Surgical Treatment of Nontumoral Stenoses of the Upper Airway

Transcription

1 Surgical Treatment of Nontumoral Stenoses of the Upper Airway Louis Couraud, MD, Jacques B. Jougon, MD, and Jean-Francois Velly, MD Thoracic Surgery Unit, Xavier Arnozan Hospital, Pessac (Bordeaux), France Background. After 1970, the widespread use of nasotracheal intubation, avoiding tracheostomy and its pitfalls, resulted in more frequent laryngeal or laryngotracheal stenoses, which required more complex and sometimes multistaged procedures. Methods. A series of 217 nontumoral stenoses of the upper airway were treated following the same therapeutic principles in the period 1978 to Two hundred one of them were iatrogenic postintubation strictures (92%); the others were posttraumatic (7), idiopathic (5), and various (4). Results, One hundred twenty (55fir) were tracheal stenoses and treated by resection and primary end-to-end anastomosis with 117 excellent or good results and three deaths. Length of the stenosis, old age, neuropsychological sequelae, and overall poor respiratory status of the patients made up the remaining difficulties in the treatment. Ninety-seven (45%) were laryngotracheal stenoses with much more complex surgical indications: 57 patients underwent tracheal and subglottic resection and anastomosis with 56 successes and one death, 7 had laryngotracheal resection and anastomosis with total cricoidectomy and consequently laryngeal stenting for 3 to 6 months (six successes, one death), 3 had supraglottic resection and anastomosis (three successes), 12 patients with glottic opening difficulties and short laryngeal stenosis underwent a laryngeal enlargement over a T tube without resection (11 successes, one death), and 18 were subjected to a complex combination of resection and modeling with 16 successes, 2 failures, and 1 death. Final results were successful in 208 cases (96%) with seven deaths and two failures. Mild phonetic sequelae were observed after laryngeal modeling. A minimal follow-up of 1 year has shown long-term stability of most repairs. Conclusions. Despite acceptable results, the therapeutic approach remains difficult for laryngotracheal stenoses involving the glottic and the supraglottic level as well as for those that have not responded to previous attempts at repair. In a few cases, despite a meticulous preoperative assessment, the surgical strategy can only be adopted intraoperatively. The key to surgical success is undoubtedly a careful preoperative treatment of infection and inflammation as well as a meticulous muco-mucosal approximation of healthy margins at the anastomosis. (Ann Thorac Surg 1995;60:250-60) A fter 1965 the number of nontumoral stenoses of the upper airway requiring surgical treatment has increased dramatically due to the increasing number of patients receiving respiratory support by tracheostomy or laryngotracheal intubation, now the leading cause of such stenoses. Less commonly they follow external laryngotracheal trauma, chemical burns, mismanaged cervicomediastinal radiotherapy, or idiopathic causes. The first attempts at surgical repair proposed various surgical approaches dependent on the speciality of the surgeon: mainly laryngotracheal enlargement with ear, nose, and throat surgeons, skin or cartilage plasties with plastic surgeons, and as often as possible resection anastomosis in the hands of thoracic surgeons. After that pioneers such as Grillo and Pearson [1-3[ described the indications and technique for safe tracheal resection followed by end-to-end reanastomosis, and it was generally accepted that tracheal stenosis should be treated through resection and anastomosis whereas laryngeal stenosis should be treated through laryngoplasty. Presented at the Thirty-first Annual Meeting4 of The Society ol Thoracic Surgeons, Palm Springs, CA, Jan 30 Feb 1, Address reprint requests to Dr Couraud, Xavier Arnozan Hospital, Pessac, France by The Society of Thoracic Surgeons After our first 15 cases (Feb 1967 to Oct 1968) we described the conditions and the local preparation necessary for good healing of the tracheal anastomosis avoiding leakages or restenosis [4]. We emphasized the necessity of palliative clearing of the obstruction of the airway through endoscopic electrocoagulation of obstructive granulomata and the importance of preoperative antibiotic and antiinflammatory therapy, delaying the surgical procedure to allow the tracheal resection anastomosis in healthy margins. After 1970 the widespread use of nasotracheal intubation avoided tracheostomy and its pitfalls, but resulted in more frequent laryngeal or laryngotracheal stenoses requiring long and multistaged procedures for their treatment. Since Gerwat and Bryce [5] and Pearson and colleagues [6] demonstrated that resection anastomosis could be extended to the subglottic larynx, surgery of laryngotracheal stenoses evolved into more aggressive and effective techniques for single-stage procedures up to the subglottic area and later up to the upper larynx [7]. After 1978 our philosophy was well established with a consistant protocol for investigations, preoperative assessment, preparation, and therapeutic options. The object of this article is to report our experience of treatment of upper airway stenoses in 217 patients in the period /95/$ (95)00464-V

2 Ann Thorac Surg COURAUD ET AL ;60: AIRWAY STENOSIS 1978 to 1992 and a minimal follow-up of 1 year. We excluded from the series patients in whom surgical treatment was unnecessary (eg, patients with inflammatory and noncircumferential obstructive lesions such as granuloma and partially stenosing fibrous arches, which were cured by antiinflammatory drugs and endoscopic laser photocoagulation). We also excluded patients with contraindications to operation who underwent endoluminal stenting (Gianturco or Dumont) [8], T tubes, or permanent tracheostomy. Furthermore, we excluded long-segment congenital tracheal stenoses with O rings as well as tracheomalacia, the treatment of the latter with expanding wire stents being nowadays well established [8]. Material and Methods Patients The average age of patients was 41 years (range, 3 to 88 years); there were almost equal numbers of male and female patients. The cause of airway stenosis was iatrogenic in 201, who had undergone an intubation for respiratory, resuscitation. That was indicated for one of the diseases that can result in respiratory failure: the main causes were trauma, frequently of brain, chest, or polytrauma, medical diseases including metabolic and neurologic problems, acute failure in chronic respiratory insufficiency, and, finally resuscitation after operation followed by metabolic or postinfective multiorgan failure. The other causes were external cervical trauma with mismanaged total laryngotracheal separation in 7 patients and various in 4 including chemical burn in 1, postinfectious fibrous scar in 1, and irradiation in 2. Five stenoses were idiopathic, always affecting the subglottic area with alteration of the cricoid cartilage. Lesions Because the location of the lesions will have an important influence on the surgical options, two types of stenotic lesions are defined depending on the presence or absence of concomitant laryngeal damage. Tracheal stenoses were present in 120 patients. They make up 55% of our series and by definition are located below the cricoid ring. The lesions were located at the cervical level in 78 patients (65%), at the cervicothoracic junction level in 22 patients (18%), and intrathoracically in 12 patients (10%). Eight patients had distant multifocal lesions at different levels (7%). Laryngeal or laryngotracheal stenoses were present in 97 patients (45%): 91 were secondary to nasotracheal intubation, 1 to chemical burn, and 5 to idiopathic causes. Additional contributing factors were prolonged intubation followed by tracheostomy, improper siting of a tracheostorny involving the cricoid cartilage, traumatic laryngeal intubation, and failed previous attempts at surgical repair (9 cases). The laryngeal lesions were located at the subglottic level in 93 patients, reached the glottis in 34, and went up to the supraglottic level in 4. Most were multifocal at the laryngeal level, and they were associated with varying degrees of tracheal lesions or stenosis in 69 patients (71%). Therapeutic options varied accordingly. Treatment and Results All patients were subject to the same principles of evaluation and treatment preoperatively, intraoperatively, and postoperatively. 1. Meticulous anatomic assessment, through flexible and rigid fibroscopy and plain or computed tomographic examination, was performed. Basal functional respiratory status and neuropsychological ability to cooperate also were evaluated. 2. Bacteriology and preoperative treatment of local infection and inflammation was performed with oral or parenteral appropriate antibiotics, inhaled antibiotics, and steroid therapy, and endoscopic laser photocoagulation was used for temporary palliative airway disobstruction. 3. Complete resection of the fibrotic or residual inflammatory lesions was performed between cartilages to allow anastomosis in healthy resection margins. 4. Careful mucomucosal approximation at the anastomosis was done with interrupted sutures with knots tied outside the lumen using a resorbable monofilament nonreactive suture material such as polydioxanone (PDS; Ethicon, Neuilly sur Seine, France) or polyglyconate (Maxon; Davis & Geck, Rungis, France). 5. Postoperative treatment included adapted antibiotics and steroids for 5 to 7 days through general and inhaled routes. 6. Oral feeding was reintroduced between the 4th and the 7th day according to the level of cooperation of the patient and after deglutition testing to avoid aspiration and frequent postoperative endoscopic tracheobronchial toilet. 7. Flexible endoscopic control of healing was performed between the 15th and the 20th day. A rigid endoscopy under general anesthesia sometimes was necessary for laser coagulation of granulomata at this time or at the first check-up on follow-up at 3 months. The long-term follow-up included an endoscopic examination at 1 and 3 years. 8. The respiratory result was classified as excellent when a patient was able to perform usual activities with an endoscopic examination showing an airway patency of 90% to 100% of the normal size at the anastomotic level. The result was classified as good when the patient had a normal functional capacity although airway patency was reduced from 70% to 80% of normal. The result was poor when the patient had dyspnea on exertion and an airway anastomotic caliber less than 70%. Failure was characterized by the necessity of stenting the airway or retracheostomy. The phonetic results were more difficult to classify because we did not assess these results systematically with physiologic testing of laryngeal function. Therefore, we report a subjective classification considering the result as good when the patient was able to speak aloud, to shout, and to sing; as fair when the patient could speak

3 252 COURAUD ET AL Ann Thorac Surg AIRWAY STENOSIS 1995;60: aloud and shout but with an altered voice; and as acceptable when the patient could speak aloud but with an evident weakness. We never observed a voiceless patient at the 1-year follow-up. Tracheal Stenosis All 120 patients underwent tracheal resection anastomosis after the above-mentioned preparation. Incision was cervical in 91 patients (76%), which allows resection at the level of the cervical trachea and the cervicothoracic junction. The cervical approach was combined with a sternotomy in 28 patients (23"/,), and the route was a right thoracotomy in 1. The results were excellent to good in 117 patients (97%) with reestablishment of a normal respiratory function and an airway patency from 90% to 100% (excellent) to 80% (good) without clinical consequences. The phonation was equally restored, with good to fair results allowing normal professional and social communication. Nevertheless, the recovery of a normal voice sometimes was impaired by some laryngeal sequelae such as hemilaryngeal palsy or fibrotic arches anchoring the vocal cords, the result of brain trauma consequences or of a long and traumatic transglottic intubation, respectively. There were three deaths, one of total anastomotic dehiscence with fatal hemorrhage, one cerebrovascular accident in an 80-year-old patient, and one respirato D, failure in a patient with chronically impaired respiratory function. Tracheal resection and anastomosis therefore can achieve good vital and functional results provided the above-mentioned principles are respected. Laryngeal and Laryngotracheal Stenosis Therapeutic strategy varied according to the site and the length of the lesions, the length of tracheal involvement, and the result of a meticulous evaluation of glottic opening and vocal cord mobility. Thereafter treatment of these complex lesions required various types of surgical procedures, which are as follows: 1. Fifty-seven patients had subglottic resection and anastomosis with resection of the cricoid ring and more or less of the lower edge of the cricoid plate (Gerwatt, Bryce and Pearson procedure). There were 56 successes but one death caused by anastomotic leakage followed by innominate artery ulceration at day 10. This patient had prior unsuccessful resection and was not compliant with postoperative care, including smoking in the immediate postoperative period. The functional respiratory results were excellent in the 56 survivors. The phonetic results were good to fair. One patient had altered phonation due to operative traumatic recurrent nerve lesion, and a few others had altered phonation as a result of a preceding failed operation or of brain trauma or cordal injury. 2. Seven patients underwent a subglottic laryngotracheal resection and anastomosis with subperichondrial total cricoid resection followed bv a laryngeal molding for 3 to 6 months over a Montgomery T tube, until the laryngeal posterior wall had recovered enough rigidity to support the arytenoid cartilage abduction and the opening of the glottis. We used a perichondrial and more recently a periostial free graft in the cricoidectomy bed to accelerate such recovery. One death occured after anastomotic leakage. All survivors had a good respiratory result. The phonetic results are less good, with five fair results (slightly hoarse voice) and only one normal voice. 3. Three patients underwent extended laryngeal resection with supraglottic laryngotracheal anastomosis for total glottic, subglottic, and tracheal stenosis associated with cricoid skeleton destruction. All survived. The respiratory results were good. The phonetic results were surprisingly acceptable, good enough to allow normal everyday communication, phoning, and speaking but not shouting. One patient had temporary swallowing difficulties with aspirations, which were controlled easily after a week of rehabilitation. 4. Twelve patients underwent a laryngeal enlargement of the glotto-subglottic segment as well as one or two rings of the upper trachea. That was carried out through anterior and posterior vertical cricotomy and stented and molded over a T tube for 3 months. One patient died of an unrelated cause. The other 11 had good functional respiratory results and fair to good phonetic results. 5. Seven patients who suffered from stenoses of the laryngotracheal junction as a result of mismanaged total posttraumatic laryngotracheal separation associated with recurrent nerve injury underwent a complex but wellestablished procedure including resection of the stenotic segment, horizontal partial cricoid ring resection up to the level of healthy mucosa, glottic enlargement through vertical cricoid plate division according to Rethi's fashion, and laryngotracheal reanastomosis over a T tube stent for 1 to 3 months. All procedures succeeded with excellent to good respiratory results (1 patient needed a secondary laser cordectomy) and a surprisingly good to fair phonation. 6. Eleven patients underwent complex and unclassifiable procedures: they combined various degree of resection, cordectomy, or cordopexy frequently followed by modeling over stents to work out important and complex stenoses that frequently were the result of previous attempts at surgical repair or inadequate palliative procedures. Eight of them were successful, 1 patient died with anastomotic leakage, and 2 failed again, (1 of them having at present an endoluminal prosthesis, the other a tracheostomy). Four patients had temporary swallowing difficulties resulting from important extraperichondrial resections or from recurrent nerve lesions. The final result was good functional respiratory ability and progressively improving from acceptable up to fair phonetic capacity in all 8. Late Results at Follow-up Overall results are successful in 208 cases (96%); there were seven deaths and two failures. Follow-up from 1 to 10 years has demonstrated long-term stability of the results acquired at 3 months after tracheal resection and acquired at 1 year after complex repairs in laryngotracheal stenoses. Indeed, in few cases with total or subtotal resection of the cricoid, we observed at 3 to 6 months postoperatively some degree of instability of the posterior supraglottic

4 Ann Thorac Surg COURAUD ET AL ;60: AIRWAY STENOSIS structures resulting in prolapse of the arytenoid or supraarytenoid cartilages and mucosa into the glottic lumen with subsequent stridor on exertion or night-time snoring and dyspnea. Laser photocoagulation of these structures and of any redundant mucosa universally has been easy to achieve and successful in relieving the symptoms. Comment In spite of major improvements in material and conditions of resuscitation in intensive care units, tracheal and laryngotracheal stenoses still represent an important group of iatrogenic complications after prolonged tracheal intubation. Consequently they remain a problem facing the surgeon, with difficult decisions at all stages of their evolution. Clinical Course and Emergency Treatment Airway stenosis usually becomes symptomatic soon after extubation. Nevertheless, in a previous study we found that 22 of 74 stenoses occurred later than the first month after extubation and caused severe difficulties when occurring in nonspecialized medical environment. Systematic endoscopic examination of all patients who underwent respiratory resuscitation reduced the necessity of emergency treatment. However, the ability to provide such care without pitfalls or inappropriate measures remains important. Acute distress from such a lesion not responding to antibiotic and steroid therapy requires immediate rigid bronchoscopy. This allows ventilation and oxygenation of the patient, dilation of the stenosis, and distal toilet of the airway. Laser coagulation of granuloma and endoluminal scars immediately reestablishes adequate tracheal caliber, allows comfortable breathing, and gives time for preoperative assessment and convenient preparation for definitive operation. If rigid endoscopy is not available, we recommend forceful orotracheal intubation using a small-caliber tube, which will permit adequate oxygenation and ventilation and should allow distal bronchial toilet, that being one of the most important steps in palliative treatment. A few hours later, the small tube can be replaced by a larger one in the molded bed of the stenosis and then the stabilized patient may be transferred to a specialized center for further treatment. Two pitfalls must be avoided at this moment: first, the creation of a tracheostomy below the stenotic segment, which augments the length of the tracheal injury and renders the final surgical treatment much more difficult, and second, an emergency surgical resection anastomosis in an infected, inflamed, and unprepared field, which significantly increases the risk of failure of any type of surgical correction. Limits and Contraindications to Operation Important difficulties may arise from the general status of the patient and from the local aspect of the lesions. Each of these difficulties is by itself a large field of discussion, where our answer was not always the best one. Old age, per se, is not a contraindication to tracheal resection. On the contrary, the aged should have as good oxygenation as anyone through a good airway and they suffer more than anyone from multiple manipulations and the problems that can result from permanent intubation and prosthesis. Three patients were more than 80 years old in our series, the oldest being 88 years. One of them died of a cerebral accident, underlining that old age is certainly a risk factor. After brain trauma, residual neuropsychological sequelae may be a problem when considering the patient for tracheal resection and closing tracheostomy. We require that the patient at least be able to swallow without aspirating cough and expectorate on command. These requirements will facilitate the postoperative course and avoid too frequent tracheobronchial suctions. The timing of resection should be coordinated with the rehabilitation physician. It should be remembered that the reestablishment of normal airway and voice in these patients often will be followed by a rapid improvement in their neurologic status. Another difficult decision is whether to resect or permanently tracheotomize patients with severe chronic respiratory insufficiency. It is difficult to predict which patients will be at risk for early relapse of acute respiratory failure. One patient in our series died in the postoperative period with pyocyanic reinfection, and 2 others with an excellent surgical result were retracheotomized in the following months for a recurrent postinfectious respiratory failure episode. When that can be anticipated these patients are best managed with endoluminal stents or permanent tracheostomy, which allows night-time oxygenation or ventilation when necessary. Another very difficult problem may arise with very long stenoses or bifocal stenoses. These raise the question of the maximal length of trachea that can be resected and reanastomosed safely. Because the tension at the anastomotic level appears to be an important prognostic factor after tracheal anastomosis, Mulliken and Grillo [9] have meticulously measured the average amount of trachea that can be resected and the resulting tension required for reapproximation of the tracheal ends as well as the effect of the right hilar dissection and carinal mobilization. The maximal length of resection generally is taken to be 5 to 6 cm or nine to ten tracheal rings. We find these indications of limited practical value because fibrotic stenosis may retract the trachea and "telescope" or completely destroy the rings. Moreover, the resectable length varies with patient morphology and age. A longer resection is possible in taller patients with a long neck because the neck flexion permits lowering of the tension at reaproximation of the tracheal ends more than in shorter patients. Furthermore, the elasticity of the trachea in younger patients permits easier tension-free approximation than the corresponding airway in older patients with fibrous and sometimes calcified tracheal structure. Prior attempts at repair also render reapproximation more difficult and dangerous. All the above factors must be taken into account when the feasibility of resection and anastomosis is evaluated. After Mulliken and Grillo we proposed that half of the trachea could represent the length of trachea that is safely resectable. Appropriate allowance to do a little less or more depends on circumstances. In some cases, lengthening the time

5 254 COURAUD ET AL Ann Thorac Surg AIRWAY STENOS1S 1995;60: Fig 1. (A) Plain tomography qf a very long tracheal stenosis in a 67- year-old man. (B) Resection and anastomosis combined with thyrohyoid release and carinal mobilization resulted in success. Take notice of the important carinal ascension. of preparation permits increasing fibrosis and retraction of the stenosed segment, which brings it within the limits of resectability. Various lengthening procedures are available to allow extended resections and tension-free repair. Laryngeal releases proposed by Dedo and Fishman [10] and by Montgomery [11] give 2 to 3 cm of additional length for end to end approximation. However, both techniques may cause significant problems of aspiration, and whenever a sternotomy is necessary we prefer hilar and carinal mobilization, which may give 2 to 4 cm (Fig 1). Flexion of the neck is used commonly in our experience, maintaining the chin fixed to the presternal skin with two strong sutures. This is used in all lengthy resections, in pediatric patients, and in patients with head injury whenever we are not sure that they will be fully cooperative. Technical Strategy and Indications Tracheal resection and end-to-end anastomosis is now a well-established technique under weu-codified indications, which requires no commentary. Laryngotracheal surgery is much more difficult. Indications for various techniques may vary from millimeter to millimeter according to the extension of the lesions, and the technical execution must pay attention to both anatomic and functional results. Two factors must be considered in the preoperative assessment to define the surgical procedure. First, the precise topography of the injury must be defined with exact location and length of the stenosis, the length of the tracheal involvement, and the amount of cricoid destruction. Second, functional evaluation of the glottis is of capital importance, with reference to glottis patency as well as mobility of the vocal cords that may be affected by scarring, recurrent nerve injury, or ankylosis of the cricoarytenoid joints. To define the lesions, preoperative assessment should include awake-flexible as well as anesthetized-rigid endo- scopic evaluation, laryngotracheal tomography, and finecut computed tomography of the larynx. Despite all the information obtained from these investigations the extent of involvement of the various structures may sometimes only be fully assessed during operation, so the planned interventions often have to be modified intraoperatively. These considerations are the basis of our policy in regard to treatment modalities. 1. Tracheal and laryngeal sublottic senoses with open and mobile glottis are best managed by resection of the involved trachea and of the cricoid ring according to the technique described by Gerwat and Bryce [5] and Pearson and associates [6] (Fig 2). More or less of the cricoid plate can be excised until the resection reaches the margin of healthy mucosa. A posterior tracheal mucosal flap may be fashioned as described by Grillo [12] to cover with mucosa the anterior part of the cricoid plate up to the basis of the vocal cords if it is bared after resection of fibrous and inflammatory tissues. Our technique of subglottic laryngotracheal resection has evolved to appreciable modifications of the original technique. We avoid identification of the recurrent nerves, plication of the posterior membranous trachea, and subanastomotic tracheostomy. We find that subperichondrial cricoid resection by curette protects the laryngeal nerves from injury and obviates the need for dissection in fibrotic scar tissue, which may itself compromise the nerves either directly or by interference with their blood supply. Furthermore, the perichondrium left in place provides a strong support to the tracheal reimplantation. We avoid tracheostomy to prevent further tracheal injury and possible tracheal ischemia between the stoma and the anastomosis. Early postoperative extubation is possible in most patients without subsequent respiratory problems provided laryngeal edema is prevented by the administration of corticosteroids (methylprednisolome, 40 to 60 mg]day for 5 to 7 days). If at extubation there is concern about the

6 Ann Thorac Surg COURAUD ET AL ;60: AIRWAY STENOSIS Fig 2. Tracheal and subglottic stenosis in a 23-year-old man (A) was cured with an excellent result after tracheal and subglottic resection and anastomosis (B). degree of laryngeal edema or vocal cord dysfunction we opt to leave a silicone nasotracheal tube with a noninflated cuff in situ for 2 to 4 days. In general, when a tracheal tube has to be left in place a short time for general or local indications after tracheal resection and anastomosis we recommend a deflated cuff to avoid mucosal and submucosal ischemia and necrosis. 2. Tracheal and subglottic stenosis associated with total destruction of the cricoid posterior plate by chondritis and necrosis justifies extended resection and anastomosis and total ablation of the cartilage with conservation of the vocal cords provided that the laryngeal patency and the vocal cord opening are maintained with a Montgomery T tube for 3 to 6 months, until the cricoid posterior wall is rigid enough to be capable of supporting the arytenoids and abduction of the vocal cords (Fig 3). Total circoid resection generally is not undertaken deliberately; the aim should be to preserve a rim or bridge of superior cricoid cartilage that will serve to support the arytenoids and thus the glottic opening. However, previous chondritis or cartilage necrosis may result in the entire rigid structure being removed during the curettage. When this happens the perichondrium is preserved and cleaned carefully, a periostial or a perichondrial graft is inserted in the bed of the cricoid plate, and the laryngotracheal anastomosis with mucomucosal apposition is carried out as previously mentioned. The graft is taken from the anterior extremi~ of the fifth rib, which is easy to reach in the large operative field including the neck and the thorax that we prepare for tracheal operations. The graft can be slightly fixed with three to four thin knots or not because the cricoid resection is fashioned subperichondrally, leaving good tissue ahead of the graft, covering it and allowing a strong and reinforced rnuco-mucosal anastomosis. For lack of covering with mucosa, a posterior tracheal flap such as Grillo's can provide a good substitute. The laryngeal collapse is prevented by the insertion of a Montgomery T tube arising just above the level of the vocal cords. This will be removed after 3 to 6 months according to the importance of the perichondrial structures left in place and consequently of the estimated delay required to achieve competent rigidity of the laryngeal posterior wall (Fig 4). When removing the tube at 3 months, the findings of a residual laryngeal softness and collapse led us to put in a new T tube for 3 further months in 2 patients; the insertion is easy to do endoscopically. 3. Total laryngeal glottic and subglottic obstruction with destruction of the cricoid skeleton cannot be managed other than by extended laryngeal resection followed Fig 3. Subglottic stenosis with total cricoid resection: schematic view qf the lesions, the specimen, and the operative technique with anastomosis and stenting.

7 256 COURAUD ET AL Ann Thorac Surg AIRWAY STENOSIS 1995;60: Fig 4. (A) Specimen of a total cricoid resection in a 55-year-old woman with idiopathic laryngotracheal stenosis. (B) Postoperative plain tomogram 6 months after resection and anastomosis and laryngeal stenting. The patient is stable at 8 years of follow-up. by supraglottic laryngotracheal anastomosis or by definitive tracheostomy. Such an aggressive surgical policy for a benign disease was justified in 3 patients in our series, allowing the removal of the tracheostomy tube of a 13-year-old girl who was tracheotomized since the age of 1 year. The laryngeal resection is managed up to the level of a patent laryngeal lumen, where the trachea is anastomosed as usual to the supraglottic wall. Subperichondrial resection of the posterior wall of the larynx preserves posterior muscle insertions, and this prevents the important swallowing difficulties often seen after extended laryngeal resection and anastomosis for malignancy. Rehabilitation of an acceptable voice was not a problem in our short series (Fig 5). 4. Short subglottic stenoses that do not involve more than one or two tracheal rings but are associated with glottic opening difficulties are not subject to resection but to laryngeal enlargement and molding. For these patients, who often have palsy or rigidity of one or two vocal cords with a well-preserved laryngeal skeleton, we perform laryngeal and glottic enlargement of the glottosubglottic segment through a vertical posterior and anterior cricoidotomy (Fig 6). The separation of the edges is widened to 3 to 5 mm and a thin slice of periostial free graft is inserted between them in the bed of the posterior cricoidotorny. The vocal cord abduction and the adequacy of the glottic passage are checked and, if they are inadequate, similar insertion between the margins of the anterior section can be achieved giving a complementary enlargement. The section of a fibrous scar between the anterior part of the vocal cords enlarging the glottic slot may be necessary also. There is a balance to find between breathing and speaking because a restricted laryngeal enlargement leaves the patient dyspneic on exertion and too wide a glottic opening makes speech rehabilitation more difficult. The posterior graft is fixed with some thin knots to the edges of the cricoidotomy until -I- tube stent positioning maintains it in place with an appropriate shape. The anterior graft, when necessary, is similarly sewn to the anterior slot over the tube. The width of the larynx is maintained over the -I- tube for 3 months until the rigidity of the posterior graft supports the posterior diastasis of the cricoidotomy (Fig 7). Large fibrotic scars must be resected as completely as possible and the wound must be covered with mucosa as fully as possible, especially in children, in whom conservation of an important block of fibrosis may be the cause of impaired laryngeal growth [13]. Many similar techniques were proposed by the ear, nose, and throat surgeons (Cotton [14], Montgomery [15]), and this approach was very popular before the description of laryngeal and subglottic resection for the treatment of tracheal and subglottic stenosis in children as well as in adults. In our initial experience before 1978 it represents 16 cases among 26 stenoses in comparison with 11 of 97 in the present series. Now we consider this technique to be indicated only when a sclerotic or functional glottic stenosis necessitates glottic enlargement. 5. Traumatic total laryngotracheal break of continuity after blunt trauma on the anterior aspect of the neck results in stenosis of the laryngotracheal junction when mismanaged in the acute period [16]. Later the patients have either a nasotracheal tube, which is impossible to remove without causing asphyxia, or an evident total obstruction of the airway above a tracheostomy. The diagnosis of total laryngotracheal separation can be based on the previous history, confirmed by plain or computed tomography, or at the operative findings as a discontinuity of the laryngeal and tracheal structures and a retraction of the lower part of the trachea into the mediastinum. The treatment has to take into account that a disruption of one or both recurrent nerves generally is associated with large laryngotracheal separations. A previously nonevident functional glottic closure therefore may compromise the success of a simple subjacent anatomic restoration of the airway continuity.

8 Ann Thorac Surg COURAUD ET AL ;60: AIRWAY S]TENOSIS Fig 5. Laryngeal endoscopic view 3 months after laryngeal resection and tracheal supraglottic anastomosis in a 13-year-old girl who had a tracheostomy since the age of 8 months. She had a bifocal stenosis. The tracheal stenosis was resected at the age of I year. The laryngeal stenosis later underwent two failed molding repairs, which resulted in total laryngeal destruction and permanent tracheostomy. Take notice ql: a slight prolapse of the posterior mucosa in the tracheal lumen, successfully treated through photocoagulation at 6 months. The result is stable at 10 years of follow-up. Consequently, the resection of the stenosed laryngotracheal segment must be complemented by glottic enlargement. In general, the laryngotracheal approximation is relatively easy after mobilization and elevation of the trachea. At the laryngeal level, the retraction of the torn mucosa into the subglottic lumen makes it necessary partially to resect the cricoid ring subperichondrally to permit a convenient muco-mucosal apposition at the laryngotracheal anastomosis. Fig 7. Plain tomography before (A) and after (B) laryngeal molding in an 18-year-old girl with subglottic and glo~'c stenosis. Fig 6. Glottic and subglottic stenosis: schematic view of the lesions (A) and the surgical enlargement with stenting (B). B The glottic enlargement can be achieved through a vertical posterior section of the cricoid plate according to Rethi's fashion, then the vocal cord abduction and the laryngeal molding are stented with a nasotracheal or T tube for 1 to 3 months (Fig 8). When the vocal cord abduction is slightly insufficient at extubation, laser cord resection may secondarily complete the operation as performed once in our experience. In our hands that combined technique provided good respiratory and fair phonetic primary successes, in contrast with many disappointing results and reoperations reported elsewere after such trauma (Fig 9). 6. There are very complex laryngeal lesions, most frequently as a result of one or more attempts at surgical repair of laryngotracheal stenoses. These unclassifiable lesions generally include chondritis, deep fibrosis, vocal

9 258 COURAUD ET AL Ann Thorac Surg AIRWAY STENOSIS 1995;60: Fig 8. Therapeutic options after posttraumatic total laryngotracheal separation and disruption of recurrent nerves. Cricofd margin resection up to the healthy mucosa was used for reanastomosis and glot~c enlargement over a stent palliated the double vocal cord palsy. cord palsy, and more or less laryngotracheal collapse and obstruction. Their surgical repair cannot be standardized but generally combines resection and anastomosis, aryngeal plasty, and remodeling over a T-tube stent (Fig 10). Covering the molding bed with mucosa as completely as possible is important, and resorting to every trick for that must be advised. In some cases one or two paralized vocal cords were fixed to the lateral wall of the thyroid cartilage or sutured into the laryngotracheal anastomosis to palliate a definitive vocal cord adduction. The high level of laryngeal alterations often necessitates very high stenting with the upper arm of the T tube open up to the laryngeal structures. Consequently, the patient is exposed to permanent saliva aspiration and impossibility of feeding orally. The closing of the upper end of the tube obviates these complications. The patient can breathe through the external branch, which has to be left open. Because meticulous cleaning of the lower branch, ultrasonic nebulizations maintaining humidity in the airways, and avoiding obstructive fibrinous scabs can be kept up, the technique may be used with acceptable comfort and security. 7. Multifocal stenoses may be the cause of very difficult problems. When bifocal distant tracheal stenosis could not be resected in the same tracheal segment, we never did both segmental resections at the same stage but gave priority to the tightest. If that was not convenient we considered that the long tracheal destruction was an indication for long stenting. Fortunately the most frequently associated stenoses are laryngeal and tracheal. In our initial experience, before 1978, we did multistaged procedures, the upper stenosis generally being cured first. We have modified our policy in the present series. In such cases combining tracheal resection and laryngeal enlargement and molding in a single-stage procedure was achieved frequently. The method is safe because the laryngeal stenting tube is cuffiess and therefore does not threaten the vascular supply of the lower tracheal segment through compressive mucosal ischemia. Conclusion The strategy for treatment of airway stenoses is now well established and leads to a high level of success with minimal or no sequelae [17]. Tracheal stenoses often are very easy to cure. The treatment of laryngo- Fig 9. (A) Plain tomogram of a total stenosis of the laryngotracheal junction after cervical trauma in a 53-year-old man. Take notice of the loss of continuity and alignment of the airway. (13) Postoperative view 3 months after the above-mentioned operation.

10 Ann Thorac Surg COURAUD ET AL ;60: AIRWAY STENOS1S Fig 10. (A) Total complex laryngotracheal stenosis in a 26-year-old man after previous attempts at operagon. (B) Postoperat4ve plain tomogram after tracheal and laryngeal resection with 3 months of stenting, Take notice of the laryngeal rigidly, which resulted in a definitive hoarseness, tracheal stenoses is less simple, but the most difficult cases arise from failure of previous surgical attempts. These add extensive destruction and sometimes sequelae to the original iatrogenic lesions. Although a solution usually could be found with an acceptable morbidity, as seen in this series, where surgical efforts were rewarded in most patients by restoration of breathing and speaking through the normal route, the necessity of a per primam recovery must be emphasized. In our opinion, meticulous preoperative assessment and preparation are as important as appropriate surgical options and a good surgical technique. References 1. Grillo HC. Circumferiential resection and reconstruction of mediastinal and cervical trachea. Ann Surg 1965;162: Grillo HC. Notes on the windpipe. Ann Thorac Surg 1989; 47: Pearson FG, Andrews MJ. Detection and management of tracheal stenosis following cuffed tube tracheostomy. Ann Thorac Surg 1971;12: Couraud L, Chevalier P, Bruneteau A, Dupont P. Le traitement des st6noses trach6ales apr6s trach6otomie. Indications th6rapeutiques, pr6paration ~ l'intervention. A propos de 9 cas de r6section pour 15 cas en insuffisance respiratoire aigu~. Ann Chir Chir Thorac Cardiovasc 1969;8: Gerwat J, Bryce DP. The management of subglottic laryngeal stenosis by resection and direct anastomosis. Laryngoscope 1974;84: Pearson FGr Cooper JD, Nelems JM, Van Nostrand AWP. Primary tracheal anastomosis after resection of the cricoid cartilage with preservation of recurrent laryngeal nerves. J Thorac Cardiovasc Surg 1975;70: Couraud L, Martigne C, Houdelette P, Dumas PJ, Morales F. Int6r6t de la r6section cricoidienne dans le traitement des st6noses cricotrach6ales apr6s intubation. Ann Chir Chir Thorac Cardiovasc 1979;33: Nosher SAM, Dromer C, Velly JF, Labrousse L, Couraud L. Expanding wire stents in benign tracheobronchial disease: indications and complications. Ann Thorac Surg 1992;54: Mulliken JB, Grillo HC. The limits of tracheal resection with primary anastomosis: further anatomical studies in man. J Thorac Cardiovasc Surg 1968;55: Dedo HH, Fishman NH. Laryngeal release and sleeve resection for tracheal stenosis. Ann Otol Rhinol Laryngol 1969;78: Montgomery WW. Suprahyoid release for tracheal anastomosis. Arch Otolaryngol 1974;99: Grillo HC. Primary reconstruction of airway after resection of subglottic laryngeal and upper tracheal stenosis. Ann Thorac Surg 1982;33: Couraud L, Moreau JM, Velly JF. The growth of circumferential scars of the major airways from infancy to adulthood. Eur J Cardiothorac Surg 1990;4: Cotton RT. Management of subglottic stenosis in infancy and childhood. Ann Otol Rhinol Laryngol 1978;87: Montgomery WW. T-tube tracheal stent. Arch Otolaryngol 1965;82:320-1, 16. Couraud L, Velly JF, Martigne C, N'Diaye M. Post-traumatic disruption of the laryngo-tracheal junction. Eur J Cardiothorac Surg 1989;3: Grillo HC, Donahue DM, Mathisen DJ, Wain JC, Wright CD. Postintubation tracheal stenosis: treatment and results. J Thorac Cardiovasc Surg 1995;109: DISCUSSION DR F. GRIFFITH PEARSON (Toronto, Ont, Canada): Doctor Couraud, I congratulate you on a clear presentation of a very important experience. I have known Dr Couraud since he organized an international meeting in his hometown of Bordeaux in That meeting was the first combined meeting of thoracic surgeons and otolaryngologists that I ever attended. As you see from Dr Couraud's presentation, he has been working for a long time in an area in which few thoracic surgeons have experience. 1 visited Dr Couraud in Bordeaux and know that he is a pragmatic, experienced clinician and an extremely skilled technical surgeon, and his observations are critical and original. He learns from his own observations, and has made some very original contributions, which have been presented here today. He also has carefully followed up these patients. Endoscopic follow-up was done early, and then again at I year and at 3 years, so this is a well-documented experience. He emphasizes the principles for success: These are (1) waiting until all the inflammatory process has subsided, (2} removal

11 260 COURAUD ET AL Ann Thorac Surg AIRWAY STENOS1S 1995;60: of all of the chronically, irreversibly damaged airway, (3) preserving circulation, (4) careful mucosal apposition, and (5) minimizing tension. The results speak for themselves, and are outstanding. Some of the very original contributions in this work are those cases where the larynx is damaged or destroyed. I first heard Dr Couraud's report of some of these cases 5 or 6 years ago. Where the entire cricoid was lost, (it was a sequestrum I think at operation), he has sewn the distal trachea to the larynx. Seven cases are reported with seven successful results. All 7 required a prolonged period of stenting with a T tube, but the end result has been, without doubt, very much better than that of an open tracheostomy. He also identifies that evaluation of the results of these high laryngotracheal reconstructions requires waiting for up to a year or so before maximal recovery is achieved. I wish to ask a couple of questions, Professor Couraud: You describe in your manuscript the routine use of postoperative steroids, something that many of us have been hesitant to use, although the experience with lung transplantation has probably encouraged this practice. Could you comment on or describe what tissue you are using as a free periosteal graft when you do the posterior cricoid split or Rethi type operation? We have been using a mucosal flap fashioned from the distal trachea to reconstruct posterior defects, and do not usually split the cricoid cartilage behind. I enjoyed the report. DR HERMES C. GRILLO (Boston, MA): 1 too am enormously impressed by the excellent results that Professor Couraud demonstrated to you today and 1 really have little to add to Dr Pearson's extensive and very appropriate discussion and plaudits for this report. Thoracic surgeons in North America, including Dr Pearson and myself, have not ventured above the glottis. We usually turn that terrain over to the otolaryngologists. It is not entirely jurisdictional; it probably is partly a lack of courage. But Dr Couraud has tackled supraglottic lesions and glottic lesions, which may necessitate total removal of the cricoid, lesions that we do not tackle. Perhaps parenthetically it is also a warning to our colleagues who produce these lesions, that maybe prolonged use of the endotracheal tube rather than a tracheostomv is indeed hazardous. The one good thing you can say about a tracheostomy lesion is you almost always can fix it. Just a couple of questions. When you do perform a supraglottic anastomosis of the trachea to what is left of the larynx, have your patients been troubled by aspiration on deglutition? The protective reflexes are gone. How do they manage? I believe you prefer Dr Pearson's technique of suturing the ends of the first cartilage together and attaching that ring to the larynx. Why does that not provide a substitute for the cricoid? I might also note that Professor Couraud's overall results are almost identical with those that we reported recently in a cumulative series of 502 patients with postintubation lesions [1]. DR COURAUD: Thank you, Dr Pearson and Dr Grillo, for your kind comments. Your first question, Dr Pearson, relates to our routine use of postoperative steroids. The main advantage of the medication is to prevent laryngeal edema or suture line edema postoperatively, which might compromise airway patency and initiate severe respiratory problems. The goal of the treatment was to avoid postoperative tracheostomy, and it was almost always effective. We have used steroids postoperatively since the beginning of our experience, 1962, and we do not think that thev impaired or delayed tracheobronchial healing. The widespread experience in use of steroids after lung transplantation indicates, as you suggest, Dr Pearson, that the medication is useful and harmless. The indication for a posterior cricoid split, as in the Rethi type operation, is to ensure glottic enlargement as necessary after bilateral vocal cord paralysis in adduction or after impairment of the glottis opening by posterior fibrous scars. The two parts of the posterior cricoid plate must be separated to maintain the interarytenoid space and vocal cord abduction. With this intention a periostal graft from a rib is inserted in the posterior split, providing support for vocal cord abduction and accelerating recovery of posterior laryngeal wall ridigity. After glottic or supraglottic resection and anastomosis for nontumoral stenosis, no patient experienced problems with aspiration or deglutition in the long term, Dr Grillo. Nevertheless a few patients had temporary problems with liquid food, mainly after brain trauma and hemilaryngeal paralysis. Only 2 patients needed, respectively, one and three sessions of swallowing reeducation before total recovery. None needed gastrostomy. I believe that the low incidence of swallowing problems after resection for benign stricture is the result of cartilaginous resection subperichondrially without alteration of the laryngeal muscles and nerves. Operations for malignancies, where resection was carried out extraperichondrially and sometimes with sacrifice of muscles and recurrent nerve, induced frequent problems with swallowing and aspiration. A large number of such patients needed gastrostomy at least during the time of postoperative irradiation as well as many sessions of rehabilitation for weeks or months. Nevertheless, at 1-year follow-up all patients had recovered a normal capacity for feeding orally. After tracheal and subglottic resection and anastomosis according to Pearson's type of operation, we never suture the posterior end of the first tracheal cartilage to form a ring. I am afraid that is not useful because a rim of cricoid plate is left in place. Our policy is to extubate the patient immediately in the operating room and never to add tracheostomy except for laryngeal molding purpose after laryngeal enlargement or total cricoid resection. In fact, immediate extubation was almost always possible. Nevertheless some patients awakened very slowly, mainly after brain trauma, and needed to have tracheobronchial suctioning in the postoperative hours. Others had some degree of laryngeal edema after very highly extended Pearson procedures with anastomosis up to the vocal cords or after mucosal detachment higher than the vocal cord attachments. In these few cases we were afraid that the glottic lumen could be compromised during the first postoperative night and that the ward team might have to perform what could be a damaging reintubation. In this limited number of cases we prefered to avoid all risk of reintubation by leaving in place a nasotracheal tube for 24 to 48 hours. The decision is taken in the operating room after routine postoperative endoscopic control when flexible fibroscopy shows a major narrowing of the larynx by edema. We can insert a nasotracheal tube without extending the neck, sliding it over the flberscope. We recommend a cuffless tube to avoid mucosal and submucosal compression and ischemia, because the submucosal circulation contributes importantly to ensure the blood supply to the trachea at the anastomosis, where tracheal dissection and mobilization deletes the external vascularization. Reference 1. Grillo HC, Donahue DM, Mathisen DJ, Wain JC, Wright CD. Postintubation tracheal stenosis. Treatment and results. J Thorac Cardiovasc Surg 1995;109: