- Email: [email protected]
A new model of tracheal stenosis and its repair with free periosteal grafts
J
THoRAc CARDIOVASC SURG
92:296-304, 1986
A new model of tracheal stenosis and its repair with free periosteal grafts Stenosis of 50 % to 80 % of the cervical or thoracic tracheal lumen was created in 17 piglets by encircling the trachea with silicone rubber sheet for 3 to 19 days (median 14 days). In 13 animals thoracic tracheal stenosis was repaired by longitudinal incision and insertion of a free tibial periosteal graft into the defect. An omental pedicle graft was applied to the surface of the free periosteal graft to augment its blood supply in three of these animals. An endotracheal silicone rubber splint was left in place for 2 to 3 weeks after the operations. Four animals served as untreated controls. In 12 of 13 piglets respiratory obstruction was relieved. One repair failed because of graft necrosis. In two animals put to death 1 month and 2 months later, stenosis was absent and the grafts showed variable ossification. The luminal surface had epithelialized. In nine of 10 animals put to death 3 months after the operation, tracheal cross-sectional areas had doubled, stenosis and granulation tissue were absent, and aU the grafts had ossified and epithelialized. A mild stenosis (24 %) developed over a 2 mm length in one animal and the graft had not ossified. The omental pedicle graft did not improve vascularization of the free periosteal graft. These studies describe a model of tracheal stenosis that can be used to assess various methods of repair. The results indicate that tracheal stenosis can be successfuUy treated by incision and insertion of a free periosteal graft into the defect.
Ralph C. Cohen, M.D., Robert M. Filler, M.D., Kunio Konuma, M.D., Andre Bahoric, M.D., Geraldine Kent, M.D., and Charles Smith, M.D., Toronto. Ontario, Canada
L
exte~sive
date, most attempts at reconstruction of tracheal stenosis in children have generally been unsuccessful. 1-3 Although resection of the stenotic segment and end-to-end anastomosis is recognized as the ideal treatment for tracheal stenosis," it is often not feasible because of the extent of the stenosis. Resection of the stenosis and prosthetic replacement of the trachea are unsuitable, because the foreign material becomes infected and granulation tissue forms causing obstruction.' Recently, successful repair of tracheal stenosis has been achieved by longitudinal incision of the stenotic segment and enlargement of the diameter by insertion of a free graft of costal cartilage in one case" and pericarFrom the Departments of Surgery and Pathology and the Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada M5G IX8. Supported by a grant from the Physician's Services Inc. Foundation (P.S.I.) and Smith Kline and French Canada Ltd. Received for publication July 10, 1985. Accepted for publication Oct. 23, 1985. Address for reprints: Dr. R. M. Filler, The Hospital for Sick Children, 555 University Ave., Toronto, Ontario, Canada M5G IX8.
296
dium in three others.' In a previous pilot study, we8 successfully repaired thoracic tracheal defects in pigs with free periosteal grafts. In the current study, the same grafting technique was used to treat an experimental model of tracheal stenosis that both produces symptoms of respiratory obstruction and histopathologically resembles congenital tracheal stenosis seen clinically. Materials and methods Tracheal stenosis model. Tracheal stenosis was produced in 17 piglets by encircling the thoracic trachea in 15 piglets (7.7 ± l.l kg) and the cervical trachea in two piglets (2 kg) with an appropriate sized piece of silicone rubber sheeting. All animals received humane care in compliance with the "Principles of Laboratory Animal 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 and published by the National Institutes of Health (NIH Publication No. 80-23, revised 1978). When severe respiratory obstruction developed from 3 to 19 days later (median 14 days), all animals were
Volume 92 Number 2 August, 1986
Tracheal stenosis
297
Fig. 1. A, Silicone rubber sheet used to encircle the trachea. B, Silicone rubber cuff (5) encircling the thoracic trachea (T). H, Heart. V. Vagus nerve.
reoperated on and the silicone rubber cuff was removed. In four control animals no further treatment was given. When it was established that respiratory obstruction was not relieved, the pigs were put to death. In 13 piglets operative repair of the thoracic tracheal stenosis was undertaken immediately after removal of the cuff. Operative techniques. All animals were premedicated with 0.3 mg of atropine sulfate intramuscularly and anesthetized with nitrous oxide, oxygen, and halothane through an endotracheal tube or an anesthetic mask when cervical tracheal stenosis was created. The skin was shaved and scrubbed with chlorhexidine soap; 70% alcohol and 10% povidone-iodinewere applied just before the incision was made. Cervical tracheal stenosis was created in this way: The entire cervical trachea was exposed by a midline incision with the strap muscles being retracted laterally and the thyroid gland inferiorly. The circumference and diameter of the trachea were measured. The trachea was encircled with a silicone rubber sheet 1 mm thick and 2 em long (nine tracheal rings). The sheet circumference was 3 mm greater than the tracheal circumference so that it did not constrict the trachea. The free margins of the silicone rubber sheet were sutured to
each other with four 4-0 Prolene sutures that also included a superficial "bite" of the underlying trachea to prevent migration of the cuff. The wound was closed in layers with 4-0 Vicryl sutures and continuous 3-0 Prolene sutures to the skin. The thoracic tracheal stenosis was created by exposing the entire thoracic trachea proximal to the right upper lobe bronchus via a right lateral extrapleural thoracotomy through the third intercostal space. This gave complete access to the entire thoracic trachea proximal to the right upper lobe bronchus. The tracheal diameter and circumference were measured and a silicone rubber sheet was applied to the circumference of the trachea as described earlier (Fig. 1). The ribs were approximated with 2-0 Vicryl pericostal sutures. The muscles were closed in layers with 3-0 Vicryl sutures, and 3-0 Prolene sutures were used for skin. No chest drain was used. When the second operation became necessary because of airway obstruction, the previous thoracotomy incision was opened. The pleura was entered and the lung was retracted inferiorly to reveal the pseudocapsule surrounding the silicone rubber cuff. This was incised longitudinally and the cuff was removed. The stenosis was repaired through a longitudinal
The Journal of
298
Cohen et al.
Thoracic and Cardiovascular Surgery
Fig. 2. A, Endotracheal silicone rubber splint. B. Stenotic trachea incised longitudinally and silicone rubber splint (S) inserted.
incision in the right anterolateral wall of the trachea over the stenotic segment. A silicone rubber endotracheal splint was inserted through the tracheal defect to seal the airway (Fig. 2). The external diameter of this tube was 9.5 mm, the same as the internal diameter of the normal trachea above the stenosis. The splint was sutured in place with a 3-0 Vicryl suture so that its lower end did not occlude the origin of the right upper lobe bronchus. The right lung was inflated to confirm the patency of the right upper lobe bronchus. A vertical incision was made over the left tibia and a 3 by 1.5 em rectangular flap of tibial periosteum was excised. The periosteum was sutured to the margins of the tracheal defect with interrupted 5-0 Vicryl sutures, the osteogenic layer facing inward. The chest was closed as already described. In three of these piglets a vertical midline upper abdominal incision was made to mobilize an omental pedicle graft from the greater curve of the stomach. The omental pedicle graft
was rotated upward into the chest via the retrosternal route and sutured to the surface of the free periosteal graft with 5-0 Vicryl sutures. A size 16 chest tube was inserted through a separate stab incision, and after the chest was closed and the lungs were inflated, the chest tube was removed. All piglets were given cefazolin (70 mgjkg) intramuscularly 1 hour preoperatively and for 7 days after each operation. Three weeks after the repair, the piglets were anesthetized with 75 to 150 mg of sodium pentobarbital intravenously to remove the endotracheal splint. Under x-ray image intensifier control, the splint was grasped and removed with a peanut forceps that had been passed down the lumen of a 4.5 mm endotracheal tube. The trachea was then examined with a flexible fiberoptic brochoscope and a lateral chest x-ray film was obtained. Assessment of trachea. Lateral cervical or chest
Volume 92 Number 2
Tracheal stenosis 2 9 9
August, 1986
Fig. 3. Normal trachea above stenosis (left). Permanent stenosis created by invagination of membranous trachea (right). F, Fibrous tissue.
radiographs were taken when respiratory obstruction became evident. Tracheal stenosis was calculated by comparing the internal sagittal diameter of the normal trachea to the diameter at mid-stenosis. Accuracy of the radiographic assessment of tracheal stenosis was checked by comparing radiographic measurements to actual measurements made in the four control (untreated) animals when they were put to death. Transverse sections of the tracheas above the stenosis and at mid-stenosis were fixed, stained, and mounted on histopathology slides, and the internal cross-sectional areas were measured with the IBAS image analyzer (Carl Zeiss, Inc., Thornwood, N. Y.). Animals that underwent repair of the tracheal stenosis were put to death 3 months (n = 10), 2 months (n = 1), and 1 month (n = 1) after repair. The trachea was removed and radiographs were obtained with a Faxitron 43805 x-ray system (Hewlett-Packard Company, Andover, Mass.) with high-resolution Kodak XOmat TL film at 35 kvP for 35 msec, Transverse sections taken from the normal trachea 1 to 2 em above the repair site and at mid-repair level were fixed and stained with hematoxylin-phloxinesaffron and studied with a light microscope. Crosssectional area measurements were made from the histopathologic slides with the IBAS image analyzer. Sections were also examined by a scanning electron microscope. The internal cross-sectional area of the stenotic trachea at the time of repair was determined by indirect measurements calculated by the following equation: Cross-sectional area of stenotic trachea = cross-sectional area of normal trachea X percent stenosis (from x-ray film)
Radiographic measurements of the airway were used to determine the percent stenosis. The cross-sectional area of the normal trachea was determined from a linear
Fig. 4. Postmortem radiograph of trachea 14 days after application of silicone rubber cuff (70% stenosis).
regression equation that has been shown to relate external tracheal circumference of the normal pig trachea to its internal cross-sectional area." Area (mm) = -77.8
+ 3.79 X circumference (mm) : (R = 0.94)
The Journal of Thoracic and Cardiovascular
3 0 0 Cohen et al.
Surgery
Table I. Comparison of radiographic and cross-sectional area measurements of tracheal stenosis in control pigs Pig 1 2 3 4
(thoracic) (thoracic) (cervical) (cervical)
Days to death
Weight at death (kg)
15 14 13
7
Cross-sectional areas* (mnr] Above stenosis 39.8 48.8 13.5 18.8
7 2 2
13
± 0.3 ± 0.9 ± 0.3 ± 0.1
I
Mid-stenosis 16.5 14.6 4.7 5.6
Percent stenosis measured
Percent stenosis from x-ray films
60 70 65 70
70 80 70 70
± 0.8 ± 0.7 ± 0.1 ± 0.1
* Measured from transverse sections by IBAS image analyzer.
Table II. Tracheal cross-sectional areas at stenosis and when animal was put to death Cross-sectional areas (mm'} Death
Stenosis Days to death
Percent stenosis from x-ray films
Above stenosis*
10
92
II
82 90 90
70 80 70 70 80 50 80 70 70 60 70 70 70
66 55 62 40 32 32
8 9
26 90 90 90 64 22 90 87 88
Pig I 2
3 4
5 6 7
12 13
I
Midstenosisi
55 40 36
55 55 55 55
20 II 19 12 6 16 II 12 11
22 17 17 17
Above repair (control) 71.3 112.8 101.7 116.8 76.1 53.8 132.1 129.1 118.6 120.4 94.3 107.6 114.8
± 0.6 ± 1.6 ± 0.5 ± 0.9 ± 0.7 ± 0.7 ± 0.4 ± 1.2 ± 1.0 ± 0.6 ± 0.5 ± 1.3 ± 1.2
I
Midrepair 22.5 86.3 104.2 136.1 139.8 73.6 170J 184.8 189.8 191.5 114.4 150.7 147.2
± 0.2 ± 0.5 ± 0.9:\: ±0.7:\: ± 0.7:1: ± 0.3:\: ± 0.6:\: ± 0.7:\: ± 0.6:\: ± 0.8:\: ± 1.2:1: ± 0.9:\: ± OJ:\:
Ossification of graft NIl (FPG) Nil (FPG) Partial (FPG) Partial (FPG) Partial (FPG) Complete (FPG) Complete (FPG) Complete (FPG) Complete (FPG) Complete (FPG) Complete (FPGjOPG) Complete (FPGjOPG) Complete (FPGjOPG)
Legend: FPG. Free periosteal graft. OPG. Omental pedicle graft. "Derived from measurement of circumference and linear regression equation. tDcrived from radiographic measurement of percent stenosis and calculated cross-sectional area of trachea above stenosis. :j:p < 0.05.
Tracheal growth 3 months after repair was determined by comparing the cross-sectional area of the trachea before creation of stenosis with the crosssectional areas of trachea above the repair and at mid-repair. A paired t test was performed to assess significant differences. Results The stenosis model. The silicone rubber cuff caused overlap and infolding of the tracheal rings along the posterior aspect of the trachea. Tracheal stenosis became permanent when this invaginated membranous trachea became surrounded by fibrous tissue (Fig. 3). Airway obstruction became apparent a median of 14 days after application of the silicone rubber cuff. Radiographic assessment demonstrated 60% to 70% reduction in the airway diameter in the stenotic segment (Fig. 4). Removal of the cuff in four animals did not
relieve airway obstruction and these piglets were put to death. The percent airway reduction obtained by actual measurement of tracheal cross-sectional areas above the stenosis compared favorably with that determined by radiographic assessment (Table I). The repair model. The pigs were ambulatory and eating within 24 hours after the stenosis was repaired. There were no signs of airway obstruction at any stage after repair in 12 pigs. Problems with the repair were noted in two of the 13 animals. A persistent cough developed 3 weeks after the operation in one pig, and the animal died a week later. Postmortem examination revealed a Staphylococcus aureus chest wound infection and a concentric narrowing of the trachea to 70% of the lumen because of granulation tissue at the lower end of the graft extending for 3 mm. In another pig an unsuspected short segment of concentric stenosis (24% of the normal tracheal
Volume 92 Number 2
Tracheal stenosis
August, t 986
30 1
Fig. 5. Ossified graft with central marrow space (M). Area of previous infolding (I).
cross-sectional area over a 2 mm length) was noted when the animal was put to death 3 months after repair. In both animals the grafts had not calcified but were entirely replaced with fibrous tissue. The cut ends of the cartilaginous rings were in close proximity and contributed to the stenosis. However, in both, the crosssectional areas of the repaired segments above and below the stenotic area were greater than the normal trachea above the area of repair. Mean body weight increased from 7.7 ± 1.1 kg atthe original operation to 56.0 ± 9.3 kg when the animals were put to death 90 ± 2 days later. The nonoperated tracheal cross-sectional area more than doubled from an initial mean of 49 ± 10 rnm' to 114.8 ± 11.5 mm' 3 months later. At autopsy the internal cross-sectional areas of the trachea at the mid-repair site were significantly greater than the normal trachea above the repair, and this was evident as early as 22 days postoperatively (Table 11). The mean internal cross-sectional area of the stenosed trachea increased approximately tenfold from 14.1 ± 4.5 mm' to 147.5 ± 37.3 mm- 3 months after repair. The grafts retained their original dimensions and the increased size of the trachea was due to growth of the tracheal rings. Histologic review demonstrated ossification in 11 of the grafts. In eight there was complete ossification (Fig. 5) and in three ossification was patchy arid incomplete. An outer ring of cortical bone surrounded central marrow spaces with hematopoietic elements. Beneath the interposed osseous tissue and the lumen was a layer of well-organized fibrous tissue as thin as 0.04 mm in
places. Patches of ciliated epithelium were demonstrated by the scanning electron microscope 22 and 64 days after repair. After 3 months the grafts were completely lined with ciliated columnar epithelium in seven pigs (FIgs. 6A and 6B). In three animals both cuboidal and ciliated epithelium were present. In the pig that died, only cuboidal epithelium lined the luminal surface of the graft. The area of infolding that was responsible for the original tracheal stenosis was represented by a "knot" of cartilage at the posterior junction of the graft and tracheal cartilage (Fig. 5). The trachea above the graft demonstrated normal respiratory epithelium.
Discussion Congenital tracheal stenosis is a lethal condition characterized by an absence of the membranous portion of the trachea with fusion of each tracheal cartilage posteriorly throughout the length of the stenotic segment. The cartilaginous rings are abnormal in shape. Histologic sections show mild chronic inflammation in the submucosa and increased amounts of dense collagen in the wall of the trachea." In a previous experiment in pigs, we created thoracic tracheal defects similar to those that would be produced when congenital tracheal stenosis is treated by a longitudinal incision. These defects were successfully repaired with free tibial periosteal grafts. However, these operations were performed on normal tracheas. 8 This model of tracheal stenosis can be created in a short and predictable period of time and has features similar to congenital tracheal stenosis: The tracheal rings were permanently fused posteriorly by dense
The Journal of
3 0 2 Cohen et al.
Thoracic and Cardiovascular Surgery
Fig. 6A. Scanning electron micrograph of epithelium on luminal surface of the graft 22 days after repair (original magnification X4,800).
collagen at the point of infolding, obliterating the membranous portion; an inflammatory reaction was present around the stenotic segment of trachea and in the submucosa, and irreversible signs of airway obstruction occurred that resulted in the animal's death. The pig trachea resembles the human trachea anatomically and functionally.'? II Posteriorly, the tracheal muscle is attached to the slightly overlapping cartilaginous rings that normally separate as the trachea expands with inspiration and is compressed during "grunting." The silicone rubber cuff appears to induce stenosis by limiting tracheal growth and altering the normal dynamic activity of the trachea so that expanding and compressing forces are repeatedly applied to the immobile tracheal rings. The weak membranous segment infolds with resultant stenosis and intraluminal obstruction. There was nothing to suggest that the remnant of the original infolding impaired or distorted tracheal growth. Tibial periosteum that is used to repair the tracheal defect is readily available, tough, pliable, and can be
sutured to the margins of the tracheal defect to create an airtight seal. It is oriented so that the osteogenic layer faces the lumen, because previous experiments demonstrated excellent ossification with the graft oriented this way."Ossification of the graft occurred by postoperative day 22 and seemed to be an important factor for prevention of stenosis. Neither of the two pigs with stenosis had evidence of graft ossification. The pig that died had a wound infection that may have affected graft vascularization and ossification so that it was unable to support the repaired segment adequately. The approximate tenfold increase in mean crosssectional area of the repaired stenotic segment over 3 months was due to tracheal cartilage growth, possibly induced by the periosteal graft. This finding was similar to that reported in our previous studies," It suggests that if the complete rings in congenital tracheal stenosis are incised longitudinally and a periosteal graft inserted, growth of the tracheal rings may be induced by the graft. Areas of mature ciliated respiratory epithelium were
Volume 92 Number 2 August, 1986
Tracheal stenosis
303
Fig. 68. Scanning electron micrograph of ciliated epithelium on luminal surface of graft 3 months after repair (original magnification X4,800). seen over the graft surface as early as postoperative day 22. Further maturation and proliferation occurred, so that by 3 months the majority of the grafts were completely covered with mature respiratory epithelium. Previously, we8 found that the application of an omental pedicle graft is not necessary for the vascularization of the periosteal graft used to repair defects in the normal trachea. In this study complete ossification of the free periosteal graft occurred when the omental pedicle graft was applied to it, whereas ossification was incomplete or absent in five of 10 animals in which an
omental pedicle graft was not used. However, these differences are not statistically significant in this small series. These experiments describe a simple, reproducible technique for creating a model of tracheal stenosis. The stenosis occurs in a short time and has similarities to congenital tracheal stenosis seen in children. The successful repair by longitudinal incision, temporary internal splinting, and insertion of a free periosteal graft into the defect may be a suitable option for correction of congenital tracheal stenosis in children.
304
2
3 4 5
6
Cohen et al.
REFERENCES Fonkalsrud EW, Sumida S: Tracheal replacement with autologous esophagus for tracheal stricture. Arch Surg 102:139-142,1971 Kufaas T, Pasila M: Free periosteal transplants in tracheal reconstruction. Report on three cases. Z Kinderchir 15:371-387,1974 Benjamin B, Pitkin J, Cohen D: Congenital tracheal stenosis. Ann Otol Rhinol Laryngol 90:364-371, 1981 Grillo HC: Tracheal reconstruction. Indications and techniques. Arch Otolaryngol 93:31-39, 1972 Bailey BJ, Kosoy J: Observations in the development of the tracheal prosthesis and tracheal transplantation. Laryngoscope 80: 1553-1565, 1970 Kimura K, Mukohara N, Tsugawa C, Matsumoto Y, Sugimura C, Murata H, Itoh H: Tracheoplasty for congenital stenosis of the entire trachea. J Pediatr Surg 17:869-871, 1982
The Journal of Thoracic and Cardiovascular Surgery
7 Idriss FS, Serafin SY, I1bawi MN, Gerson CR, Tucker GF, Holinger L: Tracheoplasty with pericardial patch for extensive tracheal stenosis in infants and children. J THORAC CARDIOVASC SURG 88:527-536, 1984 8 Cohen RC, Filler RM, Konuma K, Bahoric A, Kent G, Smith C: The successful reconstruction of thoracic tracheal defects with free periosteal grafts. J Pediatr Surg 20:852-858, 1985 9 Cantrell JR, Guild HG: Congenital stenosisof the trachea. Am J Surg 108:297-305, 1964 10 Munoz NM, Cera LM, Leff AR: Distribution of innervation and circulation in porcine cervical trachea. Am J Vet Res 45:1937-1940, 1984 11 Hare WCD: Respiratory systems, The anatomy of the domestic animal, ed 5, R Getty, ed., Philadelphia, 1975, W. B. Saunders Company, pp 1283-1296.
THoRAc CARDIOVASC SURG
92:296-304, 1986
A new model of tracheal stenosis and its repair with free periosteal grafts Stenosis of 50 % to 80 % of the cervical or thoracic tracheal lumen was created in 17 piglets by encircling the trachea with silicone rubber sheet for 3 to 19 days (median 14 days). In 13 animals thoracic tracheal stenosis was repaired by longitudinal incision and insertion of a free tibial periosteal graft into the defect. An omental pedicle graft was applied to the surface of the free periosteal graft to augment its blood supply in three of these animals. An endotracheal silicone rubber splint was left in place for 2 to 3 weeks after the operations. Four animals served as untreated controls. In 12 of 13 piglets respiratory obstruction was relieved. One repair failed because of graft necrosis. In two animals put to death 1 month and 2 months later, stenosis was absent and the grafts showed variable ossification. The luminal surface had epithelialized. In nine of 10 animals put to death 3 months after the operation, tracheal cross-sectional areas had doubled, stenosis and granulation tissue were absent, and aU the grafts had ossified and epithelialized. A mild stenosis (24 %) developed over a 2 mm length in one animal and the graft had not ossified. The omental pedicle graft did not improve vascularization of the free periosteal graft. These studies describe a model of tracheal stenosis that can be used to assess various methods of repair. The results indicate that tracheal stenosis can be successfuUy treated by incision and insertion of a free periosteal graft into the defect.
Ralph C. Cohen, M.D., Robert M. Filler, M.D., Kunio Konuma, M.D., Andre Bahoric, M.D., Geraldine Kent, M.D., and Charles Smith, M.D., Toronto. Ontario, Canada
L
exte~sive
date, most attempts at reconstruction of tracheal stenosis in children have generally been unsuccessful. 1-3 Although resection of the stenotic segment and end-to-end anastomosis is recognized as the ideal treatment for tracheal stenosis," it is often not feasible because of the extent of the stenosis. Resection of the stenosis and prosthetic replacement of the trachea are unsuitable, because the foreign material becomes infected and granulation tissue forms causing obstruction.' Recently, successful repair of tracheal stenosis has been achieved by longitudinal incision of the stenotic segment and enlargement of the diameter by insertion of a free graft of costal cartilage in one case" and pericarFrom the Departments of Surgery and Pathology and the Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada M5G IX8. Supported by a grant from the Physician's Services Inc. Foundation (P.S.I.) and Smith Kline and French Canada Ltd. Received for publication July 10, 1985. Accepted for publication Oct. 23, 1985. Address for reprints: Dr. R. M. Filler, The Hospital for Sick Children, 555 University Ave., Toronto, Ontario, Canada M5G IX8.
296
dium in three others.' In a previous pilot study, we8 successfully repaired thoracic tracheal defects in pigs with free periosteal grafts. In the current study, the same grafting technique was used to treat an experimental model of tracheal stenosis that both produces symptoms of respiratory obstruction and histopathologically resembles congenital tracheal stenosis seen clinically. Materials and methods Tracheal stenosis model. Tracheal stenosis was produced in 17 piglets by encircling the thoracic trachea in 15 piglets (7.7 ± l.l kg) and the cervical trachea in two piglets (2 kg) with an appropriate sized piece of silicone rubber sheeting. All animals received humane care in compliance with the "Principles of Laboratory Animal 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 and published by the National Institutes of Health (NIH Publication No. 80-23, revised 1978). When severe respiratory obstruction developed from 3 to 19 days later (median 14 days), all animals were
Volume 92 Number 2 August, 1986
Tracheal stenosis
297
Fig. 1. A, Silicone rubber sheet used to encircle the trachea. B, Silicone rubber cuff (5) encircling the thoracic trachea (T). H, Heart. V. Vagus nerve.
reoperated on and the silicone rubber cuff was removed. In four control animals no further treatment was given. When it was established that respiratory obstruction was not relieved, the pigs were put to death. In 13 piglets operative repair of the thoracic tracheal stenosis was undertaken immediately after removal of the cuff. Operative techniques. All animals were premedicated with 0.3 mg of atropine sulfate intramuscularly and anesthetized with nitrous oxide, oxygen, and halothane through an endotracheal tube or an anesthetic mask when cervical tracheal stenosis was created. The skin was shaved and scrubbed with chlorhexidine soap; 70% alcohol and 10% povidone-iodinewere applied just before the incision was made. Cervical tracheal stenosis was created in this way: The entire cervical trachea was exposed by a midline incision with the strap muscles being retracted laterally and the thyroid gland inferiorly. The circumference and diameter of the trachea were measured. The trachea was encircled with a silicone rubber sheet 1 mm thick and 2 em long (nine tracheal rings). The sheet circumference was 3 mm greater than the tracheal circumference so that it did not constrict the trachea. The free margins of the silicone rubber sheet were sutured to
each other with four 4-0 Prolene sutures that also included a superficial "bite" of the underlying trachea to prevent migration of the cuff. The wound was closed in layers with 4-0 Vicryl sutures and continuous 3-0 Prolene sutures to the skin. The thoracic tracheal stenosis was created by exposing the entire thoracic trachea proximal to the right upper lobe bronchus via a right lateral extrapleural thoracotomy through the third intercostal space. This gave complete access to the entire thoracic trachea proximal to the right upper lobe bronchus. The tracheal diameter and circumference were measured and a silicone rubber sheet was applied to the circumference of the trachea as described earlier (Fig. 1). The ribs were approximated with 2-0 Vicryl pericostal sutures. The muscles were closed in layers with 3-0 Vicryl sutures, and 3-0 Prolene sutures were used for skin. No chest drain was used. When the second operation became necessary because of airway obstruction, the previous thoracotomy incision was opened. The pleura was entered and the lung was retracted inferiorly to reveal the pseudocapsule surrounding the silicone rubber cuff. This was incised longitudinally and the cuff was removed. The stenosis was repaired through a longitudinal
The Journal of
298
Cohen et al.
Thoracic and Cardiovascular Surgery
Fig. 2. A, Endotracheal silicone rubber splint. B. Stenotic trachea incised longitudinally and silicone rubber splint (S) inserted.
incision in the right anterolateral wall of the trachea over the stenotic segment. A silicone rubber endotracheal splint was inserted through the tracheal defect to seal the airway (Fig. 2). The external diameter of this tube was 9.5 mm, the same as the internal diameter of the normal trachea above the stenosis. The splint was sutured in place with a 3-0 Vicryl suture so that its lower end did not occlude the origin of the right upper lobe bronchus. The right lung was inflated to confirm the patency of the right upper lobe bronchus. A vertical incision was made over the left tibia and a 3 by 1.5 em rectangular flap of tibial periosteum was excised. The periosteum was sutured to the margins of the tracheal defect with interrupted 5-0 Vicryl sutures, the osteogenic layer facing inward. The chest was closed as already described. In three of these piglets a vertical midline upper abdominal incision was made to mobilize an omental pedicle graft from the greater curve of the stomach. The omental pedicle graft
was rotated upward into the chest via the retrosternal route and sutured to the surface of the free periosteal graft with 5-0 Vicryl sutures. A size 16 chest tube was inserted through a separate stab incision, and after the chest was closed and the lungs were inflated, the chest tube was removed. All piglets were given cefazolin (70 mgjkg) intramuscularly 1 hour preoperatively and for 7 days after each operation. Three weeks after the repair, the piglets were anesthetized with 75 to 150 mg of sodium pentobarbital intravenously to remove the endotracheal splint. Under x-ray image intensifier control, the splint was grasped and removed with a peanut forceps that had been passed down the lumen of a 4.5 mm endotracheal tube. The trachea was then examined with a flexible fiberoptic brochoscope and a lateral chest x-ray film was obtained. Assessment of trachea. Lateral cervical or chest
Volume 92 Number 2
Tracheal stenosis 2 9 9
August, 1986
Fig. 3. Normal trachea above stenosis (left). Permanent stenosis created by invagination of membranous trachea (right). F, Fibrous tissue.
radiographs were taken when respiratory obstruction became evident. Tracheal stenosis was calculated by comparing the internal sagittal diameter of the normal trachea to the diameter at mid-stenosis. Accuracy of the radiographic assessment of tracheal stenosis was checked by comparing radiographic measurements to actual measurements made in the four control (untreated) animals when they were put to death. Transverse sections of the tracheas above the stenosis and at mid-stenosis were fixed, stained, and mounted on histopathology slides, and the internal cross-sectional areas were measured with the IBAS image analyzer (Carl Zeiss, Inc., Thornwood, N. Y.). Animals that underwent repair of the tracheal stenosis were put to death 3 months (n = 10), 2 months (n = 1), and 1 month (n = 1) after repair. The trachea was removed and radiographs were obtained with a Faxitron 43805 x-ray system (Hewlett-Packard Company, Andover, Mass.) with high-resolution Kodak XOmat TL film at 35 kvP for 35 msec, Transverse sections taken from the normal trachea 1 to 2 em above the repair site and at mid-repair level were fixed and stained with hematoxylin-phloxinesaffron and studied with a light microscope. Crosssectional area measurements were made from the histopathologic slides with the IBAS image analyzer. Sections were also examined by a scanning electron microscope. The internal cross-sectional area of the stenotic trachea at the time of repair was determined by indirect measurements calculated by the following equation: Cross-sectional area of stenotic trachea = cross-sectional area of normal trachea X percent stenosis (from x-ray film)
Radiographic measurements of the airway were used to determine the percent stenosis. The cross-sectional area of the normal trachea was determined from a linear
Fig. 4. Postmortem radiograph of trachea 14 days after application of silicone rubber cuff (70% stenosis).
regression equation that has been shown to relate external tracheal circumference of the normal pig trachea to its internal cross-sectional area." Area (mm) = -77.8
+ 3.79 X circumference (mm) : (R = 0.94)
The Journal of Thoracic and Cardiovascular
3 0 0 Cohen et al.
Surgery
Table I. Comparison of radiographic and cross-sectional area measurements of tracheal stenosis in control pigs Pig 1 2 3 4
(thoracic) (thoracic) (cervical) (cervical)
Days to death
Weight at death (kg)
15 14 13
7
Cross-sectional areas* (mnr] Above stenosis 39.8 48.8 13.5 18.8
7 2 2
13
± 0.3 ± 0.9 ± 0.3 ± 0.1
I
Mid-stenosis 16.5 14.6 4.7 5.6
Percent stenosis measured
Percent stenosis from x-ray films
60 70 65 70
70 80 70 70
± 0.8 ± 0.7 ± 0.1 ± 0.1
* Measured from transverse sections by IBAS image analyzer.
Table II. Tracheal cross-sectional areas at stenosis and when animal was put to death Cross-sectional areas (mm'} Death
Stenosis Days to death
Percent stenosis from x-ray films
Above stenosis*
10
92
II
82 90 90
70 80 70 70 80 50 80 70 70 60 70 70 70
66 55 62 40 32 32
8 9
26 90 90 90 64 22 90 87 88
Pig I 2
3 4
5 6 7
12 13
I
Midstenosisi
55 40 36
55 55 55 55
20 II 19 12 6 16 II 12 11
22 17 17 17
Above repair (control) 71.3 112.8 101.7 116.8 76.1 53.8 132.1 129.1 118.6 120.4 94.3 107.6 114.8
± 0.6 ± 1.6 ± 0.5 ± 0.9 ± 0.7 ± 0.7 ± 0.4 ± 1.2 ± 1.0 ± 0.6 ± 0.5 ± 1.3 ± 1.2
I
Midrepair 22.5 86.3 104.2 136.1 139.8 73.6 170J 184.8 189.8 191.5 114.4 150.7 147.2
± 0.2 ± 0.5 ± 0.9:\: ±0.7:\: ± 0.7:1: ± 0.3:\: ± 0.6:\: ± 0.7:\: ± 0.6:\: ± 0.8:\: ± 1.2:1: ± 0.9:\: ± OJ:\:
Ossification of graft NIl (FPG) Nil (FPG) Partial (FPG) Partial (FPG) Partial (FPG) Complete (FPG) Complete (FPG) Complete (FPG) Complete (FPG) Complete (FPG) Complete (FPGjOPG) Complete (FPGjOPG) Complete (FPGjOPG)
Legend: FPG. Free periosteal graft. OPG. Omental pedicle graft. "Derived from measurement of circumference and linear regression equation. tDcrived from radiographic measurement of percent stenosis and calculated cross-sectional area of trachea above stenosis. :j:p < 0.05.
Tracheal growth 3 months after repair was determined by comparing the cross-sectional area of the trachea before creation of stenosis with the crosssectional areas of trachea above the repair and at mid-repair. A paired t test was performed to assess significant differences. Results The stenosis model. The silicone rubber cuff caused overlap and infolding of the tracheal rings along the posterior aspect of the trachea. Tracheal stenosis became permanent when this invaginated membranous trachea became surrounded by fibrous tissue (Fig. 3). Airway obstruction became apparent a median of 14 days after application of the silicone rubber cuff. Radiographic assessment demonstrated 60% to 70% reduction in the airway diameter in the stenotic segment (Fig. 4). Removal of the cuff in four animals did not
relieve airway obstruction and these piglets were put to death. The percent airway reduction obtained by actual measurement of tracheal cross-sectional areas above the stenosis compared favorably with that determined by radiographic assessment (Table I). The repair model. The pigs were ambulatory and eating within 24 hours after the stenosis was repaired. There were no signs of airway obstruction at any stage after repair in 12 pigs. Problems with the repair were noted in two of the 13 animals. A persistent cough developed 3 weeks after the operation in one pig, and the animal died a week later. Postmortem examination revealed a Staphylococcus aureus chest wound infection and a concentric narrowing of the trachea to 70% of the lumen because of granulation tissue at the lower end of the graft extending for 3 mm. In another pig an unsuspected short segment of concentric stenosis (24% of the normal tracheal
Volume 92 Number 2
Tracheal stenosis
August, t 986
30 1
Fig. 5. Ossified graft with central marrow space (M). Area of previous infolding (I).
cross-sectional area over a 2 mm length) was noted when the animal was put to death 3 months after repair. In both animals the grafts had not calcified but were entirely replaced with fibrous tissue. The cut ends of the cartilaginous rings were in close proximity and contributed to the stenosis. However, in both, the crosssectional areas of the repaired segments above and below the stenotic area were greater than the normal trachea above the area of repair. Mean body weight increased from 7.7 ± 1.1 kg atthe original operation to 56.0 ± 9.3 kg when the animals were put to death 90 ± 2 days later. The nonoperated tracheal cross-sectional area more than doubled from an initial mean of 49 ± 10 rnm' to 114.8 ± 11.5 mm' 3 months later. At autopsy the internal cross-sectional areas of the trachea at the mid-repair site were significantly greater than the normal trachea above the repair, and this was evident as early as 22 days postoperatively (Table 11). The mean internal cross-sectional area of the stenosed trachea increased approximately tenfold from 14.1 ± 4.5 mm' to 147.5 ± 37.3 mm- 3 months after repair. The grafts retained their original dimensions and the increased size of the trachea was due to growth of the tracheal rings. Histologic review demonstrated ossification in 11 of the grafts. In eight there was complete ossification (Fig. 5) and in three ossification was patchy arid incomplete. An outer ring of cortical bone surrounded central marrow spaces with hematopoietic elements. Beneath the interposed osseous tissue and the lumen was a layer of well-organized fibrous tissue as thin as 0.04 mm in
places. Patches of ciliated epithelium were demonstrated by the scanning electron microscope 22 and 64 days after repair. After 3 months the grafts were completely lined with ciliated columnar epithelium in seven pigs (FIgs. 6A and 6B). In three animals both cuboidal and ciliated epithelium were present. In the pig that died, only cuboidal epithelium lined the luminal surface of the graft. The area of infolding that was responsible for the original tracheal stenosis was represented by a "knot" of cartilage at the posterior junction of the graft and tracheal cartilage (Fig. 5). The trachea above the graft demonstrated normal respiratory epithelium.
Discussion Congenital tracheal stenosis is a lethal condition characterized by an absence of the membranous portion of the trachea with fusion of each tracheal cartilage posteriorly throughout the length of the stenotic segment. The cartilaginous rings are abnormal in shape. Histologic sections show mild chronic inflammation in the submucosa and increased amounts of dense collagen in the wall of the trachea." In a previous experiment in pigs, we created thoracic tracheal defects similar to those that would be produced when congenital tracheal stenosis is treated by a longitudinal incision. These defects were successfully repaired with free tibial periosteal grafts. However, these operations were performed on normal tracheas. 8 This model of tracheal stenosis can be created in a short and predictable period of time and has features similar to congenital tracheal stenosis: The tracheal rings were permanently fused posteriorly by dense
The Journal of
3 0 2 Cohen et al.
Thoracic and Cardiovascular Surgery
Fig. 6A. Scanning electron micrograph of epithelium on luminal surface of the graft 22 days after repair (original magnification X4,800).
collagen at the point of infolding, obliterating the membranous portion; an inflammatory reaction was present around the stenotic segment of trachea and in the submucosa, and irreversible signs of airway obstruction occurred that resulted in the animal's death. The pig trachea resembles the human trachea anatomically and functionally.'? II Posteriorly, the tracheal muscle is attached to the slightly overlapping cartilaginous rings that normally separate as the trachea expands with inspiration and is compressed during "grunting." The silicone rubber cuff appears to induce stenosis by limiting tracheal growth and altering the normal dynamic activity of the trachea so that expanding and compressing forces are repeatedly applied to the immobile tracheal rings. The weak membranous segment infolds with resultant stenosis and intraluminal obstruction. There was nothing to suggest that the remnant of the original infolding impaired or distorted tracheal growth. Tibial periosteum that is used to repair the tracheal defect is readily available, tough, pliable, and can be
sutured to the margins of the tracheal defect to create an airtight seal. It is oriented so that the osteogenic layer faces the lumen, because previous experiments demonstrated excellent ossification with the graft oriented this way."Ossification of the graft occurred by postoperative day 22 and seemed to be an important factor for prevention of stenosis. Neither of the two pigs with stenosis had evidence of graft ossification. The pig that died had a wound infection that may have affected graft vascularization and ossification so that it was unable to support the repaired segment adequately. The approximate tenfold increase in mean crosssectional area of the repaired stenotic segment over 3 months was due to tracheal cartilage growth, possibly induced by the periosteal graft. This finding was similar to that reported in our previous studies," It suggests that if the complete rings in congenital tracheal stenosis are incised longitudinally and a periosteal graft inserted, growth of the tracheal rings may be induced by the graft. Areas of mature ciliated respiratory epithelium were
Volume 92 Number 2 August, 1986
Tracheal stenosis
303
Fig. 68. Scanning electron micrograph of ciliated epithelium on luminal surface of graft 3 months after repair (original magnification X4,800). seen over the graft surface as early as postoperative day 22. Further maturation and proliferation occurred, so that by 3 months the majority of the grafts were completely covered with mature respiratory epithelium. Previously, we8 found that the application of an omental pedicle graft is not necessary for the vascularization of the periosteal graft used to repair defects in the normal trachea. In this study complete ossification of the free periosteal graft occurred when the omental pedicle graft was applied to it, whereas ossification was incomplete or absent in five of 10 animals in which an
omental pedicle graft was not used. However, these differences are not statistically significant in this small series. These experiments describe a simple, reproducible technique for creating a model of tracheal stenosis. The stenosis occurs in a short time and has similarities to congenital tracheal stenosis seen in children. The successful repair by longitudinal incision, temporary internal splinting, and insertion of a free periosteal graft into the defect may be a suitable option for correction of congenital tracheal stenosis in children.
304
2
3 4 5
6
Cohen et al.
REFERENCES Fonkalsrud EW, Sumida S: Tracheal replacement with autologous esophagus for tracheal stricture. Arch Surg 102:139-142,1971 Kufaas T, Pasila M: Free periosteal transplants in tracheal reconstruction. Report on three cases. Z Kinderchir 15:371-387,1974 Benjamin B, Pitkin J, Cohen D: Congenital tracheal stenosis. Ann Otol Rhinol Laryngol 90:364-371, 1981 Grillo HC: Tracheal reconstruction. Indications and techniques. Arch Otolaryngol 93:31-39, 1972 Bailey BJ, Kosoy J: Observations in the development of the tracheal prosthesis and tracheal transplantation. Laryngoscope 80: 1553-1565, 1970 Kimura K, Mukohara N, Tsugawa C, Matsumoto Y, Sugimura C, Murata H, Itoh H: Tracheoplasty for congenital stenosis of the entire trachea. J Pediatr Surg 17:869-871, 1982
The Journal of Thoracic and Cardiovascular Surgery
7 Idriss FS, Serafin SY, I1bawi MN, Gerson CR, Tucker GF, Holinger L: Tracheoplasty with pericardial patch for extensive tracheal stenosis in infants and children. J THORAC CARDIOVASC SURG 88:527-536, 1984 8 Cohen RC, Filler RM, Konuma K, Bahoric A, Kent G, Smith C: The successful reconstruction of thoracic tracheal defects with free periosteal grafts. J Pediatr Surg 20:852-858, 1985 9 Cantrell JR, Guild HG: Congenital stenosisof the trachea. Am J Surg 108:297-305, 1964 10 Munoz NM, Cera LM, Leff AR: Distribution of innervation and circulation in porcine cervical trachea. Am J Vet Res 45:1937-1940, 1984 11 Hare WCD: Respiratory systems, The anatomy of the domestic animal, ed 5, R Getty, ed., Philadelphia, 1975, W. B. Saunders Company, pp 1283-1296.