treatment of esophagorespiratory fistulas with a polyurethane-covered self-expanding metallic mesh...

0016-5107/95/4105-049053.00 + .0 GASTROINTESTINAL ENDOSCOPY Copyright �9 1995 by the American Society for Gastrointestinal Endoscopy

Treatment of esophagorespiratory fistulas with silicone-coated self-expanding metal stents Norbert Weigert, MD, Horst Neuhaus, MD, Thomas ROsch, MD Werner Hoffmann, MD, Hans J. Dittler, MD, Meinhard Classen, MD Munich, Germany

Nine self-expanding silicone-coated modified Gianturco metal stents were inserted in 8 patients (mean age, 58.2 years) for palliation of malignant esophagorespiratory fistulas caused by esophageal (n = 5) or bronchial (n = 3) carcinoma. One patient with a fistula above a 12-cm-long malignant stenosis received two overlapping stents. The implantation procedure was well tolerated by all patients under intravenous sedation and analgesia. After release, the stents expanded to their full diameter, leading to complete occlusion of the fistulas and bridging of the concomitant stenoses. Two patients with lung cancer received an additional tracheobronchial stent before esophageal stent insertion. Failure to maintain complete contact between the proximal stent margin and the esophageal wall led to insufficient sealing of the fistula of 1 patient and recurrent aspiration, manifested 6 days after stent implantation (overall success, 87.5%). The other patients could swallow semisolid food until death. Seven patients died of advanced metastatic disease after 21 to 121 days (mean, 54 days) and 1 patient of massive hemoptysis 10 days after stent placement, which could be regarded as a complication (procedure-related mortality rate, 12.5%). These preliminary results suggest that peroral insertion of the modified silicone-coated Gianturco stent is a rapid, reasonably safe, and effective procedure for palliation of malignant esophagorespiratory fistulas. (Gastrointest Endosc 1995;41:490-6.)

The incidence of esophagorespiratory fistulas as shown in a large survey was 4.9 % in 1943 patients with esophageal cancer, 0.16% in 5714 patients with lung cancer, and 14.75 % in 41 patients with tracheal cancer. 1 Curative resection of malignant esophagorespiratory fistulas is usually impossible because of advanced tumor stage, and palliative surgery (including a combination of cervical esophagostomy and feeding gastrostomy, bypass operations, and surgical intuba-

Received March 8, 1994. For revision April 11, 1994. Accepted July 6, 1994. From the Department o[ Internal Medicine H and the Department of Surgery, Technical University of Munich, Germany. Reprint requests: Horst Neuhaus, MD, Department of Internal Medicine II, Technical University of Munich, Ismaninger Strasse 22, 81675 Munich, Germany. 3 7 / 1 / 5 9 2 2 3

tion) is associated with a mortality rate up to 50 %. For these reasons, endoscopic intubation is nowadays considered the method of choice for the management of these patients, 2-7 as it offers rapid relief of symptoms in most cases, which is considered crucial in view of their poor prognosis and quality of life. However, implantation of conventional plastic prostheses is incon- venient for the patient and associated with a mortality rate of 2 % to 16 %. Furthermore, early complications, such as perforation of the esophagus, pressure necrosis, tube obstruction, tube dislodgment, and bleeding, have been reported to occur in approxi- mately 15% to 40% of the patients, s-14

Self-expanding metal stents have been developed to overcome these limitations of plastic tubes. They have the advantage of a smaller insertion diameter, and a lumen of up to 2 cm can be attained after release owing to the pressure they exert while expanding. In a

4 9 0 GASTROINTESTINAL ENDOSCOPY VOLUME 41, NO. 5, 1995

Table 1. Clinical data of 8 patients with malignant esophagorespiratory fistulas

Diagnosis Previous Patient no. Age (y)/sex (histologic treatment

findings)

Fistula

Location* Site

Esophageal stenosis

Length* Dilationt

1 57/F Lung cancer (ADC)

2 47/M Esophageal cancer (scc)



5 51/M Esophageal cancer (SCC)

6 51/M Lung cancer (ADC) 7 56/M Esophageal cancer

(SCC) 8 71/M Lung cancer

(ADC)

Pneumonectomy, radiation, laser, bronchial stent

Laser

Radiation, chemotherapy, laser

Radiation, laser, tracheal stent Radiation, chemotherapy

Radiation

24 Le~MB

26 Le~MB

23 Trachea

26 Trachea

32 Right MB

25 Trachea 20 Trachea

25 Le~MB 31

26-34

23-35

27-35 15

35-42 14

20-24 13

25-33 15

ADC, adenocarcinoma; SCC, squamous cell carcinoma; MB, main bronchus. *Position of the fistula and/or stenosis; number of centimeters from the incisor teeth. tDiameter in millimeters of endoscopic dilation before stent placement.

recent randomized prospective study in patients with esophageal cancer, complications after implantation of uncovered self-expandable metal stents were reported to be significantly fewer than after insertion of plastic tubes. 12 However, uncovered stents are inap- propriate for sealing of fistulas and for preventing tumor ingrowth that may occur through the wire mesh.

Recently, Song and coworkers 15, 16 have designed a silicone-covered modified Gianturco stent for palliation of malignant esophageal strictures and fistulas. We report on our experience with this type of stent in 8 consecutive patients with malignant esophagorespiratory fistulas.

PATIENTS AND METHODS

Patients

From February to December 1993, 8 consecutive pat ients (1 woman and 7 men) with a mal ignant esophagorespiratory fistula were included in this prospective study. Their mean age was 58.2 years (range, 47 to 79 years). All pat ients gave their informed consent following detai led explanat ion of the risks of s tent implanta t ion and the possible therapeut ic al- ternatives. The malignant nature of the fistula was confirmed by esophagoscopic or bronchoscopic biopsy. Five pa- t ients had a squamous cell carcinoma of the esophagus, and 3 had an adenocarcinoma of the lung with an infil tration into the esophageal wall. Before endoscopic t reatment , the pa- t ients could not swallow food or water because of aspiration. The fistula was esophagotracheal in 4 pat ients and esophagobronchial in 4. To evaluate the extent of airway involve- ment and the risk of t racheal compression during esophageal di lat ion or s tent placement, bronchoscopy was performed in all pat ients with lung cancer and also in all pat ients with

esophagotracheal fistula before esophageal di lat ion or stenting.

All pat ients were staged according to the Karnofsky index. 17 This index is based on the in terpre ta t ion of quali ty of life in terms of physical ability. A score of 100 % represents a self-supporting pat ient without any symptoms, whereas a score of 10% indicates a tota l ly bedr idden patient . Mean Karnofsky score of all pat ients was 45 % (range, 20 % to 70 % ) before t rea tment . Fur ther details of the pat ients and their tumors and fistulas are shown in Tables 1 and 2.

After successful s tent insertion, the pat ients were followed up by telephone interview 1 week after endoscopic t rea tment and then monthly unti l death. Control endoscopy or contrast esophagography was performed only in cases of clinical problems, such as recurrent dysphagia or aspiration. A grading system was used to quantify dysphagia as follows: 0, normal swallowing; 1, able to swallow some but not all solids; 2, able to swallow semisolids; 3, able to swallow fluids only; 4, unable to swallow fluids.

Stent and implantation technique

A modification of a self-expanding Gianturco metal s tent tube (Myung Sung Med-tech, Seoul, Korea) was used. Details of this type of s tent have been previously described by Song and coworkers. 15' 16 The s tent is constructed of stainless steel monofi laments arranged in a cylindrical zigzag pa t t e rn of eight or ten bends (Fig. 1). The outside of this barbless metal s tent is wrapped with a nylon mesh and covered with silicone. The proximal and distal ends of the tube are 4 mm wider in diameter than the middle par t to prevent dislodgment. The stents we used were 10, 12, or 14 cm long and 18 or 20 mm in internal d iameter in their middle par t when fully expanded.

The introducer set consists of a s tent tube, a balloon catheter, an introducing tube, and a pusher catheter. First , the

VOLUME 41, NO. 5, 1995 GASTROINTESTINAL ENDOSCOPY 491

Figure 2. Endoscopy demonstrating the proximal end of a fully opened covered stent with an excellent adaption to the esophageal wall.

Figure 1. A, Fully opened 10-cm-long covered Gianturco stent (modification by Song et al. is, 16). Note the wider seg- ments at both ends to prevent migration. B, A balloon catheter inserted into the stent, which is loaded into the introducer tube.

balloon catheter (Olbert Catheter System, Meadox Sur- gimed, Stenlose, Denmark) is passed through the polyethylene introducing tube (external diameter, 11 or 13 mm; Myung Sung Med-tech). The stent is then passed over the deflated balloon, manually compressed, and loaded into the distal end of the introducing tube (Fig. 1). The balloon catheter is partially withdrawn into the distal end of the loaded stent and inflated to keep the stent firmly in place in the introducing tube.

Before stent implantation, the location and extent of the fistula between the esophagus and the respiratory tract were determined by contrast esophagography using water-soluble contrast media. Premedication for endoscopic interventions included sedation with intravenous midazolam (2.5 to 15 mg) and analgesia with intravenous meperidine (25 to 50 rag). In 4 patients, malignant stenosis of the esophagus was significant and nontraversable to standard endoscopes (Olympus JF 100; Olympus America Inc., Lake Success, N.Y.). In these cases, the stenoses were dilated to a diameter of 13 to 15 mm using Savary-Gilliard bougies over a 300-cm-long, 0.038-in semistiff guidewire (Boston Scientific Corp., Watertown, Mass.). The location of the fistula and stenosis was marked on the patient's skin under fluoroscopic and endoscopic guidance immediately before stent placement. To achieve complete bridging of both fistula and stenosis, it was intended to position the stent 2 cm beyond the proximal and distal ends of the tumor.

The prepared set for stent insertion was passed over the guide wire into the esophagus and positioned under fluoro-

scopic control at least 2 cm distal to the previously marked end of the fistula and stenosis. Thereafter, the balloon catheter was deflated and removed together with the guide wire. A polyethylene pusher catheter (Myung Sung Med-tech) was inserted through the introducing tube until it reached the proximal end of the loaded stent tube. Under fluoroscopic guidance, the introducing tube was slowly withdrawn while the stent was kept in place with the pusher catheter. The stent was thereby released and allowed to expand in the esophagus, thus occluding the esophagorespiratory fistula. After release of the stent, the introducing tube and the pusher catheter were removed. At the end of the procedure and again 4 to 6 hours later, esophagography using water- soluble contrast was performed to ascertain correct location of the stent, occlusion of the fistula, and absence of esophageal perforation.

RESULTS

Correct imp lan ta t ion of the modif ied se l f -expanding si l icone-coated Gian turco prosthesis was successfully achieved in all pa t ien ts (Table 2). One pa t i en t with an esophageal cancer had a fistula above a 12-cm-long stenosis; he received two over lapping s tents with lengths of 12 and 14 cm. The proximal end of these two s tents was placed 2 cm below the uppe r esophageal sphincter , whereas the distal margin of the s tents was located 5 cm above the distal esophageal sphincter . Thereaf te r , this pa t i en t could ma in ta in a relat ively normal diet unti l he died 15 weeks later of diffuse me tas t a t i c disease.

T h e endoscopic p rocedure was well to le ra ted by all pat ients . No early compl ica t ion was noted. The im- p lan ted s tents expanded immedia te ly af ter release and all could be passed with the endoscope thereaf ter . Figure 2 demons t r a t e s the comple te adap t ion of the expanded proximal s ten t margin to the esophageal wall. None of the pa t i en t s exper ienced increased chest pa in af ter s ten t insertion. The tolerance of the s ten t was excellent, none of the pa t i en t s repor t ing any for-

492 G A S T R O I N T E S T I N A L ENDOSCOPY VOLUME 41, NO. 5, t995

Figure 3. A, Esophagogram demonstrating large esophagotracheal fistula with contrast inflow into the bronchial tree; an endotracheal stent had previously been implanted for palliative treatment of a malignant tracheal stenosis (patient No. 6). B, Closure of the fistula after insertion of a covered esophageal metal stent.

eign body sensation. Two patients with tracheobronchial malacia after external and internal radiotherapy of a lung cancer were additionally treated, the first with a 7.0-cm-long plastic prosthesis stiffened by internal metal clasps (Dynamic stent, Rfisch, Weib- lingen, Germany) inserted into the trachea (Fig. 3) and the second with a 4.0-cm-long metal stent (Strecker stent, Boston Scientific Corp., Watertown, Mass.) reaching from the distal trachea into the left main bronchus. This was done before esophageal stenting, as bronchoscopy demonstrated a significant airway compression during esophageal dilation by a bougie with a diameter similar to that of the stent. Thereaf- ter, implantation of the esophageal prosthesis was achieved without ensuing airway compression, which was confirmed by bronchoscopy after esophageal stenting. The other patients also showed no clinical symptoms, suggesting absence of tracheobronchial obstruction by the esophageal stent. Therefore, these patients were not examined by bronchoscopy.

Complete occlusion of the fistula and complete bridging of a concomitant esophageal stenosis were achieved in all patients, as demonstrated by control esophagography after implantation of the stent. Fig- ure 4 indicates the closure of a large esophagobronchial fistula, with rapid passage of the contrast medium after stenting. Improvement of dysphagia was observed in all patients. Mean dysphagia grade after stenting was 1.7, and the mean Karnofsky score increased from 45% to 68% 1 week after treatment (Table 2). Four patients were able to eat some but not all solids (grade

Figure 4. Esophagogram showing (A) a large esophagorespiratory fistula and (B) complete closure after stent implantation (patient No. 2).

1), and 3 patients could swallow semisolid food (grade 2).

One patient could swallow semisolids only for 6 days, after which time he experienced recurrent aspiration. Endoscopic control revealed a correct position of the inserted tube. A fistula could not be detected endoscopically, and the stent bridging the esophageal stenosis could be passed. The stent tube had com- pletely expanded, but because of the massive dilation of the proximal esophagus the proximal stent margin was not in close contact with the esophageal wall. Therefore, fluid could pass between the stent and the esophageal wall into the fistula, as demonstrated by control esophagography. Further endoscopic treatment was not at tempted owing to the patient's poor general condition. He then received parenteral nutri- tion and died 3 weeks later of an aspiration pneumo- nia despite antibiotic therapy.

One patient with an esophageal cancer, who had been treated by chemotherapy plus external and in- tracavitary radiotherapy 8 months before stenting, died of a sudden hemoptysis 10 days after placement of the tube that led to hypovolemic shock and acute respiratory failure. We were not able to clarify whether bleeding was the result of malignant infiltration alone or of ulceration caused by the stent, as an autopsy was refused by the patient's relatives.

The other patients died of advanced metastatic disease. They could swallow at least semisolid food until death. No food impaction was reported and no endoscopic reintervention was necessary.

The survival rate of all patients was 10 to 121 days (mean, 49 days). The overall 30-day mortality rate of all patients was 50 % (4/8). One of the deaths was pos- sibly related to stent implantation (procedure-related mortality, 12.5%) and 3 to disease progression. The

VOLUME 41, NO. 5, 1995 G A S T R O I N T E S T I N A L ENDOSCOPY 493

Table 2. Results after placement of silicone-coated self-expanding esophageal prostheses

Patient

Stent Dysphagia grade* Karnofsky score (Diameter

[mm]/ Before After Before After length [cm]) treatment treatment treatment treatment

Complications Outcome (survival [days])

1 20/10 4 2 20 40 -- 22 2 20/12 4 1 50 70 -- 121 3 20/14 and 4 1 50 80 -- 104

20/12 4 18/14 4 4 60 60 Recurrent aspiration~ 28 5 20/12 4 2 70 70 -- 34 6 18/10 4 2 20 60 -- 21 7 20/12 4 1 60 80 Massive hemoptysis 10 8 18/12 4 1 30 80 -- 49

*0, normal swallowing; 1, able to swallow some but unable to swallow fluids.

tReopening of the fistula 6 days after stenting.

not all solids; 2, able to swallow semisolids; 3, able to swallow fluids only; 4,

overall success ra te of metal s tent p lacement for pal- l iation of fistulas was 87.5%.

DISCUSSION

Currently, endoscopic in tubat ion is considered the most effective me thod for palliation of mal ignant esophagorespiratory fistulas. 2, 5, 6, 9,11, 13 The conven- t ional prosthesis used for this purpose is a plain plastic tube or, in the case of larger fistulas, a modified device equipped with an inflatable cuff along the shaft, ls-21 Posit ioning of these tubes is f requent ly performed under general anesthesia and may be associated with complications, such as perfora t ion of the esophagus, pressure- induced necrosis, hemorrhage, food impaction, and dislocation of the tubes in 20 % to 40% of cases. Morta l i ty rates range from 2% to 16 %.s-14

An al ternat ive to plastic tubes became available with the deve lopment of self-expandable meta l stents. 2226 In general, three different types of metal prostheses are clinically used for t r ea tmen t of esophageal stenoses. The d iameter of wire mesh stents can be reduced by elongation, and these stents can be re- s t rained on a delivery device of small d iameter by a membrane (Wallstent, Schneider Co., Bfilach, Swit- zerland) or a hydrophil ic gel (Ultraflex, Boston Scien- tific Corp., Water town, Mass.). In the Z-grid s tent (Gianturco), the steel monofi laments are arranged in a cylindrical zigzag pa t t e rn of several bends. This s tent can be compressed and loaded into an introducing tube (external diameter , 11 or 13 mm) for p lacement in the esophagus. One of the main advantages of the self-expanding stents is their smaller diameter . Dila- t ion before insert ion is necessary to a lesser degree with self-expanding stents (up to 13 mm) than with plastic tubes (18 to 20 mm). In our study, dilation of the tu-

mor stenosis was necessary in only 4 pat ients (50 % ). This compares favorably with di lat ion rates in studies using plastic tubes, in which dilat ion before tube insert ion was per formed in all patients. 11-13

These theoret ical advantages of expandable metal s tents over conventional plastic prostheses, which may lead to a less invasive and less risky implanta t ion procedure and a be t te r s tent fixation, have recent ly been confirmed by several uncontrol led studies and one controlled trial. 12, 22-26 A disadvantage of meta l stents, however, is a ra te of t umor infi l tration through the spaces between the wires of 20% to 30% .23-26 More- over, because of their interstices, conventional metal s tents are not appropr ia te for t r ea tmen t of esophagorespi ra tory fistula. Song and coworkers 15' 16 designed a modified Gianturco s tent by wrapping a nylon mesh around the f ramework of the s tent and coating it with silicone rubber. This covered s tent was used effectively in 30 pat ients for sh0r t - term palliative t r e a tm en t of mal ignant esophageal str ictures with a low complica- t ion rate. S ten t migrat ion occurred in only 1 pa t ien t and a blockade of the s tent by bolus obstruct ion in 2 patients. Fur thermore , the same group has shown in a total of 10 pat ients tha t the insert ion of the silicone- covered s tent is also an effective t r ea tmen t of malig- nan t esophagorespiratory fistulas. 16' 27 In their studies the s tent was placed wi thout technical failure or pro- cedural complications and occluded the fistulas in all patients. Wu et al. 2s developed a similar covered Gianturco-RSsch Z s tent bu t omi t ted the wider distal ends and placed it in 32 pat ients for palliation of ma- l ignant esophageal obstruct ion (n = 24) and esophago- respira tory fistulas (n = 8).2s Complications, including s ten t migration, food impaction, membrane disrupt ion with tumor ingrowth, tumor overgrowth, hemorrhage, and sepsis result ing from pressure necrosis, were

494 G A S T R O I N T E S T I N A L ENDOSCOPY VOLUME 41, NO. 5, 1995

repor ted in 10 pat ients . Comple te occlusion of malig- nan t esophagoresp i ra tory fistulas was achieved in 6 of 8 pat ients ; the other 2 pa t i en t s had a par t ia l seal and were able to eat solids wi thout s y m p t o m s of aspirat ion.

Our resul ts are in ag reement with these previous papers . Successful pal l ia t ion of esophagoresp i ra to ry fistulas was achieved in 7 of 8 pa t i en t s (87.5 % ); in an- o ther case di lat ion of the proximal esophagus pre- ven ted sufficient s tent contac t with the esophageal wall, which suggests the use of a wider proximal flange. P rocedure - re la ted mor ta l i ty was 12.5% (1 of 8 pa- t ients) , a l though it is not ent i rely clear whether the fatal compl ica t ion of hemoptys i s in this pa t i en t with an advanced esophageal cancer previously t r ea t ed by c h e m o t h e r a p y and radia t ion was caused by s ten t p lacement , progression of the disease itself, or both.

Our s tudy and the resul ts of Wu et al. 2s demons t r a t e t ha t covered s tent ing can be used in pal l iat ion of esophagoresp i ra tory fistulas of bo th esophageal and lung cancer. Two pa t ien t s with lung cancer developed t racheobronchia l malacia af ter external and internal r ad io the rapy and received an addi t ional s ten t into the t racheobronchia l t ree to p reven t a i rway obs t ruc t ion by the esophageal tube. I t r emains to be de te rmined whether double s tent ing is necessary in all these cases or whether only pa t ien ts with significant t racheobron- chial s tenoses should undergo this approach. In the o ther pat ients , who had no resp i ra tory s tent ing, no clinical s y m p t o m s of a irway obs t ruc t ion developed dur ing or af ter esophageal di lat ion and s ten t placement ; therefore , no bronchoscopic control was per- fo rmed af ter esophageal s tenting.

As it is e i ther impossible (Wallstent) or very difficult (Gianturco s tent) to ex t rac t the meta l prosthesis f rom the esophagus, a i rway obs t ruc t ion by these esophageal s tents m a y lead to serious problems. Tracheobronch ia l obs t ruc t ion dur ing pall iat ive di lat ion of ma l ignan t esophageal s tenoses and af ter esophageal s tent ing has been previously repor ted. 9, 29 Colt et al. 3~ pe r fo rmed rigid bronchoscopy in all pa t i en t s immedia t e ly af ter inser t ion of plast ic esophageal pros theses for pallia- t ion of advanced esophageal cancer and observed a significant a i rway obs t ruc t ion in 10 of 39 pat ients , result ing in the addi t ional p l acemen t of s tents into the t racheobronchia l sys tem. Al though the feasibil i ty and safe ty of such an approach are still largely unknown, bronchoscopy should be r e c o m m e n d e d in pa t i en t s with esophagoresp i ra tory fistulas before esophageal s ten t ing , especially in those pa t ien ts in whom respiratory p rob lems develop dur ing dilation.

In conclusion, our pre l iminary da ta suggest t ha t modif ied si l icone-covered Gianturco me ta l s tents rep- resent a safe and effective procedure for t rea t ing ma- l ignant esophagoresp i ra tory fistula, thus enabl ing a rapid pal l iat ion of aspi ra t ion and dysphagia. T h e im-

p lan ta t ion procedure seems to be less invasive t han the inser t ion of rigid plast ic pros theses and can be accom- pl ished with litt le d iscomfor t for the pat ients . These p re l iminary resul ts war ran t a randomized prospect ive trial to compare covered se l f -expanding meta l s tents with convent ional plast ic pros theses in pa t ien ts with mal ignan t esophageal stenoses, esophagoresp i ra tory fistulas, or both.

REFERENCES 1. Martini N, Goodner JT, D'Angio GJ, Beattie EJ. Tracheo-

esophageal fistula due to cancer. J Thorac Cardiovasc Surg 1970;59:319-24.

2. Boyce HW. Medical management of esophageal obstruction and esophageal-pulmonary fistula. Cancer 1982;50:2597-600.

3. Campion JP, Bourdelat D, Launois B. Surgical treatment of malignant esopbagotracheal fistulas. Am J Surg 1983;146:641-6.

4. Duranceau A, Jamieson GG. Malignant tracheoesophageal fistula. Ann Thorac Surg 1984;37:346-54.

5. Graham DY, Dobbs SM, Zubler M. What is the role of prosthesis insertion in esophageal carcinoma? Gastrointest Endosc 1983;29:1-5.

6. Lolley DM, Ray JF, Ransdell HT, Razzuk MA, Urschel HC. Management of malignant esophagorespiratory fistula. Ann Thorac Surg 1978;25:516-20.

7. Wong J, Lam KH, Wei WI, Ong GB. Results of the Kirschner operation. World J Surg 1981;5:547-52.

8. Buset M, des Marez B, Baize M, et al. Palliative endoscopic management of obstructive esophagogastric cancer: laser or prosthesis? Gastrointest Endosc 1987;33:357-61.

9. Buset M, Des MB, Cremer M. Endoscopic palliative intubation of the esophagus invaded by lung cancer. Gastrointest Endosc 1990;36:357-9.

10. den Hartog Jager FCAI Bartalsman JFWM, Tytgat GN. Pallia- tive treatment of obstructing esophagogastric malignancy by endoscopic positioning of a plastic prosthesis. Gastroenterology 1979;77:1008-14.

11. Gasparri G, Casalegno PA, Camandona M, et al. Endoscopic insertion of 248 prostheses in inoperable carcinoma of the esophagus and cardia: short-term and long-term results. Gastrointest Endosc 1987;33:354-6.

12. Knyrim K, Wagner HJ, Bethge N, Keymling M, Vakil N. A controlled trial of an expansile metal stent for palliation of esophageal obstruction due to inoperable cancer. N Engl J Med 1993;329:1302-7.

13. Ogilvie AL, Dronfield MW, Ferguson R, Atkinson M. Palliative intubation of esophagogastric neoplasms at fibreoptic endoscopy. Gut 1982;23:1060-7.

14. Tytgat GNJ. Endoscopic therapy of esophageal cancer: possi- bilities and limitations. Endoscopy 1990;22:263-7.

15. Song HY, Choi KC, Cho BH, Ahn DS, Kim KS. Esophagogas- tric neoplasms: palliation with a modified Gianturco stent. Ra- diology 1991;180:349-54.

16. Song HY, Choi KC, Kwon HC, Yang DH, Cho BH, Lee ST. Esophageal strictures: treatment with a new design of modified Gianturco stent. Radiology 1992;184:729-34.

17. Karnofsky D, Burchenal JH. Clinical evaluation of chemotherapeutic agents in cancer. In: McCleod CM, ed. Evaluation of chemotherapeutic agents. New York: Columbia University Press, 1949:74-89.

18. Sargeant IR, Thorpe S, Bown SG. Cuffed esophageal prosthesis: a useful device in desperate situations in esophageal malignancy. Gastrointest Endosc 1992;38:669-75.

19. Smith JL, Michaletz PA, Tabibian N, Schwartz JT, Graham DY. Improved palliation of a respiratory-esophageal fistu|a with a cuffed esophageal prosthesis. Am J Gastroenterol 1987;11: 1175-6.

20. Hordijk ML, Dees J, van Blankenstein M. The management of malignant esophago-respiratory fistulas with a cuffed prosthesis. Endoscopy 1990;22:214-44.

VOLUME 41, NO. 5, 1995 G A S T R O I N T E S T I N A L ENDOSCOPY 4 9 5

21. Lux G, Wilson D, Wilson J, Demling L. A cuffed tube for the treatment of oesophago-bronchial fistulae. Endoscopy 1987;19: 28-30.

22. Binmoeller KF, Maeda M, Lieberman D, Katon RM, Ivancev K, Rosch J. Silicone-covered expandable metallic stents in the esophagus: an experimental study. Endoscopy 1992;24:416- 20.

23. Bethge N, Knyrim K, Wagner HJ, Starck E, Pausch J, Kleist DV. Self-expanding metal stents for palliation of malignant esophageal obstruction--a pilot study of eight patients. Endos- copy 1992;24:411-5.

24. Cwikiel W, Stridbeck H, Tranberg KG, et al. Malignant esophageal strictures: treatment with a self-expanding nitinol stent. Radiology 1993;187:661-5.

25. Neuhaus H, Hoffmann W, Dittler HJ, Niedermeyer HP, Clas- sen M. Implantation of self-expanding esophageal metal stents for palliation of malignant dysphagia. Endoscopy 1992;24:405- 10.

26. Schaer J, Katon RM, Ivancev K, Uchida B, Rosch J, Binmoeller K. Treatment of malignant esophageal obstruction with silicone-coated metallic self-expanding stents. Gastrointest En- dosc 1992;38:7-11.

27. Do YS, Song HY, Lee BH, et al. Esophagorespiratory fistula associated with esophageal cancer: treatment with a Gianturco stent tube. Radiology 1993;187:673-7.

28. Wu WC, Katon RM, Saxon RR, et al. Silicone-covered self-expanding metallic stents for the palliation of malignant esophageal obstruction and esophagorespiratory fistulas: experience in 32 patients and a review of the literature. Gastrointest Endosc 1994;40:22-33.

29. Utts SJ, Di Palma JA, Wytock DH. Tracheobronchial obstruction complicating peroral prosthesis insertion. Gastrointest En- dosc 1985;31:383-4.

30. Colt HG, Meric B, Dumon JF. Double stents for carcinoma of the esophagus invading the tracheo-bronchiat tree. Gastrointest Endosc 1992;38:485-9.

496 G A S T R O I N T E S T I N A L ENDOSCOPY VOLUME 41, NO. 5, 1995

treatment of esophagorespiratory fistulas with a polyurethane-covered self-expanding metallic mesh...

Documents