Objectives To describe a consecutive series of children with laryngeal mobility disorders assessed by laryngeal electromyography (LEMG), to propose a grading system for LEMG findings, and to determine whether the LEMG grades correlate with requirement for tracheostomy.
Design Retrospective, observational, uncontrolled study.
Setting A single pediatric otolaryngology practice.
Patients Children who had LEMG performed and a minimum follow-up of 3 months.
Main Outcome Measures Demographic characteristics, diagnoses, surgical procedures, number of LEMG procedures, and complications were obtained. The LEMG results from the thyroarytenoid and posterior cricoarytenoid muscles were graded 0 to 4 according to amplitude and relation to the phase of respiration. A correlation analysis between the need for tracheostomy and the baseline LEMG score as well as a multivariable analysis to determine the predictors of requirement for tracheostomy were performed.
Results Between April 28, 2008, and November 2, 2011, 43 LEMG procedures were performed on 23 patients (13 girls; mean [SD] age, 1.5 [2.85] years). Eight required tracheostomy. Among the 23 patients, 16 had laryngeal paralysis (11 bilateral, 5 unilateral), 4 had laryngeal dyskinesia, and 3 had miscellaneous conditions. Fourteen had secondary large airway lesions, and 14 had a nonairway diagnosis that affected respiration. The overall LEMG results correlated negatively with requirement for tracheostomy (r = −0.4; P < .05) and were 86.36% accurate compared with endoscopy. No predictors for tracheostomy were identified.
Conclusions The LEMG grading was accurate and correlated with the requirement for tracheostomy. Combined with endoscopy, the grading may help better characterize laryngeal mobility disorders.
Hirano and Ohala1 first introduced laryngeal electromyography (LEMG) in 1969 for the diagnosis of movement disorders in the larynx. In adults, it is commonly performed in the outpatient clinic while the patient is awake. In this setting, it is mainly used to confirm the presence of laryngeal paralysis (LP), assess the degree of specific muscle function, determine topognosis (vagal branches of the larynx), and differentiate true paralysis from mechanical dysfunction, potentially affecting the decision for the therapeutic intervention.2
Despite some interest in applying this test in children, it remains rarely used. As such, the precise clinical indications for LEMG in this age group are not agreed on, nor is there a standardized technique to execute the test or a valid method to rate its findings.3
We therefore performed a retrospective chart review of all children diagnosed as having laryngeal mobility disorders and attempted to determine the utility of LEMG. Specifically, our objectives are to describe a reproducible technique of LEMG in a consecutive series of children with laryngeal mobility disorders, propose a grading system for the findings, and determine whether the grades of the system correlate with requirement for tracheostomy.
This is a retrospective, observational, uncontrolled study. The local ethics review board granted permission (registration number Po00026714).
All the children diagnosed as having laryngeal mobility disorders who were managed in a single pediatric otolaryngology practice (based at the Stollery Children's Hospital) and underwent LEMG were eligible for inclusion. The patients were identified from a prospectively collected surgical database.
The standard of care in this practice has been to subject all children with suspected laryngeal mobility disorders to LEMG while performing diagnostic endoscopy under general anesthesia. Monopolar electrodes were used to sample the thyroarytenoid and posterior cricoarytenoid muscles bilaterally and sequentially. If no potentials were detected, the muscle was sampled twice again. The chest wall movement was monitored using surface electrodes that were direct-current coupled. This allowed synchronization of the respiratory cycle with LEMG recordings. The LEMG results and chest wall movements were recorded using a Cadwell Cascade Elite intraoperative neuromonitoring machine. Chest wall movements were low-pass filtered (3 Hz) to reduce interference from the electrocardiogram. The LEMG results were band-pass filtered (3-3000 Hz) and amplified 2000 times. At least 5 respiratory cycles were recorded and the time base was adjusted to allow for the visualization of at least 3 respiratory cycles on a single screen, irrespective of the respiratory rate.
Irrespective of the method of induction of general anesthesia, the children were kept spontaneously breathing while the airway was examined and the LEMG results were recorded. Inhalational agents were avoided during LEMG and judgment of laryngeal mobility; instead, total intravenous anesthesia (propofol and remifentanil hydrochloride) and topical lidocaine, 1%, were used. The endoscopic examination included flexible transnasal laryngoscopy, rigid bronchoscopy, and microlaryngoscopy. During the latter, palpation of the cricoarytenoid joint was performed to rule out fixation and subluxation.
Only children whose LEMG results were available and who had a minimum follow-up of 3 months were included. The following variables were collected for each patient: demographic data (age in years, sex), diagnoses, surgical procedures, number of LEMG procedures performed and interval between them, grading of each muscle and overall grade of the larynx (5 ordinal grades), endoscopic diagnosis (normal or abnormal), whether a tracheostomy had been required, and complications.
The LEMG findings of each of the 4 muscles were allocated a score from 0 to 4 based on the presence or absence of action potentials (APs), their amplitude, and their timing in relation to the appropriate phase of the respiratory cycle. Therefore, grade 4 indicates an AP in the correct phase more than 75% of the time and of excellent amplitude, whereas grade 0 is allocated to the finding of a similar AP but paradoxical in timing. For grade 2, the AP is essentially absent or less than 25% of that expected. Grade 1, on the other hand, is haphazardly distributed between appropriate and inappropriate phases of respiration, and the amplitude would be around 50% of the expected normal amplitude. Lastly, in grade 3, the AP is recorded during the normal phase of the respiratory cycle and around 50% of normal activity (Table 1, Figure 1, and Figure 2). The graded LEMG recording was always the initial one performed.
We also documented the presence or absence of a secondary airway lesion and the presence or absence of another diagnosis that affects the respiratory reserve. Secondary airway lesions included laryngomalacia, tracheomalacia, subglottic stenosis, webs, etc. Other diagnoses that affect the respiratory reserve included chronic lung disease, hypotonia, central respiratory disorders, and others.
The clinical diagnoses assignment was dependent on the clinical and endoscopic features. Bilateral abductor LP was diagnosed if the patient presented clinically with stridor, with or without increased work of breathing, swallowing dysfunction, and sleep-disordered breathing. The endoscopic features had to be similar in the awake and sedated states, demonstrate lack of abduction, and show normal passive movement on palpation. Patients with unilateral abductor LP presented with dysphonia and features of aspiration on oral feeds. The endoscopic features were unilateral findings similar to those described for the bilateral diagnosis. Laryngeal dyskinesia was diagnosed as per Denoyelle et al.4 Specifically, infants with laryngeal dyskinesia present with a picture similar to that of bilateral LP but their endoscopic abnormalities are exhibited only in the awake state. Finally, arytenoid dislocation was diagnosed as described by Rubin et al5 during full endoscopic assessment.
Descriptive statistics were determined. Correlation (Spearman coefficient) was determined between the requirement for tracheostomy and the following: (1) mean overall grade of LEMG results, and (2) mean abductor and adductor muscles LEMG grades. Multiple linear regression analysis was performed, considering the requirement for tracheostomy (score of 0 or 1) as the dependent variable. The independent variables were the LEMG score (score of 0-4), the presence or absence of a secondary airway lesion (score of 0 or 1), the presence or absence of a diagnosis that may affect the respiratory drive (score of 0 or 1), age in months, and sex (male being scored as 1 and female being scored as 0). Significance was held at P < .05. The sensitivity and specificity were determined. This analysis tested the finding of a fully normal LEMG grade of 4 against the finding of a fully normal movement on endoscopy (this was performed for each hemilarynx individually). The statistical analyses were performed using SigmaStat and SigmaPlot version 3 statistical software (Systat Software, Inc).
Between April 28, 2008, and November 2, 2011, 43 LEMG procedures were performed on 23 consecutive patients, 15 of whom had only 1 recording (mean [SD], 1.9 [1.7]; range, 1-8). There were 13 females. The mean age was 1.5 years at the first LEMG (SD, 2.85 years; range, 10 days to 12.6 years).
Sixteen patients had a confirmed diagnosis of LP (11 bilateral, 5 unilateral), 4 were diagnosed as having laryngeal dyskinesia, and 3 were normal (1 with suspicion of vincristine sulfate toxic effects who was diagnosed eventually as having laryngitis, 2 with suspected LP who turned out to have posterior glottic stenosis [1 patient] or left arytenoid dislocation secondary to intubation [1 patient]).
Among the cases of bilateral LP, 9 were congenital and 2 were secondary to cardiac surgery. As for the unilateral group, 4 were left sided and 1 was right sided. The latter was secondary to tracheoesophageal fistula repair, while 2 occurred after repair of persistent ductus arteriosus and 2 occurred after first-stage surgery for left hypoplastic heart syndrome.
Fourteen children had secondary large airway lesions (other than LP), 5 of whom had more than 1. The secondary airway lesions were subglottic stenosis (5 patients: 3 with grade 1, 1 with grade 2, and 1 with grade 3), posterior glottic stenosis (1 patient with grade 3), laryngomalacia (3 patients: 2 with type 1 and 1 with type 2), laryngeal cleft (3 patients: 2 with type 1 and 1 with type 3), and tracheomalacia (3 patients: all severe, 1 secondary to tracheoesophageal fistula). Two patients had iatrogenic arytenoid dislocation, 1 had laryngeal web, and 1 had a laryngeal granuloma secondary to intubation.
Fourteen children had a nonairway diagnosis that affected respiration. These were chronic lung disease (2 patients, who were extremely premature at birth), congenital heart disease (1 patient had ventricular septal defect, 4 had left hypoplastic heart syndrome), hemidiaphragmatic paralysis (1 patient), hypotonia (3 secondary to CHARGE syndrome, 1 secondary to velocardiofacial syndrome, and 1 secondary to trisomy 13 syndrome), and central sleep apnea (1 patient).
Eight children in total required tracheostomy. Two were decannulated at the time of writing this article, and 1 died of complications related to cardiac surgery and multiorgan system failure.
The mean interval between the first and second LEMG recordings was 4.5 months. The mean interval between the second and third recordings was 9 months.
The mean (SD) LEMG overall score at baseline was 2.8 (1.0), the mean (SD) adductors score was 2.6 (1.0), and the mean (SD) abductors score was 2.8 (0.9). Further, the same calculation was performed for each diagnosis: bilateral LP, unilateral LP, laryngeal dyskinesia, and normal (Table 2). The mean scores were 4 in the normal patients, 3.05 in those with unilateral LP, 2.63 in those with laryngeal dyskinesia, and 2.32 in those with bilateral LP.
The requirement for tracheostomy at the time of LEMG correlated negatively with the overall mean LEMG score (r = −0.4; P = .04) and with the abductor LEMG score (r = −0.5; P = .04). On performing multivariable analysis, no independent variables significantly predicted the requirement for tracheostomy (Table 3). The power of the test was less than 50%; therefore, these results should be considered with caution.
The sensitivity of LEMG performed against the endoscopic findings was 100%, while the specificity was 89.7%. The positive predictive value was 63.6%, and the negative predictive value was 100%. The overall accuracy was 86.36%. These results are shown in Table 4.
Four patients with bilateral LP who had a tracheostomy were treated with partial cricotracheal resection, tracheal resection and end-to-end anastomosis, arytenoidectomy, and suture lateralization. The first 2 were decannulated as a result, but the other 2 were not. None of the rest received any other modalities of treatment.
The status of decannulation at a later date was taken into consideration. If the patient required a tracheostomy shortly after the diagnosis, then the patient was accounted for as such. Hence, whether the patient had a later procedure should not affect this analysis.
Of the 8 patients who required tracheostomy, 4 had subglottic stenosis, 1 had tracheomalacia, 1 had tracheal stenosis, 2 with glottic webs, 1 with grade 3 laryngeal cleft, and 1 with arytenoid dislocation. Five patients had more than 1 airway lesion. In the 15 patients who did not require tracheostomy, 1 had posterior glottic stenosis, 1 had subglottic stenosis, 3 had type 1 laryngeal cleft, 3 had laryngomalacia, 1 had arytenoid dislocation, and 1 had postintubation granuloma. Only 1 patient had more than 1 associated airway lesion. Once the patients were subgrouped in this manner, no meaningful statistical comparison was feasible owing to the sample size.
In this retrospective study, we used a uniform technique for LEMG in a series of consecutively managed neonates and children. We used a novel grading system based on the amplitude and timing of the AP in relation to the respiratory cycle, which, to our knowledge, is a first in the literature. The results showed a significant correlation between the mean grades of the LEMG results and requirement for tracheostomy. It was also interesting to see that the mean LEMG grades attached to various diagnoses mirrored their expected clinical impact on the patients and demonstrated a change from the normal to the bilaterally affected larynx with decreasing values for both the abductor and adductor muscles. Judged against the reference standard test (endoscopy), the LEMG grading proved to be a highly accurate test (86.36%).
Despite the modest sample size (although one of the largest in the literature), the series included a variation of mobility disorders that represents the full spectrum the clinician might encounter and affords a range that helps in validating a grading system. None of our patients developed any complications related to the procedures.
We performed a literature search on PubMed using the search words “laryngeal electromyography,” “laryngeal mobility disorders,” “voice,” “vocal folds,” and “laryngeal paralysis.” The search spanned from 1966 through the present and was limited to English language and children. On reviewing the titles and abstracts produced, only 8 articles were found suitable to evaluate.3,6-12 We reviewed these references critically with respect to sample size, study design, technique, muscles sampled, use of a scale/grading for the findings, diagnoses assessed, conditions of anesthesia, and whether serial measurements took place (Table 5). We will comment on one of these studies9 separately because the conditions of the experiment were very different.
Historically, the first use of LEMG for evaluation of mobility disorders in children was in 1987 by Koch et al6; since then, the technology has strived to improve and to reduce interference with other equipment and sources in the operating room. Indeed, there are many differences regarding technique and conditions among the subsequent studies (Table 6). For example, while most used monopolar electrodes, only 2 used the bipolar variety8,9; aside from the studies by Berkowitz7 and Jacobs and Finkel,8 the remaining works sampled either thyroarytenoid or posterior cricoarytenoid muscle only, leaving out the posterior cricoarytenoid more commonly.
From the perspective of the conditions of testing, with the exception of the study by Ysunza et al,11 who tested cooperative older children in the awake state, all studies used LEMG in the operating room under general anesthesia. Regarding the specific anesthesia protocol, the latest 3 studies (by Scott et al,3 Jacobs and Finkel,8 and Maturo et al12) abandoned the previously used inhalational technique, which potentially can affect laryngeal movement, and used intravenous agents, allowing better control of the duration and depth.
All the cited studies were noncontrolled and retrospective (aside from those by Ysunza et al11 and Maturo et al12). The sample sizes varied between 4 and 32 patients (median, 10 patients). The study by Koch et al6 remains the only one to propose a 5-point scale of the LEMG findings; however, it was never validated or used again. None of the studies took to testing the diagnostic accuracy of LEMG except the analysis by Ysunza et al,11 which purported near-perfect sensitivity (100%) and specificity (92%). Further, serial measurements over time were performed only in the latest 2 studies by Scott et al3 (4 patients) and Maturo et al12 (8 patients), but there was no longitudinal follow-up at equivalent times for all patients included. Ultimately, the series in these studies were nonconsecutive and largely limited to patients with LP, aside from the study by Wohl et al.10 This along with the absence of a validated grading system highlight the understandable exploratory nature of these investigations and the limitations on attempts of correlating the findings with the natural history of the disease.
Our study shares some limitations with past works, namely the relatively small sample size, retrospective nature, and lack of a blinded neurophysiologist. However, we used a uniform technique of LEMG (sampling both the main adductor and abductor muscles) and the same general anesthetic protocol throughout. Despite the lack of a matched control group, there were 3 patients who had no movement abnormalities and their recordings were completely normal, and there were 5 hemilarynges among those who had unilateral LP and normal recordings. These patients and the narrow standard deviations should be a reassuring factor for the reliability of the measurements.
The grading system judges the timing of APs recorded in addition to their amplitude and thus considers the probability that synkinetic activity (grades 0 and 1) can adversely affect the clinical picture more than total neurological silence (grade 2). Despite drawing our sample from a single center, the patients included were consecutive and exhibited a varied spectrum of pathological findings. This allowed traditional decisions of management (based on differentiating fixation from paralysis) to be reconfirmed as a utility for the test. It also helped exclusion of vincristine neurotoxic effects, avoided tracheostomy in 3 patients, and bolstered a decision to decannulate another 2. Although we are not in a position as yet to comment on the results of serial LEMG owing to the currently small number of patients with more than 1 evaluation, the potential exists for changing management if appropriate longitudinal studies with more patients are undertaken. In one of the larger studies to date, we have described a uniform protocol for performing LEMG to investigate laryngeal mobility disorders in children. The novel grading system has a high overall accuracy and caters to both the adduction and abduction functions.
Correspondence: Hamdy El-Hakim, MD, Pediatric Otolaryngology Service, Division of Otolaryngology–Head and Neck Surgery, Stollery Children's Hospital, 8400 112th St NW, Edmonton, AB T6G 2P4, Canada (hamdy.elhakim@albertahealthservices.ca).
Submitted for Publication: April 20, 2012; final revision received June 28, 2012; accepted August 15, 2012.
Author Contributions: Dr El-Hakim 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: Norton and El-Hakim. Acquisition of data: AlQudehy, Norton, and El-Hakim. Analysis and interpretation of data: AlQudehy, Norton, and El-Hakim. Drafting of the manuscript: AlQudehy and El-Hakim. Critical revision of the manuscript for important intellectual content: Norton and El-Hakim. Statistical analysis: AlQudehy and El-Hakim. Administrative, technical, and material support: Norton and El-Hakim. Study supervision: Norton and El-Hakim.
Financial Disclosure: None reported.
Previous Presentation: This article was presented at the American Society of Pediatric Otolaryngology 2012 Annual Meeting; April 21, 2012; San Diego, California.
1.Hirano M, Ohala J. Use of hooked-wire electrodes for electromyography of the intrinsic laryngeal muscles.
J Speech Hear Res. 1969;12(2):362-3735808863
PubMedGoogle Scholar 2.Guindi GM, Payne JK, Higenbottam TW. Clinical electromyography in ear, nose and throat practice.
J Laryngol Otol. 1981;95(4):407-4137229523
PubMedGoogle ScholarCrossref 3.Scott AR, Chong PS, Randolph GW, Hartnick CJ. Intraoperative laryngeal electromyography in children with vocal fold immobility: a simplified technique.
Int J Pediatr Otorhinolaryngol. 2008;72(1):31-4018006083
PubMedGoogle ScholarCrossref 4.Denoyelle F, Garabedian EN, Roger G, Tashjian G. Laryngeal dyskinesia as a cause of stridor in infants.
Arch Otolaryngol Head Neck Surg. 1996;122(6):612-6168639291
PubMedGoogle ScholarCrossref 5.Rubin AD, Hawkshaw MJ, Moyer CA, Dean CM, Sataloff RT. Arytenoid cartilage dislocation: a 20-year experience.
J Voice. 2005;19(4):687-70116301111
PubMedGoogle ScholarCrossref 6.Koch BM, Milmoe G, Grundfast KM. Vocal cord paralysis in children studied by monopolar electromyography.
Pediatr Neurol. 1987;3(5):288-2933508076
PubMedGoogle ScholarCrossref 7.Berkowitz RG. Laryngeal electromyography findings in idiopathic congenital bilateral vocal cord paralysis.
Ann Otol Rhinol Laryngol. 1996;105(3):207-2128615584
PubMedGoogle Scholar 8.Jacobs IN, Finkel RS. Laryngeal electromyography in the management of vocal cord mobility problems in children.
Laryngoscope. 2002;112(7, pt 1):1243-124812169907
PubMedGoogle ScholarCrossref 9.Gartlan MG, Peterson KL, Luschei ES, Hoffman HT, Smith RJH. Bipolar hooked-wire electromyographic technique in the evaluation of pediatric vocal cord paralysis.
Ann Otol Rhinol Laryngol. 1993;102(9):695-7008373093
PubMedGoogle Scholar 10.Wohl DL, Kilpatrick JK, Leshner RT, Shaia WT. Intraoperative pediatric laryngeal electromyography: experience and caveats with monopolar electrodes.
Ann Otol Rhinol Laryngol. 2001;110(6):524-53111407843
PubMedGoogle Scholar 11.Ysunza A, Landeros L, Pamplona MC, Prado H, Arrieta J, Fajardo G. The role of laryngeal electromyography in the diagnosis of vocal fold immobility in children.
Int J Pediatr Otorhinolaryngol. 2007;71(6):949-95817418427
PubMedGoogle ScholarCrossref 12.Maturo SC, Braun N, Brown DJ, Chong PST, Kerschner JE, Hartnick CJ. Intraoperative laryngeal electromyography in children with vocal fold immobility: results of a multicenter longitudinal study.
Arch Otolaryngol Head Neck Surg. 2011;137(12):1251-125722183907
PubMedGoogle ScholarCrossref