Smith LP, Zur KB, Jacobs IN. Single- vs Double-Stage Laryngotracheal Reconstruction. Arch Otolaryngol Head Neck Surg. 2010;136(1):60-65. doi:10.1001/archoto.2009.201
Copyright 2010 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2010
To compare single-stage laryngotracheal reconstruction (ssLTR) and double-stage LTR (dsLTR).
Retrospective medical record review.
Tertiary care children's hospital.
Seventy-one patients underwent 84 procedures (22 ssLTRs and 62 dsLTRs).
Review of preoperative disease severity and surgical outcomes for patients who underwent ssLTR vs dsLTR.
Main Outcome Measure
Operation-specific and overall decannulation rates.
Regarding ssLTRs, the mean grade of subglottic stenosis was 2.1 and the overall and operation-specific decannulation rates were 100% and 91%, respectively. The mean grade of subglottic stenosis for double-stage procedures was 2.9, and the overall and operation-specific decannulation rates were 93% and 68%, respectively. Patients who underwent ssLTR and dsLTR were further divided into early and late groups based on whether the posterior graft was sutured in place (early) or not (late). Overall and operation-specific decannulation rates were 100% and 89%, respectively, for the early single-stage group and 100% and 92% for the late group. Regarding the dsLTR group, overall and operation-specific decannulation rates were 88% and 42%, respectively, for the early group and 95% and 79% for the late group. Among all groups, there was no significant difference in overall decannulation rates (P > .05). Single-stage LTR offered an increased rate of operation-specific decannulation over dsLTR (P < .05). However, that difference was not significant between the late ssLTR and the late dsLTR groups (P > .05).
Careful assessment of preoperative disease severity and overall medical status will help surgeons choose between ssLTR and dsLTR, maximizing patient outcomes for both modalities.
From its earliest descriptions in the 1970s, expansion laryngotracheal reconstruction (LTR) with cartilage grafting was performed as a 2-stage, or double-stage, procedure.1 By definition, in double-stage LTR (dsLTR), the tracheostomy tube is kept in place at the conclusion of the procedure. If the procedure is successful, the patient is decannulated during a separate hospital admission, usually several weeks or months later. In 1988, Prescott2 presented perhaps the first description of single-stage LTR (ssLTR) when he described the results of 5 children who had their tracheostomy tube removed at the time of LTR. Prescott2 advocated stenting with an endotracheal tube for approximately 10 days postoperatively. Since then, multiple researchers3- 10 have confirmed the feasibility of ssLTR, outlining their results and refining the indications. The advantage of ssLTR is immediate decannulation at the time of reconstruction (or avoidance of a tracheotomy altogether). This removes a potential source of infection and addresses the stoma site concurrent with LTR. In addition, prolonged stenting, with its potential complications, is also avoided. These advantages must be weighed against the potential for airway complications in the perioperative period and the complications that may result from prolonged endotracheal intubation.11,12
Although good data regarding surgical outcomes for ssLTR vs dsLTR remain elusive, a consensus has emerged in the literature regarding when to perform each procedure. Most researchers6,8,11 advocate dsLTR in children with complex multilevel stenosis, significant neurologic deficits, significant lung disease, or anatomy that makes reintubation technically difficult (craniofacial or vertebral anomalies). There is also a trend toward performing dsLTR in patients with higher grades of stenosis (Cotton-Myer grade III or IV),3,5,8 although we and others13- 16 would advocate for partial cricotracheal resection (pCTR) in select children with severe grade III or IV subglottic stenosis (SGS) whose disease is separated from the larynx by an adequate margin. In this study, we attempt to carefully define our surgical results for ssLTR and dsLTR to help surgeons choose between these surgical modalities.
Institutional review board approval was obtained to perform a retrospective medical record review of all patients who underwent LTR with cartilage grafting at a tertiary care children's hospital (The Children's Hospital of Philadelphia) between February 1, 2000, and April 15, 2008. All the operations were performed by 1 of 2 senior surgeons (K.B.Z. or I.N.J.). Patients who were treated endoscopically or who underwent cricotracheal or tracheal resection were excluded from this study. In an effort to compare the outcomes for ssLTR vs dsLTR, details relating to preoperative disease severity and location, surgical procedure performed, stent type, and duration of stenting or intubation were recorded. As previously defined by Hartnick et al,5 operation-specific and overall decannulation rates were recorded. Briefly, operation-specific decannulation refers to the rate at which an open surgical procedure is associated with subsequent decannulation or extubation without the need for further open airway surgery. The overall decannulation rate includes patients who subsequently receive an additional open surgical reconstruction.5 The number of postoperative endoscopic procedures and significant complications were also identified.
A reconstruction was defined as single stage if the patient did not have a tracheostomy tube in place at the conclusion of the procedure. This included patients who had their tracheostomy tube removed at the time of surgery and patients who did not have a tracheostomy tube present preoperatively. All patients who underwent ssLTR remained nasotracheally intubated after their procedure. Patients who underwent dsLTR had a tracheostomy tube present at the conclusion of their procedure. In all cases, these patients had a tracheostomy tube preoperatively.
Patients who underwent ssLTR and dsLTR were further subdivided into early and late groups. In late 2003, we began inserting and securing posterior grafts without sutures, as previously described.17 Patients operated on before this technical advancement were included in the early group and after in the late group regardless of whether they had a posterior graft placed. All patients who underwent dsLTR in the early group had an Aboulker stent, and almost all (93%) in the late group had a cylindrical Silastic (Dow Corning Corp, Midland, Michigan) stent placed. As we previously described,18 this stent was fashioned by cutting one limb off of a Montgomery T-tube (Hood Laboratories, Pembroke, Massachusetts). Single-stage LTRs and dsLTRs were otherwise performed similar to previous descriptions.11
Eighty-four (22 single- and 62 double-stage) LTRs with cartilage grafting were performed on 71 patients during the study (Table 1). Mean patient age at the time of surgery was 51.6 months, and median age was 41 months (age range, 5-212 months). Twenty revision procedures were performed (8 ssLTRs and 12 dsLTRs). Five patients who underwent revision had their first procedure performed at another institution. One child was unavailable for follow-up 3 months after reconstruction. This child was in the early dsLTR group and was considered a surgical failure (operation specific and overall). Three patients from the late dsLTR group required major endoscopic procedures before decannulation. Two patients required a carbon dioxide laser partial right arytenoidectomy, and one of them also received carbon dioxide laser treatment of the lingual tonsils. One patient required multiple postoperative balloon dilations, with microdebrider removal of a portion of the anterior costal cartilage graft that had prolapsed into the airway. These 3 patients were considered operation-specific surgical successes. One patient was excluded from the analysis. This child acutely lost her airway and expired 1 day after dsLTR.
Two patients from the dsLTR group had a postoperative pneumothorax (2.4%); one resulted from pleural injury at the rib graft harvest site and resolved without a chest tube. The other patient had subcutaneous emphysema and a pneumothorax identified on postoperative day 2 that required neck exploratory surgery with replacement of an extruded Penrose drain and a chest tube. Two patients from the dsLTR group had significant postoperative neck infections (2.4%). One of them required incision and drainage early in the postoperative period, and the other had resorption of an anterior costal cartilage graft. Both of these patients failed their respective surgical reconstructions, and one of them was decannulated after a revision procedure. The other patient remains dependent on a tracheotomy tube 3 years postoperatively. One patient who underwent ssLTR required urgent reintubation for symptomatic stridor 2 days after extubation. This patient responded well to corticosteroid therapy and remains decannulated and asymptomatic 2 years after his LTR. The remainder of the patients who underwent ssLTR had no significant complications.
Table 2 outlines the preoperative characteristics of patients who underwent ssLTR or dsLTR in this series. Eight of the 22 ssLTRs (36%) and 12 of the 62 dsLTRs (19%) were revisions, a difference that was not significant (P > .05). Mean age at the time of surgery was 57 months for patients who underwent ssLTR and 50 months for those who underwent dsLTR, a difference that was not significant (P > .05). For patients who had SGS, the mean Cotton-Myer grade for the ssLTR group was 2.1 and for the dsLTR groups was 2.9, a difference that was significant (P < .05). In contrast to patients who underwent ssLTR, most patients (58%) who underwent dsLTRs had multilevel stenosis. Twenty-eight patients (45%) had stenosis at 2 different levels of the airway, and 8 patients (13%) were actually affected at 3 different levels. Seven patients (32%) in the single-stage group were affected at 2 levels, and none had disease at more than 2 levels. χ2 Analysis revealed that patients in the double-stage group were significantly more likely to have multilevel stenosis (P < .05). Sixty-eight percent of patients who underwent single-stage surgery had normal vocal cord mobility preoperatively vs 32% of patients who had dsLTR, a difference that was significant (P < .05). Based on grade of SGS, number of airway subsites affected, and vocal cord function, patients who underwent dsLTR had significantly worse preoperative disease than did those who underwent ssLTR.
The surgical results for ssLTR vs dsLTR are outlined in Table 3. Regarding ssLTRs, overall and operation-specific decannulation rates were 100% and 91%, respectively (100% and 89% for early patients and 100% and 92% for late patients). Two patients were considered operation-specific failures (1 each from the early and late groups); both were successfully treated with a vocal cord lateralization procedure, only one of whom actually required interval replacement of the tracheostomy tube. Overall and operation-specific decannulation rates were 93% and 68%, respectively, for the dsLTR group (88% and 42% for early patients and 95% and 79% for late patients). Among all groups, there was no significant difference in overall decannulation rates (P > .05). Single-stage reconstruction offered an increased rate of operation-specific decannulation compared with dsLTR (P < .05). However, no significant difference in operation-specific decannulation was seen when patients in the late ssLTR group were compared with those in the late dsLTR group (P > .05). Although patients who underwent late dsLTR had a significantly higher operation-specific decannulation rate compared with those with early dsLTR (P < .05), no such difference was seen between the 2 ssLTR groups (P > .05). Patients in the ssLTR and dsLTR groups underwent approximately 4 postoperative airway endoscopies with no statistically significant differences between the groups. The dsLTR group had significantly fewer inpatient perioperative hospital days than did the ssLTR group (P < .05).
Surgical results were further subdivided by grade of SGS. Twenty ssLTRs and 55 dsLTRs were performed on patients who had SGS. Table 4 illustrates operation-specific and overall decannulation rates for these patients subdivided by Cotton-Myer grade of SGS. The overall decannulation rate for all patients who underwent ssLTR was 100%. The operation-specific decannulation rate for grade I was 100%, for grade II was 91%, and for grade III was 100%. No ssLTRs were performed on patients with grade IV SGS. Regarding patients with dsLTR, the overall decannulation rate for grade II was 89%, for grade III was 93%, and for grade IV was 100%. The operation-specific decannulation rate was 67% for grade II, 70% for grade III, and 0% for grade IV. No double-stage procedures were performed on patients with grade I SGS. The Fisher exact test was used to compare operation-specific and overall decannulation rates for patients with grade II or grade III SGSs, and no significant difference between patients who underwent ssLTR and dsLTR was identified (P > .05).
Children with airway stenosis may present with disease of varying severity and locations in the airway. These patients also often have significant comorbid conditions including neurologic, pulmonary, cardiac, and gastrointestinal disorders. In addition to endoscopic techniques, there are multiple options for airway reconstruction including ssLTR, dsLTR, ss-pCTR, and ds-pCTR. Pediatric airway surgeons must carefully consider which of these modalities is optimal for achieving safe and expeditious surgical reconstruction for each patient.
In recent years, several studies3,5,8 have looked at surgical outcomes after ssLTR and dsLTR. Saunders et al,8 in 1999, provided the first comparison of ssLTR vs dsLTR. In their article, the overall decannulation rate was 91.4% for the ssLTR group and 61.8% for the dsLTR group. Although acknowledging that the severity of disease seemed to be worse in the dsLTR group, the authors suggest that ssLTR is the procedure of choice for uncomplicated pediatric SGS.8 Hartnick et al5 recognized the problem of comparing surgical results in patients with different disease severity. In their article, they identified a cohort of patients with isolated SGS, divided them by Cotton-Myer grade of stenosis, and presented surgical results for ssLTR, dsLTR, and pCTR. Although formal statistical analysis was not performed, their data suggest that ssLTR offers higher operation-specific and overall decannulation rates than does dsLTR, even when grade of SGS is considered. The authors acknowledge the unavoidable confounder with this retrospective review that the very decision to perform ssLTR over dsLTR suggests that the surgeon believes that the patient has a higher likelihood of success.5 More recently, Agrawal et al3 examined their results with 53 patients who underwent ssLTR and dsLTR. Similar to previous studies, overall decannulation rates were higher in the ssLTR group, although formal statistical analysis was not performed and disease severity was worse in patients who underwent dsLTR.3
Similar to previously published studies,3,5,8 children undergoing dsLTR herein had significantly worse disease based on vocal cord function, grade of SGS, and number of involved airway subsites. This makes direct comparisons regarding success rates challenging. Nevertheless, we found no statistically significant differences in overall decannulation rates between patients who underwent ssLTR vs dsLTR. Similar to previous studies,5,8 ssLTR offered increased operation-specific decannulation rates compared with dsLTR. Although the dsLTR group had significantly worse disease, operation-specific decannulation rates were similar between the late ssLTR and late dsLTR groups. Furthermore, when we compare patients with similar grades of SGS, only a statistically insignificant trend toward improved outcomes was noted for patients who underwent ssLTR vs dsLTR (Table 4).
Substantially more posterior grafts were performed in patients who received dsLTR (Table 1). This is an unavoidable confounder in a retrospective review of patients treated in our practice. Placement of a posterior graft is indicated when LTR with anterior graft alone does not adequately expand the airway and in cases of isolated posterior glottic stenosis or SGS.19 In our practice, patients with higher-grade stenosis and glottic disease are preferentially treated in a double-stage manner. Thus, patients undergoing dsLTR are more likely to have a posterior graft.
Findings from previous studies3,5,8 suggest improved results with ssLTR compared with dsLTR. Recently, incremental technical advances, such as the sutureless posterior graft, shorter stenting times, use of cylindrical Silastic stents, and use of fibrin glue have become more widely used during LTR.17 We believe that some of these advances have more significantly affected the results for patients undergoing dsLTR, specifically, shorter stenting time and use of a cylindrical Silastic stent. Indeed, although this study demonstrated improved outcomes in patients who underwent late compared with early dsLTR, no such difference was seen between the single-stage groups. This study is the first, to our knowledge, to demonstrate similar outcomes for ssLTR and dsLTR.
The present operation-specific and overall decannulation rates compare favorably with those of previously published studies.3,5,6,8 Aside from meticulous surgical technique, there are several possible explanations. First, some other large medical centers have a disproportionate number of patients who are seen after failing previous reconstruction at an outside institution. Thus, a significant portion of their patients may be more challenging revision cases.5 Indeed, only 5 of 71 patients (7%) in the present series received a previous open airway procedure at an outside institution. Second, before performing LTR, we are careful to ensure that the larynx and trachea are prepared to accept a cartilage graft. We will not perform an airway reconstruction on an “active larynx,” and we have cancelled operations on the day of surgery because of significant laryngeal inflammation noted during microlaryngoscopy. Although we do not restrict LTR procedures to the summer months, we make every effort to ensure that patients do not have a viral upper respiratory tract infection at the time of reconstruction and have all patients evaluated by the gastroenterology service for laryngopharyngeal reflux disease and eosinophilic esophagitis. Although not universally agreed on by all researchers,20 we believe that routine preoperative screening for gastroesophageal disease significantly improves outcomes. Recently, we have become even more fastidious about our preoperative gastrointestinal tract evaluation, performing routine blind biopsies of the esophagus and impedance probe testing with the patient taking laryngopharyngeal reflux medications. Anecdotally, we have noticed an increase in the number of patients diagnosed as having eosinophilic esophagitis and have identified several patients with clinically significant nonacidic laryngopharyngeal reflux. As this article suggests, we are careful to select which operation is appropriate for each individual patient. Finally, at The Children's Hospital of Philadelphia, we are fortunate to be part of an outstanding team of airway specialists, including but not limited to nurses, nurse practitioners, speech language pathologists, intensivists, gastroenterologists, and pulmonologists.
Our inability to demonstrate significantly better outcomes for patients who underwent late ssLTR may result from the relatively small sample size in this study. Nevertheless, we believe that these data can be used to promote an individualized approach to open airway reconstruction that includes ssLTR, dsLTR, and pCTR. Even if outcomes are similar, ssLTR offers the advantage of immediate decannulation and allows for surgeons to address the stoma site concurrent with LTR.11,12 The obvious disadvantages with ssLTR are the complications associated with prolonged endotracheal intubation and the need to rely on the newly reconstructed airway immediately after extubation, increasing the potential for airway complications during the perioperative period.11,12 In addition, the perioperative hospital admission was approximately 6.5 days longer for the single-stage group (Table 3). Thus, ssLTR should be used for moderate-grade SGS (grade II and mild grade III) in patients without significant neurologic or pulmonary disease. Single-stage LTR is particularly useful for stenosis involving the tracheostomy site. Isolated suprastomal abnormalities, however, usually can be handled with a simple tracheostomal revision procedure before a decannulation trial and should not be an indication for ssLTR. Significant vocal cord involvement (with or without vocal cord immobility) or tracheomalacia may be relative contraindications to ssLTR because some authors have suggested that this leads to worse outcomes.6,8 Partial CTR should be considered for severe grade III and grade IV SGS, especially in patients whose disease is separated from the larynx by an adequate margin.13- 16 Indeed, 2 of the present patients with grade IV SGS who failed dsLTR were decannulated after pCTR. Double-stage LTR may be considered for patients with grade II and mild grade III stenosis whose disease is separated from the tracheostomy site by an adequate margin, particularly if they have significant disease at the level of the vocal cords. Double-stage LTR should be preferentially performed in children with complex multilevel stenosis, significant neurologic deficits, significant lung disease, or anatomical features that make reintubation technically difficult.6,8,11 Surgeons should also consider dsLTR for patients with more severe grade III or grade IV disease involving the vocal cords. To facilitate expeditious decannulation, surgeons should consider dsLTR in complex patients with significant comorbidities rather than waiting for the comorbid conditions to resolve before proceeding with ssLTR.
In conclusion, pediatric LTR may be approached in a single- or double-stage manner. Although patients in this series who underwent dsLTR had significantly worse disease, decannulation rates were similar to those of patients who received ssLTR. These results suggest that an individualized approach to the treatment of pediatric airway stenosis may maximize decannulation rates and facilitate expeditious decannulation.
Correspondence: Ian N. Jacobs, MD, Division of Pediatric Otolaryngology, The Children's Hospital of Philadelphia, Richard D. Wood Center, First Floor, 34th and Civic Center Boulevard, Philadelphia, PA 19104-4399 (JacobsI@email.chop.edu).
Submitted for Publication: March 29, 2009; final revision received May 18, 2009; accepted September 21, 2009.
Author Contributions: All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Smith, Zur, and Jacobs. Acquisition of data: Smith. Analysis and interpretation of data: Smith, Zur, and Jacobs. Drafting of the manuscript: Smith. Critical revision of the manuscript for important intellectual content: Smith, Zur, and Jacobs. Statistical analysis: Smith. Administrative, technical, and material support: Smith, Zur, and Jacobs. Study supervision: Zur and Jacobs.
Financial Disclosure: None reported.
Previous Presentation: This study was presented at the American Society of Pediatric Otolaryngology Annual Meeting; May 30, 2009; Seattle, Washington.