagopleural fistula (EPF) is very rare and is associated with high morbidity and mortality because of ensuing empyema and nutritional debilitation (1,2). EPF is an uncommon complication of pneumonectomy and is associated with advanced esophageal carcinoma, tuberculosis, diverticulum, Boerhaave syndrome, surgical procedures,
From the Department of Radiological Science (T.-H.K.), College of Health Science, Kangwon National University, Gangwon Province, South Korea; Department of Radiology and Research Institute of Radiology (J.H.S., J.-H.P., J.H.K., H.-Y.S.), University of Ulsan College of Medicine, Asan Medical Center, 388-1, Pungnap-2dong, Songpa-gu, Seoul 138-736, Korea; and Department of Radiology (K.R.K.), University of North Carolina at Chapel Hill, Chapel Hill, North Carolina. Received October 7, 2013; final revision received December 13, 2013; accepted December 15, 2013. Address correspondence to J.H.S.; E-mail: [email protected] & SIR, 2014 J Vasc Interv Radiol 2014; 25:623–629 http://dx.doi.org/10.1016/j.jvir.2013.12.015
endoscopic examinations, chemical injury after ingestion of corrosive substances, and radiation therapy (1–8). Treatment of EPF begins with control of the empyema, repair of the esophagus, and nutritional support (9,10). Eradication of empyema by direct repair combined with a cover of viable tissue and thoracoplasty should be considered as the treatment of choice; however, because of its invasiveness, it is associated with high mortality rates (2,10). Endoscopic obliteration of the fistulous tract using electrocoagulation, a clip, or a suturing device has been reported to be successful (9,11– 13). To our knowledge, there have been only a few reports describing the placement of covered retrievable expandable metallic stents for the treatment of EPF (14). The purpose of this study is to investigate the clinical efficacy of treating EPF with covered retrievable expandable metallic stents.
M/69
’
JVIR
ca. ¼ cancer; EPF ¼ esophagopleural fistula; Eso. ¼ esophageal; Explo. ¼ exploratory; LE ¼ lower esophagus; Lt. ¼ left; M ¼ male; ME ¼ midesophagus; NA ¼ not available; Postop. ¼ postoperative; Rt. ¼ right; RT ¼ radiation therapy; SP ¼ stent placement; UE ¼ upper esophagus; ↓ ¼ decreased; ↑ ¼ increased. *Empyema developed after primary esophageal repair for Boerhaave syndrome.
NA Empyema, EPF tract Thoracoplasty for empyema LE—Lt. lung
NA
9
Postop. empyema*
NA Hydropneumothorax, EPF tract
NA Chemotherapy and RT; explo. thoracotomy UE—Rt. lung
M/53 M/55 7 8
Eso. ca. M/55 6
(left lung)
NA Pneumonia ↑ Hydropneumothorax, esophagopulmonary fistula Aspiration pneumonia Chemotherapy and RT Chemotherapy and RT; explo. thoracotomy ME—Rt. lung ME—Rt. lung M/58 M/55 4 5
Eso. ca. Eso. ca.
Pleural effusion, EPF tract Chemotherapy and RT LE—Rt. lung M/79 3
Eso. ca.
Hydropneumothorax, EPF tract Unremarkable
Before SP Treatment before SP
Chemotherapy and RT Chemotherapy and RT LE—Lt. lung UE—Rt. lung
Fistula Site Diagnosis or Cause
Eso. ca. Eso. ca. M/76 M/44
Sex/Age (y)
1 2
No.
Table 1 . Baseline Clinical and Radiologic Information
Baseline characteristics of the patients are detailed in Table 1. From September 1997 to June 2013, 91 patients underwent esophageal or airway stent placement for esophagorespiratory fistula; 9 of these patients were treated using covered retrievable expandable metallic stents for EPF. All nine patients were men with a mean age of 60 years (range, 44–79 y). None of the patients was considered a surgical candidate at the time of stent placement. The diagnosis of EPF was established with esophagography or computed tomography (CT) examination in all patients. The diagnostic esophagogram finding was a direct communication between the esophagus and pleura without visualization of the trachea or bronchial trees. The diagnostic CT scan finding was a direct communication between the esophagus and pleural space. The underlying causes of EPF were esophageal carcinoma with or without exploratory thoracotomy (n ¼ 6), non–small cell lung cancer with right pneumonectomy or right lower lobectomy (n ¼ 2), and open thoracoplasty for chronic empyema from Boerhaave syndrome (n ¼ 1). Chemotherapy with (n ¼ 7) or without (n ¼ 1) radiation therapy was performed in all eight patients with a diagnosis of malignancy. The diagnosis of malignancy was established using endoscopic biopsy in all eight patients. The location of the EPF was evaluated on esophagography before stent placement; fistulas were located in the upper (n ¼ 2), middle (n ¼ 3), or lower thoracic (n ¼ 4) esophagus. The EPF was to the right and left pleura in seven and two patients, respectively. Patients were evaluated for symptoms of dysphagia before and after stent placement. The grade of dysphagia at the time of clinical presentation was defined using a 0–4 grading system and with a previously published grading system (15): grade 0, normal swallowing; grade 1, ability to swallow semisolids; grade 2, ability to swallow soft foods; grade 3, ability to swallow liquids only; and grade 4, absolute dysphagia. All patients reported coughing while swallowing food or saliva. Dysphagia was grade 4 in six patients, grade 3 in one patient, and grade 2 in two patients. The average dysphagia score was 3.44. Seven patients underwent contrast-enhanced CT before stent placement, and CT scan abnormalities were seen in six patients (Table 1). In one (patient no. 4) of the three patients with hydropneumothorax, there was another esophagopulmonary fistula with a resulting lung
CT Findings
Patient Population
(chest tube), Pneumonia
Effusion ↓
After SP
This retrospective study was performed with the approval of our institutional review board. Written informed consent was obtained from all patients at the time of their enrollment.
NA Unremarkable
MATERIALS AND METHODS
NA Hydropneumothorax ↓
Kim et al
Chemotherapy and RT; Rt. lower lobectomy Chemotherapy; Rt. pneumonectomy
Covered Retrievable Expandable Metallic Stent to Treat EPF
ME—Rt. lung LE—Rt. lung
’
Lung ca. Lung ca.
624
Volume 25
’
Number 4
’
April
’
2014
abscess. The EPF tract was visible in four patients (patient nos. 1, 3, 8, 9).
Techniques of Stent Placement We used polyurethane-covered or polytetrafluoroethylene (PTFE)-covered self-expandable esophageal stents (Niti-S esophageal covered stent; Taewoong Medical, Ilsan, Korea). PTFE was used as a covered material beginning in 2001; we have used a polyurethane-covered stent in five patients and a PTFE-covered stent in four patients. A stent at least 4 cm longer than the stricture was placed so that its proximal and distal parts rested on the upper and lower margins of the stricture as well as on the fistula opening (15). The body of each stent was 18 mm in diameter and 60–120 mm long when fully expanded. Both ends of the stents were 4–6 mm wider in diameter than the body of the stents to prevent stent migration. Before the procedure, a topical anesthetic, lidocaine hydrochloride (Dai Han Scientific Co, Seoul, Korea), was administered to the pharynx via aerosol spray. Neither sedatives nor general anesthesia was used. Under fluoroscopic guidance, a 0.035-inch, angled exchange guide wire (Radifocus Guide wire M; Terumo, Tokyo, Japan) was inserted through the patient’s mouth, across the stricture and fistula, and into the stomach. After measuring the stricture length, a stent was placed so that its proximal and distal parts rested on the upper and lower margins of the stricture and on the fistula. Stents were removed if stent-related complications occurred. The nylon drawstrings attached to the stent were grasped with the hook wire, the hook wire was withdrawn into the sheath to collapse the proximal stent, and the entire assembly was removed together. An esophagogram was obtained immediately after stent placement. If the stent completely sealed the fistula, the patient was allowed to consume a liquid diet 1 hour after the procedure and was encouraged to resume a tolerable diet gradually. If the esophagogram showed persistent leakage through the fistula and that resulted from incomplete stent expansion, a follow-up esophagogram was obtained 1–3 days after stent placement to verify the stent expansion before food intake was allowed.
Follow-up An esophagogram or a CT scan was obtained 3–7 days after stent placement to evaluate the closure of the fistula and again every 1 or 2 months until the patient died or whenever stent-related problems were suspected. Followup information was obtained from each patient, either in the outpatient clinic or by telephone interviews, concerning possible coughing while swallowing, dysphagia, or complications such stent migration or pain. Technical success was defined as accurate stent placement in the target area so as to cover the fistula or
625
stricture or both immediately after stent placement. Clinical success was defined as complete closure of the fistula and with dysphagia improvement within 7 days after stent placement. Clinical failure was defined as incomplete closure of the fistula or reopening of a completely closed fistula and without dysphagia improvement within 7 days after stent placement. Dysphagia was evaluated 1 week after stent placement in patients with clinical success because only the patients with complete fistula closure were encouraged to swallow. Placement of an additional stent or percutaneous feeding gastrostomy was performed in cases of clinical failure or fistula reopening during follow-up. Final information regarding patient survival and the cause of a patient’s death was obtained.
RESULTS Stent placement outcomes are summarized in Table 2. Technical success was achieved in eight of nine patients (88.9%) using a single stent. In one patient with technical failure (patient no. 1), the esophagogram obtained immediately after stent placement demonstrated fistula persistence (ie, incomplete closure of the EPF), probably owing to incomplete expansion of the stent. The stent was removed 2 days later, and a second stent of the same size was reinserted at the same location to solve the problem. Esophagograms obtained within 7 days after stent placement showed complete occlusion of the EPF and with no further symptoms of aspiration, achieving clinical success in five patients (55.6%) (Fig 1a–f). The remaining four patients (44.4%) had persistent aspiration secondary to fistula persistence (n ¼ 1) or fistula reopening (n ¼ 3) within 7 days after successful stent placement (ie, clinical failure). In three patients with fistula reopening, the fistula was caused by contrast material passage through the gap between the apparently well-expanded stent and the esophagus in two patients (patient nos. 6, 8) (Fig 2a–f) and by stent migration in one patient (patient no. 9). In one patient (patient no. 5), fistula reopening developed 15 days after stent placement and was caused by contrast material passage through the gap between the apparently wellexpanded stent and the esophagus. The overall fistula persistence or reopening occurred in five patients (55.6%) 0–15 days after stent placement. Four of five patients with clinical success were evaluated for dysphagia 1 week after stent placement. The average dysphagia score in these four patients improved from 3.3 to 1.5 after stent placement. Follow-up CT scans were available in three of the six patients in whom an abnormality was demonstrated on CT scans before stent placement (Table 1). A complication occurred in one (patient no. 9) of the nine patients (ie, stent migration occurred 1 day after
eso. ¼ esophageal; NA ¼ not available; PTFE ¼ polytetrafluoroethylene; PU ¼ polyurethane; SP ¼ stent placement; SR ¼ stent removal.
PTFE 9
gastrostomy
1 d, stent migration/spontaneous stent passage 567 d: alive with thoracocutaneous fistula NA 4 Failure Success
3 d/myocardial infarction 372 d/ Disease progression None 2 d/SR and eso. diversion with surgical 4 4 Success Failure Success Success 18 mm, 8 cm 18 mm, 12 cm PU PTFE 7 8
18 mm, 6 cm
63 d/pneumonia 30 d/pneumonia 15 d/gastrostomy 5 d/gastrostomy
NA NA
21 d/disease progression None 2
2 NA 2 4
3 Success
Success Failure Success Success
Success
PU PU 5 6
18 mm, 12 cm 18 mm, 10 cm
PTFE 4
18 mm, 10 cm
64 d/disease progression 15 d/pneumonia initial SP
None None 1 1 4 4 Success Success Success Success PU PTFE
18 mm, 10 cm 18 mm, 12 cm
62 d/cachexia
2 3
Fistula Reopening/Management
Persistent fistula/SR and 2nd SP 2 d after NA 2 Failure Failure 18 mm, 10 cm
Dysphagia score Stent
No. Covering Diameter, Length Technical Success Clinical Result Before SP After SP
PU
Survival Days/Cause of Death
Covered Retrievable Expandable Metallic Stent to Treat EPF
1
’
Table 2 . Clinical Outcome after Stent Placement
626
Kim et al
’
JVIR
stent placement and caused fistula reopening). The migrated stent, which was 18 mm in diameter and 6 cm long, spontaneously passed through this patient’s rectum 5 days after stent placement. This patient had a thoracocutaneous fistula, which had developed before stent placement. Although exclusion of the EPF failed, he has remained alive as a result of percutaneous jejunostomy, which had been performed before stent placement. Four of the five patients with fistula persistence or reopening underwent percutaneous gastrostomy (n ¼ 2, patient nos. 5, 6), stent removal and reinsertion of the same size 18-mm-diameter stent at the same site (n ¼ 1, patient no. 1), or stent removal and esophageal diversion with surgical gastrostomy (n ¼ 1, patient no. 8) to ensure enteral nutritional support. Eight patients died during the follow-up period (mean survival, 78.8 d ⫾ 120.9; median survival, 46 d; range, 3–372 d). Causes of death included pneumonia (n ¼ 3), disease progression (n ¼ 3), myocardial infarction (n ¼ 1), and cachexia (n ¼ 1).
DISCUSSION Because of the close anatomic relationship of the esophagus and pleura, abnormal communications or fistulas may develop between these structures secondary to various benign and malignant processes. Early diagnosis is important because of the life-threatening consequences of these fistulas (4). EPF is an unusual but potentially adverse complication observed in 0.5%–0.65% of patients undergoing pneumonectomy (2,10,12). Anatomically, EPF occurs more frequently after a right pneumonectomy because the thoracic esophagus is directly covered by the parietal pleura on the right, whereas on the left the thoracic aorta is interposed between the mediastinal pleura and the esophagus (7). We treated five patients who developed EPF after chest surgery (ie, two patients with right pneumonectomy or right lower lobectomy, two with exploratory thoracotomy, and one with open thoracoplasty performed for empyema of the left lung). In this study, the predisposition of the fistula between the esophagus and right pleura seems comparable to the predilection for EPF after right pneumonectomy. Occasional reports of EPF after radiation are described in the literature (7,16). In this study, all eight patients with malignancy underwent chemoradiation (n = 7) or chemotherapy (n = 1), which contributed to the development of EPF because malacia and ulceration induced by radiation therapy and chemotherapy can produce EPF. Because many patients with EPF are not ideal surgical candidates owing to advanced disease and poor performance status, less invasive methods to achieve fistula closure are preferable. Several clinical investigators have reported success after electrocoagulation; clips; suturing
Volume 25
’
Number 4
’
April
’
2014
627
Figure 1. A 79-year-old man (patient no. 3) with EPF caused by esophageal cancer. (a) Chest CT image shows esophageal thickening (arrows) with a fistulous tract (arrowheads) between the esophagus and the right pleural space and a large amount of bilateral pleural effusion. (b) Esophagogram obtained at the time of stent placement shows fistulous tract (arrows) from the esophagus to the right pleural space (arrowhead). (c) An 18-mm-diameter, 12-cm PTFE-covered esophageal stent was used. (d, e) Esophagograms obtained immediately (c) and 2 days (d) after esophageal stent placement show gradual stent expansion (arrows) and occluded fistula. (f) Chest CT image obtained 10 days after stent placement shows the tight apposition of the stent (arrows) and the esophagus. Right pleural effusion decreased after stent placement, although the right minor fissure effusion increased.
device; or combination of vascular plug, coils, and n-butyl cyanoacrylate (Histoacryl; B. Braun, Sempach, Switzerland) (11,13,17). However, these interventional procedures are not always possible in patients with a large fistula or short fistulous tract because a large fistula could not be covered with clips or suturing device and a short fistulous tract is difficult to close with embolic materials. Covered expandable metallic stents are considered a safe and effective treatment for various types of malignant esophagorespiratory fistulas, including esophagotracheal, esophagobronchial, and esophagopulmonary fistulas (14,15,18). There are only a few published
reports describing the use of covered metallic stents for the treatment of EPF. Kang et al (14) reported a successful case of temporary esophageal stent placement for a 12-mm-long fistula. They implanted an 18-mm-diameter, covered retrievable metallic stent to treat a benign spontaneous EPF and removed the stent endoscopically 3 months later. The fistula seemed to be well remodeled and obliterated based on the 6-month follow-up endoscopy. However, in this study, clinical success was achieved in only five of the nine patients (55.6%) owing to frequent fistula persistence or reopening. The fistula
628
’
Covered Retrievable Expandable Metallic Stent to Treat EPF
Kim et al
’
JVIR
Figure 2. A 55-year-old man (patient no. 8) with EPF caused by right pneumonectomy performed to treat lung cancer. (a) Chest CT image shows a fistulous tract (arrow) between the esophagus and the right pneumonectomy space. (b, c) Anteroposterior (b) and lateral (c) views of esophagograms obtained immediately after esophageal stent placement (18 mm diameter, 12 cm in length) show complete closure of the fistula. (d) Esophagogram obtained 2 days after esophageal stent placement shows reopening of the fistula (arrows). Contrast material passed through the gap between the esophagus and the apparently well-expanded stent. (e) Chest CT image obtained immediately after the esophagogram (ie, 2 days after esophageal stent placement) shows that contrast material remained in both the right pneumonectomy space and the stent lumen. (f) After the stent was removed using a retrievable hook and under fluoroscopic guidance, esophageal diversion with surgical gastrostomy was performed. The collapsed proximal portion (arrows) of the stent was seen during stent removal.
reopening was caused by the gap between the stent and the esophagus or by stent migration. Incomplete covering of a fistula secondary to the gap between the stent and the esophagus or secondary to stent migration is also a common cause of fistula reopening in esophagotracheal, esophagobronchial, or esophagopulmonary fistulas (15,18). Looseness of the esophagus surrounding
the stent, possibly attributed to past history such as exploratory thoracostomy, pneumonectomy, or thoracoplasty, may have resulted in failure of approximation between the stent and esophageal wall and, ultimately, in sealing of the fistula. Chemotherapy and radiation therapy might contribute to the weakness of the esophagus and the soft tissue around the fistula and the stent.
Volume 25
’
Number 4
’
April
’
2014
Although the clinical success rate in the present study was not high, the dysphagia scores decreased in patients with clinical success, and secondary complications of respiratory diseases could be prevented. For fistula persistence or reopening, interventional management using percutaneous feeding gastrostomy or stent removal with secondary stent placement or surgical gastrostomy was useful to allow oral intake. The deaths of eight patients during this study were caused by malignancies, and mean survival time (ie, 78.8 d), was shorter than the 93.8 days seen in patients with esophagotracheal or esophagobronchial fistulas (15) or the 100.9 days seen in patients with esophagopulmonary fistula (18). The high clinical failure rate in our study might explain the worse patient prognosis. Failure to control pulmonary contamination can decrease the ability to achieve palliation and reduce patient survival and quality of life (15,19). This worse prognosis can also be attributed to the higher proportion of study patients who underwent chemoradiation or surgery or both. For patients with simple esophageal strictures, a 16-mmdiameter stent is generally used (20), but an 18-mmdiameter stent could enhance the sealing effect in patients with various types of esophagorespiratory fistula (14,15,18). Nevertheless, in our study, fistula persistence or the reopening rate was high when using 18-mm-diameter stents, and further research using a stent 4 18 mm in diameter is necessary for determining the successful treatment of EPF. The covered retrievable expandable metallic stents used in this study had the advantage of allowing stent removal when a stent-related complication occurred or if elective removal was necessary. In a report by Kang et al (14), temporary esophageal stent placement was possible secondary to being able to pull the removal snare with an endoscopic forceps in a patient with benign EPF. In our clinical experience, two stents were removed when the fistula persisted or reopened, and secondary stent placement was achieved for one of the patients. The present study has some limitations. First, the retrospective nature presented some inherent limitations, including limited availability of data and lack of a common protocol with respect to management and follow-up. Second, the number of patients in this series is small. Third, the use of two types of esophageal stents could be a potential confounder of the study results. In conclusion, placement of a covered expandable metallic esophageal stent for the palliative treatment of EPF is technically feasible, although our clinical success rate was poor secondary to fistula persistence or reopening. Fistula reopening was caused by the gap between the
629
stent and the esophagus or by stent migration, and additional interventional treatment was useful to ensure enteral nutritional support.
REFERENCES 1. Liu PS, Levine MS, Torigian DA. Esophagopleural fistula secondary to esophageal wall ballooning and thinning after pneumonectomy: findings on chest CT and esophagography. AJR Am J Roentgenol 2006; 186:1627–1629. 2. Sethi GK, Takaro T. Esophagopleural fistula following pulmonary resection. Ann Thorac Surg 1978; 25:74–81. 3. Agarwal V, Singh SK, Siddiqi MS, Joshi LM, Tandon S. Esophagopleural fistula following spontaneous rupture of traction diverticulum. Asian Cardiovasc Thorac Ann 2003; 11:344–345. 4. Gimenez A, Franquet T, Erasmus JJ, Martinez S, Estrada P. Thoracic complications of esophageal disorders. Radiographics 2002; 22:S247–S258. 5. Guerra JM, Zuil M, Garcia I, Moreno E. Epiphrenic diverticula, esophageal carcinoma and esophagopleural fistula. Hepatogastroenterology 2001; 48:718–719. 6. Inoue M, Nakanishi R, Osaki T, Yoshimatsu T, Yasumoto K. Esophagopleural fistula originating from diverticulum after pneumonectomy: a case report and review of the literature. J Cardiovasc Surg (Torino) 1999; 40:761–763. 7. Oliaro A, Filosso PL, Casadio C, et al. Right post pneumonectomy pleural empyema caused by an esophagopleural fistula due to an esophageal carcinoma. J Cardiovasc Surg (Torino) 1995; 36:607–609. 8. Prasad H, Poddar U, Thapa BR, Narashiman KL, Singh K. Esophagopleural fistula after endoscopic sclerotherapy in a child. Gastrointest Endosc 2000; 52:804–806. 9. Dosios T, Karavokyros I, Felekouras E, et al. Presternal gastric bypass for late postpneumonectomy esophagopleural fistula. Dis Esophagus 2005; 18:202–203. 10. Massard G, Ducrocq X, Hentz JG, et al. Esophagopleural fistula: an early and long-term complication after pneumonectomy. Ann Thorac Surg 1994; 58:1437–1440, discussion 1441. 11. Adler DG, McAfee M, Gostout CJ. Closure of an esophagopleural fistula by using fistula tract coagulation and an endoscopic suturing device. Gastrointest Endosc 2001; 54:652–653. 12. Massard G, Wihlm JM. Early complications. Esophagopleural fistula. Chest Surg Clin N Am 1999; 9:617–631. 13. van Bodegraven AA, Kuipers EJ, Bonenkamp HJ, Meuwissen SG. Esophagopleural fistula treated endoscopically with argon beam electrocoagulation and clips. Gastrointest Endosc 1999; 50:407–409. 14. Kang GH, Yoon BY, Kim BH, et al. A case of spontaneous esophagopleural fistula successfully treated by endoscopic stent insertion. Clin Endosc 2013; 46:91–94. 15. Shin JH, Song HY, Ko GY, Lim JO, Yoon HK, Sung KB. Esophagorespiratory fistula: long-term results of palliative treatment with covered expandable metallic stents in 61 patients. Radiology 2004; 232:252–259. 16. Seaman WB, Ackerman LV. The effect of radiation on the esophagus: a clinical and histologic study of the effects produced by the betatron. Radiology 1957; 68:534–541. 17. Koo JH, Park KB, Choo SW, Kim K, Do YS. Embolization of postsurgical esophagopleural fistula with AMPLATZER vascular plug, coils, and Histoacryl glue. J Vasc Interv Radiol 2010; 21:1905–1910. 18. Kim KR, Shin JH, Song HY, et al. Palliative treatment of malignant esophagopulmonary fistulas with covered expandable metallic stents. AJR Am J Roentgenol 2009; 193:W278–W282. 19. Burt M, Diehl W, Martini N, et al. Malignant esophagorespiratory fistula: management options and survival. Ann Thorac Surg 1991; 52: 1222–1228, discussion 1228–1229. 20. Park JH, Song HY, Kim JH, Jung HY, Kim SB, Lee H. Polytetrafluoroethylene-covered retrievable expandable nitinol stents for malignant esophageal obstructions: factors influencing the outcome of 270 patients. AJR Am J Roentgenol 2012; 199:1380–1386.
