- Email: [email protected]
tissue complex
J
THoRAc CARDIOVASC SURG
86:800-808, 1983
Original Communications
Neovascularity of a tracheal prosthesis/tissue complex Permanent bioincorporation of a microporous tracheal prosthesis will require a stable blood supply to connective tissue supporting an epithelial surface. In experience with over 80 tracheal implants in dogs, we have observed that end-on ingrowth and epithelialization does not occur in the absence of lateral ingrowth, epithelialization is marked by the appearance of a subepithelial network of vessels, and this process must be well advanced by 6 to 8 weeks for long-term stability. These observations were extended by using microangiography to delineate the blood supply of the prosthesis/tissue complex. Six implants of bioelectric polyurethane with 10 % gentamicin (3 em length, 2 cm diameter, I to 1.25 mm wall thickness, 60 to 120 J.L micropore diameter) were interposed in the dog thoracic trachea and wrapped with an omental pedicle. The aorta was perfused with a barium suspension at elective sacrifice between 10 weeks and 21 months. Radiographs of specimens were correlated with bronchoscopic, gross, and histopathological findings. Neovascularity of the prosthesis/tissue complex can be described in three categories: outer capsule, prosthetic wall, and inner lining. Outer capsule vessels were oriented circumferentially immediately adjacent to the prosthetic wall. They resembled arteries up to 75 J.L diameter on microscopy and appeared to develop from the omentum with connections developing to the bronchial circulation. Prosthetic wall vessels up to 75 J.L with thin muscular walls were noted to traverse the porous prosthetic wall. The inner lining had a network of subepithelial vessels that connected to the lamina propria vasculature of the native trachea across the anastomoses with vessels up to 120 J.L in diameter .. We conclude that the omentum provides an immediate blood supply and a base for early connective tissue ingrowth. Epithelialization occurs as early as 3 weeks on the favorable bed, accompanied by vascular connections to the existing lamina propria tracheal vessels. This dual organization of blood supply with connections across the prosthetic wall is probably important to long-term stability of healing.
Ronald J. Nelson, M.D., Lise Goldberg, M.D. (by invitation), Rodney A. White, M.D. (by invitation), Edwin Shors, Ph.D. (by invitation), and Frank M. Hirose, M.D. (by invitation), Torrance and Los Angeles, Calif.
Reconstruction of the trachea following extensive resection continues to challenge surgeons. Ingenious From the Departments of Surgery and Pathology, Harbor/UCLA Medical Center, Torrance, California and the UCLA School of Medicine, Los Angeles, Calif. Supported in part by Grant No. GM-24300 from the National Institutes of Health. Presented in part at the Sixty-eighth Annual Clinical Congress of the American College of Surgeons, Chicago, Ill., 1982. Read at the Sixty-third Annual Meeting of The American Association for Thoracic Surgery, Atlanta, Ga., April 25-27, 1983. Address for reprints: Ronald J. Nelson, M.D., Harbor/UCLA Medical Center, 1000 West Carson St., Torrance, Calif. 90509.
800
methods of mobilization have been devised to allow primary anastomosis in lieu of a satisfactory prosthesis.I, 2 The most successful clinical device for circumferential replacement has been a nonporous silicone rubber tube.' However, long-term healing of porous prostheses has been limited by chronic infection at the contaminated interface with progressive granulation tissue and stenosis. We4.5 have pursued the development of a prosthetic trachea with a unique, interconnecting microporous structure fabricated by using the microporous exoskeleton of marine corals and echinoderms as a template for biomedical polymers. The versatility of the model may
Volume 86 Number 6 December, 1983
Tracheal prosthesis
80 1
Fig. 1. Photographs taken at bronchoscopy through a Stortz telescope of a bioelectric polyurethane prosthesis/ tissue complex (A) at II weeks and (B) at 22 months, There is a subepithelial network of vessels connecting across the anastomoses with vessels of the lamina propria of the native trachea. Stable healing over the interval is evident. allow the gradual evolution of a successful replacement as favorable factors are identified and problems solved. The design allows independent control of major factors such as polymer, pore diameter, configuration, and methods of reinforcement. In addition, substances can be dispersed in the polymer or coated on its surface and temporary fillers can be used in the porous phase.
Previous experience Significant ingrowth was first shown into the outer surface (lateral ingrowth) of experimental bioelectric polyurethane* prostheses with a wall thickness of 3 to 5 rom at I month but was not present at 2 months, the pores being filled instead with inflammatory cells. Silicone rubber] grafts at the same configuration and pore diameter failed to show ingrowth at either time period. Both silicone rubber and bioelectric polyurethane grafts with 18 to 25 fJ, pores and 2 rom wall thickness showed no ingrowth. Complete incorporation of the prosthetic wall at the tracheal site was first achieved by wrapping an omental pedicle about the test prothesis (wall thickness 1.25 to 2 'Bioelectric polyurethane, Goodyear Tire and Rubber Co., Akron, Ohio. tSilicone rubber, MDX-4 4210, Dow Coming Corp., Midland, Mich.
rom) within a silicone rubber envelope that was positioned against the mediastinum and prevented the formation of adhesions. The bioincorporated graft was later rotated on its pedicle into a circumferential tracheal defect as a second procedure. 6 When the second procedure was done between 29 and 81 days after the first, extensive epithelialization occurred in all of the bioelectric polyurethane prostheses. Experience with silicone rubber was not as consistent. The appearance of a subepithelial network of vessels at bronchoscopy was noted to be a marker for epithelialization. This network connected with that of the native lamina propria across the anastomoses. When extensive epithelialization failed to occur, circumferential granulation tissue progressed to stenosis within 1 to 3 months. At the same 1.25 to 2 rom wall thickness in bioelectric polyurethane prostheses with a 60 to 120 fJ, pore size, lateral ingrowth with extensive epithelialization was achieved at direct implantation when the prosthesis was wrapped with omentum in six of 10 dogs.' In the absence of lateral ingrowth in all experimental series, proliferative granulation tissue at the anastomoses has led to stenosis without significant ingrowth along the inner wall of the prosthesis. A selected animal in this series, followed up 2Vz years, was found to have stable healing. Photographs taken during bronchoscopy at 11 weeks and 29 months are shown in Fig. 1. At 11 weeks,
The Journal of Thoracic and Cardiovascular Surgery
8 0 2 Nelson et al.
intervals and were electively put to death at 10, 10, 15, and 15 weeks and 18 and 21 months. At the time of death, the segment of the aorta including the arch vessels and the bronchial and celiac arteries was perfused with a suspension of micronized barium sulfate in normal saline until the tissues turned white," The segment of the trachea containing the prostheses was resected and examined grossly. The prosthesis and adjacent trachea was divided longitudinally. One half was further sectioned transversely in its midportion, Roentgenograms of these sections were taken in a Faxitron unit at 30 kvp for I minute. After formalin fixation, sections were stained for histologic study with hematoxylin and eosin. .
Results Fig. 2. Experimental tracheal prosthesis of microporous bioelectricpolyurethane (length 3 em) with a porediameter range of 60 to 120 IJ.. complete epithelialization is shown with blood vessels crossing the anastomoses and connecting the lamina propria of the native trachea with the subepithelial network of the prosthesis. Though blood vessels are not as prominent, stable healing is seen at 29 months. At 2Y2 years, discoloration of the lining suggested breakdown of the polyester urethane containing a dispersion of carbon black (bioelectric polyurethane). This success with direct implantation led to the series reported here, in which the wall thickness was reduced to I to 1.25 mm to favor bioincorporation and 10% gentamicin by weight was dispersed in the polymer to suppress bacterial growth."
Material and methods Bioelectric polyurethane prostheses with a micropore size of 60 to 120 JL (3 em length, 2 em internal diameter, I to 1.25 mm wall thickness), reinforced with stainless steel rings were implanted into three-ring thoracic tracheal defects in mongrel dogs (Fig 2). Polyglactin 910* interrupted sutures were used for the anastomoses. Omentum was mobilized through an upper midline abdominal incision, brought through a peripheral incision in the diaphragm and wrapped about the implants. All animals received humane animal care in compliance with the "Principles of Laboratory Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences. Six animals were followed up by bronchoscopy at *Vicryi, Ethicon, Inc., Somerville, N. J.
The neovascularity of the' prosthesis/tissue complex will be described in three categories: outer capsule, prosthetic wall, and inner lining. The outer capsule was provided at the time of implantation by wrapping the omentum circumferentially about the prosthesis. The predominant circumferential orientation of the outer capsule vessels is demonstrated on x-ray films of both longitudinal and transverse sections of the specimens (Fig. 3). On the longitudinal roentgenogram well-developed connections across the anastomoses are seen between the bronchial arterial supply of the native trachea and the omentally derived arterial supply of the prosthesis. On transverse sections, the larger outer capsule vessels are closely applied to the outer surface of the prosthetic wall and extend for long segments of the circumference. These vessels are muscular arteries up to 75 JL in diameter on histologic study (Fig. 4). This pattern of outer capsule vascular supply to the prosthesis/tissue complex was seen at each of the time intervals. Microangiography of sections demonstrates tortuous vessels traversing the prosthetic wall at multiple areas (Fig. 3, B). These vessels appear to feed a capillary-like network in the inner lining. On microscopy, vessels with thin muscular walls are noted in the collagenous soft tissue filling the pores. These can be observed on some sections connecting with the outer capsule, as shown in Fig. 5, or the inner lining. Extensive lateral ingrowth of granulation tissue was noted on bronchoscopy as early as 13 days (Fig. 6). This surface generally became epithelialized between 3 to 4 weeks with stratified squamous epithelium. With epithelialization the surface became smooth and translucent and the pink-red color intensity of the granulation tissue faded. The change was also marked by the appearance of a subepithelial network of vessels very similar to those of the lamina propria of the native trachea (Fig. 7).
Volume 86 Number 6
Tracheal prosthesis 8 0 3
December, 1983
Fig. 3. Faxitron x-ray films (30 kYP, 1 minute) of a prosthesis/tissue complex in which the arterial supply was perfused with a barium suspension at the time of sacrifice at 10 weeks. A, A longitudinal half with the inner lining not present to show the arterial supply of the native trachea and the outer capsule with predominant circumferential orientation of the latter (derived from the omental pedicle wrap) and well-developed connections (c) to the native bronchial circulation (wire reinforcing rings are approximately 0.89 mm in diameter). B, A transverse section showing the circumferential arteries of the outer capsule (OC) to be positioned next to the prosthetic wall (PW), with smaller arteries trav.ersing the prosthetic wall in a tortuous course to supply a capillary-like network in the inner lining (IL).
Extensive connections were noted .across both anastomoses, some vessels extending up to 1 to 2 em in length. On microscopy, thin-walled vessels up to 120 JL in diameter were seen. Normal appearing respiratory epithelium was noted to extend up to 7 mm into the prostheses at later time intervals (Fig. 8). Each of the six prostheses showed complete or near complete lateral ingrowth with epithelialization. However, the relatively thin wall (1 to 1.25 mm) and possible weakening of the polyurethane by the dispersion of 10% gentamicin led to accordian collapse of the prosthetic wall and thickening of scar tissue in the ''valleys'' of the collapse. Although the lumen was narrowed by this
accordian collapse, none of the animals had clinical problems related to stenosis even when followed up for 18 and 21 months. Construction of the anastomoses with absorbable sutures to avoid the problem of suture granuloma resulted in late protrusion of the ends of the prostheses into the lumen with subsequent regression of the inner lining at those areas.
Discussion Complete bioincorporation with epithelialization of an experimental tracheal prosthesis has been achieved with an interconnecting microporous structure with a pore diameter range of 60 to 120 JL. This pore structure
The Journal of Thoracic and Cardiovascular Surgery
8 04 Nelson et al.
,
f
•
Fig. 4. Photomicrograph of a muscular artery of the outer capsule in a specimen at 18 months. (Hematoxylin and eosin stain; original magnification X40.)
Fig. 5. Photomicrograph of a specimen at 15 weeks showing the fibrocollagenous tissue filling the pores of the prosthetic wall with a muscular artery at the junction of the prosthetic wall with the outer capsule. (Hematoxylin and eosin stain; original magnification X40.)
Volume 86 Number 6 December. 1983
Tracheal prosthesis
805
Fig. 6. Photograph taken at bronchoscopy through a Stortz telescope of a prosthesis 13 days following implantation. There is extensive lateral ingrowth of granulation tissue. Fig. 7. Photograph taken through a Stortz telescope at bronchoscopy showing complete epithelialization of a prosthesis at 16 weeks, with a subepithelial vascular network connecting with that of the lamina propria of the native trachea across the anastomoses.
allows the development of arteries up to 75 Jl in diameter that traverse the wall connecting the blood supply of the outer capsule and the inner lining. Soft tissue invasion into the prosthetic wall has been found to regress without the presence of a stable inner lining. The more profuse soft tissue ingrowth associated with polyurethanes as compared to silicone rubber seems important to the successful bioincorporation of tracheal prostheses. lO The omental pedicle used as an external wrap supplies an immediate segmental blood supply to the newly inserted prosthesis and is a rich source of neovascularity and soft tissue ingrowth . I I The omental blood vessels position themselves adjacent to the prosthetic wall, and extensive through-and-through lateral ingrowth is seen as early as 13 days on bronchoscopy. This early bioincorporation provides a favorable bed for epithelialization, which is initially stratified squamous epithelium presumably migrating from the cut ends of the native tracheal mucosa, although we have not ruled out a metaplastic source. This early epithelialization appears to stabilize the inner lining. If epithelialization does not occur by 6 to 8 weeks, progression of granulation tissue to significant stenosis occurs in 8 to 12 weeks. The epithelial lining is associated with the develop-
ment of a subepithelial vascular network similar in pattern to that of the native trachea. Simultaneous vascular connections across the anastomoses develop with the vessels of the lamina propria. Respiratory epithelium migrates over the inner lining to a distance of about 7 mm. This end-on ingrowth has not been seen without favorable lateral ingrowth. Thus there is a vascular network of both the outer capsule and inner lining. These networks connect with the similar vascular supplies of the native trachea and with each other across the prosthetic wall. In this experimental series, the balance between factors favoring early ingrowth and those affecting the strength of the prostheses was upset in favor of ingrowth at the expense of longitudinal accordian collapse. Adequate longitudinal as well as radial reinforcement will be required to prevent contracture and thickening of relatively avascular scar tissue. The problem of protrusion associated with absorbable suture material further emphasizes the importance of the dual blood supply. With loss of vascular connections across the anastomosis, regression of the inner lining was noted. We conclude that the dual organ ization of neovascularity to the prosthesis/tissue complex is important to long-term stability of healing. This degree of tissue
The Journal 01 Thoracic and Cardiovascular Surgery
8 0 6 Nelson et al.
5
6
7
8
9
10
II
Fig. 8. Photomicrograph showing respiratory epithelium with a small vessel in the underlying connective tissue. (Hematoxylin and eosin stain ; original magnification XIOO.)
organization may be necessary to avoid hypoxic tissue gradients that are associated with proliferation of granulation tissue." Design factors favoring this dual organization need to be refined in future experimental prostheses. The late time course of healing, bioincorporation of longer prostheses, factors favoring respiratory versus stratified squamous epithelium, and quality control of the biologically derived wall are among the areas for further investigation. REFERENCES Grillo HC: Congenital lesions, neoplasms and injuries of the trachea, Gibbon's Surgery of the Chest, DC Sabiston Jr, FC Spencer eds., Philadelphia, 1976, W. B. Saunders Company, p 256 2 Grillo HC: Carinal reconstruction. Ann Thorac Surg 34:356-373, 1982 3 Neville WE, Bolanowski PJP, Soltanzadeh H : Prosthetic reconstruction of the trachea and carina. J Thorac Cardiavase Surg 72:525-538, 1976 4 White RA, Weber IN, White EW: Replamineform. A
12
new process for preparing porous ceramic, metal, and polymer prosthetic materials. Science 176:922-924, 1972 Nelson RJ, White RA, Lawrence RS, Walkinshaw MD, Fiaschetti FL, White EW, Hirose FM: Development of a microporous tracheal prosthesis. Trans Am Soc Artif Intern Organs 25:8-12, 1979 Holwick JL, Klein SR, Miranda RM, Hirose FM , White RA , Nelson RJ: Heterotopic bioincorporation of a porous tracheal prosthesis with second stage transfer to the trachea. Proc Eur Soc Artif Organs 7:200-204, 1980 Holwick JL, Klein SR, Sproat RW, Miranda RM, Hi~e FM, White RA, Nelson RJ: Healing of a new microporous tracheal prosthesis. Surg Forum 31:541-543, 1980 Goldberg L, White RA, Shors E, Hirose FM, Nelson RJ: Experience with dispersion of gentamicin in a microporous polymeric tracheal prosthesis, Surg Forum 33:40-42, 1982 Thompson DP, Moore TC, Fingerhut AG: A method for high resolution renal and splenic microangiography. Surg Gynecol Obstet 132:101-108, 1971 White RA , Hirose FM , Sproat RW, Lawrence RS, Nelson RJ: Histopathologic observations after short-term implantation of two porous elastomers in dogs. Biomaterials 2:171-176,1981 Morgan E, Lima 0 , Goldberg M, Ferdman A, Luk SK, Cooper JD: Successful revascularization of totally ischemic bronchial autografts with omental pedicle flaps in dogs, J THORAC CARDIOVASC S URG 84:204-210, 1982 Knighton Dr, Silver lA, Hunt TK : Regulation of woundhealing angiogenesis. Effect of oxygen gradients and inspired oxygen concentration. Surg 90:262-270, 1981
Discussion DR. WILLIAM E. NEVILLE N ewark , N.J.
I have followed Dr. Nelson's work and wholeheartedly have supported his endeavors over the years. His report is an excellent and sophisticated piece of work. Without question there is a need for a tracheal prosthesis . An artificial heart has been developed, as have a large variety of good heart valves which are continually being improved, and with Dr. Nelson and his associates' work we can look forward to another artificial trachea. I must admit this is not a profitable enterprise but it is intriguing. When we presented our results with a tracheal silicone prosthesis before this Association in 1976, I outlined what I thought were the prerequisites for a suitable implant. It should be airtight, have adequate consistency, be well accepted by the host and cause minimal inflammatory reaction but still be incorporated into the surrounding tissue, be impervious to fibroblastic and bacterial invasion of the lumen, and pennit ingrowth of epithelium along the inner surface. The silicone model has all of these features with the exception of an eventual epithelial inner surface. However, observations exper-
Volume 86 Number 6 December. 1983
imentally and clinically over the past 20 years convince me that this probably does not make any difference. The nonwettable, highly polished, smooth inner surface of the silicone prosthesis permits the easy removal of distal secretions. I have never seen encrustations intraluminally. However, Dr. Nelson's prosthesis satisfies all the requirements for an ideal tracheal substitute when it is wrapped with omentum. This does take time and it may be difficult for the patient to raise his secretions initially prior to the formation of the respiratory epithelial surface. Nevertheless, it is a step in the right direction, and that is what we as surgeons strive for. We have implanted a straight silicone prosthesis in 46 patients to date. Twenty-eight had a long primary stricture or recurrent stricture, two had tracheoesophageal fistula plus stricture, five had tracheal malacia, and II a malignancy. We have seen four anastomotic granulomas in the past 10 years which were around nonabsorbable sutures; two patients had subglottic granulomas which were fulgurated. Recently a proximal anastomosis dehisced when I used Dexon as the suture material. This suture material should never be used for a tracheal anastomosis. There was one early death from a cerebrovascular accident. The late deaths have been the result of cardiac arrhythmia, drug overdose, innominate artery erosion, and myocardial infarction. Four individuals with adenocystic carcinoma are still alive I to 6 years postoperatively, and one with adenocarcinoma is alive at I year. Additionally we placed three prostheses intraluminally as a stent. One of those patients is still alive 3 years following irradiation. In 14 patients with a bifurcated prosthesis four suture line granulomas occurred around nonabsorbable sutures, two shifts of the prosthesis occurred from recurrent carcinoma, there were nine late deaths from disseminated residual recurrent carcinoma, and five patients are living and well after I to II years. DR. F. GRIFFITH PEARSON Toronto, Ontario, Canada
I had the opportunity to read the full manuscript of Dr. Nelson's work. I think it is a very well designed study and provides important new information. I would like to comment on our past experience with the use of a porous prosthesis for tracheal replacement. More than 20 years ago, Dr. Arthur Beall from Houston, Texas, reported on his experimental work and clinical experience with a porous prosthesis of heavy Marlex mesh. This is not the Marlex that is used for abdominal wall and chest wall defects, but a specially prepared heavier fiber. The material was initially developed by an engineer named Unger. The use of a porous prosthesis at that time was an ingenious and innovative introduction for tracheal replacement. I believe Dr. Beall has used it in several patients, some of whom have been long-term survivors. We first used this Marlex prosthesis in 1963 and in the first two patients had successful results. The prosthesis remained for 5V2 years in one
Tracheal prosthesis
807
patient and for 2V2 years in the second. Subsequently, however, we used this prosthesis in four consecutive patients, all of whom died of a fatal innominate artery erosion and tracheainnominate fistula within 3 weeks of operation. Our experimental studies were done in dogs and were evaluating the same potential for epithelialization of the inner surface of the prosthesis. We found that epithelialization progressed from each of the tracheal ends, and as Dr. Nelson has observed, epithelialization occurred only after a vascularized bed of granulation tissue had formed within the endoluminal aspect of the Marlex mesh. The entire inner surface of the Marlex mesh became covered with vascularized granulation tissue within 4 to 6 weeks of prosthetic replacement. Epithelialization then proceeded from each of the cut tracheal ends and slowly extended toward the central part of the graft. As long as there was no epithelial cover on the granulation tissue, the granulations became progressively thicker, and since the central part of the prosthesis was the last to be covered with epithelium, the animals developed a thick ring of scar at this level resulting in a degree of concentric stenosis. We were using a prosthesis 6 ern long in these animals. We observed the same changes in granulation tissue as those described by Dr. Nelson: Once an epithelial cover was achieved, the inflammatory reaction in the granulations subsided and left a relatively thin layer of mature collagen and scar. The current challenge is to identify mechanisms which would lead to early epithelial cover throughout the entire length of the prosthesis. In this regard, the work which Dr. John Burke of the Massachusetts General Hospital is doing with the production of "artificial skin" in burn patients may find an application for early epithelialization of the inner surface of a porous prosthesis. Dr. Burke's artificial skin consists of a dermal layer of synthetic collagen having the physical structure of dermis. Recently, a suspension of individual epidermal cells has been placed on the surface of the artificial dermis and covered with a thin surface layer of silicone, and an organized epidermal layer has covered the surface. It might be possible to apply this type of cover to the inner surface of the prosthesis once complete permeation with granulation tissue had occurred. I agree with the observation of Dr. Neville that the experimental studies using a 3 ern long prosthesis are not optimal. The indications for prosthetic replacement of the trachea are rare and, when they do occur, imply removal of very long tracheal segments. Indeed, if tracheal replacement were as common a problem as coronary heart disease, there would likely be a concentration of effort and resources which might result in the development of much more successful prostheses-that would end up with names like Starr, Bjork, and Hancock. The persistent problems with prosthetic replacement relate to the rigidity of the material used and the potential risk of innominate artery erosion. I think the concept of a porous prosthesis which can be epithelialized is attractive. Pore size may be an important factor and warrants further evaluation. I wish to ask two questions of the authors. First, I do not
808
Nelson et al.
clearly understand the term "bioelectric." Could you define that for me? Second, in the manuscript Dr. Nelson describes some decay or deterioration in the prosthetic material. Could you expand on that observation for us? DR. NELSON (Closing) This is not a potentially lucrative area. However, it is a project that meets a real need, and advances will have other potential applications where healing occurs in a contaminated environment and where tubular porous prostheses are used. Dr. Neville has alluded to our dependence upon spinoff from other areas. We are using a new polymer, Tecoflex, in our latest series of prostheses. Bronchoscopy at 4 weeks of a Tecoflex prosthesis shows excellent epithelialization and an extensive subepithelial network of vessels.
The Journal of Thoracic and Cardiovascular Surgery
Bioelectric polyurethane is a polyester urethane that contains a dispersion of carbon black. We have noticed degeneration at about 2 years in the tracheal site. This is a problemof polyurethanes in biological applications, and we are working with a very high surface area-to-mass ratio. We are hopeful that advances in polymer chemistry will surmount this problem. We are pursuing a stepwise process to a new tracheal prosthesis and have used 3 cm lengths to work out some of the basic problems. We are now beginning to use 6 em lengths experimentally. Erosion of the innominate artery may be preventable by using a similar porous wrap of polyurethane or silicone rubber to protect it.
THoRAc CARDIOVASC SURG
86:800-808, 1983
Original Communications
Neovascularity of a tracheal prosthesis/tissue complex Permanent bioincorporation of a microporous tracheal prosthesis will require a stable blood supply to connective tissue supporting an epithelial surface. In experience with over 80 tracheal implants in dogs, we have observed that end-on ingrowth and epithelialization does not occur in the absence of lateral ingrowth, epithelialization is marked by the appearance of a subepithelial network of vessels, and this process must be well advanced by 6 to 8 weeks for long-term stability. These observations were extended by using microangiography to delineate the blood supply of the prosthesis/tissue complex. Six implants of bioelectric polyurethane with 10 % gentamicin (3 em length, 2 cm diameter, I to 1.25 mm wall thickness, 60 to 120 J.L micropore diameter) were interposed in the dog thoracic trachea and wrapped with an omental pedicle. The aorta was perfused with a barium suspension at elective sacrifice between 10 weeks and 21 months. Radiographs of specimens were correlated with bronchoscopic, gross, and histopathological findings. Neovascularity of the prosthesis/tissue complex can be described in three categories: outer capsule, prosthetic wall, and inner lining. Outer capsule vessels were oriented circumferentially immediately adjacent to the prosthetic wall. They resembled arteries up to 75 J.L diameter on microscopy and appeared to develop from the omentum with connections developing to the bronchial circulation. Prosthetic wall vessels up to 75 J.L with thin muscular walls were noted to traverse the porous prosthetic wall. The inner lining had a network of subepithelial vessels that connected to the lamina propria vasculature of the native trachea across the anastomoses with vessels up to 120 J.L in diameter .. We conclude that the omentum provides an immediate blood supply and a base for early connective tissue ingrowth. Epithelialization occurs as early as 3 weeks on the favorable bed, accompanied by vascular connections to the existing lamina propria tracheal vessels. This dual organization of blood supply with connections across the prosthetic wall is probably important to long-term stability of healing.
Ronald J. Nelson, M.D., Lise Goldberg, M.D. (by invitation), Rodney A. White, M.D. (by invitation), Edwin Shors, Ph.D. (by invitation), and Frank M. Hirose, M.D. (by invitation), Torrance and Los Angeles, Calif.
Reconstruction of the trachea following extensive resection continues to challenge surgeons. Ingenious From the Departments of Surgery and Pathology, Harbor/UCLA Medical Center, Torrance, California and the UCLA School of Medicine, Los Angeles, Calif. Supported in part by Grant No. GM-24300 from the National Institutes of Health. Presented in part at the Sixty-eighth Annual Clinical Congress of the American College of Surgeons, Chicago, Ill., 1982. Read at the Sixty-third Annual Meeting of The American Association for Thoracic Surgery, Atlanta, Ga., April 25-27, 1983. Address for reprints: Ronald J. Nelson, M.D., Harbor/UCLA Medical Center, 1000 West Carson St., Torrance, Calif. 90509.
800
methods of mobilization have been devised to allow primary anastomosis in lieu of a satisfactory prosthesis.I, 2 The most successful clinical device for circumferential replacement has been a nonporous silicone rubber tube.' However, long-term healing of porous prostheses has been limited by chronic infection at the contaminated interface with progressive granulation tissue and stenosis. We4.5 have pursued the development of a prosthetic trachea with a unique, interconnecting microporous structure fabricated by using the microporous exoskeleton of marine corals and echinoderms as a template for biomedical polymers. The versatility of the model may
Volume 86 Number 6 December, 1983
Tracheal prosthesis
80 1
Fig. 1. Photographs taken at bronchoscopy through a Stortz telescope of a bioelectric polyurethane prosthesis/ tissue complex (A) at II weeks and (B) at 22 months, There is a subepithelial network of vessels connecting across the anastomoses with vessels of the lamina propria of the native trachea. Stable healing over the interval is evident. allow the gradual evolution of a successful replacement as favorable factors are identified and problems solved. The design allows independent control of major factors such as polymer, pore diameter, configuration, and methods of reinforcement. In addition, substances can be dispersed in the polymer or coated on its surface and temporary fillers can be used in the porous phase.
Previous experience Significant ingrowth was first shown into the outer surface (lateral ingrowth) of experimental bioelectric polyurethane* prostheses with a wall thickness of 3 to 5 rom at I month but was not present at 2 months, the pores being filled instead with inflammatory cells. Silicone rubber] grafts at the same configuration and pore diameter failed to show ingrowth at either time period. Both silicone rubber and bioelectric polyurethane grafts with 18 to 25 fJ, pores and 2 rom wall thickness showed no ingrowth. Complete incorporation of the prosthetic wall at the tracheal site was first achieved by wrapping an omental pedicle about the test prothesis (wall thickness 1.25 to 2 'Bioelectric polyurethane, Goodyear Tire and Rubber Co., Akron, Ohio. tSilicone rubber, MDX-4 4210, Dow Coming Corp., Midland, Mich.
rom) within a silicone rubber envelope that was positioned against the mediastinum and prevented the formation of adhesions. The bioincorporated graft was later rotated on its pedicle into a circumferential tracheal defect as a second procedure. 6 When the second procedure was done between 29 and 81 days after the first, extensive epithelialization occurred in all of the bioelectric polyurethane prostheses. Experience with silicone rubber was not as consistent. The appearance of a subepithelial network of vessels at bronchoscopy was noted to be a marker for epithelialization. This network connected with that of the native lamina propria across the anastomoses. When extensive epithelialization failed to occur, circumferential granulation tissue progressed to stenosis within 1 to 3 months. At the same 1.25 to 2 rom wall thickness in bioelectric polyurethane prostheses with a 60 to 120 fJ, pore size, lateral ingrowth with extensive epithelialization was achieved at direct implantation when the prosthesis was wrapped with omentum in six of 10 dogs.' In the absence of lateral ingrowth in all experimental series, proliferative granulation tissue at the anastomoses has led to stenosis without significant ingrowth along the inner wall of the prosthesis. A selected animal in this series, followed up 2Vz years, was found to have stable healing. Photographs taken during bronchoscopy at 11 weeks and 29 months are shown in Fig. 1. At 11 weeks,
The Journal of Thoracic and Cardiovascular Surgery
8 0 2 Nelson et al.
intervals and were electively put to death at 10, 10, 15, and 15 weeks and 18 and 21 months. At the time of death, the segment of the aorta including the arch vessels and the bronchial and celiac arteries was perfused with a suspension of micronized barium sulfate in normal saline until the tissues turned white," The segment of the trachea containing the prostheses was resected and examined grossly. The prosthesis and adjacent trachea was divided longitudinally. One half was further sectioned transversely in its midportion, Roentgenograms of these sections were taken in a Faxitron unit at 30 kvp for I minute. After formalin fixation, sections were stained for histologic study with hematoxylin and eosin. .
Results Fig. 2. Experimental tracheal prosthesis of microporous bioelectricpolyurethane (length 3 em) with a porediameter range of 60 to 120 IJ.. complete epithelialization is shown with blood vessels crossing the anastomoses and connecting the lamina propria of the native trachea with the subepithelial network of the prosthesis. Though blood vessels are not as prominent, stable healing is seen at 29 months. At 2Y2 years, discoloration of the lining suggested breakdown of the polyester urethane containing a dispersion of carbon black (bioelectric polyurethane). This success with direct implantation led to the series reported here, in which the wall thickness was reduced to I to 1.25 mm to favor bioincorporation and 10% gentamicin by weight was dispersed in the polymer to suppress bacterial growth."
Material and methods Bioelectric polyurethane prostheses with a micropore size of 60 to 120 JL (3 em length, 2 em internal diameter, I to 1.25 mm wall thickness), reinforced with stainless steel rings were implanted into three-ring thoracic tracheal defects in mongrel dogs (Fig 2). Polyglactin 910* interrupted sutures were used for the anastomoses. Omentum was mobilized through an upper midline abdominal incision, brought through a peripheral incision in the diaphragm and wrapped about the implants. All animals received humane animal care in compliance with the "Principles of Laboratory Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences. Six animals were followed up by bronchoscopy at *Vicryi, Ethicon, Inc., Somerville, N. J.
The neovascularity of the' prosthesis/tissue complex will be described in three categories: outer capsule, prosthetic wall, and inner lining. The outer capsule was provided at the time of implantation by wrapping the omentum circumferentially about the prosthesis. The predominant circumferential orientation of the outer capsule vessels is demonstrated on x-ray films of both longitudinal and transverse sections of the specimens (Fig. 3). On the longitudinal roentgenogram well-developed connections across the anastomoses are seen between the bronchial arterial supply of the native trachea and the omentally derived arterial supply of the prosthesis. On transverse sections, the larger outer capsule vessels are closely applied to the outer surface of the prosthetic wall and extend for long segments of the circumference. These vessels are muscular arteries up to 75 JL in diameter on histologic study (Fig. 4). This pattern of outer capsule vascular supply to the prosthesis/tissue complex was seen at each of the time intervals. Microangiography of sections demonstrates tortuous vessels traversing the prosthetic wall at multiple areas (Fig. 3, B). These vessels appear to feed a capillary-like network in the inner lining. On microscopy, vessels with thin muscular walls are noted in the collagenous soft tissue filling the pores. These can be observed on some sections connecting with the outer capsule, as shown in Fig. 5, or the inner lining. Extensive lateral ingrowth of granulation tissue was noted on bronchoscopy as early as 13 days (Fig. 6). This surface generally became epithelialized between 3 to 4 weeks with stratified squamous epithelium. With epithelialization the surface became smooth and translucent and the pink-red color intensity of the granulation tissue faded. The change was also marked by the appearance of a subepithelial network of vessels very similar to those of the lamina propria of the native trachea (Fig. 7).
Volume 86 Number 6
Tracheal prosthesis 8 0 3
December, 1983
Fig. 3. Faxitron x-ray films (30 kYP, 1 minute) of a prosthesis/tissue complex in which the arterial supply was perfused with a barium suspension at the time of sacrifice at 10 weeks. A, A longitudinal half with the inner lining not present to show the arterial supply of the native trachea and the outer capsule with predominant circumferential orientation of the latter (derived from the omental pedicle wrap) and well-developed connections (c) to the native bronchial circulation (wire reinforcing rings are approximately 0.89 mm in diameter). B, A transverse section showing the circumferential arteries of the outer capsule (OC) to be positioned next to the prosthetic wall (PW), with smaller arteries trav.ersing the prosthetic wall in a tortuous course to supply a capillary-like network in the inner lining (IL).
Extensive connections were noted .across both anastomoses, some vessels extending up to 1 to 2 em in length. On microscopy, thin-walled vessels up to 120 JL in diameter were seen. Normal appearing respiratory epithelium was noted to extend up to 7 mm into the prostheses at later time intervals (Fig. 8). Each of the six prostheses showed complete or near complete lateral ingrowth with epithelialization. However, the relatively thin wall (1 to 1.25 mm) and possible weakening of the polyurethane by the dispersion of 10% gentamicin led to accordian collapse of the prosthetic wall and thickening of scar tissue in the ''valleys'' of the collapse. Although the lumen was narrowed by this
accordian collapse, none of the animals had clinical problems related to stenosis even when followed up for 18 and 21 months. Construction of the anastomoses with absorbable sutures to avoid the problem of suture granuloma resulted in late protrusion of the ends of the prostheses into the lumen with subsequent regression of the inner lining at those areas.
Discussion Complete bioincorporation with epithelialization of an experimental tracheal prosthesis has been achieved with an interconnecting microporous structure with a pore diameter range of 60 to 120 JL. This pore structure
The Journal of Thoracic and Cardiovascular Surgery
8 04 Nelson et al.
,
f
•
Fig. 4. Photomicrograph of a muscular artery of the outer capsule in a specimen at 18 months. (Hematoxylin and eosin stain; original magnification X40.)
Fig. 5. Photomicrograph of a specimen at 15 weeks showing the fibrocollagenous tissue filling the pores of the prosthetic wall with a muscular artery at the junction of the prosthetic wall with the outer capsule. (Hematoxylin and eosin stain; original magnification X40.)
Volume 86 Number 6 December. 1983
Tracheal prosthesis
805
Fig. 6. Photograph taken at bronchoscopy through a Stortz telescope of a prosthesis 13 days following implantation. There is extensive lateral ingrowth of granulation tissue. Fig. 7. Photograph taken through a Stortz telescope at bronchoscopy showing complete epithelialization of a prosthesis at 16 weeks, with a subepithelial vascular network connecting with that of the lamina propria of the native trachea across the anastomoses.
allows the development of arteries up to 75 Jl in diameter that traverse the wall connecting the blood supply of the outer capsule and the inner lining. Soft tissue invasion into the prosthetic wall has been found to regress without the presence of a stable inner lining. The more profuse soft tissue ingrowth associated with polyurethanes as compared to silicone rubber seems important to the successful bioincorporation of tracheal prostheses. lO The omental pedicle used as an external wrap supplies an immediate segmental blood supply to the newly inserted prosthesis and is a rich source of neovascularity and soft tissue ingrowth . I I The omental blood vessels position themselves adjacent to the prosthetic wall, and extensive through-and-through lateral ingrowth is seen as early as 13 days on bronchoscopy. This early bioincorporation provides a favorable bed for epithelialization, which is initially stratified squamous epithelium presumably migrating from the cut ends of the native tracheal mucosa, although we have not ruled out a metaplastic source. This early epithelialization appears to stabilize the inner lining. If epithelialization does not occur by 6 to 8 weeks, progression of granulation tissue to significant stenosis occurs in 8 to 12 weeks. The epithelial lining is associated with the develop-
ment of a subepithelial vascular network similar in pattern to that of the native trachea. Simultaneous vascular connections across the anastomoses develop with the vessels of the lamina propria. Respiratory epithelium migrates over the inner lining to a distance of about 7 mm. This end-on ingrowth has not been seen without favorable lateral ingrowth. Thus there is a vascular network of both the outer capsule and inner lining. These networks connect with the similar vascular supplies of the native trachea and with each other across the prosthetic wall. In this experimental series, the balance between factors favoring early ingrowth and those affecting the strength of the prostheses was upset in favor of ingrowth at the expense of longitudinal accordian collapse. Adequate longitudinal as well as radial reinforcement will be required to prevent contracture and thickening of relatively avascular scar tissue. The problem of protrusion associated with absorbable suture material further emphasizes the importance of the dual blood supply. With loss of vascular connections across the anastomosis, regression of the inner lining was noted. We conclude that the dual organ ization of neovascularity to the prosthesis/tissue complex is important to long-term stability of healing. This degree of tissue
The Journal 01 Thoracic and Cardiovascular Surgery
8 0 6 Nelson et al.
5
6
7
8
9
10
II
Fig. 8. Photomicrograph showing respiratory epithelium with a small vessel in the underlying connective tissue. (Hematoxylin and eosin stain ; original magnification XIOO.)
organization may be necessary to avoid hypoxic tissue gradients that are associated with proliferation of granulation tissue." Design factors favoring this dual organization need to be refined in future experimental prostheses. The late time course of healing, bioincorporation of longer prostheses, factors favoring respiratory versus stratified squamous epithelium, and quality control of the biologically derived wall are among the areas for further investigation. REFERENCES Grillo HC: Congenital lesions, neoplasms and injuries of the trachea, Gibbon's Surgery of the Chest, DC Sabiston Jr, FC Spencer eds., Philadelphia, 1976, W. B. Saunders Company, p 256 2 Grillo HC: Carinal reconstruction. Ann Thorac Surg 34:356-373, 1982 3 Neville WE, Bolanowski PJP, Soltanzadeh H : Prosthetic reconstruction of the trachea and carina. J Thorac Cardiavase Surg 72:525-538, 1976 4 White RA, Weber IN, White EW: Replamineform. A
12
new process for preparing porous ceramic, metal, and polymer prosthetic materials. Science 176:922-924, 1972 Nelson RJ, White RA, Lawrence RS, Walkinshaw MD, Fiaschetti FL, White EW, Hirose FM: Development of a microporous tracheal prosthesis. Trans Am Soc Artif Intern Organs 25:8-12, 1979 Holwick JL, Klein SR, Miranda RM, Hirose FM , White RA , Nelson RJ: Heterotopic bioincorporation of a porous tracheal prosthesis with second stage transfer to the trachea. Proc Eur Soc Artif Organs 7:200-204, 1980 Holwick JL, Klein SR, Sproat RW, Miranda RM, Hi~e FM, White RA, Nelson RJ: Healing of a new microporous tracheal prosthesis. Surg Forum 31:541-543, 1980 Goldberg L, White RA, Shors E, Hirose FM, Nelson RJ: Experience with dispersion of gentamicin in a microporous polymeric tracheal prosthesis, Surg Forum 33:40-42, 1982 Thompson DP, Moore TC, Fingerhut AG: A method for high resolution renal and splenic microangiography. Surg Gynecol Obstet 132:101-108, 1971 White RA , Hirose FM , Sproat RW, Lawrence RS, Nelson RJ: Histopathologic observations after short-term implantation of two porous elastomers in dogs. Biomaterials 2:171-176,1981 Morgan E, Lima 0 , Goldberg M, Ferdman A, Luk SK, Cooper JD: Successful revascularization of totally ischemic bronchial autografts with omental pedicle flaps in dogs, J THORAC CARDIOVASC S URG 84:204-210, 1982 Knighton Dr, Silver lA, Hunt TK : Regulation of woundhealing angiogenesis. Effect of oxygen gradients and inspired oxygen concentration. Surg 90:262-270, 1981
Discussion DR. WILLIAM E. NEVILLE N ewark , N.J.
I have followed Dr. Nelson's work and wholeheartedly have supported his endeavors over the years. His report is an excellent and sophisticated piece of work. Without question there is a need for a tracheal prosthesis . An artificial heart has been developed, as have a large variety of good heart valves which are continually being improved, and with Dr. Nelson and his associates' work we can look forward to another artificial trachea. I must admit this is not a profitable enterprise but it is intriguing. When we presented our results with a tracheal silicone prosthesis before this Association in 1976, I outlined what I thought were the prerequisites for a suitable implant. It should be airtight, have adequate consistency, be well accepted by the host and cause minimal inflammatory reaction but still be incorporated into the surrounding tissue, be impervious to fibroblastic and bacterial invasion of the lumen, and pennit ingrowth of epithelium along the inner surface. The silicone model has all of these features with the exception of an eventual epithelial inner surface. However, observations exper-
Volume 86 Number 6 December. 1983
imentally and clinically over the past 20 years convince me that this probably does not make any difference. The nonwettable, highly polished, smooth inner surface of the silicone prosthesis permits the easy removal of distal secretions. I have never seen encrustations intraluminally. However, Dr. Nelson's prosthesis satisfies all the requirements for an ideal tracheal substitute when it is wrapped with omentum. This does take time and it may be difficult for the patient to raise his secretions initially prior to the formation of the respiratory epithelial surface. Nevertheless, it is a step in the right direction, and that is what we as surgeons strive for. We have implanted a straight silicone prosthesis in 46 patients to date. Twenty-eight had a long primary stricture or recurrent stricture, two had tracheoesophageal fistula plus stricture, five had tracheal malacia, and II a malignancy. We have seen four anastomotic granulomas in the past 10 years which were around nonabsorbable sutures; two patients had subglottic granulomas which were fulgurated. Recently a proximal anastomosis dehisced when I used Dexon as the suture material. This suture material should never be used for a tracheal anastomosis. There was one early death from a cerebrovascular accident. The late deaths have been the result of cardiac arrhythmia, drug overdose, innominate artery erosion, and myocardial infarction. Four individuals with adenocystic carcinoma are still alive I to 6 years postoperatively, and one with adenocarcinoma is alive at I year. Additionally we placed three prostheses intraluminally as a stent. One of those patients is still alive 3 years following irradiation. In 14 patients with a bifurcated prosthesis four suture line granulomas occurred around nonabsorbable sutures, two shifts of the prosthesis occurred from recurrent carcinoma, there were nine late deaths from disseminated residual recurrent carcinoma, and five patients are living and well after I to II years. DR. F. GRIFFITH PEARSON Toronto, Ontario, Canada
I had the opportunity to read the full manuscript of Dr. Nelson's work. I think it is a very well designed study and provides important new information. I would like to comment on our past experience with the use of a porous prosthesis for tracheal replacement. More than 20 years ago, Dr. Arthur Beall from Houston, Texas, reported on his experimental work and clinical experience with a porous prosthesis of heavy Marlex mesh. This is not the Marlex that is used for abdominal wall and chest wall defects, but a specially prepared heavier fiber. The material was initially developed by an engineer named Unger. The use of a porous prosthesis at that time was an ingenious and innovative introduction for tracheal replacement. I believe Dr. Beall has used it in several patients, some of whom have been long-term survivors. We first used this Marlex prosthesis in 1963 and in the first two patients had successful results. The prosthesis remained for 5V2 years in one
Tracheal prosthesis
807
patient and for 2V2 years in the second. Subsequently, however, we used this prosthesis in four consecutive patients, all of whom died of a fatal innominate artery erosion and tracheainnominate fistula within 3 weeks of operation. Our experimental studies were done in dogs and were evaluating the same potential for epithelialization of the inner surface of the prosthesis. We found that epithelialization progressed from each of the tracheal ends, and as Dr. Nelson has observed, epithelialization occurred only after a vascularized bed of granulation tissue had formed within the endoluminal aspect of the Marlex mesh. The entire inner surface of the Marlex mesh became covered with vascularized granulation tissue within 4 to 6 weeks of prosthetic replacement. Epithelialization then proceeded from each of the cut tracheal ends and slowly extended toward the central part of the graft. As long as there was no epithelial cover on the granulation tissue, the granulations became progressively thicker, and since the central part of the prosthesis was the last to be covered with epithelium, the animals developed a thick ring of scar at this level resulting in a degree of concentric stenosis. We were using a prosthesis 6 ern long in these animals. We observed the same changes in granulation tissue as those described by Dr. Nelson: Once an epithelial cover was achieved, the inflammatory reaction in the granulations subsided and left a relatively thin layer of mature collagen and scar. The current challenge is to identify mechanisms which would lead to early epithelial cover throughout the entire length of the prosthesis. In this regard, the work which Dr. John Burke of the Massachusetts General Hospital is doing with the production of "artificial skin" in burn patients may find an application for early epithelialization of the inner surface of a porous prosthesis. Dr. Burke's artificial skin consists of a dermal layer of synthetic collagen having the physical structure of dermis. Recently, a suspension of individual epidermal cells has been placed on the surface of the artificial dermis and covered with a thin surface layer of silicone, and an organized epidermal layer has covered the surface. It might be possible to apply this type of cover to the inner surface of the prosthesis once complete permeation with granulation tissue had occurred. I agree with the observation of Dr. Neville that the experimental studies using a 3 ern long prosthesis are not optimal. The indications for prosthetic replacement of the trachea are rare and, when they do occur, imply removal of very long tracheal segments. Indeed, if tracheal replacement were as common a problem as coronary heart disease, there would likely be a concentration of effort and resources which might result in the development of much more successful prostheses-that would end up with names like Starr, Bjork, and Hancock. The persistent problems with prosthetic replacement relate to the rigidity of the material used and the potential risk of innominate artery erosion. I think the concept of a porous prosthesis which can be epithelialized is attractive. Pore size may be an important factor and warrants further evaluation. I wish to ask two questions of the authors. First, I do not
808
Nelson et al.
clearly understand the term "bioelectric." Could you define that for me? Second, in the manuscript Dr. Nelson describes some decay or deterioration in the prosthetic material. Could you expand on that observation for us? DR. NELSON (Closing) This is not a potentially lucrative area. However, it is a project that meets a real need, and advances will have other potential applications where healing occurs in a contaminated environment and where tubular porous prostheses are used. Dr. Neville has alluded to our dependence upon spinoff from other areas. We are using a new polymer, Tecoflex, in our latest series of prostheses. Bronchoscopy at 4 weeks of a Tecoflex prosthesis shows excellent epithelialization and an extensive subepithelial network of vessels.
The Journal of Thoracic and Cardiovascular Surgery
Bioelectric polyurethane is a polyester urethane that contains a dispersion of carbon black. We have noticed degeneration at about 2 years in the tracheal site. This is a problemof polyurethanes in biological applications, and we are working with a very high surface area-to-mass ratio. We are hopeful that advances in polymer chemistry will surmount this problem. We are pursuing a stepwise process to a new tracheal prosthesis and have used 3 cm lengths to work out some of the basic problems. We are now beginning to use 6 em lengths experimentally. Erosion of the innominate artery may be preventable by using a similar porous wrap of polyurethane or silicone rubber to protect it.