- Email: [email protected]
Laryngeal Electromyography for Prognosis of Vocal Fold Paralysis
ARTICLE IN PRESS Laryngeal Electromyography for Prognosis of Vocal Fold Paralysis *Adriana Pardo-Maza, *Isabel García-Lopez, †Susana Santiago-Pérez, and *Javier Gavilán, *†Madrid, Spain Summary: Objective. This study aimed to determine the value of laryngeal electromyography in the prognosis of vocal fold paralysis. Study Design. This is a retrospective descriptive study. Materials and Methods. This study included 80 patients diagnosed with unilateral or bilateral vocal fold paralysis on flexible laryngoscopy between 2002 and 2014 in a tertiary medical center. Laryngeal electromyography using a standardized protocol was performed; the outcome measures were classified and analyzed into two groups according to the degree of injury. Group 1 included patients with mild to moderate injury, and group 2 included patients with severe to complete injury. Prognosis was correlated with vocal fold motion recovery status with a minimum of 6 months of follow-up since the symptoms onset using positive and negative predictive values. Results. Sixty patients showed acute or chronic recurrent laryngeal neuropathy in laryngeal electromyography. Twelve of 41 patients included in group 1 recovered motion, and 30 of 35 patients included in group 2 did not recover, resulting in 88.2% of positive predictive value and 35.7% of negative predictive value. Conclusions. Our data confirm that laryngeal electromyography is a useful clinical tool in predicting poor recovery in patients with vocal fold paralysis. It allows identification of candidates for early intervention. Key Words: laryngeal electromyography–vocal fold paralysis–vocal cord paralysis–laryngeal synkinesis– electromyography.
INTRODUCTION The impairment of laryngeal nerve function may be the result of heterogeneous conditions. According to the majority of authors, idiopathic paralysis and iatrogenic injury are the most frequent causes. Data for spontaneous recovery based on cause have been reported in some studies, but statistical analysis of prognosis from these two common causes is still lacking.1,2 Laryngeal electromyography (LEMG) was introduced in 1944 by Weddel, and their principles and techniques were established in the 1950s.3 Since then, it has been used traditionally for the evaluation of vocal fold motion abnormalities and as a guide for therapeutic injections into the intrinsic muscles of the larynx. In the evaluation of vocal fold paralysis (VFP), LEMG is particularly helpful in differentiating paralysis from other causes of vocal fold immobility.4,5 The EMG examination is evaluated in four parts: Insertional activity is the burst of electrical signal that is produced as the needle is introduced into the muscle. The needle produces a mechanical stimulus in the muscle membrane, causing a local and transitory change of potential. Early nerve and muscle injuries produce prolonged insertional activity, whereas late nerve and muscle injuries cause a decrease in the insertional activity. The second one is the spontaneous activity, which refers to the presence of electrical activity in a resting muscle. The presence of
Accepted for publication February 22, 2016. This paper has been presented during the Annual Meeting of the Spanish Society of Otolaryngology Head and Neck Pathology (Madrid, Spain, October 24–27, 2014). No other people contributed to the study. From the *Department of Otolaryngology, Voice Pathology Unit, La Paz University Hospital, IdiPaz Research Institute, Madrid, Spain; and the †Department of Neurophysiology, Voice Pathology Unit, La Paz University Hospital, IdiPaz Research Institute, Madrid, Spain. Address correspondence and reprint requests to Adriana Pardo-Maza, Paseo de la Castellana 261, 28046 Madrid, Spain. E-mail: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2016 The Voice Foundation. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvoice.2016.02.018
spontaneous activity implies that the muscle has an acute/ inflammatory disease or that the nerve has been injured and the process that caused the injury is ongoing. Another parameter is the waveform morphology, which refers to the shape, amplitude, and duration of the motor unit potentials (MUPs), which are the electrical signals produced by the activation of the motor unit (spinal motoneuron, motor axon, and muscle fibers) and registered by EMG. The recruitment refers to the activation of motor units during increased voluntary muscle contraction.6,7 Synkinesis is another parameter that can be assessed by EMG: as a result of aberrant reinnervation, a voluntary contraction produces inappropriate muscular activation and MUPs are registered in several muscles not related to the desired activity. Recruitment reduction pattern during the thyroarytenoid (TA) maximum voluntary contraction allows determination of the degree of severity of the recurrent laryngeal nerve injury.5,6 Based on this and with the aim of answering the question most frequently asked by patients with VFP about whether laryngeal function will return to normal, we performed this study to assess the use of LEMG for the prediction of the outcome in cases of VFP. MATERIALS AND METHODS A retrospective descriptive study was performed, including all patients evaluated in the voice clinic (n = 80) between 2002 and 2014 with unilateral or bilateral vocal fold immobility who had a LEMG evaluation. Diagnosis of vocal fold immobility was made by videoendoscopic evaluation. Recovery of motion was defined as ≥70% of normal motion compared with the healthy side, with obvious abduction during inspiration, and an increase of ≥50% in glottis space during abduction in cases of bilateral VFP. All the examinations were recorded and revised by two different otolaryngologists. Patients with a normal LEMG were excluded from the study, as were those with less than 6 months’ follow-up.
ARTICLE IN PRESS 2
Journal of Voice, Vol. ■■, No. ■■, 2016
The LEMG was performed by an otolaryngologist and a neurophysiologist at least 3 weeks after the endoscopic diagnosis (mean 8 months with a standard deviation of 2.1).
TABLE 1. Etiology of Vocal Fold Immobility Etiology
Electromyographic characteristics and record The EMG was performed with a two-channel electromyograph (Medelec Synergy, VYASIS Healthcare, Surrey, UK). Coaxial needle electrodes are used and the acquisition parameter screens are: – Rest: Sweep 100 ms and gain of 100 μV/division – Voluntary contraction (MUPs and recruitment): Sweeping 300 ms and gain of 100 μV–2 mV/division Our protocol for LEMG has been described previously by García-López et al.5 The cricothyroid and thyroarytenoid (TA) muscles were both approached transcutaneously without local anesthesia. Only the recurrent laryngeal nerve function obtained from the TA muscle activity was used to determine prognosis of vocal fold palsy. The four categories of LEMG parameters registered were spontaneous activity, morphologic characteristics of MUPs, recruitment pattern, and presence of synkinesis. First, we asked the patient to breathe easy to see if there was any spontaneous activity at rest such as fibrillation potentials or positive sharp waves to determine if there was an acute (presence of spontaneous activity) or chronic injury. After this, we recorded the morphologic characteristics of MUPs and the motor unit recruitment by asking the patient to phonate the vowel /i/ with unit reach maximum voluntary contraction. Finally, we asked the patient to do a sniff maneuver to detect the presence of synkinesis defined by the presence of MUPs during respiration. The evaluation of the degree of the injury was made by the neurophysiologist as follows: – – – –
Rich mixed pattern: mild axonal injury Poor mixed pattern: moderate axonal injury Simple pattern: severe axonal injury Absence of voluntary activation: very severe axonal injury (apparently complete).
For the purpose of analysis, we classified patients according to the degree of injury into high-grade injury (severe and complete) and low-grade injury (mild and moderate). Statistical analysis The outcome measurements of vocal fold motion were dichotomized into persistent VFP or recovered vocal fold mobility. Recovery was defined as return of appropriate gross motion of the paralyzed vocal fold on subsequent laryngoscopy and revised by two different laryngologists separately. A multivariate logistic regression analysis was used to test the statistical significance of association between recovered vocal fold motion with different variables included. The positive predictive value (PPV) was calculated by dividing the number of patients who had positive results and disease outcome by the total number of patients with positive results. This quotient determines the percentage of patients with positive test results (poor
N (%)
Iatrogenic Idiopathic Tumor in the thoracic cavity Skull base lesion Trauma CREST syndrome (X cranial nerve mononeuritis)
36 (60%) 13 (21%) 4 (6.6%) 3 (1.8%) 3 (1.8%) 1 (1.65)
prognosis) who actually had a persistent VFP. The negative predictive value (NPV) was calculated by dividing the number of patients who had a negative test result and normal outcome by the total number of patients with a negative test result. This percentage determines the amount of patients with negative test results (good prognosis) who actually recovered vocal fold motion. RESULTS Twenty patients were excluded for statistical analysis because they did not fulfill the inclusion criteria: those who recovered mobility before doing the LEMG (10 patients), patients with normal LEMG (six patients), and lost to follow-up (four patients). Of the remaining 60 patients, 36 (60%) were women. The average age was 59 years (18–83 years). The videoendoscopic diagnosis was 13 patients with right vocal fold immobility, 33 (55%) with left vocal fold immobility, and 14 with bilateral vocal fold immobility. Regarding the etiology, 36 cases (60%) were iatrogenic and 13 (21%) were idiopathic. Other causes are summarized in Table 1. Complete recovery of vocal fold motion was achieved in 18 (23.6%) of 76 vocal folds, whereas 58 vocal folds (76.3%) remained immobile. The timing of vocal fold motion recovery after the onset of the paralysis ranged from 2 to 12 months, with a median of 5 months. The multivariate logistic regression analysis did not show statistical significance in the association between recovering vocal fold motion and any of the studied variables (Table 2).
TABLE 2. Multivariate Logistic Regression Analysis Recovery Motion Sex Age Etiology Side Severity lesion degree RRLN Severity lesion degree LRLN Synkinesis
Odds Ratio
P > |z|
[95% Conf. Interval]
0.5 0.9 0.7 0.5 0.9
0.368 0.670 0.587 0.362 0.801
0.16–1.95 0.94–1.03 0.27–2.07 0.19–1.82 0.57–1.52
1.1
0.667
0.66–1.87
1.1
0.847
0.27–4.74
Abbreviations: LRLN, left recurrent laryngeal nerve; RRLN, right recurrent laryngeal nerve.
ARTICLE IN PRESS Adriana Pardo-Maza et al
Laryngeal Electromyography for Prognosis of VFP
TABLE 3. Laryngeal Electromyography Outcome Severity of Injury
RRLN
LRLN
Synkinesis
Recovered
Mild Moderate Severe Complete
6 10 13 3
10 16 13 5
5 14 8 0
6 9 4 0
Abbreviations: LRLN, left recurrent laryngeal nerve; RRLN, right recurrent laryngeal nerve.
LEMG outcome (Table 3) Analysis of the outcomes showed an association between presence of synkinesis and the moderate degree injury in the LEMG. This association was statistically significant, P = 0.03 (CI 95% 1.0–2.3). Regarding the LEMG findings, 34 vocal folds (44.7%) had a high-grade injury (severe and complete), with the expectation of poor recovery, whereas 42 vocal folds (55.2%) had MUPs and recruitment consistent with low-grade injury (mild and moderate) with the expectation of good recovery. After determining whether the patient recovered mobility, we found that 30 of 34 vocal fold with a high-grade nerve injury had no recovery (PPV 88.2% CI 95% 0.65–0.94), whereas 4 of 34 (11.7%) had recovery. In recurrent laryngeal nerve with LEMG, findings deemed consistent with low-grade injury: 15 of 42 (NPV 35.7% CI 95% 0.65–0.94) achieved good recovery, whereas 27 of 42 (64.2%) showed no recovery. DISCUSSION True VFP results from denervation secondary to injury to the laryngeal or vagus nerve. VFP may be unilateral or bilateral, and depending on this, the patient will complain of dysphonia, hoarseness, choking episodes, dysphagia or dyspnea, and stridor. Endoscopic laryngeal findings can also vary. In cases of unilateral VFP (UVFP), the most common laryngoscopic findings include bowing, incomplete glottal closure, and phase asymmetry on videostroboscopy. The paralyzed side tends to be shortened and the arytenoid is commonly anteriorly rotated.8 Bilateral VFP findings include lack of movement of both vocal folds resulting in a reduced glottis space. Uniformly, in UVFP, the left vocal fold is affected more often than the right (approximately 60% or more), due to the greater length and more profound descent into the mediastinum of the left-sided nerve, and consequent greater vulnerability to disease and surgical injury. The impression that the left fold is less likely to be paralyzed by surgery, on the other hand, is probably an artifact of the relatively high number of cases accounted for by malignancy.9 The etiologies of VFP have evolved through the years. Rosenthal and colleagues10 performed a large comparative retrospective analysis spanning a 20-year period by including 827 patients who were seen with vocal fold immobility. In the first decade, from 1985 to 1995, the most common cause was malignancy (mostly lung). By contrast, in the second decade, from 1996 to 2005,
3
the most common cause was non-thyroid surgery (including anterior cervical spine and carotid surgery). Consistent with these findings, Merati and colleagues11 reported a greater proportion of anterior cervical spine surgery than thyroid, thoracic, or cranial procedures among patients with iatrogenic vocal fold motion impairment in a 1-year retrospective study. In our series, the most frequent cause of VFP was thyroidectomy, and in only one case was the lesion secondary to an anterior cervical spinal surgery. In our study, 20 patients have been excluded because of different reasons. We are aware that this can constitute a limitation by reducing the number of cases. Anyway, because of the low number of studies with high number of patients, our study has a significant value in assessing the importance of LEMG in predicting the recovery of motion in VFP. The potential for spontaneous motion recovery depends on the clinical scenario, but even reports addressing VFP from the same cause can be difficult to reconcile because of differences in data collection and reporting. The rates of return of vocal fold motion reported in the literature vary significantly, from 5% to 83%.12 In this work, the definition of vocal fold motion recovery has been chosen according to the literature. The metaanalysis of Rickert et al13 does not specify the percentage of recovery that is published in all the studies; Wang et al14 indicate a recovery of 75%, which is very similar to our value of 70%. There is a lack of information about the exact relation between the different parameters involved in the VFP and the vocal fold motion recovery rates. One of these parameters is the Voice Handicap Index. There is no correlation between the handicap index and the recovery motion, as demonstrated in the study published by Young and colleagues in which, despite recovery of vocal fold motion, 16% of patients had significant voice handicap. In contrast, 21% of patients without motion recovery had normal Voice Handicap Index-10 scores.12 In this context, clinical neurophysiological testing would appear to be well suited in resolving some of these ambiguities. LEMG is able to identify normal innervation, absence of innervation, reinnervation, and even synkinesis by characteristic electrical signals. One important issue is the time when the EMG has to be done. Wang et al14 recommend 2 months, but the time classically established by neurologists for the EMG is 3 weeks. At 3 weeks, you can establish the type and the severity of the lesion, whereas at 2 months you can see the presence of the reinnervation, which actually is related with the prognosis. The objective of our study is to relate the degree of the lesion with the prognosis of vocal motion recovery, that is why we perform the LEMG at a minimum of 3 weeks after the injury. Statham and colleagues15 demonstrated that the presence of synkinesis in patients with VFP improved the NPV of LEMG from 53% to 100% and the sensitivity from 56% to 100%. The presence of synkinesis in our study was 30%, more frequent in the low-grade injury, and its presence is not associated with worse prognosis. This can be attributed to the fact that in our study, the prognosis was determined only by the grade of injury. Rickert and colleagues13 published a meta-analysis in 2012 of studies reporting LEMG results and clinical outcomes, which
ARTICLE IN PRESS 4 include 503 patients, reporting a PPV of 90.9% and an NPV of 55.6%. Similar results were published in 2014 by Wang and colleagues14 who performed a prospective study using LEMG to predict the long-term prognosis of UVFP including 85 patients, reporting a PPV of 93% and an NPV of 40%. In our study, the follow-up is longer, because it starts in the moment of the laryngoscopic diagnosis, instead of in the symptoms onset. Moreover, the accuracy of the recovery motion of the paralyzed vocal fold is supported by two different laryngologists. Our results are comparable with those published in the literature. They allow us to conclude that LEMG is a good predictor of no recovery of vocal fold motion, but it is not useful to predict recovery. The choice of treatment in patients with laryngeal paralysis should be influenced by the prognosis, the severity of symptoms, and the vocal demand of the patient. In those cases where the LEMG indicates a favorable prognosis, we can choose a conservative treatment or a temporary surgical procedure. On the other side, patients with severe symptoms or high professional vocal demands in which laryngeal EMG shows a bad prognosis can be treated with permanent surgical procedure without waiting too much.
CONCLUSIONS Laryngeal EMG is useful for the prediction of negative outcomes, but it is not so helpful in predicting positive outcomes. This fact can aid in deciding when to carry out a definitive treatment such as thyroplasty in cases of lack of compensation.
REFERENCES 1. Munin MC, Rosen CA, Zullo T. Utility of laryngeal electromyography in predicting recovery after vocal fold paralysis. Arch Phys Med Rehabil. 2003;84:1150–1153.
Journal of Voice, Vol. ■■, No. ■■, 2016 2. Wang C-C, Chang M-H, Wang C-P, et al. Prognostic indicators of unilateral vocal fold paralysis. Arch Otolaryngol Head Neck Surg. 2008;134:380–388. doi:10.1001/archotol.134.4.380. 3. Smith LJ, Rosen CA, Niyonkuru C, et al. Quantitative electromyography improves prediction in vocal fold paralysis. Laryngoscope. 2012;122:854– 859. doi:10.1002/lary.21884. 4. Heman-Ackah YD, Barr A. The value of laryngeal electromyography in the evaluation of laryngeal motion abnormalities. J Voice. 2006;20:452–460. doi:10.1016/j.jvoice.2005.04.009. 5. García-López I, Santiago-Pérez S, Peñarrocha-Teres J, et al. [Laryngeal electromyography in diagnosis and treatment of voice disorders]. Acta Otorrinolaringol Esp. 2012;63:458–464. doi:10.1016/j.otorri.2012.05.004. 6. Heman-Ackah YD, Mandel S, Manon-Espaillat R, et al. Laryngeal electromyography. Otolaryngol Clin North Am. 2007;40:1003–1023, vi-vii. doi:10.1016/j.otc.2007.05.007. 7. Volk GF, Hagen R, Pototschnig C, et al. Laryngeal electromyography: a proposal for guidelines of the European Laryngological Society. Eur Arch Otorhinolaryngol. 2012;269:2227–2245. doi:10.1007/s00405-012-2036-1. 8. Misono S, Merati AL. Evidence-based practice: evaluation and management of unilateral vocal fold paralysis. Otolaryngol Clin North Am. 2012;45:1083– 1108. doi:10.1016/j.otc.2012.06.011. 9. Sulica L, Blitzer A. Vocal Fold Paralysis. New York: Springer Berlin Heidelberg; 2006. 10. Rosenthal LHS, Benninger MS, Deeb RH. Vocal fold immobility: a longitudinal analysis of etiology over 20 years. Laryngoscope. 2007;117:1864–1870. doi:10.1097/MLG.0b013e3180de4d49. 11. Merati AL, Shemirani N, Smith TL, et al. Changing trends in the nature of vocal fold motion impairment. Am J Otolaryngol. 2006;27:106–108. doi:10.1016/j.amjoto.2005.07.020. 12. Young VN, Smith LJ, Rosen C. Voice outcome following acute unilateral vocal fold paralysis. Ann Otol Rhinol Laryngol. 2013;122:197–204. 13. Rickert SM, Childs LF, Carey BT, et al. Laryngeal electromyography for prognosis of vocal fold palsy: a meta-analysis. Laryngoscope. 2012;122:158– 161. doi:10.1002/lary.22354. 14. Wang C-C, Chang M-H, De Virgilio A, et al. Laryngeal electromyography and prognosis of unilateral vocal fold paralysis—a long-term prospective study. Laryngoscope. 2015;125:898–903. doi:10.1002/lary.24980. 15. Statham MM, Rosen CA, Smith LJ, et al. Electromyographic laryngeal synkinesis alters prognosis in vocal fold paralysis. Laryngoscope. 2010;120:285–290. doi:10.1002/lary.20629.
INTRODUCTION The impairment of laryngeal nerve function may be the result of heterogeneous conditions. According to the majority of authors, idiopathic paralysis and iatrogenic injury are the most frequent causes. Data for spontaneous recovery based on cause have been reported in some studies, but statistical analysis of prognosis from these two common causes is still lacking.1,2 Laryngeal electromyography (LEMG) was introduced in 1944 by Weddel, and their principles and techniques were established in the 1950s.3 Since then, it has been used traditionally for the evaluation of vocal fold motion abnormalities and as a guide for therapeutic injections into the intrinsic muscles of the larynx. In the evaluation of vocal fold paralysis (VFP), LEMG is particularly helpful in differentiating paralysis from other causes of vocal fold immobility.4,5 The EMG examination is evaluated in four parts: Insertional activity is the burst of electrical signal that is produced as the needle is introduced into the muscle. The needle produces a mechanical stimulus in the muscle membrane, causing a local and transitory change of potential. Early nerve and muscle injuries produce prolonged insertional activity, whereas late nerve and muscle injuries cause a decrease in the insertional activity. The second one is the spontaneous activity, which refers to the presence of electrical activity in a resting muscle. The presence of
Accepted for publication February 22, 2016. This paper has been presented during the Annual Meeting of the Spanish Society of Otolaryngology Head and Neck Pathology (Madrid, Spain, October 24–27, 2014). No other people contributed to the study. From the *Department of Otolaryngology, Voice Pathology Unit, La Paz University Hospital, IdiPaz Research Institute, Madrid, Spain; and the †Department of Neurophysiology, Voice Pathology Unit, La Paz University Hospital, IdiPaz Research Institute, Madrid, Spain. Address correspondence and reprint requests to Adriana Pardo-Maza, Paseo de la Castellana 261, 28046 Madrid, Spain. E-mail: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2016 The Voice Foundation. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jvoice.2016.02.018
spontaneous activity implies that the muscle has an acute/ inflammatory disease or that the nerve has been injured and the process that caused the injury is ongoing. Another parameter is the waveform morphology, which refers to the shape, amplitude, and duration of the motor unit potentials (MUPs), which are the electrical signals produced by the activation of the motor unit (spinal motoneuron, motor axon, and muscle fibers) and registered by EMG. The recruitment refers to the activation of motor units during increased voluntary muscle contraction.6,7 Synkinesis is another parameter that can be assessed by EMG: as a result of aberrant reinnervation, a voluntary contraction produces inappropriate muscular activation and MUPs are registered in several muscles not related to the desired activity. Recruitment reduction pattern during the thyroarytenoid (TA) maximum voluntary contraction allows determination of the degree of severity of the recurrent laryngeal nerve injury.5,6 Based on this and with the aim of answering the question most frequently asked by patients with VFP about whether laryngeal function will return to normal, we performed this study to assess the use of LEMG for the prediction of the outcome in cases of VFP. MATERIALS AND METHODS A retrospective descriptive study was performed, including all patients evaluated in the voice clinic (n = 80) between 2002 and 2014 with unilateral or bilateral vocal fold immobility who had a LEMG evaluation. Diagnosis of vocal fold immobility was made by videoendoscopic evaluation. Recovery of motion was defined as ≥70% of normal motion compared with the healthy side, with obvious abduction during inspiration, and an increase of ≥50% in glottis space during abduction in cases of bilateral VFP. All the examinations were recorded and revised by two different otolaryngologists. Patients with a normal LEMG were excluded from the study, as were those with less than 6 months’ follow-up.
ARTICLE IN PRESS 2
Journal of Voice, Vol. ■■, No. ■■, 2016
The LEMG was performed by an otolaryngologist and a neurophysiologist at least 3 weeks after the endoscopic diagnosis (mean 8 months with a standard deviation of 2.1).
TABLE 1. Etiology of Vocal Fold Immobility Etiology
Electromyographic characteristics and record The EMG was performed with a two-channel electromyograph (Medelec Synergy, VYASIS Healthcare, Surrey, UK). Coaxial needle electrodes are used and the acquisition parameter screens are: – Rest: Sweep 100 ms and gain of 100 μV/division – Voluntary contraction (MUPs and recruitment): Sweeping 300 ms and gain of 100 μV–2 mV/division Our protocol for LEMG has been described previously by García-López et al.5 The cricothyroid and thyroarytenoid (TA) muscles were both approached transcutaneously without local anesthesia. Only the recurrent laryngeal nerve function obtained from the TA muscle activity was used to determine prognosis of vocal fold palsy. The four categories of LEMG parameters registered were spontaneous activity, morphologic characteristics of MUPs, recruitment pattern, and presence of synkinesis. First, we asked the patient to breathe easy to see if there was any spontaneous activity at rest such as fibrillation potentials or positive sharp waves to determine if there was an acute (presence of spontaneous activity) or chronic injury. After this, we recorded the morphologic characteristics of MUPs and the motor unit recruitment by asking the patient to phonate the vowel /i/ with unit reach maximum voluntary contraction. Finally, we asked the patient to do a sniff maneuver to detect the presence of synkinesis defined by the presence of MUPs during respiration. The evaluation of the degree of the injury was made by the neurophysiologist as follows: – – – –
Rich mixed pattern: mild axonal injury Poor mixed pattern: moderate axonal injury Simple pattern: severe axonal injury Absence of voluntary activation: very severe axonal injury (apparently complete).
For the purpose of analysis, we classified patients according to the degree of injury into high-grade injury (severe and complete) and low-grade injury (mild and moderate). Statistical analysis The outcome measurements of vocal fold motion were dichotomized into persistent VFP or recovered vocal fold mobility. Recovery was defined as return of appropriate gross motion of the paralyzed vocal fold on subsequent laryngoscopy and revised by two different laryngologists separately. A multivariate logistic regression analysis was used to test the statistical significance of association between recovered vocal fold motion with different variables included. The positive predictive value (PPV) was calculated by dividing the number of patients who had positive results and disease outcome by the total number of patients with positive results. This quotient determines the percentage of patients with positive test results (poor
N (%)
Iatrogenic Idiopathic Tumor in the thoracic cavity Skull base lesion Trauma CREST syndrome (X cranial nerve mononeuritis)
36 (60%) 13 (21%) 4 (6.6%) 3 (1.8%) 3 (1.8%) 1 (1.65)
prognosis) who actually had a persistent VFP. The negative predictive value (NPV) was calculated by dividing the number of patients who had a negative test result and normal outcome by the total number of patients with a negative test result. This percentage determines the amount of patients with negative test results (good prognosis) who actually recovered vocal fold motion. RESULTS Twenty patients were excluded for statistical analysis because they did not fulfill the inclusion criteria: those who recovered mobility before doing the LEMG (10 patients), patients with normal LEMG (six patients), and lost to follow-up (four patients). Of the remaining 60 patients, 36 (60%) were women. The average age was 59 years (18–83 years). The videoendoscopic diagnosis was 13 patients with right vocal fold immobility, 33 (55%) with left vocal fold immobility, and 14 with bilateral vocal fold immobility. Regarding the etiology, 36 cases (60%) were iatrogenic and 13 (21%) were idiopathic. Other causes are summarized in Table 1. Complete recovery of vocal fold motion was achieved in 18 (23.6%) of 76 vocal folds, whereas 58 vocal folds (76.3%) remained immobile. The timing of vocal fold motion recovery after the onset of the paralysis ranged from 2 to 12 months, with a median of 5 months. The multivariate logistic regression analysis did not show statistical significance in the association between recovering vocal fold motion and any of the studied variables (Table 2).
TABLE 2. Multivariate Logistic Regression Analysis Recovery Motion Sex Age Etiology Side Severity lesion degree RRLN Severity lesion degree LRLN Synkinesis
Odds Ratio
P > |z|
[95% Conf. Interval]
0.5 0.9 0.7 0.5 0.9
0.368 0.670 0.587 0.362 0.801
0.16–1.95 0.94–1.03 0.27–2.07 0.19–1.82 0.57–1.52
1.1
0.667
0.66–1.87
1.1
0.847
0.27–4.74
Abbreviations: LRLN, left recurrent laryngeal nerve; RRLN, right recurrent laryngeal nerve.
ARTICLE IN PRESS Adriana Pardo-Maza et al
Laryngeal Electromyography for Prognosis of VFP
TABLE 3. Laryngeal Electromyography Outcome Severity of Injury
RRLN
LRLN
Synkinesis
Recovered
Mild Moderate Severe Complete
6 10 13 3
10 16 13 5
5 14 8 0
6 9 4 0
Abbreviations: LRLN, left recurrent laryngeal nerve; RRLN, right recurrent laryngeal nerve.
LEMG outcome (Table 3) Analysis of the outcomes showed an association between presence of synkinesis and the moderate degree injury in the LEMG. This association was statistically significant, P = 0.03 (CI 95% 1.0–2.3). Regarding the LEMG findings, 34 vocal folds (44.7%) had a high-grade injury (severe and complete), with the expectation of poor recovery, whereas 42 vocal folds (55.2%) had MUPs and recruitment consistent with low-grade injury (mild and moderate) with the expectation of good recovery. After determining whether the patient recovered mobility, we found that 30 of 34 vocal fold with a high-grade nerve injury had no recovery (PPV 88.2% CI 95% 0.65–0.94), whereas 4 of 34 (11.7%) had recovery. In recurrent laryngeal nerve with LEMG, findings deemed consistent with low-grade injury: 15 of 42 (NPV 35.7% CI 95% 0.65–0.94) achieved good recovery, whereas 27 of 42 (64.2%) showed no recovery. DISCUSSION True VFP results from denervation secondary to injury to the laryngeal or vagus nerve. VFP may be unilateral or bilateral, and depending on this, the patient will complain of dysphonia, hoarseness, choking episodes, dysphagia or dyspnea, and stridor. Endoscopic laryngeal findings can also vary. In cases of unilateral VFP (UVFP), the most common laryngoscopic findings include bowing, incomplete glottal closure, and phase asymmetry on videostroboscopy. The paralyzed side tends to be shortened and the arytenoid is commonly anteriorly rotated.8 Bilateral VFP findings include lack of movement of both vocal folds resulting in a reduced glottis space. Uniformly, in UVFP, the left vocal fold is affected more often than the right (approximately 60% or more), due to the greater length and more profound descent into the mediastinum of the left-sided nerve, and consequent greater vulnerability to disease and surgical injury. The impression that the left fold is less likely to be paralyzed by surgery, on the other hand, is probably an artifact of the relatively high number of cases accounted for by malignancy.9 The etiologies of VFP have evolved through the years. Rosenthal and colleagues10 performed a large comparative retrospective analysis spanning a 20-year period by including 827 patients who were seen with vocal fold immobility. In the first decade, from 1985 to 1995, the most common cause was malignancy (mostly lung). By contrast, in the second decade, from 1996 to 2005,
3
the most common cause was non-thyroid surgery (including anterior cervical spine and carotid surgery). Consistent with these findings, Merati and colleagues11 reported a greater proportion of anterior cervical spine surgery than thyroid, thoracic, or cranial procedures among patients with iatrogenic vocal fold motion impairment in a 1-year retrospective study. In our series, the most frequent cause of VFP was thyroidectomy, and in only one case was the lesion secondary to an anterior cervical spinal surgery. In our study, 20 patients have been excluded because of different reasons. We are aware that this can constitute a limitation by reducing the number of cases. Anyway, because of the low number of studies with high number of patients, our study has a significant value in assessing the importance of LEMG in predicting the recovery of motion in VFP. The potential for spontaneous motion recovery depends on the clinical scenario, but even reports addressing VFP from the same cause can be difficult to reconcile because of differences in data collection and reporting. The rates of return of vocal fold motion reported in the literature vary significantly, from 5% to 83%.12 In this work, the definition of vocal fold motion recovery has been chosen according to the literature. The metaanalysis of Rickert et al13 does not specify the percentage of recovery that is published in all the studies; Wang et al14 indicate a recovery of 75%, which is very similar to our value of 70%. There is a lack of information about the exact relation between the different parameters involved in the VFP and the vocal fold motion recovery rates. One of these parameters is the Voice Handicap Index. There is no correlation between the handicap index and the recovery motion, as demonstrated in the study published by Young and colleagues in which, despite recovery of vocal fold motion, 16% of patients had significant voice handicap. In contrast, 21% of patients without motion recovery had normal Voice Handicap Index-10 scores.12 In this context, clinical neurophysiological testing would appear to be well suited in resolving some of these ambiguities. LEMG is able to identify normal innervation, absence of innervation, reinnervation, and even synkinesis by characteristic electrical signals. One important issue is the time when the EMG has to be done. Wang et al14 recommend 2 months, but the time classically established by neurologists for the EMG is 3 weeks. At 3 weeks, you can establish the type and the severity of the lesion, whereas at 2 months you can see the presence of the reinnervation, which actually is related with the prognosis. The objective of our study is to relate the degree of the lesion with the prognosis of vocal motion recovery, that is why we perform the LEMG at a minimum of 3 weeks after the injury. Statham and colleagues15 demonstrated that the presence of synkinesis in patients with VFP improved the NPV of LEMG from 53% to 100% and the sensitivity from 56% to 100%. The presence of synkinesis in our study was 30%, more frequent in the low-grade injury, and its presence is not associated with worse prognosis. This can be attributed to the fact that in our study, the prognosis was determined only by the grade of injury. Rickert and colleagues13 published a meta-analysis in 2012 of studies reporting LEMG results and clinical outcomes, which
ARTICLE IN PRESS 4 include 503 patients, reporting a PPV of 90.9% and an NPV of 55.6%. Similar results were published in 2014 by Wang and colleagues14 who performed a prospective study using LEMG to predict the long-term prognosis of UVFP including 85 patients, reporting a PPV of 93% and an NPV of 40%. In our study, the follow-up is longer, because it starts in the moment of the laryngoscopic diagnosis, instead of in the symptoms onset. Moreover, the accuracy of the recovery motion of the paralyzed vocal fold is supported by two different laryngologists. Our results are comparable with those published in the literature. They allow us to conclude that LEMG is a good predictor of no recovery of vocal fold motion, but it is not useful to predict recovery. The choice of treatment in patients with laryngeal paralysis should be influenced by the prognosis, the severity of symptoms, and the vocal demand of the patient. In those cases where the LEMG indicates a favorable prognosis, we can choose a conservative treatment or a temporary surgical procedure. On the other side, patients with severe symptoms or high professional vocal demands in which laryngeal EMG shows a bad prognosis can be treated with permanent surgical procedure without waiting too much.
CONCLUSIONS Laryngeal EMG is useful for the prediction of negative outcomes, but it is not so helpful in predicting positive outcomes. This fact can aid in deciding when to carry out a definitive treatment such as thyroplasty in cases of lack of compensation.
REFERENCES 1. Munin MC, Rosen CA, Zullo T. Utility of laryngeal electromyography in predicting recovery after vocal fold paralysis. Arch Phys Med Rehabil. 2003;84:1150–1153.
Journal of Voice, Vol. ■■, No. ■■, 2016 2. Wang C-C, Chang M-H, Wang C-P, et al. Prognostic indicators of unilateral vocal fold paralysis. Arch Otolaryngol Head Neck Surg. 2008;134:380–388. doi:10.1001/archotol.134.4.380. 3. Smith LJ, Rosen CA, Niyonkuru C, et al. Quantitative electromyography improves prediction in vocal fold paralysis. Laryngoscope. 2012;122:854– 859. doi:10.1002/lary.21884. 4. Heman-Ackah YD, Barr A. The value of laryngeal electromyography in the evaluation of laryngeal motion abnormalities. J Voice. 2006;20:452–460. doi:10.1016/j.jvoice.2005.04.009. 5. García-López I, Santiago-Pérez S, Peñarrocha-Teres J, et al. [Laryngeal electromyography in diagnosis and treatment of voice disorders]. Acta Otorrinolaringol Esp. 2012;63:458–464. doi:10.1016/j.otorri.2012.05.004. 6. Heman-Ackah YD, Mandel S, Manon-Espaillat R, et al. Laryngeal electromyography. Otolaryngol Clin North Am. 2007;40:1003–1023, vi-vii. doi:10.1016/j.otc.2007.05.007. 7. Volk GF, Hagen R, Pototschnig C, et al. Laryngeal electromyography: a proposal for guidelines of the European Laryngological Society. Eur Arch Otorhinolaryngol. 2012;269:2227–2245. doi:10.1007/s00405-012-2036-1. 8. Misono S, Merati AL. Evidence-based practice: evaluation and management of unilateral vocal fold paralysis. Otolaryngol Clin North Am. 2012;45:1083– 1108. doi:10.1016/j.otc.2012.06.011. 9. Sulica L, Blitzer A. Vocal Fold Paralysis. New York: Springer Berlin Heidelberg; 2006. 10. Rosenthal LHS, Benninger MS, Deeb RH. Vocal fold immobility: a longitudinal analysis of etiology over 20 years. Laryngoscope. 2007;117:1864–1870. doi:10.1097/MLG.0b013e3180de4d49. 11. Merati AL, Shemirani N, Smith TL, et al. Changing trends in the nature of vocal fold motion impairment. Am J Otolaryngol. 2006;27:106–108. doi:10.1016/j.amjoto.2005.07.020. 12. Young VN, Smith LJ, Rosen C. Voice outcome following acute unilateral vocal fold paralysis. Ann Otol Rhinol Laryngol. 2013;122:197–204. 13. Rickert SM, Childs LF, Carey BT, et al. Laryngeal electromyography for prognosis of vocal fold palsy: a meta-analysis. Laryngoscope. 2012;122:158– 161. doi:10.1002/lary.22354. 14. Wang C-C, Chang M-H, De Virgilio A, et al. Laryngeal electromyography and prognosis of unilateral vocal fold paralysis—a long-term prospective study. Laryngoscope. 2015;125:898–903. doi:10.1002/lary.24980. 15. Statham MM, Rosen CA, Smith LJ, et al. Electromyographic laryngeal synkinesis alters prognosis in vocal fold paralysis. Laryngoscope. 2010;120:285–290. doi:10.1002/lary.20629.