- Email: [email protected]
Subglottic tracheal resection and synchronous laryngeal reconstruction
General Thoracic Surgery
Subglottic tracheal resection and synchronous laryngeal reconstruction Postintubation injury of the upper airway commonly results in stenotic lesions of the larynx, subglottis, and adjacent trachea. The traditional approach to surgical correction is laryngofissure for the laryngeal component and staged plastic reconstruction of the subglottic stenosis. Reported results are variable and unpredictable, and permanent extubation is impossible in a significant number of patients. We report experience with 15 patients with combined laryngeal, subglottic, and tracheal stenosis who were managed by a one-stage operation: circumferential resection of the subglottis and trachea with primary thyrotracheal anastomosis, combined with laryngofissure and laryngeal reconstruction. These procedures required the coUaboration of the Departments of Otolaryngology and Thoracic Surgery of the Toronto General Hospital. Between 1972 and 1991, our thoracic surgical division did 53 circumferential subglottic tracheal resections with primary thyrotracheal anastomosis for benign disease. There were no operative deaths and 51 of 53 patients were successfuUy extubated. In 15 of these patients, a concomitant laryngofissure for laryngeal reconstruction was required. Laryngeal repair included excision or incision of interarytenoid scar (n = 13), interarytenoid mucosal graft (n = 6), or mobilization of cricoarytenoid joint (n = 3). A temporary laryngotracheal stent (usuaUy a Montgomery T tube) was maintained after the operation in aU cases (duration 3 to 42 months). Thirteen of these 15 patients are now permanently extubated and none has functionaUy significant restenosis. Vocal function is satisfactory to good in these patients. The approach described for these combined laryngotracheal lesions provides better results than those reported with traditional staged and plastic techniques of reconstruction. The coUaboration of the departments of otolaryngology and thoracic surgery was essential to achieve these results. (J THORAC CARDIOVASC SURG 1992;104:1443-50)
Michael A. Maddaus, MD (by invitation), Julius L. R. Toth, MD (by invitation), Patrick J. Gullane, MD, FRCS(C) (by invitation), and F. Griffith Pearson, MD, FRCS(C), Toronto, Ontario, Canada
Qstructing lesions in the subglottic airway remain a challenging problem for surgeons. Because of the anatomic relationships in this region, surgical procedures risk From the Division of Thoracic Surgery and Department of Otolaryngology, Toronto General Hospital and University of Toronto, Toronto, Ontario, Canada. Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Address for reprints: F. G. Pearson, MD, Eaton North 10-233, Toronto General Hospital, 200 Elizabeth St., Toronto, Ontario M5G 2C4, Canada.
12/6/40691
injury to the vocal cords, the recurrent laryngeal nerves, and the stabilizing framework of the subglottic airway. Although thoracic surgeons are experienced with techniques of circumferential tracheal resection and primary reconstruction, they may be hesitant to extend resections above the levelofthecricotrachealjunction because many are unfamiliar with the anatomy and function at this level. With few exceptions, otolaryngologists have developed a variety of plastic techniques to widen the subglottic airway, which involve a vertical division of the cricoid ring in front or behind and interposition of cartilage, bone, or skin to maintain the enlarged lumen. These procedures
1443
The Journal of
1444
Maddaus et al.
Thoracic and Cardiovascular Surgery
Fig. 1. Operative photograph. The cervicaltrachea has been transected just beyond the lowerend of the stenotic lesion. The tracheal segmenthas been mobilized circumferentially to the inferiorborder of the cricoidcartilage.The lowerborder ofthecricoid plate has been exposed subperichondrially and a rim of white posteriorcartilage is clearly seen.The subperichondrial plane has beendeveloped upward into the subglotticspacein front of the posteriorcricoid plate. Dissection in this plane avoids injury to the recurrent laryngeal nerves.
Table I. Circumferential, subglottic resection with primary thyrotracheal anastomosis: 53 patients (January 1973 to March 1992)
Cause Postintubation injury Blunt trauma Inhalation injury Idiopathic Amyloid disease Total
Subglottic resection only 18
8 2 9 I 38
Synchronous reconstruction of subglottis and glottis 12 I I I 15
fail to remove the underlying lesion, a fact that is reflected in the variable and often poor success rates reported. Techniques of circumferential subglottic resection and reconstruction by primary anastomosis were reported by Ogura and Biller! in 1971, Gerwat and Bryce/ in 1974, and Pearson and associates' in 1975. Subsequent reports by groups headed by Grillo,4, 5 Pearson," Couraud.? and Savary (personal communication, 1992) substantiate the effectiveness of this approach. Savary has demonstrated the successful application of this operation in infants and children and has shown that such resections do not interfere with the normal growth and development of the lar-
ynx and subglottis in long-term follow-up. These techniques remove the lesion and reconstruct the subglottic airway with a mucosal lining. The subglottic mucosa and submucosa can be resected at any level below the inferior margin of the vocal folds with preservation of recurrent nerve function and airway stability. Postintubation injury of the upper airway commonly results in stenotic lesions of the larynx, subglottis, and adjacent trachea. The traditional approach to surgical correction has been laryngofissure for the laryngeal component and plastic reconstruction of the subglottic stenosis. Reported results are variable and unpredictable, and permanent extubation cannot be achieved in a significant number of patients. This article updates our experience with isolated subglottic resection and reports on 15 patients with combined laryngeal, subglottic, and tracheal stenosis who were managed by a one-stage operation: circumferential resection of the subglottis and trachea and primary thyrotracheal anastomosis, combined with laryngofissure and laryngeal repair. These latter operations were done in collaboration with our otolaryngology colleagues.
Patients and methods Since January of 1973,64 patients have undergone subglottic resection. Eleven patients with subglottic neoplas.ns are excluded from the review. The causes of 53 benign lesions are
Volume 104 Number 5 November 1992
Tracheal resection and laryngeal reconstruction
I445
Fig. 2. Left, The diseased segment of trachea has been mobilized circumferentially up to the inferior border of the cricoid cartilage. The vertical dotted line indicates the site of incision for laryngofissure through the midline of the thyroid cartilage and carried inferiorly to the lower border of the stenotic segment. The distal resection margin is indicated by the transverse dotted line through normal trachea. Right, The airway has been opened in the midline anteriorly from the thyroid notch to the inferior limit of the stenosis in the trachea. Posterior glottic interarytenoid scar and subglottic and tracheal stenosis are now clearly visualized from within. summarized in Table I. These 53 patients are divided into two groups: 38 patients with subglottic lesionsamenable to isolated subglottic resection and 15 patients with combined glottic and subglottic lesions managed by synchronous laryngofissure, laryngeal repair, and subglottic resection. Isolated subglottic resection. Ages of the 38 patients in this group ranged from 16 to 72 years. In 18 patients the stenosis followed a period of translaryngeal endotracheal intubation, with or without subsequent tracheostomy. In all eight patients with lesions resulting from blunt trauma there was complete transection of the airway at the cricotracheal junction, a fracture of the cricoid arch, and permanent disruption of both recurrent laryngeal nerves. In four of the eight trauma cases there was a vertical fracture of the posterior cricoid plate and a tracheoesophageal fistula. Eight of the nine idiopathic lesions occurred in young women. Presentation. Twenty-seven patients had dyspnea on exertion withor without obviousrespiratory stridor. Six patients had an acute life-threatening obstruction, and five patients already had a distal tracheostomy established to secure a safe airway. Thirty of 38 patients had some form of prior treatment: one or more dilatations in 13, temporary tracheostomy in II, endoluminal stent in 6, laser resection in 3, laryngofissure and glottic surgery in 2, and previous subglottic resection in 5. Operative procedure. A collar incisionin the neck sufficedin all 38 cases. An airway segment of 2 to 7 em was resected. The tracheoesophageal fistula that complicated four cases of blunt trauma was repaired by a two-layer closure of the defect in the anterior esophageal wall, without the addition of an interposed flap of muscle or other tissue. A release procedure was added in
nine patients: thyrohyoid release in five and suprahyoid release in four. After the operation the airway was supported with a T tube or a distal tracheostomy in 29 patients. Nine patients were managed without any form of postoperative airway protection.
Results There were no operative deaths. Those patients requiring postoperative support with a T tube or distal tracheostomy were all decannulated between I week and 24 months after the operation. Only six patients required more than 2 months before decannulation was achieved. The need for prolonged airway support was the result of persistent glottic edema in two patients and delayed healing at the anastomosis in four. After decannulation, restenosis occurred in two patients. In one of these patients the problem was successfully managed by a second subglottic resection that included an excision of chronically infected cartilage in the posterior cricoid plate. The other patient with a short anastomotic stricture has been satisfactorily managed by laser resection and dilatation. There have been three instances of injury to a previously intact recurrent laryngeal nerve. Two of these injuries are permanent, and the other was transient, resolving after 2 months. Patency of the airway was categorized as follows:
The Journal of
1446
Maddaus et at.
Thoracic and Cardiovascular Surgery
Fig. 3. Left, The stenosed subglottic and tracheal segments have been resected circumferentially, including the anterior cricoidarch. The upper diagram illustratesthe mucosal defect created after subglottic resection and posterior excision of the interarytenoid scar. In the lower diagram, a well-vascularized pedicle of tracheal mucosa has been prepared from the distal trachea by resection of one or two cartilaginous rings with preservation of the membranous component. A well vascularized pedicle of tracheal mucosa has been prepared from the distal trachea by resection of oneor twocartilaginous ringswith preservation of the membranouscomponent. Right, Appearanceafter suture of the mucosalpedicle from the distal trachea into the interarytenoid defect.The laryngofissure willbe closed to complete the thyrotracheal anastomosis. Good: Normal exercise tolerance or else dyspnea only on severe exertion Satisfactory: Dyspnea with moderate exertion such as walking uphill or climbing stairs; capable of ordinary activity Poor: Dyspnea with minimal exertion or else intubated In 29 of the 38 patients airway patency was categorized as good and in seven as satisfactory, although all seven of these patients were in much better condition than before the operation. Poor results were recorded in the two patients, both of whom had functionally significant restenosis. A quantitative evaluation of vocal function has yet to be clearly defined and was not done in these patients. Voice was graded as good in 24 patients with normal or minimally altered function. Minimal change in function in these patients was usually registered as a lowering of pitch and slight huskiness. In 14 patients the voice was satisfactory for normal conversation but limited in strength (they were unable to shout or sing and had fatigue with sustained use), husky, and lower in pitch. Subglottic resection with laryngofissure and laryngotracheal reconstruction. Since 1984, 15 patients with
laryngeal and subglottic stenosis were managed by this combined approach. There were 14 women and I man, with ages ranging from 16 to 62 years. The cause of stenosis was postintubation injury in 12, trauma in 1, inhalation burn in 1, and idiopathic in 1. The laryngeal lesion was limited to the posterior glottis in 10 patients and to the anterior glottis in 2; there was fixation of the entire glottis in 3 patients. All 15 patients were referred from other institutions, and in some it was not possible to obtain complete or precise details of prior treatment. Three patients had undergone prior cardiac operations with laryngofissure and various laryngoplastic procedures, 4 had been managed by one dilatations or more (which was combined with laser resection in 3), and 6 patients had a distal tracheostomy on admission to maintain a safe airway. The preoperative evaluation included direct and indirect laryngoscopy to assess the laryngeal lesion and the mobility of the vocal cords. Before the operation 12 of the 15 patients had a significant restriction of vocal cord movement. In the 3 patients with the most severe injuries, this fixation was complete. All patients had some degree of dysphonia. Operative management. With a bolster behind the
Volume 104 Number5 November 1992
Tracheal resection and laryngeal reconstruction
I447
Fig. 4. Computed tomographic scan, October 1985, showing severe subglottic stenosis beforesynchronous laryngofissure, resection of interarytenoid scar, and resection of subglottic stenosis with thyrotracheal anastomosis. Table II. Benign subglottic resection with thyrotracheal anastomosis: Reportedresults Result First author Ogura' Gerwat-
Pearson! Grillo4 , 5
Couraud? Savary (infants and children)
Year 1972 1974 1992 · 1982, 1992 1988 1992
patient's shoulders and the neck extended, exposure is obtained through a standard collar incision in the neck. The strap muscles and thyroid isthmus are divided in the midline to expose the anterior aspect of the airway from the thyroid notch to the manubrium. The cervical trachea is mobilized circumferentially beginning at the lower end of the disease process. Dissection is maintained as close as possible to the tracheal wall to avoid injury to the recurrent nerves. No deliberate effort is made to identify the nerve at any stage in the operation, because normal anatomy is frequently obscured in these patients by inflammation and scar. Dissection is continued upward to the inferior border of the cricoid ring. The perichondrium on the inferior border of the ring is incised and can be freed completely from the anterior two thirds of the cricoid arch. Posteriorly, the perichondrium is elevated only from the inner surface of the cricoid plate, which ensures preservation of the recurrent laryngeal nerves (Fig. I). The
No. ofpatients
Satisfactory-good
Failed
17 4 38 80 27 15
14 4 36 77 27 14 172
3 0 2 3 0 1 9
subperichondrial planes is avascular and can easily be developed up to the level of the inferior glottis if the underlying cartilage is normal. The dissection is more difficult when the cartilage has already been damaged by the disease process or by a previous operation. The mobilized segment of the diseased airway is opened in the midline anteriorly with a vertical incision in the proximal trachea and cricoid. The glottis is displayed by a laryngofissure (Fig. 2). The cartilage of the an terior two thirds of the cricoid ring is removed . Posteriorly, a variable part of the inner aspect of the posterior cricoid plate is removed with a rongeur, small currette, or burr. This will widen the diameter of the rigid margins of the airway at the level of the subsequent anastomosis. Any diseased cricoid is removed. The mucosa, submucosa, and perichondrium can now be divided proximally above the level of the subglottic lesion. The laryngeal stenosis is now addressed. In six patients
I 44 8
The Journal of Thoracic and Cardiovascular Surgery
M addaus et al.
Fig. 5. Computed tomographic scan, May 1989, showing Montgomery T tube positioned with the upper limb above the level of the vocal cords. Four previous attempts at intubation had failed. The patient was successfully extubated in May 1989 and has a widely patent airway with a normal exercise tolerance and good voice.
with isolated posterior glottic stenosis, the interarytenoid scar was excised and the mucosal defect reconstructed with a pedicled flap of posterior membranous trachea prepared by revision of the distal tracheal resection margin (Fig. 3). Preparation of the flap necessitates resection of one or two additional normal tracheal rings, which does not preclude a safe thyrotracheal anastomosis. This pedicle of membranous trachea is relatively thin and pliable and very well vascularized. During end-to-end thyrotracheal anastomosis, the pedicle is advanced into the posterior defect and sutured in place. The remaining three patients with pure posterior glottic stenosis were managed by scar excision alone, which included mobilization of the cricoarytenoid joint in two instances. The two patients with anterior commissure stenosis were managed by incision alone. Three patients with both posterior and anterior glottic stenosis were managed by incision anteriorly and by scar excision in the posterior glottis. An end-to-end anastomosis between the distal trachea and the remaining subglottic mucosa was done with interrupted sutures of 35-gauge stainless steel wire or 4-0 Vicryl sutures (Ethicon, Inc., Somerville, N.J.). Details of the technique of subglottic resection and primary thyrotracheal anastomosis are provided in an earlier publication.' The anastomosis and glottis were supported with an internal stent in all IS patients. In 12 of the IS, a No. II or No. 12 Montgomery T tube was inserted with the proximal limb lying
about 0.5 cm above the vocal cords. In three patients, a solid laryngeal stent was used. Results Complications. There were no operative deaths. Seven of the IS patients had transient difficulty with swallowing and aspiration. This was most pronounced for liquids and resolved in all patients within I to 3 weeks. This complication is undoubtedly related to the position of the proximal limb of the T tube above the cord level. Two patients had painful ulcera tion on the anterior wall of the base of the epiglottis caused by impingement of the T tube. In each case the tube was removed, shortened, and replaced, and the problem resolved. Airway. Nine of 10 patients with posterior glottic stenosis have had the cannula removed-seven between I Yz and 7 months, one at 14 months, and one at 42 months. One patient still has a T tube in place because of persistent glottic malacia; extubation is pending. The two patients with anterior glottic stenosis had the cannula removed at 2 and 5 months. Two of the three patients with complete glottic stenosis were decannulated at 5 and 9 months. The third patient has severe, cicatricial restenosis at the level of the vocal cords and still has a distal tracheostomy 4 years after undergoing laryngotracheal reconstruction. The tracheostomy will probably be permanent. Thirteen of IS patients have a satisfactory to good glottic and subglottic airway. This has
Volume 104 Number 5 November 1992
Tracheal resection and laryngeal reconstruction
I 449
Fig. 6. Endoscopic photographof the same patient shown in Figs. 5 and 6, illustratingnormal appearance of vocal cords in June 1991 ,2 years after extubation.
Fig. 7. Endoscopic photograph of the same patient shown in Figs. 5, 6, and 7, illustratinga widely patent, well-healed subglottic anastomosis 2 years after extubation.
been confirmed in each case by one postoperative bronchoscopic examination or more after successful decannulation . In II of the 15 patients, clinical exercise tolerance either is completely normal or is limited only on severe exertion. In two patients, limitation of exercise tolerance is due to associated abnormalities. One patient has a stricture in the left main bronchus, a nd the other has significant cardiac disease. In each of these patients the lumen of the laryngeal aperture a nd subglottis appear satisfactory.
selected cases. To our knowledge, this is the first report on a group of patients with this condition who were managed by a one-stage, synchronous correction of the problem at both levels. Twelve of the 15 lesions in this report were the result of translaryngeal endotracheal intubation, with or without subsequent tracheostomy. The commonest site of mucosal ulceration in the glottis after endotracheal intubation is in the posterior compartment and results in an interarytenoid scar with limitation of abduction of the vocal cords . In the narrowest part of the upper airway, which lies within the cricoid ring, postintubation ulceration is usually circumferential and results in concentric subglottic stenosis. In patients with lesions at both levels, the upper limit of the circumferential subglottic stricture frequently extends to within a few millimeters of the inferior margin of the vocal folds. The combination of a high thyrotracheal anastomosis and excision or incision of granulation tissue or scar in the glottis will require some form of postoperative laryngeal stent. In three early cases, a solid, molded stent with distal tracheostomy was used. The remaining 12 patients were all managed by insertion of a Montgomery T tube during the operation. The proximal limb of the tube was positioned a few millimeters above the top of the vocal cords , and a tube diameter was selected that would lie loosely within the anastomosis and glottis after closure of the laryngofissure. The cervical arm of the T tube is corked postoperatively, so that the patient is breathing through the nose and mouth with normal humidification of the airway. A soft whispered voice is retained, and most patients are able to clear secretions through the closed airway. When necessary, secretions can be aspirated by opening the cervical
Discussion Circumferential subglottic resection and reconstruction by primary thyrotracheal anastomosis is a safe and effective one-stage operation for patients with benign subglottic stenosis. A high proportion of successful results have been reported in a series of communications since 1971. The data from these reports are summarized in Table II. Ninet y-five percent of a group of 172 patients have been successfully decannulated, and the results achieved have been variously reported as satisfactory to excellent. The material in these publications ha s been reviewed in detail in a recent article by Grillo , Mathisen, and Wain.s Benign stenosis that involves both the glottic and subglottic levels presents a more complicated management problem. In some of these cases, the glottic stenosis may be best managed with a lar yngofissure and some type of open operative repair. When a concomitant subglottic lesion would be most appropriately managed by circumferential resection and thyrotracheal anastomosis, operative management may be achieved at a single stage in
The Journal of Thoracic and Cardiovascular Surgery
1450 Maddaus et al.
arm of the tube. Using a No. 11 or 12 diameter tube, the airway is functionally normal and exercise tolerance is not limited. This form of airway support is much more comfortable and functional for the patient than a solid laryngeal stent. We have found it impossible to predict the length of time during which the stent must be maintained. We have usually elected, on an impirical basis, to leave the T tube in place for the first 2 or 3 months. When the T tube is removed, a small, distal tracheostomy (No.4 or 5) is inserted through the cervical stoma and maintained for at least 2 weeks until it is certain that the upper airway will remain patient. The tube or stent was finally removed in 10 of the 13 successfully decannulated patients between 1V2 and 7 months after operation. In three patients, despite attempts at earlier decannulation, a T tube was maintained for 9, 14, and 42 months before successful removal. In each of these three patients, the subglottic anastomosis appeared to be healed and satisfactorily patent, but progressive laryngeal edema developed within a few days or weeks of tube removal. In the most extreme case, the T tube was removed on four occasions (3 months, 9 months, 1'/2 years, and 2V2 years) before successful extubation after 42 months (Figs. 4, to 7). During her 3 1/ 2 years with a No. 10 Montgomery T tube in place, this woman led a normal and active life in every respect and was prepared to continue indefinitely with this arrangement. Though reluctant at this late stage, she agreed to yet another trial and was successfully extubated in May of 1989. Her voice is good and her exercise tolerance normal. Endoscopic photographs of the normal vocal cords and widely patent, well-healed subglottic anastomosis are seen in Figs. 6 and 7. It is apparent that the postoperative course in this group of patients is often complex and difficult to predict. The two patients in this series in whom this approach failed both had extensive and advanced lesions at the glottic and subglottic levels and both had unstable diabetes. In retrospect, they may have been inappropriate candidates for this operation. We conclude that selected patients with benign stenosis involving both the glottis and subglottis may be successfully managed by a synchronous correction of both lesions with good results. REFERENCES I. Ogura JH, Biller HF. Reconstruction of the larynx following blunt trauma. Ann Otol Rhinol Laryngol1971 ;80-492506. 2. Gerwat J, Bryce DP. The management of subglottic steno-
3.
4.
5.
6.
7.
sis by resection and direct anastomosis. Laryngoscope 1974;84:940-7. Pearson FG, Cooper JD, NelemsJM, Van Nostrand AWP. Primary tracheal anastomosis after resection of the cricoid cartilage with preservation of recurrent laryngeal nerves. J THORAC CARDIOVASC SURG 1975;70:806-16. Grillo He. Primary reconstructionof airway after resection of subglottic laryngeal and upper tracheal stenosis. Ann Thorac Surg 1982;33:3-18. Grillo HC, Mathisen DJ, Wain rc Laryngotracheal resection and reconstruction for subglottic stenosis. Ann Thorac Surg 1992;53:54-63. Pearson FG, Brito-Filomeno L, Cooper JD. Experience with partial cricoid resection and thyrotracheal anastomosis. Ann Otol Rhinol Laryngol 1986;95:582-5. Couraud L, Brichon PY, VellyJF. The surgical treatment of inflammatory and fibrous laryngotracheal stenosis. Eur J Cardiothorac Surg 1988;2:410-5.
Discussion Dr. F. Griffith Pearson (Toronto, Ontario. Canada). I would liketo respondbrieflyto Dr. Grillo's comment about doingthese cases in staged operations. That is the way these were handled on our servicein earlier years. The advantage that this synchronous approach has isthat it removesall of the pathologictissue and reconstructs the airway with a healthier mucosalcoverthan can be achieved when the larynn and subglottis are operated on in separate stages. Often if the laryngofissureis done and a posterior glottic lesiondealt with, that lesion is continuous with an inflammatory subglottic process,which may impair the success of the laryngeal result. I willtake this opportunity to comment on two contributions of significant importance in this area. Dr. Louis Couraud from Bordeaux,has reported on a group of about 10patients in whom he removed the entire cricoid for inflammatory injuries, usually when the cricoid was involved with an infective processand in some cases was just a sequestrum. As an alternative to an open, permanent tracheostomy, Dr. Couraud has sewn the trachea to the vocalcords.All but one of these patients was extubated after a prolonged period of stenting with a T tube or nasotracheal tube. These patients do not aspirate and have a better voiceand a healthier airway than the patient with an open tracheostomy. The other important contribution was made by a group of otolaryngologists in Lausanne-Drs. Savary and Monnier. They have reported recentlyon 15infants and children whohave had this type of circumferential subglottic resection, mostlyfor either postintubation or congenital stenosis,and they have successfully extubated all but one patient. Four of these children were under I year of age, their follow-up goes back 13 years, and, importantly, these investigators report normal growth in the infant and child larynx. Dr. Maddaus. I would like to thank Dr. Pearson for his endless support and guidance during my thoracic surgical training and for playing the role of a genuine surgical mentor.
Subglottic tracheal resection and synchronous laryngeal reconstruction Postintubation injury of the upper airway commonly results in stenotic lesions of the larynx, subglottis, and adjacent trachea. The traditional approach to surgical correction is laryngofissure for the laryngeal component and staged plastic reconstruction of the subglottic stenosis. Reported results are variable and unpredictable, and permanent extubation is impossible in a significant number of patients. We report experience with 15 patients with combined laryngeal, subglottic, and tracheal stenosis who were managed by a one-stage operation: circumferential resection of the subglottis and trachea with primary thyrotracheal anastomosis, combined with laryngofissure and laryngeal reconstruction. These procedures required the coUaboration of the Departments of Otolaryngology and Thoracic Surgery of the Toronto General Hospital. Between 1972 and 1991, our thoracic surgical division did 53 circumferential subglottic tracheal resections with primary thyrotracheal anastomosis for benign disease. There were no operative deaths and 51 of 53 patients were successfuUy extubated. In 15 of these patients, a concomitant laryngofissure for laryngeal reconstruction was required. Laryngeal repair included excision or incision of interarytenoid scar (n = 13), interarytenoid mucosal graft (n = 6), or mobilization of cricoarytenoid joint (n = 3). A temporary laryngotracheal stent (usuaUy a Montgomery T tube) was maintained after the operation in aU cases (duration 3 to 42 months). Thirteen of these 15 patients are now permanently extubated and none has functionaUy significant restenosis. Vocal function is satisfactory to good in these patients. The approach described for these combined laryngotracheal lesions provides better results than those reported with traditional staged and plastic techniques of reconstruction. The coUaboration of the departments of otolaryngology and thoracic surgery was essential to achieve these results. (J THORAC CARDIOVASC SURG 1992;104:1443-50)
Michael A. Maddaus, MD (by invitation), Julius L. R. Toth, MD (by invitation), Patrick J. Gullane, MD, FRCS(C) (by invitation), and F. Griffith Pearson, MD, FRCS(C), Toronto, Ontario, Canada
Qstructing lesions in the subglottic airway remain a challenging problem for surgeons. Because of the anatomic relationships in this region, surgical procedures risk From the Division of Thoracic Surgery and Department of Otolaryngology, Toronto General Hospital and University of Toronto, Toronto, Ontario, Canada. Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Address for reprints: F. G. Pearson, MD, Eaton North 10-233, Toronto General Hospital, 200 Elizabeth St., Toronto, Ontario M5G 2C4, Canada.
12/6/40691
injury to the vocal cords, the recurrent laryngeal nerves, and the stabilizing framework of the subglottic airway. Although thoracic surgeons are experienced with techniques of circumferential tracheal resection and primary reconstruction, they may be hesitant to extend resections above the levelofthecricotrachealjunction because many are unfamiliar with the anatomy and function at this level. With few exceptions, otolaryngologists have developed a variety of plastic techniques to widen the subglottic airway, which involve a vertical division of the cricoid ring in front or behind and interposition of cartilage, bone, or skin to maintain the enlarged lumen. These procedures
1443
The Journal of
1444
Maddaus et al.
Thoracic and Cardiovascular Surgery
Fig. 1. Operative photograph. The cervicaltrachea has been transected just beyond the lowerend of the stenotic lesion. The tracheal segmenthas been mobilized circumferentially to the inferiorborder of the cricoidcartilage.The lowerborder ofthecricoid plate has been exposed subperichondrially and a rim of white posteriorcartilage is clearly seen.The subperichondrial plane has beendeveloped upward into the subglotticspacein front of the posteriorcricoid plate. Dissection in this plane avoids injury to the recurrent laryngeal nerves.
Table I. Circumferential, subglottic resection with primary thyrotracheal anastomosis: 53 patients (January 1973 to March 1992)
Cause Postintubation injury Blunt trauma Inhalation injury Idiopathic Amyloid disease Total
Subglottic resection only 18
8 2 9 I 38
Synchronous reconstruction of subglottis and glottis 12 I I I 15
fail to remove the underlying lesion, a fact that is reflected in the variable and often poor success rates reported. Techniques of circumferential subglottic resection and reconstruction by primary anastomosis were reported by Ogura and Biller! in 1971, Gerwat and Bryce/ in 1974, and Pearson and associates' in 1975. Subsequent reports by groups headed by Grillo,4, 5 Pearson," Couraud.? and Savary (personal communication, 1992) substantiate the effectiveness of this approach. Savary has demonstrated the successful application of this operation in infants and children and has shown that such resections do not interfere with the normal growth and development of the lar-
ynx and subglottis in long-term follow-up. These techniques remove the lesion and reconstruct the subglottic airway with a mucosal lining. The subglottic mucosa and submucosa can be resected at any level below the inferior margin of the vocal folds with preservation of recurrent nerve function and airway stability. Postintubation injury of the upper airway commonly results in stenotic lesions of the larynx, subglottis, and adjacent trachea. The traditional approach to surgical correction has been laryngofissure for the laryngeal component and plastic reconstruction of the subglottic stenosis. Reported results are variable and unpredictable, and permanent extubation cannot be achieved in a significant number of patients. This article updates our experience with isolated subglottic resection and reports on 15 patients with combined laryngeal, subglottic, and tracheal stenosis who were managed by a one-stage operation: circumferential resection of the subglottis and trachea and primary thyrotracheal anastomosis, combined with laryngofissure and laryngeal repair. These latter operations were done in collaboration with our otolaryngology colleagues.
Patients and methods Since January of 1973,64 patients have undergone subglottic resection. Eleven patients with subglottic neoplas.ns are excluded from the review. The causes of 53 benign lesions are
Volume 104 Number 5 November 1992
Tracheal resection and laryngeal reconstruction
I445
Fig. 2. Left, The diseased segment of trachea has been mobilized circumferentially up to the inferior border of the cricoid cartilage. The vertical dotted line indicates the site of incision for laryngofissure through the midline of the thyroid cartilage and carried inferiorly to the lower border of the stenotic segment. The distal resection margin is indicated by the transverse dotted line through normal trachea. Right, The airway has been opened in the midline anteriorly from the thyroid notch to the inferior limit of the stenosis in the trachea. Posterior glottic interarytenoid scar and subglottic and tracheal stenosis are now clearly visualized from within. summarized in Table I. These 53 patients are divided into two groups: 38 patients with subglottic lesionsamenable to isolated subglottic resection and 15 patients with combined glottic and subglottic lesions managed by synchronous laryngofissure, laryngeal repair, and subglottic resection. Isolated subglottic resection. Ages of the 38 patients in this group ranged from 16 to 72 years. In 18 patients the stenosis followed a period of translaryngeal endotracheal intubation, with or without subsequent tracheostomy. In all eight patients with lesions resulting from blunt trauma there was complete transection of the airway at the cricotracheal junction, a fracture of the cricoid arch, and permanent disruption of both recurrent laryngeal nerves. In four of the eight trauma cases there was a vertical fracture of the posterior cricoid plate and a tracheoesophageal fistula. Eight of the nine idiopathic lesions occurred in young women. Presentation. Twenty-seven patients had dyspnea on exertion withor without obviousrespiratory stridor. Six patients had an acute life-threatening obstruction, and five patients already had a distal tracheostomy established to secure a safe airway. Thirty of 38 patients had some form of prior treatment: one or more dilatations in 13, temporary tracheostomy in II, endoluminal stent in 6, laser resection in 3, laryngofissure and glottic surgery in 2, and previous subglottic resection in 5. Operative procedure. A collar incisionin the neck sufficedin all 38 cases. An airway segment of 2 to 7 em was resected. The tracheoesophageal fistula that complicated four cases of blunt trauma was repaired by a two-layer closure of the defect in the anterior esophageal wall, without the addition of an interposed flap of muscle or other tissue. A release procedure was added in
nine patients: thyrohyoid release in five and suprahyoid release in four. After the operation the airway was supported with a T tube or a distal tracheostomy in 29 patients. Nine patients were managed without any form of postoperative airway protection.
Results There were no operative deaths. Those patients requiring postoperative support with a T tube or distal tracheostomy were all decannulated between I week and 24 months after the operation. Only six patients required more than 2 months before decannulation was achieved. The need for prolonged airway support was the result of persistent glottic edema in two patients and delayed healing at the anastomosis in four. After decannulation, restenosis occurred in two patients. In one of these patients the problem was successfully managed by a second subglottic resection that included an excision of chronically infected cartilage in the posterior cricoid plate. The other patient with a short anastomotic stricture has been satisfactorily managed by laser resection and dilatation. There have been three instances of injury to a previously intact recurrent laryngeal nerve. Two of these injuries are permanent, and the other was transient, resolving after 2 months. Patency of the airway was categorized as follows:
The Journal of
1446
Maddaus et at.
Thoracic and Cardiovascular Surgery
Fig. 3. Left, The stenosed subglottic and tracheal segments have been resected circumferentially, including the anterior cricoidarch. The upper diagram illustratesthe mucosal defect created after subglottic resection and posterior excision of the interarytenoid scar. In the lower diagram, a well-vascularized pedicle of tracheal mucosa has been prepared from the distal trachea by resection of one or two cartilaginous rings with preservation of the membranous component. A well vascularized pedicle of tracheal mucosa has been prepared from the distal trachea by resection of oneor twocartilaginous ringswith preservation of the membranouscomponent. Right, Appearanceafter suture of the mucosalpedicle from the distal trachea into the interarytenoid defect.The laryngofissure willbe closed to complete the thyrotracheal anastomosis. Good: Normal exercise tolerance or else dyspnea only on severe exertion Satisfactory: Dyspnea with moderate exertion such as walking uphill or climbing stairs; capable of ordinary activity Poor: Dyspnea with minimal exertion or else intubated In 29 of the 38 patients airway patency was categorized as good and in seven as satisfactory, although all seven of these patients were in much better condition than before the operation. Poor results were recorded in the two patients, both of whom had functionally significant restenosis. A quantitative evaluation of vocal function has yet to be clearly defined and was not done in these patients. Voice was graded as good in 24 patients with normal or minimally altered function. Minimal change in function in these patients was usually registered as a lowering of pitch and slight huskiness. In 14 patients the voice was satisfactory for normal conversation but limited in strength (they were unable to shout or sing and had fatigue with sustained use), husky, and lower in pitch. Subglottic resection with laryngofissure and laryngotracheal reconstruction. Since 1984, 15 patients with
laryngeal and subglottic stenosis were managed by this combined approach. There were 14 women and I man, with ages ranging from 16 to 62 years. The cause of stenosis was postintubation injury in 12, trauma in 1, inhalation burn in 1, and idiopathic in 1. The laryngeal lesion was limited to the posterior glottis in 10 patients and to the anterior glottis in 2; there was fixation of the entire glottis in 3 patients. All 15 patients were referred from other institutions, and in some it was not possible to obtain complete or precise details of prior treatment. Three patients had undergone prior cardiac operations with laryngofissure and various laryngoplastic procedures, 4 had been managed by one dilatations or more (which was combined with laser resection in 3), and 6 patients had a distal tracheostomy on admission to maintain a safe airway. The preoperative evaluation included direct and indirect laryngoscopy to assess the laryngeal lesion and the mobility of the vocal cords. Before the operation 12 of the 15 patients had a significant restriction of vocal cord movement. In the 3 patients with the most severe injuries, this fixation was complete. All patients had some degree of dysphonia. Operative management. With a bolster behind the
Volume 104 Number5 November 1992
Tracheal resection and laryngeal reconstruction
I447
Fig. 4. Computed tomographic scan, October 1985, showing severe subglottic stenosis beforesynchronous laryngofissure, resection of interarytenoid scar, and resection of subglottic stenosis with thyrotracheal anastomosis. Table II. Benign subglottic resection with thyrotracheal anastomosis: Reportedresults Result First author Ogura' Gerwat-
Pearson! Grillo4 , 5
Couraud? Savary (infants and children)
Year 1972 1974 1992 · 1982, 1992 1988 1992
patient's shoulders and the neck extended, exposure is obtained through a standard collar incision in the neck. The strap muscles and thyroid isthmus are divided in the midline to expose the anterior aspect of the airway from the thyroid notch to the manubrium. The cervical trachea is mobilized circumferentially beginning at the lower end of the disease process. Dissection is maintained as close as possible to the tracheal wall to avoid injury to the recurrent nerves. No deliberate effort is made to identify the nerve at any stage in the operation, because normal anatomy is frequently obscured in these patients by inflammation and scar. Dissection is continued upward to the inferior border of the cricoid ring. The perichondrium on the inferior border of the ring is incised and can be freed completely from the anterior two thirds of the cricoid arch. Posteriorly, the perichondrium is elevated only from the inner surface of the cricoid plate, which ensures preservation of the recurrent laryngeal nerves (Fig. I). The
No. ofpatients
Satisfactory-good
Failed
17 4 38 80 27 15
14 4 36 77 27 14 172
3 0 2 3 0 1 9
subperichondrial planes is avascular and can easily be developed up to the level of the inferior glottis if the underlying cartilage is normal. The dissection is more difficult when the cartilage has already been damaged by the disease process or by a previous operation. The mobilized segment of the diseased airway is opened in the midline anteriorly with a vertical incision in the proximal trachea and cricoid. The glottis is displayed by a laryngofissure (Fig. 2). The cartilage of the an terior two thirds of the cricoid ring is removed . Posteriorly, a variable part of the inner aspect of the posterior cricoid plate is removed with a rongeur, small currette, or burr. This will widen the diameter of the rigid margins of the airway at the level of the subsequent anastomosis. Any diseased cricoid is removed. The mucosa, submucosa, and perichondrium can now be divided proximally above the level of the subglottic lesion. The laryngeal stenosis is now addressed. In six patients
I 44 8
The Journal of Thoracic and Cardiovascular Surgery
M addaus et al.
Fig. 5. Computed tomographic scan, May 1989, showing Montgomery T tube positioned with the upper limb above the level of the vocal cords. Four previous attempts at intubation had failed. The patient was successfully extubated in May 1989 and has a widely patent airway with a normal exercise tolerance and good voice.
with isolated posterior glottic stenosis, the interarytenoid scar was excised and the mucosal defect reconstructed with a pedicled flap of posterior membranous trachea prepared by revision of the distal tracheal resection margin (Fig. 3). Preparation of the flap necessitates resection of one or two additional normal tracheal rings, which does not preclude a safe thyrotracheal anastomosis. This pedicle of membranous trachea is relatively thin and pliable and very well vascularized. During end-to-end thyrotracheal anastomosis, the pedicle is advanced into the posterior defect and sutured in place. The remaining three patients with pure posterior glottic stenosis were managed by scar excision alone, which included mobilization of the cricoarytenoid joint in two instances. The two patients with anterior commissure stenosis were managed by incision alone. Three patients with both posterior and anterior glottic stenosis were managed by incision anteriorly and by scar excision in the posterior glottis. An end-to-end anastomosis between the distal trachea and the remaining subglottic mucosa was done with interrupted sutures of 35-gauge stainless steel wire or 4-0 Vicryl sutures (Ethicon, Inc., Somerville, N.J.). Details of the technique of subglottic resection and primary thyrotracheal anastomosis are provided in an earlier publication.' The anastomosis and glottis were supported with an internal stent in all IS patients. In 12 of the IS, a No. II or No. 12 Montgomery T tube was inserted with the proximal limb lying
about 0.5 cm above the vocal cords. In three patients, a solid laryngeal stent was used. Results Complications. There were no operative deaths. Seven of the IS patients had transient difficulty with swallowing and aspiration. This was most pronounced for liquids and resolved in all patients within I to 3 weeks. This complication is undoubtedly related to the position of the proximal limb of the T tube above the cord level. Two patients had painful ulcera tion on the anterior wall of the base of the epiglottis caused by impingement of the T tube. In each case the tube was removed, shortened, and replaced, and the problem resolved. Airway. Nine of 10 patients with posterior glottic stenosis have had the cannula removed-seven between I Yz and 7 months, one at 14 months, and one at 42 months. One patient still has a T tube in place because of persistent glottic malacia; extubation is pending. The two patients with anterior glottic stenosis had the cannula removed at 2 and 5 months. Two of the three patients with complete glottic stenosis were decannulated at 5 and 9 months. The third patient has severe, cicatricial restenosis at the level of the vocal cords and still has a distal tracheostomy 4 years after undergoing laryngotracheal reconstruction. The tracheostomy will probably be permanent. Thirteen of IS patients have a satisfactory to good glottic and subglottic airway. This has
Volume 104 Number 5 November 1992
Tracheal resection and laryngeal reconstruction
I 449
Fig. 6. Endoscopic photographof the same patient shown in Figs. 5 and 6, illustratingnormal appearance of vocal cords in June 1991 ,2 years after extubation.
Fig. 7. Endoscopic photograph of the same patient shown in Figs. 5, 6, and 7, illustratinga widely patent, well-healed subglottic anastomosis 2 years after extubation.
been confirmed in each case by one postoperative bronchoscopic examination or more after successful decannulation . In II of the 15 patients, clinical exercise tolerance either is completely normal or is limited only on severe exertion. In two patients, limitation of exercise tolerance is due to associated abnormalities. One patient has a stricture in the left main bronchus, a nd the other has significant cardiac disease. In each of these patients the lumen of the laryngeal aperture a nd subglottis appear satisfactory.
selected cases. To our knowledge, this is the first report on a group of patients with this condition who were managed by a one-stage, synchronous correction of the problem at both levels. Twelve of the 15 lesions in this report were the result of translaryngeal endotracheal intubation, with or without subsequent tracheostomy. The commonest site of mucosal ulceration in the glottis after endotracheal intubation is in the posterior compartment and results in an interarytenoid scar with limitation of abduction of the vocal cords . In the narrowest part of the upper airway, which lies within the cricoid ring, postintubation ulceration is usually circumferential and results in concentric subglottic stenosis. In patients with lesions at both levels, the upper limit of the circumferential subglottic stricture frequently extends to within a few millimeters of the inferior margin of the vocal folds. The combination of a high thyrotracheal anastomosis and excision or incision of granulation tissue or scar in the glottis will require some form of postoperative laryngeal stent. In three early cases, a solid, molded stent with distal tracheostomy was used. The remaining 12 patients were all managed by insertion of a Montgomery T tube during the operation. The proximal limb of the tube was positioned a few millimeters above the top of the vocal cords , and a tube diameter was selected that would lie loosely within the anastomosis and glottis after closure of the laryngofissure. The cervical arm of the T tube is corked postoperatively, so that the patient is breathing through the nose and mouth with normal humidification of the airway. A soft whispered voice is retained, and most patients are able to clear secretions through the closed airway. When necessary, secretions can be aspirated by opening the cervical
Discussion Circumferential subglottic resection and reconstruction by primary thyrotracheal anastomosis is a safe and effective one-stage operation for patients with benign subglottic stenosis. A high proportion of successful results have been reported in a series of communications since 1971. The data from these reports are summarized in Table II. Ninet y-five percent of a group of 172 patients have been successfully decannulated, and the results achieved have been variously reported as satisfactory to excellent. The material in these publications ha s been reviewed in detail in a recent article by Grillo , Mathisen, and Wain.s Benign stenosis that involves both the glottic and subglottic levels presents a more complicated management problem. In some of these cases, the glottic stenosis may be best managed with a lar yngofissure and some type of open operative repair. When a concomitant subglottic lesion would be most appropriately managed by circumferential resection and thyrotracheal anastomosis, operative management may be achieved at a single stage in
The Journal of Thoracic and Cardiovascular Surgery
1450 Maddaus et al.
arm of the tube. Using a No. 11 or 12 diameter tube, the airway is functionally normal and exercise tolerance is not limited. This form of airway support is much more comfortable and functional for the patient than a solid laryngeal stent. We have found it impossible to predict the length of time during which the stent must be maintained. We have usually elected, on an impirical basis, to leave the T tube in place for the first 2 or 3 months. When the T tube is removed, a small, distal tracheostomy (No.4 or 5) is inserted through the cervical stoma and maintained for at least 2 weeks until it is certain that the upper airway will remain patient. The tube or stent was finally removed in 10 of the 13 successfully decannulated patients between 1V2 and 7 months after operation. In three patients, despite attempts at earlier decannulation, a T tube was maintained for 9, 14, and 42 months before successful removal. In each of these three patients, the subglottic anastomosis appeared to be healed and satisfactorily patent, but progressive laryngeal edema developed within a few days or weeks of tube removal. In the most extreme case, the T tube was removed on four occasions (3 months, 9 months, 1'/2 years, and 2V2 years) before successful extubation after 42 months (Figs. 4, to 7). During her 3 1/ 2 years with a No. 10 Montgomery T tube in place, this woman led a normal and active life in every respect and was prepared to continue indefinitely with this arrangement. Though reluctant at this late stage, she agreed to yet another trial and was successfully extubated in May of 1989. Her voice is good and her exercise tolerance normal. Endoscopic photographs of the normal vocal cords and widely patent, well-healed subglottic anastomosis are seen in Figs. 6 and 7. It is apparent that the postoperative course in this group of patients is often complex and difficult to predict. The two patients in this series in whom this approach failed both had extensive and advanced lesions at the glottic and subglottic levels and both had unstable diabetes. In retrospect, they may have been inappropriate candidates for this operation. We conclude that selected patients with benign stenosis involving both the glottis and subglottis may be successfully managed by a synchronous correction of both lesions with good results. REFERENCES I. Ogura JH, Biller HF. Reconstruction of the larynx following blunt trauma. Ann Otol Rhinol Laryngol1971 ;80-492506. 2. Gerwat J, Bryce DP. The management of subglottic steno-
3.
4.
5.
6.
7.
sis by resection and direct anastomosis. Laryngoscope 1974;84:940-7. Pearson FG, Cooper JD, NelemsJM, Van Nostrand AWP. Primary tracheal anastomosis after resection of the cricoid cartilage with preservation of recurrent laryngeal nerves. J THORAC CARDIOVASC SURG 1975;70:806-16. Grillo He. Primary reconstructionof airway after resection of subglottic laryngeal and upper tracheal stenosis. Ann Thorac Surg 1982;33:3-18. Grillo HC, Mathisen DJ, Wain rc Laryngotracheal resection and reconstruction for subglottic stenosis. Ann Thorac Surg 1992;53:54-63. Pearson FG, Brito-Filomeno L, Cooper JD. Experience with partial cricoid resection and thyrotracheal anastomosis. Ann Otol Rhinol Laryngol 1986;95:582-5. Couraud L, Brichon PY, VellyJF. The surgical treatment of inflammatory and fibrous laryngotracheal stenosis. Eur J Cardiothorac Surg 1988;2:410-5.
Discussion Dr. F. Griffith Pearson (Toronto, Ontario. Canada). I would liketo respondbrieflyto Dr. Grillo's comment about doingthese cases in staged operations. That is the way these were handled on our servicein earlier years. The advantage that this synchronous approach has isthat it removesall of the pathologictissue and reconstructs the airway with a healthier mucosalcoverthan can be achieved when the larynn and subglottis are operated on in separate stages. Often if the laryngofissureis done and a posterior glottic lesiondealt with, that lesion is continuous with an inflammatory subglottic process,which may impair the success of the laryngeal result. I willtake this opportunity to comment on two contributions of significant importance in this area. Dr. Louis Couraud from Bordeaux,has reported on a group of about 10patients in whom he removed the entire cricoid for inflammatory injuries, usually when the cricoid was involved with an infective processand in some cases was just a sequestrum. As an alternative to an open, permanent tracheostomy, Dr. Couraud has sewn the trachea to the vocalcords.All but one of these patients was extubated after a prolonged period of stenting with a T tube or nasotracheal tube. These patients do not aspirate and have a better voiceand a healthier airway than the patient with an open tracheostomy. The other important contribution was made by a group of otolaryngologists in Lausanne-Drs. Savary and Monnier. They have reported recentlyon 15infants and children whohave had this type of circumferential subglottic resection, mostlyfor either postintubation or congenital stenosis,and they have successfully extubated all but one patient. Four of these children were under I year of age, their follow-up goes back 13 years, and, importantly, these investigators report normal growth in the infant and child larynx. Dr. Maddaus. I would like to thank Dr. Pearson for his endless support and guidance during my thoracic surgical training and for playing the role of a genuine surgical mentor.