The author has no funding, financial relationships, or conflicts of interest to disclose.
Triological Society Candidate Thesis
A randomized study of suprastomal stents in laryngotracheoplasty surgery for grade III subglottic stenosis in children
Article first published online: 13 MAY 2013
© 2013 The American Laryngological, Rhinological and Otological Society, Inc.
Volume 124, Issue 1, pages 207–213, January 2014
How to Cite
Preciado, D. (2014), A randomized study of suprastomal stents in laryngotracheoplasty surgery for grade III subglottic stenosis in children. The Laryngoscope, 124: 207–213. doi: 10.1002/lary.24141
- Issue published online: 20 DEC 2013
- Article first published online: 13 MAY 2013
- Manuscript Accepted: 15 MAR 2013
- Manuscript Revised: 22 FEB 2013
- Manuscript Received: 3 JAN 2013
- Laryngeal and tracheal stents;
- suprastomal stents;
- Aboulker stents;
- Montgomery T-tubes;
- double stage laryngotracheoplasty
Studies comparing the relative effectiveness of suprastomal stents in laryngotracheoplasty surgery are lacking in the literature. The goal of this study was to compare the performance of two widely used short-term suprastomal stents in open airway surgery.
Prospective, randomized study of a single surgeon's outcomes for grade III subglottic stenosis (SGS) in children.
The open Aboulker Teflon stent versus the cut, closed, soft Silastic Montgomery T-tube stent for short-term stenting in double-stage laryngotracheoplasty (dsLTP) were compared.
Twenty-four patients were recruited into the study; 12 received an Aboulker (A) stent and 12 a cut, soft, Silastic (S) stent. There was no statistically significant difference in the operation specific decannulation rate between the two stent groups. Patients who received S stents were found to tolerate postoperative feeding better than those with A stents. The median time to decannulate, however, was longer for those receiving S stents (5.5 vs. 3.5 months, P = .02). Furthermore, those receiving S stents had increased need for airway dilation after stent removal (1.75 vs. 0.17 dilations per patient, P = .02) and increased presence of granulation tissue in the airway at the time of stent removal. Multivariate analysis showed the type of stent used independently correlated to time to decannulate (P = .022).
Compared to the Teflon Aboulker stents, closed, soft, Silastic stents are associated with improved feeding in the postoperative period. Their use, however, also correlates with worsened granulation tissue formation and longer time to decannulation in patients undergoing dsLTP for grade III SGS.
Level of Evidence
1b. Laryngoscope, 124:207–213, 2014
The management of laryngotracheal stenosis in children poses multiple challenges for treating clinicians. Although avoiding tracheotomy or achieving decannulation with open airway surgery can be expected at rate of approximately 90%, success rates in children with severe stenosis are lower.[1-3] A single, uniform open procedure to reconstruct the narrowed pediatric airway does not exist, but in general, open airway reconstruction is indicated when the stenosis is mature, circumferential, long, >50%, or failing to resolve with simple endoscopic techniques such as balloon dilation.
Laryngotracheoplasty (LTP) with interposition of cartilage graft was introduced by Fearon and Cotton in 1972 as a means to expand an otherwise narrowed subglottic airway segment. The principle of the procedure is to distract the cricoid cartilage either anteriorly and/or posteriorly by placing cartilaginous grafts in place over a luminal appropriately sized stent. The use of stents plays an important role in the surgical management of these patients. The stents are imperative to maintain the grafts in position, lend support to the reconstructed area, and provide a rigid luminal framework around which healing and scar contracture can occur. In single-stage LTP procedures, the stent used is an endotracheal tube, which is left in place while the child is nasotracheally intubated and sedated in the pediatric intensive care unit for a period of 5 to 14 days.
In double-stage LTP (dsLTP) procedures, a tracheotomy tube is either placed or left in place after the surgery. A sutured indwelling suprastomal stent, separate and above from the tracheotomy tube, is left in place postoperatively while the grafts heal. Usually the stents are left in place for a period of 2 to 4 weeks, unless if longer stenting is warranted (severely flaccid reconstructed tracheal walls from failed resection, poor graft stability, highly distorted anatomy, severe concomitant subglottic stenosis [SGS]). Over the recent past, little progress has been made in the introduction of new stents in postoperative dsLTP patients. Although an ideal suprastomal stent does not exist, the very useful stents that have been around for over 20 years remain the most popular today. These are primarily the Teflon Aboulker stent and a cut, soft, Silastic stent. The cut, soft, Silastic stent is comprised by a cut limb of a Montgomery T-tube (Boston Medical Products, Boston, MA). The Aboulker stent is typically open, whereas the cut, soft, Silastic tube is usually sutured shut at the top. If significant aspiration is occurring, the upper aspect of the Aboulker stent can be plugged. Others have shown that the closed, soft, Silastic stent allows for improved oral feeding postoperatively, with less aspiration compared to open stents. Little is known, however, about the relative performance of these stents in terms of surgical outcome. We hypothesized there would be little difference in surgical outcome between the open Aboulker stent and the closed, soft, Silastic stent when placed postoperatively in the surgically reconstructed airway for a short duration of time. We further hypothesized that patients with the closed stent would have improved oral feeding while the stent was in position.
To test these hypotheses, a prospective randomized study into a single surgeon's outcomes of dsLTP comparing open indwelling Aboulker stents versus closed, cut, soft, Silastic stents in dsLTP for grade III SGS in children was undertaken.
MATERIALS AND METHODS
Patients under consideration for open dsLTP surgery between January 1, 2007 and June 31, 2012 were accrued into the study. Inclusion criteria comprised the following: grade III Myer-Cotton SGS, clinical indications for a double-stage surgical approach (glottic involvement, vocal cord dysmotility, history of chronic pulmonary disease, tongue base obstruction, craniofacial syndromes, or neurologic comorbidities). Exclusion criteria included: isolated SGS under consideration for single-stage LTP, revision LTP surgery, supraglottic stenosis, SGS milder than grade III, any clinical indication for stenting longer than 3 weeks, active laryngeal inflammation, or eosinophilic esophagitis. All patients were treated for acid reflux with proton pump inhibitors (lansoprazole 1.0 mg/kg once daily) throughout the course of their surgical management and postoperatively. Patients were not routinely referred to GI for reflux or eosinophilic esophagitis. If the larynx was noted to be chronically inflamed, patients were excluded from this study and referred to the gastroenterology service. At the time of the surgery patients were randomized to receive either a Teflon Aboulker suprastomal stent (A) or a cut, soft, Silastic Montgomery T-tube (S) (Boston Medical Products, Boston, MA) suprastomal stent. Randomization was carried out by having an assistant in the room blindly pull out a paper from a presealed envelope containing two equal-sized papers labeled A or S. The outer diameter of the stent selected corresponded to the outer diameter of an endotracheal tube expected to appropriately fit the child by age/weight. Figure 1 depicts how the stents looked in situ. Institutional review board approval was obtained for this study (Pro00001414).
Surgical and Postoperative Protocol
All patients received open LTP surgery by a single surgeon (d.p.) with the use of costochondral anterior and posterior grafts to expand the narrowed subglottic airway. Posterior grafts were all placed in a sutureless fashion. After placement of the posterior graft, the stent (A vs. S) was secured in a suprastomal fashion, above the tracheotomy tube, but never attached to it, with a transairway 2.0 Prolene suture. For the Aboulker stents, the lower end was situated approximately 1 cm above the stoma. The upper, rounded, and open end was situated at the level of the petiole. For the cut, soft, Silastic stent, one of the limbs of a Montgomery T-tube (outer diameter size being equal to the outer diameter of an age appropriate endotracheal tube) was cut. This cut tube segment was then used as the stent. The machined, factory-smooth, rounded end of the cut tube was situated in the suprastomal area, also approximately 1 cm above the stoma. The cut end was placed at the level of the petiole. This cut end of the soft Silastic tube was sutured closed at the top with a 4.0 Prolene horizontal mattress stitch. After suturing the anterior graft in place, the Prolene stitch holding the stent was snugly tied over the strap musculature with a long 2.5-cm knot coming out of the corner of the skin incision on the right side. Patients were typically hospitalized for 5 to 7 days, during which time their diet was advanced to the preoperative status. If patients were orally feeding preoperatively, oral feeds were attempted on postoperative day 2. If patients were gastrostomy (G-tube) dependent preoperatively, the G-tube feeds were started on postoperative day 2. For orally feeding patients, if they did not tolerate oral feeds postoperatively because of dysphagia, aspiration, or odynophagia, a nasogastric tube was placed prior to discharge for the duration of the time the stent was in place.
Patients returned to the operating room 3 weeks after the surgery to remove the stent endoscopically after clipping the Prolene stitch on the corner of the neck incision. Direct microlaryngoscopy and bronchoscopy (DLB) was then carried out, and the amount of granulation tissue was graded from 1 to 3 (1 = minimal, 2 = moderate, 3 = severe granulation). The tracheotomy tube was the downsized a half a size. In the recovery room, patients were fit with a speaking, Passy-Muir valve and were discharged home the same day. All patients were treated with antibiotics, reflux medication, and every other day with 0.5 mg/kg enteral dexamethasone for 1 week after stent removal. All patients returned to the operating room 1 week later for DLB, and at that point and if the airway appeared to be healing adequately with minimal granulation, the child was admitted for 48 hours an initial capping trial monitoring. If there was still significant granulation, this was removed endoscopically with cold instrumentation. If the airway showed signs of edema or early restenosis, it was dilated with a high-pressure, noncompliant, airway balloon catheter. The size of the balloon selected corresponded to the outer diameter of the stent that was used.
After dilation, patients returned to the operating room 1 week later for surveillance endoscopy and reassessment of the airway. This was repeated until the child was deemed ready for capping. Once a child had tolerated capping for 48 hours as an inpatient, they were discharged for capping at home for 4 weeks, prior to returning to the operating room for DLB. If at that point the airway remained stable and the capping was tolerated for the month, they were admitted for decannulation and another 48 hours of monitoring. All patients were kept on lansoprazole enterically from the time of the reconstructive surgery until the time of decannulation. This protocol is summarized in Figure 2. Surgical success was defined as ability to decannulate. If patients were not able to decannulate because of inadequate airway healing and diameter, second open airway surgeries were considered no sooner than 6 months after the first open airway surgery. Clinical data gathered for comparison included age, demographics, clinical indication for double-stage surgery, operation-specific surgical success, time to decannulate, number of DLBs required after stent removal, number of dilations required after stent removal, need for second open airway surgery, and ability to tolerate oral feeds while the stent was in place (if feeding orally preoperatively).
Continuous variables were analyzed with Mann-Whitney tests for nonparametric data. Categorical variables were analyzed with χ2 with Fisher exact test. Time to decannulate was analyzed with Wilcoxon tests and Kaplan-Meier survival curves. Multiariable linear regression was preformed with age, stent type (A vs. S), and glottic involvement (yes or no) as independent variables.
A total of 24 patients were recruited into the study; 12 received an Aboulker (A) stent and 12 a cut, soft, Silastic tube (S). There was no statistically significant difference in the mean age of the A patients (mean, 34.1 months; standard deviation [SD] 11.1 months) and S patients (mean, 43.9 months; SD 20.5 months) (p = 0.22). Four of 12 A patients and five of 12 S patients had posterior glottic stenosis involvement (χ2, P = .8), whereas eight of 12 A patients and nine of 12 S patients had significant history of pulmonary or neurological comorbidities (χ2, P = .86). These data are summarized in Tables 1 and 2.
|Age, mo||34.1 (±11.1)||43.9 (±20.5)||.22|
|Glottic involvement||4/12 patients||5/12 patients||.8|
|Pulmonary/neurological comorbidities||8/12 patients||9/12 patients||.86|
|Age (mo)||Glottic Stenosis/Paralysis (Yes/No)||Surgery||Stent||Postoperative||Feeding|
|No. of Dilations||No. of DLBs||Granulation Tissue (Score)||Decannulated (Yes/No)||TTD (mo)||Return to Preoperative Oral Diet, d||Need for New NG Tube|
|19||No||dLTP with A/P CCG||A||0||2||1||Yes||2||N/A||G-tube fed|
|24||No||dLTP with A/P CCG||A||0||3||1||Yes||2||N/A||G-tube fed|
|20||Yes||dLTP with A/P CCG||A||0||3||2||Yes||4||3||No|
|40||No||dLTP with A/P CCG||A||0||3||2||Yes||2||N/A||G-tube fed|
|16||No||dLTP with A/P CCG||A||0||4||2||No||4||N/A||G-tube fed|
|12||No||dLTP with A/P CCG||A||2||6||1||Yes||6||Not able||Yes|
|2||No||dLTP with A/P CCG||A||0||4||1||Yes||4||Not able||Yes|
|21||Yes||dLTP with A/P CCG||A||0||3||1||Yes||4||N/A||G-tube fed|
|17||No||dLTP with A/P CCG||A||0||4||1||Yes||1||N/A||G-tube fed|
|154||No||dLTP with A/P CCG||A||0||3||1||Yes||2||6||No|
|66||Yes||dLTP with A/P CCG||A||0||3||1||No||6||N/A||G-tube fed|
|18||Yes||dLTP with A/P CCG||A||0||3||2||Yes||3||N/A||G-tube|
|50||No||dLTP with A/P CCG||S||0||3||1||No||4||N/A||G-tube fed|
|11||No||dLTP with A/P CCG||S||3||6||3||No||19||2||No|
|132||No||dLTP with A/P CCG||S||0||3||1||Yes||4||1||No|
|204||Yes||dLTP with A/P CCG||S||0||5||2||Yes||20||4||No|
|19||No||dLTP with A/P CCG||S||2||6||3||Yes||11||N/A||G-tube fed|
|12||Yes||dLTP with A/P CCG||S||4||6||2||No||5||N/A||G-tube fed|
|51||No||dLTP with A/P CCG||S||7||8||3||No||24||Not able||Yes|
|27||No||dLTP with A/P CCG||S||0||2||1||Yes||6||3||No|
|193||No||dLTP with A/P CCG||S||1||4||1||Yes||3||N/A||G-tube fed|
|15||No||dLTP with A/P CCG||S||1||4||1||Yes||3||N/A||G-tube fed|
|17||No||dLTP with A/P CCG||S||3||4||3||No||15||N/A||G-tube fed|
|36||No||dLTP with A/P CCG||S||0||3||3||Yes||2||2||No|
There was no statistically significant difference in the operation-specific eventual decannulation rate. Eleven of 12 (91.6%) A patients decannulated with one open surgery, whereas seven of 12 (58.3%) of the S patients did so (χ2, P = .06). The time taken to decannulate, however, was statistically significantly different between the two groups (Fig. 3), as the A patients had a median time to decannulate of 3.5 months, and the S patients had a median time to decannulate of 5.5 months (P = .02). Along the same lines, the mean granulation score at the time of stent removal was slightly worse for the S patients than the A patients (2.0 vs. 1.3, P = .04), and the number of dilations needed after stent removal was more for the S patients than for the A patients (1.75 vs. 0.17, P = .02). Of the five patients where S stents were used, three have undergone second surgeries and have been successfully decannulated.
Multivariate regression analysis demonstrated that stent type independently correlated with time to decannulate and number of dilations required after stent removal when controlling for age and glottic involvement (Tables 3 and 4). Those patients receiving S stents had a greater need for dilation after stent removal and took longer to decannulate regardless of age or concomitant glottic pathology status.
|Correlation Coefficients||Lower 95%||Upper 95%||Standard Error||t Statistic||P Value|
|S vs. A||7.101||1.154||13.048||2.851||2.491||.022|
|Correlation Coefficients||Lower 95%||Upper 95%||Standard Error||t Statistic||P Value|
|S vs. A||1.872||0.462||3.281||0.676||2.771||.012|
Children treated with S stents were able to return to an oral diet at a greater rate than those treated with an A stent. Overall, five of six patients receiving S stents who were on an oral diet preoperatively returned to an oral diet postoperatively. Only one of four patients receiving an A stent who were on an oral diet preoperatively were able to return to an oral diet postoperatively as shown in Figure 4 (χ2, P = .03). In both groups, those who were preoperatively orally feeding and did not tolerate oral diet while the stent was in place, were all able to return to an oral diet once the stents had been removed.
Results of our study show that important differences can be expected with the utilization of closed, cut, soft, Silastic (S) versus open Teflon (A) suprastomal stents after dsLTP for grade III SGS in children. Data analyses suggest that although operation-specific decannulation rates are not statistically different, the trend favors the usage of the A stent. Closed, soft, Silastic stents are associated with improved feeding in the postoperative period but correlate to worsened granulation tissue formation and longer lengths of time to decannulation. Of note, none of the patients with the A stents had the stent closed in the postoperative period. This is something that can physically be done with special plugs for these stents, which we did not utilize in this study. It is possible that if the A stents would have been closed, the rate of oral feeding between the two stent groups may not have been significantly different.
Ideally, stents used after dsLTP should be rigid enough to keep the reconstructed area and grafts stable in position, permit vocalization, allow for oral intake of food without aspiration, be easy to examine and remove, minimize granulation tissue formation, and be compatible with tracheotomy tube cares and changes. Both A and S stents retain a majority of these qualities; however, when sutured postoperatively in position at the reconstruction site with some distance between the stent and the tracheal stoma (i.e., a short suprastomal stent), their use is associated with the formation of granulation tissue and potential stenosis at the stent's distal end in the suprastomal location. For this reason, the stenting duration use is limited to 2 to 3 weeks. When using cut, soft, Silastic stents, it is important for this reason to position the machined factory-smooth end of the stent in the suprastomal airway, so as to minimize granulation tissue in this area. Even with the smooth end suprastomally, we still noted worse granulation with the soft Silastic material than with the Teflon material. It is possible that the polished Teflon material is minimally irritating in terms of granulation tissue formation as suggested by others. Teflon is composed of polytetrafluoroethylene (PTFE), a synthetic fluoropolymer of tetrafluoroethylene. PTFE is one of the smoothest forms of plastic, possessing one of the lowest coefficients of friction of any solid material, and as such it is minimally mechanically traumatic when implanted in the body. Silastic is also a highly smooth plastic, which is known to potentially promote less macrophage and bacterial adherence than other biomaterials. The exact reasons for the noted differences in our results are difficult to explain. Further studies into the comparative formation of biofilms on these stents may be warranted. In a study of 23 consecutive mostly Teflon stents removed from dsLTP patients, Simoni and Wiatrak found that all stents were colonized with multiple pathogens. These pathogens included Streptococcus viridians, Pseudomonas aeruginosa, Staphylococcus aureus, Haemophilus influenza, and Neisseria species. Candida species were found in 57% of their stents. Anaerobes were isolated in 26%. Of note, a separate study showed that colonization of Silastic airway stents with S aureus correlated significantly with granulation tissue formation. Whether different biomaterial types, such as Teflon versus Silastic, confer distinct bacterial contamination or biofilm formation profiles is unknown, but these potentially different profiles are likely an important factor in determining the amount of postoperative airway granulation tissue occurrence.
Differences in stent rigidity may be important as well. We noted that the more rigid Teflon material was associated with less need for airway dilation after stent removal when compared to softer and more compressible Silastic material. The less rigid Silastic potentially provides less adequate support to the cartilage grafts over the 3-week stenting period. On the other hand, the soft and compressible nature of the soft Silastic tube offers advantages well. Given its compressibility, the stent can be positioned luminally, closer to the tracheal stoma, without the fear of it getting in the way of postoperative tracheotomy tube changes. Also, its upper end can be very readily sutured shut, allowing for reduced aspiration and a higher likelihood of tolerating oral feeds while it is in place in the postoperative period. Results from our series validate what was shown by Smith et al., that in fact, most patients with the closed, soft, Silastic stents are able to return to their preoperative oral feeding status.
Other stents have been historically used after laryngotracheal reconstructive surgery. A Silastic “Swiss-roll” stent has also been reported as a short-duration suprastomal stent, but given that the support it provides is too gentle and it is associated with worsened granulation tissue formation, it is rarely used anymore. To prevent the suprastomal, distal, stent granulation tissue formation, a modification in the use of the Aboulker stent, where a tracheotomy tube was inserted through a fenestration in a longer Aboulker stent encompassing the stomal region and securely wired in position, has been classically described. The stenting duration could then be extended to >2 months. However, due to the inability to change the tracheotomy tube postoperatively and to anecdotal instances of stent fracture, the use of this fenestrated, tracheotomy-wired, long Aboulker stent has long fallen out of favor and is used less frequently. In efforts to minimize suprastomal granulation, a modification of the Aboulker stent, where the stent is positioned all the way down to (and past) the stoma, with a tracheotomy tube sliding through, but not wired to, a hole on the corresponding anterior surface of the stent, has been described but not been widely adopted or reported on.
In adults and older children, the most frequently used long-term laryngeal stent has been the full uncut Montgomery T-tube. Due to the significant potential for mucous plug obstruction associated with smaller T-tubes, their usage is limited in children under 4 years of age and requires meticulous postoperative and home care. In general, for children, the usage of full T-tubes is limited to the small group of children with the most difficult airway problems, where stenting is required for long periods of time, due to relatively flaccid laryngotracheal walls or when there is severe transglottic scarring. In the event of tube obstruction, the T-tube cannot be easily changed at home. Failure to maintain plugging of the T-tube's anterior limb invariably results in mucous obstruction. Given the young age of the majority of the children in this study, we did not propose to use full T-tubes in any patient. The aim of the study was to compare the Silastic stent to the Teflon stent. As such, we cannot comment on the comparative degree of granulation associated with the use of full uncut T-tubes in children.
Clearly, decisions as to the usage of laryngeal stents after reconstructive surgery have to be individualized to each patient depending on the severity and location of the stenosis. The stent type, length, and placement duration have to be individually tailored. During the stenting period, meticulous care of the tracheotomy tube is mandatory. This is especially the case considering the presence of a suprastomal stent above the tracheotomy tube, a fact that may further confer the child completely T-tube dependent for an airway, especially with closed, soft, Silastic stents. In this setting, accidental decannulation or tracheotomy mucous plugging are potentially lethal events without the appropriate level of nursing and monitoring. Careful cooperation with a speech and language pathologist to gauge the child's ability to swallow or aspiration risk is critical in postoperative double-stage patients with indwelling stents.
An ideal, widely used, short-duration suprastomal stent does not exist. Although operation-specific decannulation rates are similar between stents, closed, soft, Silastic stents are associated with improved feeding in the postoperative period. The usage of the cut, soft, Silastic tubes as stents, however, correlates to worsened granulation tissue formation and longer lengths of time to decannulation.
- 5Modifications apportees l'intervention de Rethi, interest dans les stenosis laryngo-tracheales et tracheales. Ann Otolaryngol (Paris) 1966;83:98–106., , .