Cricotracheal Resection With Hilar Release for Pediatric Airway Stenosis FREE

Subglottic stenosis (SGS) remains a challenging surgical problem for the pediatric otolaryngologist. Historically, SGS resulted from infection; however, the standardization of immunizations coupled with the widespread adoption of prolonged neonatal intubation incited a dramatic shift in the incidence and etiology of SGS development. In 2003, more than 2000 SGS hospital admissions were captured by Kids Inpatient Database (KID), and approximately 90% of patients reported a history of endotracheal intubation.¹

To treat SGS, pediatric airway reconstruction has evolved rapidly over the previous 30 years. Decannulation rates have improved substantially since the initial introduction of expansion laryngotracheal reconstruction in the early 1970s. More recently, cricotracheal resection (CTR) has been described as an alternative and complimentary technique. Traditionally, surgical centers reserve CTR for severe SGS with a sufficient cuff between the superior extent of the stenosis and the true vocal cords in children weighing more than 10 kg. A recent Swiss series highlights the value of this approach with its reported overall decannulation rate of 92% for SGS of grade 3 or 4 in 57 children observed over a median follow-up of 5.1 years.²

Anastomotic dehiscence remains a feared risk inherent to CTR not associated with laryngotracheal reconstruction. Even when implementing hyoid release techniques, if the extent of required tracheal resection is greater than about 5 rings of tracheal cartilage, a 2-staged procedure is typically required to achieve ultimate decannulation. In the same Swiss series by Jaquet et al,² 3 of 57 patients had dehiscence-related problems, and 19 of 57 cases required a double-staged procedure.

The creation of a tension-free anastomosis increases the rate of surgical success. To this end, suprahyoid, infrahyoid, and hilar releases have been developed.^3- 5 The risk of anastomotic dehiscence and airway complications has been shown to increase with longer resections for combined pediatric tracheal and laryngotracheal resections.⁶ Whether an increased risk of failure specifically applies to longer CTRs requires further observation, but half of decannulation failures in one series had undergone suprahyoid release.⁷ Despite its ability to increase mobilization, hilar release has not been noted to be routinely used in the thoracic surgery literature.^8- 10

To address these limitations, the present series puts forward the use of routine hilar release in high-grade, long-segment CTR as a means to diminish risk of dehiscence and make single-staged procedures permissible in most cases with an acceptable margin of added risk.

After obtaining approval from our institutional review board, we reviewed the medical records of 16 children who underwent CTR with hilar release by our pediatric thoracic airway team. The variables that were analyzed included age, sex, grade of stenosis, comorbidities, necessity of infrahyoid release, length of stay, decannulation status, and complications.

Our surgical approach for CTR with hilar release is via a right posterolateral thoracotomy as follows: After the induction of general endotracheal anesthesia and the institution of single left lung ventilation, the patient is placed in the left lateral decubitus position. A limited right posterolateral thoracotomy is performed through the fourth intercostal space. The inferior pulmonary ligament is divided to mobilize the right lower lobe and expose the pericardium below the hilum. The pericardium is then opened just below the right lower pulmonary vein and posterior to the lateral attachment of the inferior vena cava to the pericardium. The pericardium is then opened circumferentially around the hilum. In addition to the hilar structures, care is taken to avoid injury to the phrenic nerve anteriorly, to the vagus nerve and esophagus posteriorly, and when dissecting the inferior vena cava off the pericardium inferiorly.

The azygous vein is doubly ligated and divided, and the pleura overlying the trachea and right mainstem bronchus is incised. The right and left mainstem bronchi, carina, and distal trachea are mobilized on their anterior and posterior surfaces, while preserving the lateral vascular attachments. Once these maneuvers have been accomplished, it is generally possible to advance the carina several centimeters. A thoracotomy tube is placed and the wound is then closed in the usual fashion.

The patient is returned to the supine position after the hilar release. A horizontal elliptical incision that incorporates the tracheostoma is made between the sternocleidomastoid muscles. Subplatysmal flaps are elevated superiorly and inferiorly. The strap muscles are identified and followed to the hyoid, and a suprahyoid release is routinely performed. Next, the strap muscles are separated in the midline and the trachea is skeletonized, taking care to prevent injury to the recurrent laryngeal nerves. The trachea is separated from the esophagus and dissected into the thorax where the dissection will come in continuity with that obtained from the intrathoracic approach. The trachea is then divided, and the stenotic portion of the trachea is resected to the level of healthy-appearing mucosa along with the anterior portion of the cricoid. After identification of the innominate artery, two 2-0 sutures are placed para-midline to elevate and suspend the trachea to the manubrium. This prevents the divided trachea from retracting into the chest and maximizes the benefit from the thoracic release.

The anastomosis is performed with 4-0 antibiotic-impregnated suture. The knots are tied on the exterior of the lumen. The anesthesia team transnasally places an appropriately sized endotracheal tube, which is guided distally through the anastomosis site prior to the anterior sutures being placed. The integrity of the closure is tested by giving positive pressure via a mask. TISSEEL (Baxter Healthcare Corporation, Glendale, California) is placed over the closure. Jackson-Pratt drains are placed laterally and deep and are maintained on low-pressure continuous wall suction. The skin is closed with deep and superficial interrupted sutures. A Grillo stitch is placed circummandibularly to the periosteum of the manubrium to maintain chin approximation to the sternum. Flexible bronchoscopy is performed to suction all aspirated blood and confirm location of the endotracheal tube. The endotracheal tube is secured to the septum with nylon suture.

Postoperative management is a collaborative effort between the pediatric intensivists, pharmacists, physical therapists, occupational therapists, anesthesiologists, thoracic surgeons, and otolaryngologists. Patients are kept sedated in the pediatric intensive care unit for 1 week. The patient is then brought to the operating room for suture removal, Grillo stitch removal, direct laryngoscopy and bronchoscopy of the anastomosis, and extubation. Following this, the patient remains extubated with a modified Minerva brace in place to prevent neck hyperextension. Corticosteroids and racemic epinephrine are given for 48 hours, and the patient is weaned from sedatives. Weekly endoscopies are performed with intermittent laser treatment of granulation tissue as needed until the airway has stabilized.

Sixteen patients were identified who underwent CTR and intrathoracic release in a 5-year period (Table 1). The median age was 34 months, and 4 patients were male and 12, female. Ten of the patients had high grade 3 stenosis (usually 99%) on the Cotton-Meyer scale. Five patients had complete or grade 4 stenosis. All stenoses were dense, fibrotic, endoluminal, and presumably noncongenital. One patient had severe, long-segment tracheomalacia. Six patients had glottic involvement, and 2 had separate pericarinal involvement requiring treatment. Two patients were revision cases following previously failed laryngotracheal reconstruction. Prematurity and bronchopulmonary dysplasia were the most common comorbidities.

With hilar release in children, we are generally capable of advancing the trachea several centimeters, which is often sufficient to bring the carina into the lower neck if desired. For us, this mobilization has directly translated into lower anastomotic tension, a decreased need for infrahyoid releases, the capability of resecting marginal tissue that might otherwise have been spared, and the ability to excise stenoses that might otherwise be unresectable.

Of the 16 patients, 15 were decannulated successfully. The one exception was a quadriplegic child who remains ventilator-dependent and for whom speech has been restored. All patients underwent a single-stage operation with resection of the tracheostomy site and at least 4 tracheal rings. One patient had a near-total resection (all but 2.2 cm) of the trachea. One patient required a second tracheostomy for pulmonary failure but was subsequently decannulated. Eleven patients did not require an infrahyoid release. All patients with the exception of one who remains gastrostomy tube dependent returned to age-appropriate deglutition. There were no incidents of anastomotic breakdown, subcutaneous air, or wound infection. On average, each patient underwent 4 postoperative endoscopic laser procedures prior to stabilization of the anastomotic site. The complications seen in our series are listed in Table 2. Fifty percent of patients had a pulmonary complication. Six patients required nonotolaryngology operations for complications. The overall length of stay ranged from 15 to 38 days, with a mean duration of 26 days.

No new cases of laryngeal paralysis were noted. One patient had preoperative left vocal cord paralysis associated with a patent ductus arteriosus ligation. Four patients had partial paresis with less robust motion of a vocal cord (3 left, 1 right), which resolved without treatment. None of these children had problems with deglutition.

Cricotracheal resection has become well established in the surgical armamentarium for treatment of pediatric SGS. Debate continues regarding the indications for laryngotracheal reconstruction and CTR, but mounting experience favors the use of CTR for high-grade stenosis. Decannulation rates of higher than 90% for grades 3 and 4 stenosis are common.^7,11 Our results compare favorably with those of others, with decannulation of all patients except for a quadriplegic child who continues to require mechanical ventilation but is now able to talk. All of our patients had high-grade, long-segment stenosis. However, with the use of releasing techniques, long segments of the airway were able to be resected while still providing a tension-free closure. We particularly note that with this technique, compromised mucosa does not need to be spared to prevent undo tension on the anastomosis. We also note that the inferior descent of the larynx and laryngomalacia often seen with CTR are uncommon with intrathoracic release unless the resection involves most of the trachea.

A most serious surgical complication reported in other series is dehiscence of the thyrotracheal anastamosis.^2,7,12 We routinely take several measures to prevent this complication, including hilar release, suprahyoid release, and the application of several anchoring sutures to provide a tension-free closure. We now routinely avoid infrahyoid release. In addition, we have developed an anchoring stitch that tacks the distal tracheal segment to the manubrium to prevent tracheal retraction into the chest, maximizes the benefit from the intrathoracic release, and facilitates closure of the anastomosis. The use of fibrin glue provides additional insurance against anastomotic leak. With these measures, we had no anastomotic leak or dehiscence in our series.

We had a low complication rate with the use of the hilar release, though we did have several patients with postoperative pneumothorax and chylothorax, most of which were treated conservatively.

One complication that has been described by others is temporary postoperative swallowing dysfunction.² All of our patients had nasogastric tubes placed in the immediate postoperative period. However, prolonged swallowing dysfunction was uncommon, with only 1 patient requiring gastrostomy tube placement for silent aspiration of thin liquids.

In conclusion, although hilar release performed in conjunction with CTR requires dual service operative management and introduces some novel complications to the perioperative period, its application was associated with decannulation of all patients in a single stage except for 1 quadriplegic child who continues to require mechanical ventilation. There were no episodes of dehiscence of the thyrotracheal anastomosis. Overall, this technique appears safe and effective and should be considered when performing CTR for extended-length, high grade 3 and grade 4 SGS, where the thyrotracheal anastomosis might otherwise be under excessive tension or require the preservation of marginal tissue.

Correspondence: Glenn E. Green, MD, Division of Pediatric Otolaryngology, University of Michigan, F6905 Mott Children's Hospital, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0241 (gegreen@med.umich.edu).

Submitted for Publication: May 8, 2009; final revision received September 14, 2009; accepted September 29, 2009.

Author Contributions: Dr Green had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Taylor, Collar, Ohye, and Green. Acquisition of data: Taylor, Collar, Wilson, and Green. Analysis and interpretation of data: Taylor, Collar, and Green. Drafting of the manuscript: Taylor, Collar, Wilson, Ohye, and Green. Critical revision of the manuscript for important intellectual content: Taylor, Collar, and Green. Administrative, technical, and material support: Ohye and Green. Study supervision: Taylor and Green.

Financial Disclosure: None reported.

Additional Contributions: Lauren Heise, BS, aided in the acquisition of data for the revision of this manuscript.