Jabbour J, Martin T, Beste D, Robey T. Pediatric Vocal Fold ImmobilityNatural History and the Need for Long-Term Follow-up. JAMA Otolaryngol Head Neck Surg. 2014;140(5):428-433. doi:10.1001/jamaoto.2014.81
Copyright 2014 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
The clinical course and outcomes of pediatric vocal fold immobility (VFI) vary widely in the literature, and follow-up in these patients varies accordingly. A better understanding of the natural history of pediatric VFI is crucial to improved management.
To characterize the natural history of pediatric VFI, including symptoms and rates of resolution and surgical intervention.
Design, Setting, and Participants
Retrospective review at an academically affiliated private pediatric otolaryngology practice in a metropolitan area of all patients seen between July 15, 2001, and September 1, 2012, with a diagnosis of complete or partial VFI. After elimination of 92 incomplete or duplicate files, 404 patient records were reviewed for demographic characteristics, etiologies, symptoms, follow-up, resolution, and interventions. Follow-up records were available for 362 patients (89.6%).
Main Outcomes and Measures
Resolution of VFI confirmed by repeated laryngoscopy, length of follow-up, and surgical intervention rates.
Among the 404 patients, left VFI was present in 66.8%, right VFI in 7.9%, and bilateral VFI in 25.3%. Median (range) age at presentation was 2.9 (0-528.1) months. Major etiological categories included cardiac surgery in 68.8%, idiopathic immobility in 21.0%, and neurologic disease in 7.4%. At presentation, 61.4% experienced dysphonia, 54.0% respiratory symptoms, and 49.5% dysphagia. Tracheotomy was performed in 25.7% and gastrostomy in 40.8%. Median (range) duration of follow-up among the 89.6% of patients with follow-up was 17.2 (0.2-173.5) months. Resolution evidenced by laryngoscopy was found in 28.0%, with a median (range) time to resolution of 4.3 (0.4-38.7) months. In patients without laryngoscopic resolution, median follow-up was 26.0 months, and 28.9% reported symptomatic resolution.
Conclusions and Relevance
The natural history of pediatric VFI involves substantial morbidity, with lasting symptoms and considerable rates of surgical intervention. In this large database, the majority of patients did not experience resolution. This suggests a need for more regimented follow-up in these patients, a recommendation for which is proposed here.
Vocal fold immobility (VFI) in children is related to a variety of causes, including cardiac surgical procedures, neurologic conditions, and idiopathic injury.1 Complications caused by VFI include respiratory problems, voice alterations, and feeding difficulties, the distribution of which depends on whether a patient has bilateral or unilateral involvement. Despite what is known about pediatric VFI, important questions remain. Reported recovery rates range widely in the literature, from 8% to 82%2- 7; accordingly, follow-up in these patients is also quite variable, with no established guidelines or recommendations.
As children with VFI live longer as a result of advancements in pediatric cardiac surgery, as well as improved management of certain neurologic conditions, the aforementioned sequelae have the potential to substantially affect a child’s life. The possibility of therapeutic intervention in these children has also expanded, including options for injection laryngoplasty, thyroplasty, and nerve reinnervation.8- 11 The pediatric otolaryngologist has become a key caregiver for patients who do not recover their vocal fold function and manages these issues into adulthood. Yet this management—as well as counseling at the time of diagnosis—is often challenging given the questions that remain regarding the natural history of pediatric VFI and the appropriate follow-up in these patients.
This article presents, to our knowledge, the largest series of pediatric patients with VFI reported in the literature. The purpose of our retrospective review was to better characterize the natural history of pediatric VFI. We highlight the common etiologies and manifestations of VFI, the presenting symptoms and the comorbidities often associated with VFI, and the resolution rate and surgical intervention rate for these children. For those patients who do not recover, the issues of ongoing symptoms and the challenges of consistent follow-up are also discussed. A recommended follow-up schedule is proposed.
This study was a retrospective review of medical records of patients at a private pediatric otolaryngology practice in Milwaukee, Wisconsin. The practice was affiliated with the Children’s Hospital of Wisconsin, and the study was approved by that hospital’s institutional review board. Consent was waived due to the nature of the study. All patients seen between July 15, 2001, and September 1, 2012, with a diagnosis of unilateral or bilateral, partial or complete “Paralysis of Vocal Cords or Larynx” (International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes 478.31-478.35) were included. For the purposes of this study, we combined all of these patients under the heading of VFI. All diagnoses were made by means of flexible laryngoscopy by 1 of 3 attending pediatric otolaryngologists (T.M., D.B., T.R.) in the clinic or in consultation at 1 of 3 area hospitals.
A total of 496 patient files were identified. Patient records were eliminated from the database for the following reasons: insufficient information regarding diagnosis, duplicate files, or failure to confirm VFI when laryngoscopy was repeated within 1 week of diagnosis. A total of 404 records remained in the database. These records were reviewed for information regarding etiologies, laterality of immobility (left, right, or bilateral), presenting symptoms, and comorbidities. For the purposes of data analysis, common symptoms such as stridor, hoarse cry, and feeding difficulties were placed in 1 of 3 categories: respiratory difficulties, dysphonia, or dysphagia. The primary outcome studied was resolution confirmed by means of repeated flexible laryngoscopy. Secondary outcomes included length of follow-up, resolution of symptoms, and rates of surgical interventions related to VFI.
The collected data underwent simple descriptive analysis, as well as comparative analyses using the χ2 and 2-tailed Mann-Whitney tests. Independent audit of the database was undertaken to ensure accurate transcription of the available information in the medical records.
Of the 404 patients, 211 were male (52.2%). Unilateral left VFI (LVFI) was observed in 270 patients (66.8%), unilateral right VFI (RVFI) in 32 (7.9%), and bilateral VFI (BVFI) in 102 (25.2%). The median (range) age at diagnosis was 2.9 months (1 day to 528.1 months) (Table 1). There were 3 patients older than 18 years who remained in the database because they were initially evaluated in this pediatric otolaryngology clinic and their injuries were related to congenital disease. Three hundred seventeen patients (78.5%) were younger than 12 months at diagnosis, and 290 (71.8%) were younger than 6 months.
The etiologies of VFI included cardiac surgery (both patent ductus arteriosus ligation and surgery for complex heart disease) in 278 patients (68.8%), neurological disease in 30 (7.4%), mixed neurological and cardiac disease in 5 (1.2%), idiopathic etiology in 85 (21.0%), and miscellaneous causes (including trauma, thoracoscopy, and lymphoma) in another 6 (1.5%) (Table 1).
At the time of diagnosis, 248 patients (61.4%) presented with dysphonia, 218 (54.0%) with respiratory difficulties, and 192 (47.5%) with dysphagia. Patients with unilateral VFI were significantly more likely to have dysphonia, and those with BVFI were significantly more likely to have respiratory difficulties (Table 2).
Concomitant airway disease was seen in 162 patients (40.1%), most commonly in those with BVFI (60 of 102 [58.8%]). Laryngomalacia (90 of 158 [57.0%]) and subglottic stenosis (53 of 158 [33.5%]) were the most common diagnoses.
Three hundred sixty-two patients (89.6%) had at least 1 follow-up evaluation. The median (range) length of follow-up was 17.2 (0.2-173.5) months. Of those with any follow-up, duration longer than 6 months and longer than 12 months was observed in 320 patients (88.4%) and 306 patients (84.5%), respectively. Eleven patients (3.0%) died during follow-up, most of whom had substantial neurologic and/or cardiac comorbidities.
Of the total population, 113 patients (28.0%) showed resolution evidenced by laryngoscopy. Forty-two (10.4%) had insufficient follow-up data, and 249 (61.6%) had presumed ongoing immobility. Of those in the “ongoing” group, 72 (28.9%) had no symptoms at their last follow-up appointment. This group was labeled the “clinical resolution” group and had a median follow-up period of 26.0 (0.6-120.5) months. In those with laryngoscopic evidence of resolution, median (range) time to resolution was 4.3 (0.4-38.7) months. Combining the true resolution and clinical resolution groups, a total of 185 patients (45.8%) showed no symptoms related to VFI at their last follow-up appointment (Table 3).
Of the 113 patients with confirmed resolution on laryngoscopy, 72 (63.7%) had LVFI, 8 (7.1%) RVFI, and 33 (29.2%) BVFI. Thus, VFI resolved in 27% of all patients with LVFI, 25% of patients with RVFI, and 32% of patients with BVFI. These differences were not statistically significant. Median time to resolution was significantly shorter in patients with LVFI than in patients with BVFI (P = .01). When broken down by etiology, VFI resolved in 24% (68 of 278) of the cardiac surgery group, 27% (8 of 30) of the neurological group, 40% (34 of 85) of the idiopathic group, 0% (0 of 5) of the mixed group, and 50% (3 of 6) of the miscellaneous group. Details regarding median time to resolution can be found in Table 4 and Table 5. The only statistically significant difference by etiology was between the idiopathic group and cardiac surgery group (P = .005).
In our study population, 104 patients (25.7%) required tracheotomy. The proportion of patients with BVFI requiring tracheotomy (69%) was significantly higher than that of patients with LVFI (11%) or RVFI (16%) (P < .001). Patients with concomitant airway disease were also significantly more likely to require a tracheotomy than those without other airway disease (42.6% vs 14.5%; P < .001). One hundred sixty-five patients (40.8%) required gastrostomy tube placement. Once again, the proportion of patients with BVFI requiring gastrostomy tube placement (51%) was significantly higher compared with patients with LVFI (39%) or RVFI (28%) (P = .03 and .02, respectively). Patients with concomitant airway disease were also significantly more likely to require gastrostomy tube placement (50.0% vs 34.7%; P = .002). Collagen medialization was performed in 39 patients (9.7%). The indication for all collagen medializations was aspiration and feeding difficulties. The procedure was performed by means of injection of collagen into the paraglottic space until the affected cord was brought to, or just past, the midline.
To our knowledge, this is the largest cohort of pediatric patients with VFI described in the literature. It is also unique in that etiologies and injuries of any type are combined in 1 database, providing a more global picture of pediatric VFI. Along with previous studies, most of which examine a particular injury and/or etiology, this study provides important information to better characterize the natural history of pediatric VFI and guide optimal management of these patients.
Our findings are similar to those of the 15-year review conducted by Daya et al,3 in which iatrogenic injury (namely, cardiac surgery) was the leading cause of VFI in children. This contrasts with the findings of a survey of pediatric otolaryngologists in which idiopathic causes were deemed predominant.10 Differences between our study and that of Daya et al3 include a lower rate of bilateral injury, higher rate of iatrogenic etiology, and lower rates of neurologic and idiopathic etiology in our data.3 Whereas the exact reason for these differences is unclear, they are logically related because injury secondary to cardiac surgery is most frequently unilateral (in this study, 86.7%) and neurologic and idiopathic etiologies are more likely to result in bilateral VFI.12 The main difference likely stems from a higher rate of cardiac surgery in this study’s population. The data to make this comparison are not available, but this practice presumably saw a greater proportion of cardiac surgery patients because of a referral pattern that included nearly all pediatric patients with cardiac surgery–related VFI in the Milwaukee area.
Given the high rate of cardiac surgery as an etiology in our cohort, it is also worthwhile to make comparisons with the more robust body of literature on pediatric VFI in that population. These studies support the higher incidence of LVFI in our patients, as well as the relative distribution of presenting symptoms.2,4- 6,13 The reported incidence of VFI in pediatric cardiac surgery ranges from 8% to 52%, varying in different studies and with different types of surgical procedures.2,14- 16 The question of the incidence of VFI among all pediatric patients who underwent cardiac surgery was not pursued in this study.
Our rate of follow-up was higher than that of other larger-scale studies, which ranged from 64% to 82%.2- 4,6 This follow-up only included evaluations performed by the investigators themselves for any issue related to VFI; it excluded evaluations by other services at the same hospital or even otolaryngologic evaluations later in life that were exclusively for unrelated concerns. Our length of follow-up is similar to that reported by others.5,6,12,15,17,18 Duration of follow-up longer than 11 years in our study is explained by the fact that some patients were included for VFI evaluations during the course of the study but had a previous diagnosis of VFI with a known date of diagnosis.
Our overall rate of resolution evidenced by laryngoscopy (28.0%) was slightly lower than the 35.3% rate reported by Daya et al.3 The minor difference is likely a function of a higher incidence of iatrogenic etiologies, which tend to cause cases of VFI that resolve less frequently. This same explanation is the most probable cause for the dramatic difference between our recovery rate and the 64% rate reported by de Guademar et al7 and Miyamoto et al,12 both of which observed only patients with congenital VFI. The resolution among our cardiac group (24%) is almost identical to the 25.7% rate published by Truong et al.6 Our data support higher rates of resolution in idiopathic VFI, as reported by several others.1,3,7,12 It is worth noting that all percentages shown here are calculated on the basis of the number of patients who received a diagnosis of VFI and not the portion who had follow-up evaluations. As demonstrated in previous articles, looking only at those with follow-up would lead to higher reported recovery rates.3,6
The time to laryngoscopic resolution noted here was similar to that found in the literature.3,7,12,19 Of those whose condition resolved, 79.6% experienced resolution within 1 year, but it took up to 3 years in others. There were 6 patients whose condition resolved at least 30 months after diagnosis, 5 of whom had laryngoscopies between diagnosis and resolution that showed ongoing VFI. Other investigators report recovery up to 11 years after diagnosis.3 These findings call into question the conventional view that recovery that does not occur within 12 months is unlikely to occur.3,20
The majority of our pediatric patients with VFI did not recover and remained symptomatic. Those in the clinical resolution group had presumed ongoing immobility but were asymptomatic at the most recent follow-up appointment. Some of these patients may have had true resolution as well but did not undergo flexible fiberoptic laryngoscopy (FFL) because it was not believed that FFL would change their management. Importantly, when this group was added to the laryngoscopic resolution group, there were still nearly 54% who had no documented resolution of symptoms. Even if those with no follow-up were eliminated, nearly 44% of the total population had ongoing VFI with symptoms.
The importance of ongoing difficulties related to VFI is highlighted by the surgical interventions necessitated. The rate of tracheotomy in this study was comparable to the 4% to 35% reported by others.3,6,7 Not surprisingly, tracheotomy and gastrostomy tube placement were significantly more likely in patients with BVFI and those with concomitant airway disease. Because those with concomitant airway disease were also more likely to have BVFI, these potential confounders were statistically controlled for, and each condition was an independent risk factor for these surgical interventions. Daya et al3 reported a rate of 68% in their study of patients with BVFI, nearly identical to the 69% rate of tracheotomy in our BVFI cohort. Of note, there were 15 patients with BVFI in our cohort whose condition did not resolve but who did not require tracheotomy either. These patients were observed for a median (range) of 6 (1-76) months, and only 3 showed some improvement on follow-up laryngoscopy. Gastrostomy tube placement occurred in 40.8% of our patients, comparable to the 15% to 63% reported by others.5,6,17 Our rate of collagen medialization was comparable to the 6% rate reported by Truong et al.6
The strengths of this study bear highlighting. It is a larger cohort of pediatric patients with VFI than any that we found reported in the literature. The evaluation of all etiologies helps the otolaryngologist approach pediatric VFI differently in different patients. Furthermore, the robust follow-up achieved in the study is among the best reported and gives the results—particularly those involving outcomes and interventions—added weight.
The study also has several limitations. As a retrospective review, the accuracy of the data are dependent on the thoroughness of the original documentation and subsequent retrieval and interpretation of records. In particular, the study period spanned a transition from written to electronic records within this practice, adding a challenge to evaluating the medical records for each patient. Furthermore, in some cases of VFI, it can be difficult to assign etiology to 1 factor in particular. For example, in an infant whose vocal folds are not evaluated until after a prolonged intubation and cardiac surgery, is VFI related to the intubation or surgery? Finally, the data are limited at times by inconsistent follow-up—not only for those who were lost to follow-up but also for those who did not undergo FFL again because they were asymptomatic. Without laryngoscopic evidence of resolution, we could not firmly conclude that it had occurred in any of these patients.
This study offers valuable information to better characterize the natural history of pediatric VFI. Whereas the etiologies and symptoms of pediatric VFI have been well described, this large cohort highlights the low rate of resolution and high rate of surgical interventions seen in these patients. Most importantly, it shows that nearly half of patients were still symptomatic at the time of last follow-up. For these patients, it was concluded that follow-up on an as-needed basis is not sufficient, especially because the long-term effects of VFI on patients as they grow into school age are not well understood. Anecdotally, we have observed that such patients experience speech and educational challenges more frequently than their peers. To our knowledge, no studies have examined this observation, but we are currently pursuing an experimental project at our institution to investigate this question. Until that is completed, we practice and propose more regimented follow-up for all patients with VFI according to the following schedule: Within the first year after diagnosis, patients should be observed every 4 months, correlating with the median time to resolution. If at any point, their condition has clinically resolved, they should undergo FFL to ensure true resolution. If true resolution is verified, they can then follow up as needed. If their VFI has not resolved at 1 year, they should undergo follow-up FFL to reevaluate vocal fold motion. If ongoing immobility is noted, they should then follow up yearly or as needed for worsened symptoms. If they remain asymptomatic for several years, they should repeat FFL every 5 to 6 years, roughly coinciding with transitions into grade school, middle school, and adulthood. Such a schedule would allow potentially subtle effects of ongoing VFI to be identified earlier and the patient to then be offered appropriate therapies and interventions to reduce morbidity.
Pediatric VFI, often diagnosed early in life and most frequently secondary to cardiac surgery, causes substantial morbidity. This is highlighted by considerable rates of surgical intervention and a majority of patients with persistent VFI at last follow-up. These findings suggest a role for close, regimented follow-up in these patients, a recommendation for which has been proposed here. With a better understanding of the natural history of pediatric VFI, management of these patients will improve as well.
Submitted for Publication: November 1, 2013; final revision received January 2, 2014; accepted January 23, 2014.
Corresponding Author: Thomas Robey, MD, Department of Otolaryngology and Communication Sciences, Medical College of Wisconsin, 9000 W Wisconsin Ave, Ste 340, Milwaukee, WI 53201 (firstname.lastname@example.org).
Published Online: March 13, 2014. doi:10.1001/jamaoto.2014.81.
Author Contributions: Drs Jabbour and Robey had full access to all of 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: All authors.
Acquisition of data: Martin, Beste, Robey.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Jabbour, Martin, Robey.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Jabbour, Martin.
Obtained funding: Martin.
Administrative, technical, and material support: All authors.
Study supervision: Martin, Robey.
Conflict of Interest Disclosures: None reported.
Previous Presentations: This study was presented at the American Academy of Otolaryngology–Head and Neck Surgery 117th Annual Meeting; October 1, 2013; Vancouver, British Columbia, Canada; and Wisconsin Society of Otolaryngology–Head and Neck Surgery 2013 Annual Meeting; October 19, 2013; Spring Green, Wisconsin.
Additional Contributions: John Petronovich, MD, Marshfield Clinic, Marshfield, Wisconsin, assisted in compiling the data. Brent Nichols, MD, Medical College of Wisconsin, assisted in auditing the data. Nancy Ghanayem, MD, Angela Birler, MS, CCC-SLP, and Katherine Frontier, MS, CCC-SLP, Children’s Hospital of Wisconsin, assisted in establishing the database. None of these contributors was compensated for involvement in the project.