A, Endoscopic view of probing laryngeal cleft; B, laser ablation of cleft mucosa; and C, final view of larynx following suture placement. D, Isolated tracheoesophageal fistula (TEF); E, view of cleft following laser ablation; and F, view of cleft following endoscopic suture placement. G, Endoscopic view of LC type 3 (note the esophagus herniating through the cleft into the lumen of the infraglottic/subglottic area); H, laser ablation of cleft mucosa; and I, view of posterior glottis following endoscopic suture placement. J, Probing of LC type 3; K, endoscopic view during ablation of cleft mucosa; L, distant view following suture placement; and M, close-up view following suture placement. N, Endoscopic view during cleft palpation; and O, endoscopic view following endoscopic repair. P, Palpation of laryngeal cleft (note the redundant mucosa); Q, laser ablation of cleft mucosa; R, endoscopic suture placement; and S, view following endoscopic repair.
Adil E, Al Shemari H, Rahbar R. Endoscopic Surgical Repair of Type 3 Laryngeal Clefts. JAMA Otolaryngol Head Neck Surg. 2014;140(11):1051-1055. doi:10.1001/jamaoto.2014.2421
Type 3 laryngeal clefts (LC type 3) are traditionally repaired through an open approach, which requires tracheal intubation or tracheotomy placement and risks potential wound complications.
To describe the surgical technique and outcomes of endoscopic carbon dioxide laser–assisted repair in pediatric patients with LC type 3.
Design, Setting, and Participants
Retrospective medical record review of 6 patients with LC type 3, diagnosed via direct laryngoscopy and rigid bronchoscopy, from January 2007 to September 2013, at a tertiary pediatric hospital.
All patients underwent endoscopic carbon dioxide laser–assisted repair.
Main Outcomes and Measures
Patient demographics, medical comorbidity, surgical technique, swallowing outcomes, and complications were analyzed.
Median age at diagnosis was 4 months (interquartile range [IQR], 1.6 months) and at endoscopic repair, 7.5 months (IQR, 2.1 month). Congenital anomalies were found in 4 patients (67%). Five patients (83%) had gastrostomy tubes and 2 (33%) had a Nissen fundoplication prior to cleft repair. All patients aspirated preoperatively on thickened liquids as diagnosed by modified barium swallow. Median operative time was 98.2 minutes (IQR, 16.0 minutes). Five patients (83%) had no aspiration on their 3-month follow-up modified barium swallow, and no patients developed aspiration pneumonia during the follow-up period.
Conclusions and Relevance
Endoscopic carbon dioxide laser–assisted repair should be considered as an alternative to open repair for LC type 3 when an adequate level of anesthesia with spontaneous ventilation can be maintained throughout the procedure and there is sufficient posterior glottic exposure for laser ablation and suture placement.
Laryngeal clefts (LCs) are congenital anomalies characterized by a defect in the posterior laryngeal wall that can extend inferiorly into the trachea. The most common symptoms include aspiration, recurrent pneumonia, cyanotic episodes, and difficulty feeding. Diagnosis requires a high index of suspicion based on clinical presentation, interpretation of preoperative studies, and a thorough endoscopic evaluation using general anesthesia.1,2
The most commonly used classification system was described by Benjamin and Inglis3 in 1989. In this system, type 1 clefts are interarytenoid defects that extend to the level of the true vocal folds. Type 2 clefts extend below the level of the true vocal folds, through a portion of the posterior cricoid lamina. Type 3 clefts extend completely through the cricoid cartilage, with or without extension into the cervical tracheoesophageal wall, and type 4 clefts extend through the entire length of the tracheoesophageal wall into the thorax. Type 3 laryngeal clefts (LC type 3) are very rare and their repair is commonly approached through an open technique.4 The aim of this study was to describe our surgical technique with a focus on when we consider endoscopic repair feasible. The clinical outcome of endoscopic carbon dioxide–assisted cleft repair in these patients is also reviewed.
After obtaining approval from the institutional review board of Boston Children's Hospital, a retrospective medical record review study was performed on all patients with LC type 3 diagnosed via direct laryngoscopy and rigid bronchoscopy from January 2007 to September 2013 at a tertiary pediatric hospital who underwent repair by the senior author (R.R.). A total of 6 patients were identified. No patient was excluded from the study, and all underwent endoscopic repair. Two patients had been transferred from outside institutions after the diagnosis of LC type 3 was confirmed.
To determine the type of cleft, a direct laryngoscopy is performed using a Parson laryngoscope and 4-mm, 0° telescope. The interarytenoid area is exposed using a laryngeal spreader. The interarytenoid area is then palpated using a probe to confirm the diagnosis and measure the length of the cleft (Figure, A, J, N, and P). Laryngeal cleft type 3 is diagnosed when the cleft extends completely through the cricoid cartilage. The whole procedure is performed with the patient spontaneously ventilating.
All of these patients underwent endoscopic carbon dioxide laser–assisted repair of their LC. Endoscopic repair was performed using suspension microlaryngoscopy with the administration of general anesthesia with spontaneous ventilation using the anesthesia technique we previously described.5 The larynx was visualized with a Lindholm laryngoscope. An OmniGuide Surgical carbon dioxide laser at a setting of 6 W at 0.3- to 0.4-second intermittent mode was used to denude the mucosal lining beginning at the apex of the cleft and extending cranially (Figure, B, H, K, and Q). Curved fine microlaryngeal alligator forceps were used to spread the mucosa during ablation with the laser. Cotton pledgets soaked with oxymetazoline hydrochloride (Afrin; Merck) were applied to the area for hemostasis. It is important to remove the mucosa completely at the apex of the cleft to prevent development of a fistula at the distal end of the repair. Absorbable interrupted sutures (4-0, 5-0, and/or 6-0 Vicryl on P1 or P3 needles; Ethicon Inc) were used to reapproximate the mucosal edges (Figure, R). The first suture was the most important and placed at the most caudal extent of the cleft. We generally placed 3 to 4 sutures in a distal to proximal fashion, depending on the extent of the cleft (Figure, C, F, I, L, M, O, and S).
Following cleft repair, the patients were transferred to the pediatric intensive care unit (PICU) for observation. Patients are typically monitored in the PICU overnight and then transferred to floor status for observation for another 48 hours. They resume their preoperative diet and receive dexamethasone (0.5 mg/kg, up to 10 mg) for 24 hours. Following discharge, they are seen in clinic 1 week later for flexible fiberoptic laryngoscopy to examine their surgical site. A postoperative modified barium swallow (MBS) is performed 3 months following surgery to reevaluate their swallow function.
Outcome measures including demographics, medical comorbidity, swallowing outcome, and complications were collected and analyzed. Descriptive statistics were calculated to describe the overall group.
Over the study period, 6 patients were diagnosed as having LC type 3 and treated with endoscopic carbon dioxide laser–assisted repair (Table 1). The median age at diagnosis was 4 months (interquartile range [IQR], 1.6 months) and at endoscopic surgical repair, 7.5 months (IQR, 2.1 month). There was a male to female ratio of 5:1. Congenital anomalies were found in 4 patients (67%): 2 had VACTERL association (vertebral, anal, cardiac, tracheal, esophageal, renal, and limb anomalies), 1 had isolated tracheoesophageal fistula (H-type), and 1 had Opitz syndrome. In 4 patients (67%), the cleft extended through the cricoid cartilage and ended above the first tracheal ring. In 1 patient, the cleft passed to the first tracheal ring and in the last patient the cleft extended to the second tracheal ring.
At presentation, recurrent aspiration pneumonia was the chief complaint in 4 patients (67%). Two patients (33%) presented with feeding difficulty and weight loss. All patients had aspiration of thickened liquids documented by MBS before the endoscopic repair indicating severe aspiration and precluding an oral diet. Median operative time was 98.2 minutes (IQR, 16.0 minutes). Five patients (83%) underwent gastrostomy tube placement, and 1 had a nasogastric tube for feeding prior to laryngeal cleft repair. Nissen fundoplication was performed on 2 patients (33%).
Length of stay for 5 patients (83%) was 3 days. Patient 3 was admitted for respiratory distress secondary to aspiration pneumonia and underwent urgent endoscopic surgical repair during her hospitalization. Follow-up MBS following surgery showed no aspiration in 5 patients. Patient 6 aspirated on nectar-thickened liquids, but not on honey-thickened liquids or purees.
No patients developed aspiration pneumonia following surgery during the follow-up period. There was 1 patient who required a second endoscopic repair 12 months after the first surgery due to residual type 1 defect with intermittent cough during feeds.
There is a trend toward endoscopic airway management.6 The successful endoscopic repair of LC types 1 and 2 has been described in the literature, but endoscopic repair of LC type 3 has had mixed results (Table 2).1,4,11 Sandu et al7 showed the successful use of endoscopic repair in 4 patients with type 3 clefts without any complications. All 4 patients in their series achieved normal feeding with no clinical signs of aspiration and all had a good voice following surgery. However, they excluded patients with secondary airway abnormalities (laryngomalacia, tracheomalacia, tracheoesophageal fistula) and associated congenital anomalies (Pallister-Hall and Opitz-Frias syndromes).7
Garabedian et al8 also reported results of endoscopic repair in 4 patients with LC type 3. They included 2 patients with congenital anomalies (pyelocalyceal dilatation and Opitz syndrome), and both required a second endoscopic procedure. In the series by Broomfield et al,9 endoscopic repair was attempted in 4 patients with type 3 clefts. Two patients required a second endoscopic procedure to correct small, residual type 1 defects, and the other 2 patients required open revision procedures. Thiel et al10 also presented 1 case of LC type 3b in their series in whom both endoscopic and subsequent open repair approaches had failed.
The endoscopic approach for the repair of LCs has several potential advantages over open approaches. Endoscopic repair avoids some of the risks of the external approach including wound complications and laryngeal airway instability following laryngofissure. In addition, if a lateral pharyngotomy approach is selected for open repair, there is potential for injury to the recurrent laryngeal nerve. Furthermore, endoscopic repair may avoid tracheal intubation and tracheostomy.7 Finally, patients who undergo endoscopic repair potentially have a shorter hospital stay. In our series, 5 patients (83%) were discharged after 3 days.
In our experience, there are 2 factors that determine when endoscopic repair is possible. First, the ability to maintain anesthesia with spontaneous ventilation is necessary for endoscopic repair. Prior to considering an endoscopic approach, a conversation with the anesthesiologist is necessary to determine if this will be possible from a cardiopulmonary perspective. The median operative time in our series was 98.2 minutes (IQR, 16 minutes) which means the anesthesia team must be comfortable with maintaining the patient at this anesthesia level for 1.5 to 2.0 hours.
If the anesthesiologist is confident that the appropriate anesthesia level can be maintained for the duration of the procedure, then the second requisite to an endoscopic approach is adequate posterior glottic exposure. The depth of the cleft has to be visible and accessible for suture placement while the patient is in suspension. At the depth of the cleft where access is most limited, we find it useful to use 5-0 or 6-0 Vicryl sutures on a P1 needle. As repair proceeds cranially, we increase the size of the suture and needle to 4-0 or 5-0 Vicryl on a P3 needle.
In our series, the median age at diagnosis was 4 months (IQR, 1.6 months) and at endoscopic surgical repair, 7.5 months (IQR, 2.1 month). The apparent lapse between diagnosis and repair is largely to optimize their pulmonary status prior to surgery. As mentioned previously, all of these patients aspirate and have varying degrees of lung injury and/or infection as a result. These issues are challenging and can delay both endoscopic and open repair. Therefore, we often refer these patients to our pulmonary colleagues to optimize their pulmonary function prior to surgery. Bronchodilators, respiratory syncytial virus prophylaxis, and antibiotics are often necessary regardless of approach.
One of the goals of endoscopic repair is to avoid intubation and tracheotomy placement. When repairing a LC type 3, it is important to remember that as the cleft is closed, the airway is being narrowed. In addition, manipulation of the airway will cause some swelling. Therefore, we tend to be cautious with our endoscopic repairs and would prefer to place less sutures and possibly return to the operating room for endoscopic repair of a residual defect rather than risk post-operative airway compromise that may require intubation and/or tracheotomy placement. There was 1 patient in our series who required a second endoscopic procedure for repair of a symptomatic residual type 1 cleft. Aspiration completely resolved following the second procedure.
Patient number 3 in our series required a tracheotomy. This was performed 2.5 months before the diagnosis of LC type 3. The tracheostomy was not related to LC type 3 or to the endoscopic cleft repair. It was due to prolonged intubation and failed extubation after tracheoesophageal fistula repair. She was decannulated 4 months ago.
Four patients in our series had congenital anomalies or syndromes. We had 2 patients with VACTERL who also had vocal cord paralysis following tracheoesophageal fistula repair. Patient 3 had no aspiration on her first postoperative MBS. She currently takes half of her feedings by mouth and half via G-tube because she has severe oral aversion and microgastria that continue to limit her oral intake. Patient 6 was our most recent repair, and his 3-month follow-up MBS showed aspiration with nectar thickened liquids, but no aspiration with honey-thickened liquids or purees. He has oropharyngeal dysphagia characterized by reduced oral motor skills and delayed onset of the swallow as well as persistent left vocal fold immobility, which are likely the cause of his aspiration. He currently takes purees by mouth and supplements with G-tube feeds. He has not had any episodes of aspiration pneumonia since his cleft repair. This indicates that endoscopic carbon dioxide laser–assisted repair of LC type 3 is a valuable procedure even in patients with significant comorbidities.
The management of LC remains challenging, despite developments in surgical instrumentation and technique. We have noticed a trend toward fewer failed repairs possibly because of our extensive experience with endoscopic repair of LC types 1 and 2. The ability to maintain an adequate level of anesthesia and sufficient posterior glottic exposure are the 2 factors we consider when determining if a patient is a candidate for endoscopic repair. Overall, the outcomes in our series of LC type 3 have been favorable with no major complications. We believe that the endoscopic carbon dioxide laser–assisted procedure is a relatively safe and reproducible technique for repair of LC type 3.
Submitted for Publication: January 25, 2014; final revision received July 16, 2014; accepted August 29, 2014.
Corresponding Author: Reza Rahbar, DMD, MD, Department of Otolaryngology and Communication Enhancement, Boston Children’s Hospital, 300 Longwood Ave, Room LO-367, Boston, MA 02115 (firstname.lastname@example.org).
Author Contributions: Dr Rahbar had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Adil and Al Shemari contributed equally to this work.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Adil.
Statistical analysis: Adil, Al Shemari.
Administrative, technical, or material support: Al Shemari.
Study supervision: Rahbar.
Published Online: October 16, 2014. doi:10.1001/jamaoto.2014.2421.
Conflict of Interest Disclosures: None reported.