- Email: [email protected]
Prosthetic replacement of the esophagus
Prosthetic replacement of the esophagus A tubular prosthesis was initially made of silicone rubber, spot-bonded on the inside and outside to a layer of knitted Dacron. This was covered on the outside by a loose layer of knitted Dacron, attached to the rubber only at the two ends, to allow incorporation by fibrous tissue and fixation of the prosthesis. Anastomoses were made with one layer of continuous sutures. There was initial incorporation of the graft. After 29 to 44 days the tube became loose and migrated, leaving a stricture. The prosthesis was redesigned with two loose layers of wider mesh. Incorporation occurred early with abundant fibrous reaction. Separation of the tube started at 40 days and was complete within 72 to 152 days. When the sutures became disengaged, bacteria entered the space between rubber and incorporated fabric, resulting in collagen lysis and separation of the tube. There were no anastomotic leaks in the entire series.
Fayek D. Salama, F.R.C.S., * London, England
Many investigators have tried to solve the problem of bridging a gap in the esophagus with a prosthesis. The pioneer work of Berman- was followed by other determined attempts. Inconsistency of results, rather than over-all failure, has so far defeated efforts to design a reliable prosthesis. The main problems are leakage at the anastomosis"- fi-n. 1~, 1:< and migration of the prosthesis, with contraction of the fibrous sheath formed around ie,',', 10, 11 causing obstruction. The present study has been conducted with these problems in mind. Construction of the prosthesis
Various workers used materials such as methacrylate/ polythene;'- a, r" n, n Tygon,": 111 nylon,!'" silicone rubber,": J:l tantalum," Ivalon," Hydron,' Marlex," and woven Dacron,': 11 either singly or in various combinations. The prosthesis currently under study was From Westminster Hospital, London S.W.I, England. Supported by a grant from the Westminster Hospital. Received for publication March 17, 1975. • Present address: St. George's Hospital, Hyde Park Corner, London, S.W.I, England.
made of a silicone rubber tube, spot-bonded on the inside and outside to a layer of knitted Dacron. This was wrapped in a loose layer of knitted Dacron, attached to the tube only at the two ends, to allow fibrous tissue incorporation of the fabric and thus fix the prosthesis in place (Fig. 1). The prosthesis was 18 to 20 mm. in internal diameter and 7 to 10 em. long; the rubber was 1 mm. thick (see Fig. 5). Procedure
Pigs weighing between 30 and 45 kilograms were used. Through a right thoracotomy, a segment of the middle third of the esophagus, 7 to 10 em. long, was mobilized and resected. The prosthesis was then anastomosed at either end with a continuous layer of 3-0 non absorbable material. The esophagus was fixed to the prosthesis beyond its edge, so that 2 mm. of the prosthesis projected into the lumen (Fig. 2). Care was taken to pick up the rubber with every stitch to ensure watertight anastomoses. The mediastinal pleura was left open. The chest was closed and intercostal drainage was employed for 6 hours. The animals were allowed fluids by mouth
739
The Journal of
740
Salama
Thoracic and Cardiovascular Surgery
® Fig. 1. Method of construction of the prosthesis. A, Knitted Dacron tube pulled over a mandrel. Silicone rubber tube built around it and spot-bonded to it (shaded). B, One end of the Dacron tube pulled over the rubber and spot-bonded to it. C, The other end of the Dacron tube pulled back and fixed only to the other end of the prosthesis by a running stitch. This produced a single outer loose layer.
Fig. 2. Anastomoses effected with one layer of continuous sutures. The prosthesis projected into the esophagus at either end for a distance of 2 mm.
from the first postoperative day. The diet was gradually increased until by the end of the seventh day it consisted of ordinary pig meal, ground and made into thick gruel. Results
The animals were divided into three groups.
Group I comprised 17 pigs. There was one death from the anesthetic. Eight animals were put to death between 3 and 18 days postoperatively. The prostheses were found to be well incorporated and fixed in place (Fig. 3). Eight animals developed obstruction between 29 and 44 days postoperatively. At necropsy, the prostheses were found loose and were recovered from the stomach. The fibrous sheaths had contracted and caused strictures (Fig. 4). The fibrous sheaths were devoid of any epithelial lining. Comment. I did not understand why some tubes became incorporated while others did not. A possible explanation was that although the outer layer was loose, it was so tightly applied to the rest of the prosthesis that fibrous tissue was kept from completely invading the fibers (Fig. 5). The prosthesis was redesigned with knitted Dacron that had much wider pores (Fig. 6). The internal, spot-bonded layer was omitted and an extra outer loose layer was added (Figs. 7 and 8).
Volume 70
Prosthetic replacement of esophagus
Number 4 October, 1975
c t..-+-rl-r
+ rl
...:,
~ r-t-
74 1
. :,
Fig. 3. Specimen removed after I week. The loose layer was well incorporated by fibrous tissue, which securely fixed the prosthesis in place.
Fig. 4. Tube recovered from the stomach after 29 days. The fibrous sheath had contracted laterally, causing obstruction, and longitudinally, causing shortening.
Group II comprised 4 animals. There was one death from the anesthetic. One pig developed bolus obstruction after 2 days and was put to death. Two other animals developed obstruction after 7 and 12 days. At necropsy the tubes were incorporated, but they were too soft and had collapsed causing obstruction. Comment. The omission of the internal, spot-bonded layer had weakened the prosthesis, although the thickness of the rubber was the same as before. A new batch of prostheses was constructed with 2 rom. thick rubber to enhance the rigidity of the tubes. Group III comprised 19 pigs. There was
one death from the anesthetic. One animal had an obstruction from the onset and was put to death on the second day. I found that the upper esophageal segment had been half twisted during insertion of the prosthesis, an avoidable technical pitfall . Eight animals were put to death between 14 and 37 days postoperatively. The prostheses were well incorporated (Fig. 9). Four pigs were killed between 40 and 70 days, and their prostheses were found to be only partially incorporated (Fig. 10). Five other pigs were put to death between 71 and 152 days postoperatively. The prostheses were found to have loosened to variable degrees (Fig. 11). In 2 animals in which there was
The Journol of
742
Salama
Thoracic and Cardiavascular Surgery
Fig. 5. Completed prosthesis. Loose outer layer evenly stretched over rest of the prosthesis. Note internal spot-bonded layer.
11m m
I
Fig. 6. The pores of the new knit Dacron (b) are wider than those of the original (a) .
...
o
0
CD
Fig. 7. Modified method of construction of the prosthesis. A, Rubber tube built directly around the mandrel. Knitted Dacron tube pulled over it and spot-bonded to it (shaded). B, One layer of the Dacron tube pulled back as a loose layer. C, The other end of the Dacron tube pulled back as a second loose layer.
Volume 70 Number 4 October, 1975
Prosthetic replacement of esophagus
743
Fig. 8. Modified prosthesis. Note the wider mesh and the loose external layers. The bare rubber was trimmed to 2 mm. before insertion.
Fig. 9. Specimen removed after 33 days. Fibrous tissue almost completely incorporates both loose layers. Note the space between the incorporated layers and the rest of the prosthesis.
complete dehiscence of the prosthesis, the mucosa of the upper and lower segments had become approximated, although the muscle layers remained apart. A stricture occurred in both cases (Fig. 12). Fibrous tissue invasion was noticeable from the second day, especially at the suture line. There were no anastomotic leaks in the entire series.
Discussion This prosthesis was designed with two main objectives: (1) to reduce anastomotic leaks to the absolute minimum and (2) to allow incorporation by fibrous tissue, so that the prosthesis would be fixed in place and the esophageal lumen thus kept open. The first objective was fully attained, the second, only partially achieved. The absence of any anastomotic leaks was due to many factors, some technical and some inherent in the design of the prosthesis. Mobilization of the esophagus was limited to within 1 to 1.5 em. from the proposed
lines of section. No clamps were applied, and a meticulous suture technique was used for the anastomoses, rubber being picked in every stitch to ensure a watertight seal. Silicone rubber was found to have the right consistency for easy suturing. Some materials, such as polythene, do not lend themselves easily to suture. A watertight anastomosis is essential in preventing esophageal contents from escaping before the whole area has been sealed by fibrous tissue. Watanabe and Mark:" used a silicone rubber prosthesis with Dacron sewing rings four layers deep. They made the anastomosis to the ring without picking up the rubber, and the result was a high incidence of leakage. Nonsuture techniques have been advocated by some workers. Berman- passed purse-string sutures through the esophageal wall and the wall of the tube without traversing the lumen. Leaks can occur if the
The Journal of
744
Salama
Thoracic and Cardiovascular Surgery
Fig. 10. Specimen removed after 41 days. Fibrous tissue incorporating the loose layers has been resorbed in many places. (A wood stick is lying between prosthesis and fibrous sheath.) Note the mucosa pulled to line the resulting crevices .
.
"
Fig. 11. Specimen removed after 91 days illustrating a more advanced stage of fibrous tissue resorption. The prosthesis is fixed only at two points. Note the mucosa lining the sheath.
sutures are applied either too 100sely14 or too tightly, the latter causing necrosis of the esophagus beyond the prosthesis." Fryfogle and associates' designed a toothed nylon ring to fix the esophagus to the prosthesis by a special applicator. Half the animals died as a result of technical errors in applying the ring. Closure of the mediastinal pleura over the prosthesis is believed to be an important factor in preventing leaks. However, this would not be feasible in operating on patients with carcinoma. Therefore, closure of the pleura was omitted in this study with no ill effects. Early infiltration of the fabric by granu-
lation tissue started from the esophageal ends at the suture lines and played an important part in preventing leaks. This was evident from 2 days onward. Invasion of the fabric by fibrous tissue continued until the prosthesis was fixed, usually within 7 days. Within that time, the lung, the mediastinal tissues, and the sheath of the aorta were firmly adherent to the graft. The subsequent loosening and migration of the prosthesis needs explanation. This was provided during the course of the third group of experiments. Initially, all prostheses were well incorporated and fixed in place. When the anastomotic sutures eventually became disengaged by the seventh to eighth
Volume 70 Number 4
Prosthetic replacement of esophagus
745
October. 1975
Fig. 12. Specimen removed after 156 days. The prosthesis was completely loose and was recovered from the stomach. Mucosa of upper and lower esophageal segments meet, but the muscular coats stay apart.
days, the space between the incorporated loose layers and the rest of the prosthesis became open to esophageal contents (Fig. 9). The fibrous tissue was slowly eroded by a process of digestion until the prosthesis was completely loose. I was unaware of this sequence of events during the first group of experiments because the loosening process was much more rapid. The prostheses used had only one loose layer of much smaller mesh, and therefore the amount of fibrous tissue involved was necessarily smaller. So long as the prostheses remained in place, the animals swallowed normally and gained weight. Migration of the tubes was regularly followed by strictures. The longer the tube stayed in place, the shorter the fibrous sheath became, because of longitudinal contraction. Finally, the mucosa of the upper and lower esophageal segments met. However, the muscle layers always remained apart, and migration was followed by a stricture in every case (Fig. 12). It is interesting to speculate on what would happen in a similar clinical situation. It is possible that such a stricture could be regularly dilated so that a reasonable lumen could be maintained.
From this study it is clear that migration of the prosthesis is a continuing problem. Further investigation is required. I would like to thank Mr. C. E. Drew, who supervised this work, for his valuable criticisms and continuing encouragement. I would also like to thank Dr. P. Cliff, Director of the Department of Clinical Measurement, in whose laboratory this work was done, Miss Angela Weston for secretarial work, and the Department of Medical Illustration at Westminster Hospital. REFERENCES
2 3 4
5
6
Battersby, J. S., and King, H.: Esophageal Replacement With Plastic Tubes: Experimental Study, Arch. Surg. 69: 400, 1954. Berman, E. F.: Experimental Replacement of the Esophagus by a Plastic Tube, Ann. Surg, 135: 337, 1952. Braunwald, N. S., and Hufnagel, C. A.: Reconstruction of the Esophagus With Wire Mesh Prosthesis, Surgery 43: 600, 1958. Fryfogle, J. D., Cyrowski, E. A., Rothwell, D., Rheault, E., and Clark, T.: Replacement of the Middle Third of the Esophagus With a Silicone Rubber Prosthesis, Chest 43: 464, 1963. Hawk, J. c., Jr., and Jeffords, J. V.: Replacement of Esophageal Segments: Experimental and Clinical Observations (Special Reference to Various Plastic Tubes), Am. Surg, 21: 939, 1955. Kirpatovskii, I. D., and Kulik, V. P.: Porous
The Journol of
746
Salama
Thoracic and Cardiovascular Surgery
7
8
9 10
and Nonporous Prosthesis in Esophageal Alloplasty, Khirurgiia 36: 112, 1960. LaGuerre, J. N., Schoenfeld, H., Calem, W., Gould, F. E., and Levowitz, B. S.: Prosthetic Replacement of Esophageal Segments, J. THORAc. CARDIOVASC. SURG. 56: 674, 1968. Lister, J., Altman, R. P., and Allison, W. A.: Prosthetic Substitution of the Thoracic Esophagus in Puppies: Use of Marlex Mesh With Collagen or Anterior Rectus Sheath, Am. Surg, 162: 812, 1965. Moore, H. D.: Replacement of Segments of the Thoracic Esophagus by Polythene Tubes, Surg, Gynecol, Obstet. 98: 619, 1954. Rubenstein, L. H.: Experiments With Substitute Esophagus, J. THORAC. SURG. 32: 691, 1956.
11 Schuring, A. E., and Ray, J. W.: Experimental Use of Dacron as an Esophageal Prosthesis, Ann. Otol. Rhinol. Laryngol. 75: 202, 1966. 12 Smith, R. A., and Raison, J. c.: Excision of Carcinoma of the Middle Third of the Oesophagus With Aortic Graft Replacement: Preliminary Report, 44: 566, 1957. 13 Watanabe, K., and Mark, J. B. D.: Segmental Replacement of the Thoracic Esophagus With a Silastic Prosthesis, Am. J. Surg. 121: 238, 1971. 14 Wawro, N. W.: Fatal Complications After Esophageal Replacement With Plastic (Polythene) Tube, Surgery 36: 903, 1954.
Fayek D. Salama, F.R.C.S., * London, England
Many investigators have tried to solve the problem of bridging a gap in the esophagus with a prosthesis. The pioneer work of Berman- was followed by other determined attempts. Inconsistency of results, rather than over-all failure, has so far defeated efforts to design a reliable prosthesis. The main problems are leakage at the anastomosis"- fi-n. 1~, 1:< and migration of the prosthesis, with contraction of the fibrous sheath formed around ie,',', 10, 11 causing obstruction. The present study has been conducted with these problems in mind. Construction of the prosthesis
Various workers used materials such as methacrylate/ polythene;'- a, r" n, n Tygon,": 111 nylon,!'" silicone rubber,": J:l tantalum," Ivalon," Hydron,' Marlex," and woven Dacron,': 11 either singly or in various combinations. The prosthesis currently under study was From Westminster Hospital, London S.W.I, England. Supported by a grant from the Westminster Hospital. Received for publication March 17, 1975. • Present address: St. George's Hospital, Hyde Park Corner, London, S.W.I, England.
made of a silicone rubber tube, spot-bonded on the inside and outside to a layer of knitted Dacron. This was wrapped in a loose layer of knitted Dacron, attached to the tube only at the two ends, to allow fibrous tissue incorporation of the fabric and thus fix the prosthesis in place (Fig. 1). The prosthesis was 18 to 20 mm. in internal diameter and 7 to 10 em. long; the rubber was 1 mm. thick (see Fig. 5). Procedure
Pigs weighing between 30 and 45 kilograms were used. Through a right thoracotomy, a segment of the middle third of the esophagus, 7 to 10 em. long, was mobilized and resected. The prosthesis was then anastomosed at either end with a continuous layer of 3-0 non absorbable material. The esophagus was fixed to the prosthesis beyond its edge, so that 2 mm. of the prosthesis projected into the lumen (Fig. 2). Care was taken to pick up the rubber with every stitch to ensure watertight anastomoses. The mediastinal pleura was left open. The chest was closed and intercostal drainage was employed for 6 hours. The animals were allowed fluids by mouth
739
The Journal of
740
Salama
Thoracic and Cardiovascular Surgery
® Fig. 1. Method of construction of the prosthesis. A, Knitted Dacron tube pulled over a mandrel. Silicone rubber tube built around it and spot-bonded to it (shaded). B, One end of the Dacron tube pulled over the rubber and spot-bonded to it. C, The other end of the Dacron tube pulled back and fixed only to the other end of the prosthesis by a running stitch. This produced a single outer loose layer.
Fig. 2. Anastomoses effected with one layer of continuous sutures. The prosthesis projected into the esophagus at either end for a distance of 2 mm.
from the first postoperative day. The diet was gradually increased until by the end of the seventh day it consisted of ordinary pig meal, ground and made into thick gruel. Results
The animals were divided into three groups.
Group I comprised 17 pigs. There was one death from the anesthetic. Eight animals were put to death between 3 and 18 days postoperatively. The prostheses were found to be well incorporated and fixed in place (Fig. 3). Eight animals developed obstruction between 29 and 44 days postoperatively. At necropsy, the prostheses were found loose and were recovered from the stomach. The fibrous sheaths had contracted and caused strictures (Fig. 4). The fibrous sheaths were devoid of any epithelial lining. Comment. I did not understand why some tubes became incorporated while others did not. A possible explanation was that although the outer layer was loose, it was so tightly applied to the rest of the prosthesis that fibrous tissue was kept from completely invading the fibers (Fig. 5). The prosthesis was redesigned with knitted Dacron that had much wider pores (Fig. 6). The internal, spot-bonded layer was omitted and an extra outer loose layer was added (Figs. 7 and 8).
Volume 70
Prosthetic replacement of esophagus
Number 4 October, 1975
c t..-+-rl-r
+ rl
...:,
~ r-t-
74 1
. :,
Fig. 3. Specimen removed after I week. The loose layer was well incorporated by fibrous tissue, which securely fixed the prosthesis in place.
Fig. 4. Tube recovered from the stomach after 29 days. The fibrous sheath had contracted laterally, causing obstruction, and longitudinally, causing shortening.
Group II comprised 4 animals. There was one death from the anesthetic. One pig developed bolus obstruction after 2 days and was put to death. Two other animals developed obstruction after 7 and 12 days. At necropsy the tubes were incorporated, but they were too soft and had collapsed causing obstruction. Comment. The omission of the internal, spot-bonded layer had weakened the prosthesis, although the thickness of the rubber was the same as before. A new batch of prostheses was constructed with 2 rom. thick rubber to enhance the rigidity of the tubes. Group III comprised 19 pigs. There was
one death from the anesthetic. One animal had an obstruction from the onset and was put to death on the second day. I found that the upper esophageal segment had been half twisted during insertion of the prosthesis, an avoidable technical pitfall . Eight animals were put to death between 14 and 37 days postoperatively. The prostheses were well incorporated (Fig. 9). Four pigs were killed between 40 and 70 days, and their prostheses were found to be only partially incorporated (Fig. 10). Five other pigs were put to death between 71 and 152 days postoperatively. The prostheses were found to have loosened to variable degrees (Fig. 11). In 2 animals in which there was
The Journol of
742
Salama
Thoracic and Cardiavascular Surgery
Fig. 5. Completed prosthesis. Loose outer layer evenly stretched over rest of the prosthesis. Note internal spot-bonded layer.
11m m
I
Fig. 6. The pores of the new knit Dacron (b) are wider than those of the original (a) .
...
o
0
CD
Fig. 7. Modified method of construction of the prosthesis. A, Rubber tube built directly around the mandrel. Knitted Dacron tube pulled over it and spot-bonded to it (shaded). B, One layer of the Dacron tube pulled back as a loose layer. C, The other end of the Dacron tube pulled back as a second loose layer.
Volume 70 Number 4 October, 1975
Prosthetic replacement of esophagus
743
Fig. 8. Modified prosthesis. Note the wider mesh and the loose external layers. The bare rubber was trimmed to 2 mm. before insertion.
Fig. 9. Specimen removed after 33 days. Fibrous tissue almost completely incorporates both loose layers. Note the space between the incorporated layers and the rest of the prosthesis.
complete dehiscence of the prosthesis, the mucosa of the upper and lower segments had become approximated, although the muscle layers remained apart. A stricture occurred in both cases (Fig. 12). Fibrous tissue invasion was noticeable from the second day, especially at the suture line. There were no anastomotic leaks in the entire series.
Discussion This prosthesis was designed with two main objectives: (1) to reduce anastomotic leaks to the absolute minimum and (2) to allow incorporation by fibrous tissue, so that the prosthesis would be fixed in place and the esophageal lumen thus kept open. The first objective was fully attained, the second, only partially achieved. The absence of any anastomotic leaks was due to many factors, some technical and some inherent in the design of the prosthesis. Mobilization of the esophagus was limited to within 1 to 1.5 em. from the proposed
lines of section. No clamps were applied, and a meticulous suture technique was used for the anastomoses, rubber being picked in every stitch to ensure a watertight seal. Silicone rubber was found to have the right consistency for easy suturing. Some materials, such as polythene, do not lend themselves easily to suture. A watertight anastomosis is essential in preventing esophageal contents from escaping before the whole area has been sealed by fibrous tissue. Watanabe and Mark:" used a silicone rubber prosthesis with Dacron sewing rings four layers deep. They made the anastomosis to the ring without picking up the rubber, and the result was a high incidence of leakage. Nonsuture techniques have been advocated by some workers. Berman- passed purse-string sutures through the esophageal wall and the wall of the tube without traversing the lumen. Leaks can occur if the
The Journal of
744
Salama
Thoracic and Cardiovascular Surgery
Fig. 10. Specimen removed after 41 days. Fibrous tissue incorporating the loose layers has been resorbed in many places. (A wood stick is lying between prosthesis and fibrous sheath.) Note the mucosa pulled to line the resulting crevices .
.
"
Fig. 11. Specimen removed after 91 days illustrating a more advanced stage of fibrous tissue resorption. The prosthesis is fixed only at two points. Note the mucosa lining the sheath.
sutures are applied either too 100sely14 or too tightly, the latter causing necrosis of the esophagus beyond the prosthesis." Fryfogle and associates' designed a toothed nylon ring to fix the esophagus to the prosthesis by a special applicator. Half the animals died as a result of technical errors in applying the ring. Closure of the mediastinal pleura over the prosthesis is believed to be an important factor in preventing leaks. However, this would not be feasible in operating on patients with carcinoma. Therefore, closure of the pleura was omitted in this study with no ill effects. Early infiltration of the fabric by granu-
lation tissue started from the esophageal ends at the suture lines and played an important part in preventing leaks. This was evident from 2 days onward. Invasion of the fabric by fibrous tissue continued until the prosthesis was fixed, usually within 7 days. Within that time, the lung, the mediastinal tissues, and the sheath of the aorta were firmly adherent to the graft. The subsequent loosening and migration of the prosthesis needs explanation. This was provided during the course of the third group of experiments. Initially, all prostheses were well incorporated and fixed in place. When the anastomotic sutures eventually became disengaged by the seventh to eighth
Volume 70 Number 4
Prosthetic replacement of esophagus
745
October. 1975
Fig. 12. Specimen removed after 156 days. The prosthesis was completely loose and was recovered from the stomach. Mucosa of upper and lower esophageal segments meet, but the muscular coats stay apart.
days, the space between the incorporated loose layers and the rest of the prosthesis became open to esophageal contents (Fig. 9). The fibrous tissue was slowly eroded by a process of digestion until the prosthesis was completely loose. I was unaware of this sequence of events during the first group of experiments because the loosening process was much more rapid. The prostheses used had only one loose layer of much smaller mesh, and therefore the amount of fibrous tissue involved was necessarily smaller. So long as the prostheses remained in place, the animals swallowed normally and gained weight. Migration of the tubes was regularly followed by strictures. The longer the tube stayed in place, the shorter the fibrous sheath became, because of longitudinal contraction. Finally, the mucosa of the upper and lower esophageal segments met. However, the muscle layers always remained apart, and migration was followed by a stricture in every case (Fig. 12). It is interesting to speculate on what would happen in a similar clinical situation. It is possible that such a stricture could be regularly dilated so that a reasonable lumen could be maintained.
From this study it is clear that migration of the prosthesis is a continuing problem. Further investigation is required. I would like to thank Mr. C. E. Drew, who supervised this work, for his valuable criticisms and continuing encouragement. I would also like to thank Dr. P. Cliff, Director of the Department of Clinical Measurement, in whose laboratory this work was done, Miss Angela Weston for secretarial work, and the Department of Medical Illustration at Westminster Hospital. REFERENCES
2 3 4
5
6
Battersby, J. S., and King, H.: Esophageal Replacement With Plastic Tubes: Experimental Study, Arch. Surg. 69: 400, 1954. Berman, E. F.: Experimental Replacement of the Esophagus by a Plastic Tube, Ann. Surg, 135: 337, 1952. Braunwald, N. S., and Hufnagel, C. A.: Reconstruction of the Esophagus With Wire Mesh Prosthesis, Surgery 43: 600, 1958. Fryfogle, J. D., Cyrowski, E. A., Rothwell, D., Rheault, E., and Clark, T.: Replacement of the Middle Third of the Esophagus With a Silicone Rubber Prosthesis, Chest 43: 464, 1963. Hawk, J. c., Jr., and Jeffords, J. V.: Replacement of Esophageal Segments: Experimental and Clinical Observations (Special Reference to Various Plastic Tubes), Am. Surg, 21: 939, 1955. Kirpatovskii, I. D., and Kulik, V. P.: Porous
The Journol of
746
Salama
Thoracic and Cardiovascular Surgery
7
8
9 10
and Nonporous Prosthesis in Esophageal Alloplasty, Khirurgiia 36: 112, 1960. LaGuerre, J. N., Schoenfeld, H., Calem, W., Gould, F. E., and Levowitz, B. S.: Prosthetic Replacement of Esophageal Segments, J. THORAc. CARDIOVASC. SURG. 56: 674, 1968. Lister, J., Altman, R. P., and Allison, W. A.: Prosthetic Substitution of the Thoracic Esophagus in Puppies: Use of Marlex Mesh With Collagen or Anterior Rectus Sheath, Am. Surg, 162: 812, 1965. Moore, H. D.: Replacement of Segments of the Thoracic Esophagus by Polythene Tubes, Surg, Gynecol, Obstet. 98: 619, 1954. Rubenstein, L. H.: Experiments With Substitute Esophagus, J. THORAC. SURG. 32: 691, 1956.
11 Schuring, A. E., and Ray, J. W.: Experimental Use of Dacron as an Esophageal Prosthesis, Ann. Otol. Rhinol. Laryngol. 75: 202, 1966. 12 Smith, R. A., and Raison, J. c.: Excision of Carcinoma of the Middle Third of the Oesophagus With Aortic Graft Replacement: Preliminary Report, 44: 566, 1957. 13 Watanabe, K., and Mark, J. B. D.: Segmental Replacement of the Thoracic Esophagus With a Silastic Prosthesis, Am. J. Surg. 121: 238, 1971. 14 Wawro, N. W.: Fatal Complications After Esophageal Replacement With Plastic (Polythene) Tube, Surgery 36: 903, 1954.