Eur Arch Otorhinolaryngol DOI 10.1007/s00405-014-3442-3
LARYNGOLOGY
Consideration of vocal fold position in unilateral vocal fold paralyses Arno Olthoff • Julia Steinle • Thomas Asendorf Eberhard Kruse
•
Received: 8 August 2014 / Accepted: 10 December 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract The objective of this study was to improve the evaluation of unilateral vocal fold paralyses (uVFP) by means of an area measurement of the glottic plane, which describes the position of the paralysed vocal fold. The area measurements were related to electromyographic findings and clinical outcome (recovery, voice quality). In 56 patients (33 women and 23 men), uVFP were confirmed by endolaryngeal electromyography (EMG) of the paralysed vocal fold and cricothyroid muscles (CT). The EMG response was classified on a 4-point scale (from 0 to 3). Vocal fold position was divided into ‘paramedian’ and ‘intermediate’ and additionally quantified by measurement of the glottic area. An ‘area quotient’ (AQ) was calculated and related to the EMG findings and clinical outcome. Voice qualities were objectified regarding their additive noise (breathiness) and irregularity (roughness) using the ‘Go¨ttingen Hoarseness Diagram’. The majority of uVFP was due to iatrogenic lesions. The AQ of classically graduated ‘paramedian’ and ‘intermediate’ vocal fold positions was significantly different but did not correlate with objective voice quality values. There were no significant correlations regarding EMG findings, duration or recovery from paralyses. Laryngeal EMG remains the gold standard for verifying uVFP. But EMG did not correlate significantly with AQ or functional outcome of uVFP. The
A. Olthoff (&) J. Steinle E. Kruse Division of Phoniatrics and Pedaudiology, Department of Otorhinolaryngology, University of Go¨ttingen, Robert-Koch-Str. 40, 37077 Go¨ttingen, Germany e-mail: [email protected] T. Asendorf Department of Medical Statistics, University of Go¨ttingen, Humboldtallee 32, 37073 Go¨ttingen, Germany
measurement of an AQ is suitable for obtaining continuous data describing the position of paralysed vocal folds beyond the terms ‘paramedian’ or ‘intermediate’ and provides the basis for clinical evaluations of diagnostic tools and therapeutic interventions. Keywords Vocal fold paralysis Electromyography Dysphonia Paramedian
Introduction There is still controversy surrounding the origin of paralysed vocal fold positions. In the 19th century, ‘Semon’s law’ postulated a vocal fold lateralisation during the course of laryngeal paralyses, based on the fact that predominant activity of adductors in early stages decreases during persistent paralyses [1]. This ‘law’ was replaced by the ‘Wagner–Grossmann theory’, which described the cricothyroid muscle (CT) as the sole adductor of paralysed vocal folds, because of its independent innervation (independent of the recurrent laryngeal nerve; RLN) via the external branch of the superior laryngeal nerve (extSLN) [2, 3]. Several authors rejected this ‘theory’ after experimental studies on animals and clinical studies failed to prove this causal connection [4, 5]. In our department, laryngeal electromyography (EMG) and objective voice analyses are part of routine diagnostics in the case of laryngeal paralyses. The aim of this study was to evaluate the influence of laryngeal vocal fold position and EMG activity on the clinical outcome in unilateral vocal fold paralyses (uVFP). To improve the ability to classify the vocal fold position beyond the terms ‘paramedian’ and ‘intermediate’, an area measurement of the glottic plane was introduced.
123
Eur Arch Otorhinolaryngol
Patients and methods All analyses were done after routine clinical diagnostics in outpatients who had given their consent. Patients with tumour-related paralyses (e.g., lung cancer) were not considered in this study. To detect and describe vocal fold mobility impairments, a rigid endoscope (90°) (Wolf, Knittlingen, Germany) was used for laryngoscopy and videostroboscopy. A camera system was connected to the endoscope and all films were stored to a hard disc recording system (Rheder&Partner, Hamburg, Germany). From these films, two pictures were taken for further analyses: one picture from the opened glottis during respiration and one picture during phonatory closure. All films of the 56 patients and the respective pictures were presented to two laryngologists (specialised MDs). In a blinded design (both raters were unaware of each others’ analyses, as well as of patient´s personal, anamnestic and clinical data) both had to classify the position of the paralysed vocal fold into ‘paramedian’ and ‘intermediate’ types. Using a graphic feature of the recording system (Rheder&Partner, Hamburg, Germany), they also had to mark the glottic midline in the laryngeal picture with maximal vocal fold abduction (during respiration) on the video screen. Now the area between this midline and the rim of vocal folds was determined bilaterally in number of pixels, using the above-mentioned graphic feature (Fig. 1). From the two areas, an area quotient (AQ) was calculated: AQ = Ap/Am (‘‘Ap’’ represents the glottic area of the paralysed and ‘‘Am’’ the glottic area of the mobile vocal fold). Thus AQ varied continuously between 0 and 1: A low AQ described a more medial and a high AQ a more lateral vocal fold (VF) position. To demonstrate that paralysis was the cause of vocal fold mobility impairments in all patients, an EMG of the impaired thyroarytenoid muscle (TA) was performed. Hooked wire electrodes were placed transorally using a special bowed applicator (Inomed, Teningen, Germany). An EMG of both CTs was also recorded using bipolar needle electrodes (Inomed, Teningen, Germany). EMG measures were done consecutively in one outpatient treatment. The EMG response was classified on a 4-point scale (0 to 3): ‘0’ represented silence, possible fibrillation, possible single muscle unit action potentials (MUAPs); ‘1’ represented severe and ‘2’ a moderate rarefied response; ‘3’ represented a finding of interfering MUAPs during phonation. The duration of paralysis (in months) was registered for each patient and the clinical outcome considering recovery of vocal fold mobility was evaluated after 12 or more months after onset of paralysis. Voice qualities were determined perceptually using the RBH scale, which classifies roughness (R), breathiness (B) and hoarseness (H, corresponding to G in the
123
Fig. 1 Vocal fold paralysis of the right side (small glottal area). The indirect endoscopic view into the larynx is shown. The glottal midline and both glottal areas (left and right) are marked and quantified in pixels. The area quotient (AQ) was defined as AQ = Ap/Am (‘Ap’ represents the glottic area of the paralysed and ‘Am’ the glottic area of the mobile vocal fold)
international GRBAS scale) on a scale from ‘0’ to ‘3’ (‘0’ is normal, ‘3’ severely impaired) [6]. An additional objective computerised measurement was performed using the Go¨ttingen hoarseness diagram (GHD), which allows independent measures of roughness (signal irregularity) and breathiness (additive noise) [7]. Statistics Inter-rater reliability in determining the VF position (paramedian, intermediate) was assessed using Cohens Kappa. Correlation between AQ values from rater 1 and rater 2 was estimated using Pearson’s correlation coefficient and tested to be significantly unequal to 0. Differences between EMG findings between left and right side of paralysis were tested using the Wilcoxon signed-rank test. In the case of low inter-rater reliability, differences in the AQ according to the classified VF position were tested using the Wilcoxon rank-sum test for both rater 1 and rater 2 separately. Correlation between AQ measurements and voice quality values was tested for using Spearman’s correlation coefficient. All correlations were tested separately for rater 1 and rater 2. The analyses performed were conducted using the statistics software R version 3.1.1 (2014-07-10).
Results Demographic and clinical data Unilateral VFPs were diagnosed in 56 patients (33 women and 23 men). Their mean age was 52 years, with a range
Eur Arch Otorhinolaryngol
from 16 to 79 years. In 24 cases the uVFP followed thyroid, in 3 cases endothoracal, in 2 cases carotid and in 2 cases neck bone surgery. In 3 cases it was caused by neurosurgical brainstem treatments and in 3 cases intubations were the only interventions in the region of the neck and chest. Three patients were suffering from a polytrauma of the neck and chest and 1 patient from a brainstem insult. No potential origin was found in 15 (27 %) patients and they were classified as ‘idiopathic paralyses’ (Table 1). Thyroid surgery was the main cause of paralyses in women (61 %), but not in men (17 %). Idiopathic lesions were found more often in males (39 %) than in females (18 %). Considering all surgical treatments (thyroid, endothoracal, carotid, neck bone, brainstem surgery and intubation), we saw 37 iatrogenic lesions (66 %). The mean latency between onset of paralyses and EMG measures was 30 months and ranged between 7 days and 312 months. Clinical outcome regarding recovery of vocal fold mobility could be evaluated of 32 (57 %) patients after more than 12 months. No recovery was observed in 24/32 (75 %) patients. Recovery within the first 3 months was seen in 1/32 (3 %) patients, between 3 and 6 months in 5/32 (16 %) patients, between 6 and 9 months in 2/32 (6 %) patients, and between 9 and 12 months in no patients. Males and females did not differ significantly in AQ measurements of both raters, rater 1 (p = 0.205) and rater 2 (p = 0.4635).
Kappa = 0.6316, 95 % CI (0.3897, 0.8118)], differences in the AQ according to the classified VF position were tested using the Wilcoxon rank-sum test separately for both raters. The distributions of the AQ differed significantly for vocal folds with either ‘paramedian’ or ‘intermediate’ position for both raters, rater 1 (p = 0.0193) and rater 2 (p \ 0.0001). In ‘paramedian’ positions, we saw low and in ‘intermediate’ positions high AQ values (Fig. 2). AQ measurements of rater 1 and rater 2 were significantly correlated (r = 0.52, p \ 0.0001) and VF classifications were shown to be similar. There was no significant correlation between AQ and duration of paralyses. Functional data AQ only had an impact on perceptual voice quality values. Lower AQ (more medial position of the paralysed vocal fold) correlated with superior voice qualities. This result was significant considering ‘Hoarseness’ values from rater 1 (r = 0.44, p = 0.001) and showed a tendency taking rater 2 into account separately (r = 0.26, p = 0.06) (Table 2). Regarding objective measures obtained from the GHD, there was no significant correlation with AQ even though the correlation of breathiness measured with the GHD and RBH scale was significant (r = 0.56, p \ 0.001)
Endoscopic data Due to the low inter-rater reliability in determining the VF position (‘paramedian’, ‘intermediate’) [Cohens Kappa: Table 1 Demographic and clinical data Variable
Females (n = 33)
Males (n = 23)
Total (n = 56)
Age (years) Mean ± SD
54.0 ± 12.4
48.4 ± 16.9
51.7 ± 14.6
Range
25–79
16–75
16–79
Origin of paralysis, number (%) Thyroid surgery
20 (61 %)
4 (17 %)
24 (43 %)
Endothoracal surgery
0 (0 %)
3 (13 %)
3 (5 %)
Carotid surgery
2 (6 %)
0 (0 %)
2 (4 %)
Neck bone surgery
0 (0 %)
2 (9 %)
2 (4 %)
Brainstem surgery
3 (9 %)
0 (0 %)
3 (5 %)
Intubation
1 (3 %)
2 (9 %)
3 (5 %)
Polytrauma
1 (3 %)
2 (9 %)
3 (5 %)
Brainstem insult
0 (0 %)
1 (4 %)
1 (2 %)
unknown
6 (18 %)
9 (39 %)
15 (27 %)
Fig. 2 Area quotient (AQ) values from rater 1 and rater 2 and the corresponding assignment to ‘paramedian’ or ‘intermediate’ VF position. If VF position assignment of rater 1 and rater 2 differs, the VF position is referred to as ‘undecided’. Results of all patients (n = 56) are shown. The AQ describes the vocal fold position. Higher values represent a more lateral vocal fold position. The figure illustrates that higher AQ values tend to be classified as ‘intermediate’, while lower AQ values tend to be classified as ‘paramedian’
123
Eur Arch Otorhinolaryngol Table 2 Correlation between AQ measurements and voice quality values, separately for rater 1 and rater 2 Method
Variable
Rater 1
Rater 2
GHD
Irregularity
r = 0.1496, p = 0.3910
r = 0.09015, p = 0.6065
Noise component
r = 0.2460, p = 0.1544
r = 0.2770, p = 0.1072
Roughness
r = 0.2893, p = 0.03946
r = 0.1633, p = 0.2522
Breathiness
r = 0.3743, p = 0.0068
r = 0.2646, p = 0.0606
Hoarseness
r = 0.4423, p = 0.001
r = 0.2641, p = 0.0611
RBH scale
Bold values indicate statistical significance p \ 0.05
and the correlation of roughness showed a tendency (r = 0.32, p = 0.06). Electromyographic data Paralyses were confirmed by EMG of the affected TA muscle. There were no significant differences in EMG findings between left and right side paralyses (Wilcoxon signed-rank test p = 0.8537). We saw no significant correlation of CT-EMG values with AQ, not even if the synergistic TA-EMG activity was taken into account. Regarding recovery of paralyses, no significant correlation with EMG activities (CT or TA) could be found. A comparison of iatrogenic versus idiopathic lesions in terms of paralysis recovery failed due to small sample sizes.
Discussion Demographic and clinical data The fact that only 32/56 (57 %) cases could be followed up after 12 months might be the reason for the high percentage of permanent paralyses (75 %). In particular patients who had idiopathic lesions missed the follow-up diagnostics at our clinic. Many of them were free of complaints and not willing to undergo a long journey for a follow-up investigation. The majority of recoveries were evident during the first 6 months. Later recoveries were observed in some cases up to 10 months. This result is consistent with findings of other studies and supports the recommendation of permanent phonosurgical interventions after this period [8]. The high percentage of iatrogenic lesions (66 %) emphasises the importance of pre- and postoperative endoscopic evaluations of VF mobilities to evaluate the risk of various surgical treatments (thyroid, neck bone,
123
carotid, thoracic, cervical, brainstem). The majority of iatrogenic lesions are the direct consequence of thyroid surgery (24/37, 65 %). Because thyroid surgery is performed more often in women, they are affected disproportionately [9]. Endoscopic data To achieve continuous data about positions of paralysed vocal folds, we decided to introduce the AQ. Our computerised videostroboscopic system includes a function for measuring areas in the number of pixels (Rheder&Partner, Hamburg, Germany). Other authors described additional measurements of the anterior commissure´s angle for the same purpose [10]. Both methods allow the evaluation of vocal fold positions without being dependent on absolute measurements of laryngeal dimensions. Because the AQ represents a ratio of left and right glottic areas individual and gender-dependent impacts (morphologies) were neutralised. In our study, the AQ values differed significantly between the traditional classification into ‘paramedian’ and ‘intermediate’ for rater 1 (p = 0.01925) and rater 2 (p \ 0.0001). Correlation analyses of AQ and duration of uVFP could not confirm the (historical) observation of Semon, who described the change from ‘paramedian’ to ‘intermediate’ vocal fold positions over the course of time in permanent VFP [1]. The low inter-rater reliability illustrates the doubtful relevance of classified vocal fold positions in unilateral paralyses. The AQ measurements as well as the dichotomous classification into ‘paramedian’ and ‘intermediate’ depend on the identification of midlines in paralysed larynges. Our study revealed an uncertainty in identifying midlines in paralysed larynges. Alternative and more useful clinical criteria might be: Tension of the paralysed VF, ability to adduct the paralysed VF, videostroboscopic findings regarding glottis closure, mucosal wave and amplitude. This assumption should be tested in further studies. Functional data A more medial vocal fold position had a positive impact on voice qualities evaluated perceptually by the researchers. This result could not be confirmed in the objective measures of roughness and breathiness. The perceptual evaluation of voice might be influenced by the impression of the previously classified videostroboscopic findings. This finding underlines the need for objective measures of voice qualities. The AQ did not correlate with objectified functional deficits (voice quality) or clinical outcome (prognosis, recovery) or site and cause of lesion evaluated by
Eur Arch Otorhinolaryngol
anamneses, endoscopy and EMG. This raised the question whether classifications of paralysed VF positions are useful. Even for indications of phonosurgical treatments after a period of 10 months, videostroboscopic findings might contain more valuable information than documentations of VF positions. This aspect was not resolved by the present study. Electromyographic data Regarding the findings of the CT-EMG and the vocal fold position, we could not detect a significant effect of CT activity on the AQ. This might be interpreted as a contradiction of the ‘Wagner–Grossmann theory’, which stated a more medial vocal fold position accompanied by high CT activity [2, 3]. Kirchner ascribed this aspect of ‘flaccid’ paralyses of vocal folds in an intermediate position (Kirchner: ‘cadaveric position’) to ‘vagal paralyses’. He assumed an inhibited, non-paralysed CT caused by the interruption of afferent vagal stimuli. At that time he regretted the lack of CT-EMG diagnostics [11, 12]. A similar idea of an ‘Inaktivita¨tsatrophie’ (translation: atrophy caused by inactivity) was first published by Grossmann [13]. In our study, we saw increased CT activities as well as decreased CT activities without significant effect on vocal fold position (AQ). Increased CT activity might be compensatory to adduct the vocal fold. This ‘‘medializing’’ effect of the CT should work in non-paralysed vocal folds that have a normal muscle tension. If paralysed vocal folds are bowed and fluffy, even increased CT activities might not be able to cause a medialisation of the paralysed vocal fold. Additionally, there was no significant impact of EMG activities either on clinical outcome regarding recovery of vocal fold mobility or on voice quality. The group of patients who were diagnosed early (e.g., within the first 3 months) was unfortunately too small to enable sound analyses. Later diagnostics are influenced by the ‘synkinesis’ (Crumley) or ‘autoparalytic syndrome’ (Stennert) that appears in permanent paralyses beyond the 6th month [14, 15]. In our study, EMG did not provide useful prognostic data in uVFP even under consideration of continuous data for paralysed vocal fold positions. This observation is consistent with previous data of earlier investigations [16, 17]. The fact that we found iatrogenic paralyses in 66 % of cases emphasises the relevance of the laryngeal EMG for medicolegal reasons. In 27 % of cases, we saw idiopathic paralyses and used the EMG within the scope of detailed diagnostics. In the case of posttraumatic lesions (7 %), laryngeal EMGs were often required for medical reports. In all cases the laryngeal EMG was a valuable diagnostic tool to prove laryngeal paralyses. It should be performed early, within 3 months after the onset of paralyses.
Neither CT- or TA-EMG activity nor duration of uVFP had a significant influence on the position of the paralysed vocal fold (AQ) or on voice quality. EMG remains the gold standard diagnostic tool for demonstrating and classifying vocal fold paralyses. However, EMG results are not suitable for predicting their clinical outcome [18]. The classification of paralysed vocal folds into ‘paramedian’ and ‘intermediate’ positions may not be helpful. More diagnostically conclusive criteria like videostroboscopic data should be preferred instead. The AQ provides sound continuous data that can be used to resolve this issue in further studies. Conflict of interest flicts of interest.
All the authors declare that there are no con-
References 1. Clerf LH, Baltzell WH (1953) Re-evaluation of Semon’s hypothesis. Laryngoscope 63(8):693–699 2. Grossmann M (1897) Experimentelle Beitra¨ge zur Lehre von der ,,Posticusla¨hmung‘‘. Arch LaryngRhinol 6:282–360 3. Dedo HH (1970) The paralyzed larynx: an electromyographic study in dogs and humans. Laryngoscope 80(10):1455–1517 4. Woodson GE (1993) Configuration of the glottis in laryngeal paralysis. I: Clinical study. Laryngoscope 103(11):1227–1234 5. Koufman JA, Walker FO, Joharji GM (1995) The cricothyroid muscle does not influence vocal fold position in laryngeal paralysis. Laryngoscope 105(4):368–372 6. Nawka T, Anders LC, Wendler J (1994) Die auditive Beurteilung heiserer Stimmen nach dem RBH-system. Sprache-Stimme-Geho¨r 18:130–133 7. Fro¨hlich M, Michaelis D, Strube HW, Kruse E (2000) Acoustic voice analysis by means of the hoarseness diagram. J Speech Lang Hear Res 43:706–720 8. Friedman AD, Burns JA, Heaton JT, Zeitels SM (2010) Early versus late injection medialization for unilateral vocal cord paralysis. Laryngoscope 120:2042–2046 9. Sanabria A, Silver CE, Sua´rez C et al (2013) Neuromonitoring of the laryngeal nerves in thyroid surgery: a critical appraisal of the literature. Eur Arch Otorhinolaryngol 270:2383–2395 10. Inagi K, Khidr AA, Ford CN, Bless DM, Heisey DM (1997) Correlation between vocal functions and glottal measurements in patients with unilateral vocal fold paralysis. Laryngoscope 107(6):782–791 11. Kirchner JA (1966) Atrophy of laryngeal muscles in vagal paralysis. Laryngoscope 76(11):1753–1765 12. Kirchner JA (1977) Intrathoracic injury to the motor nerve supply of the larynx. Acta Otolaryngol 83(1–2):163–169 13. Grossmann M (1906) Beitrag zur Lehre von der wechselseitigen funktionellen Beziehung der Kehlkopfmuskeln untereinander. Arch Laryng Rhinol 18:463–474 14. Crumley RL (1989) Laryngeal synkinesis: its significance to the laryngologist. Ann Otol Rhinol Laryngol 98(2):87–92 15. Stennert E (1979) I. Hypoglossal facial anastomosis: its significance for modern facial surgery. II. Combined approach in extratemporal facial nerve reconstruction. Clin Plast Surg 6(3): 471–486 16. Laskawi R, Drobik C, Baaske C (1994) Prognostic value of electrodiagnosis of Bell’s palsy. Laryngorhinootologie 73:338–341
123
Eur Arch Otorhinolaryngol 17. Sataloff RT, Mandel S, Mann EA, Ludlow CL (2004) Practice parameter: laryngeal electromyography (an evidence-based review). Otolaryngol Head Neck Surg 130:770–779
123
18. Blitzer A, Crumley RL, Dailey SH et al (2009) Recommendations of the Neurolaryngology Study Group on laryngeal electromyography. Otolaryngol Head Neck Surg 140:782–793
LARYNGOLOGY
Consideration of vocal fold position in unilateral vocal fold paralyses Arno Olthoff • Julia Steinle • Thomas Asendorf Eberhard Kruse
•
Received: 8 August 2014 / Accepted: 10 December 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract The objective of this study was to improve the evaluation of unilateral vocal fold paralyses (uVFP) by means of an area measurement of the glottic plane, which describes the position of the paralysed vocal fold. The area measurements were related to electromyographic findings and clinical outcome (recovery, voice quality). In 56 patients (33 women and 23 men), uVFP were confirmed by endolaryngeal electromyography (EMG) of the paralysed vocal fold and cricothyroid muscles (CT). The EMG response was classified on a 4-point scale (from 0 to 3). Vocal fold position was divided into ‘paramedian’ and ‘intermediate’ and additionally quantified by measurement of the glottic area. An ‘area quotient’ (AQ) was calculated and related to the EMG findings and clinical outcome. Voice qualities were objectified regarding their additive noise (breathiness) and irregularity (roughness) using the ‘Go¨ttingen Hoarseness Diagram’. The majority of uVFP was due to iatrogenic lesions. The AQ of classically graduated ‘paramedian’ and ‘intermediate’ vocal fold positions was significantly different but did not correlate with objective voice quality values. There were no significant correlations regarding EMG findings, duration or recovery from paralyses. Laryngeal EMG remains the gold standard for verifying uVFP. But EMG did not correlate significantly with AQ or functional outcome of uVFP. The
A. Olthoff (&) J. Steinle E. Kruse Division of Phoniatrics and Pedaudiology, Department of Otorhinolaryngology, University of Go¨ttingen, Robert-Koch-Str. 40, 37077 Go¨ttingen, Germany e-mail: [email protected] T. Asendorf Department of Medical Statistics, University of Go¨ttingen, Humboldtallee 32, 37073 Go¨ttingen, Germany
measurement of an AQ is suitable for obtaining continuous data describing the position of paralysed vocal folds beyond the terms ‘paramedian’ or ‘intermediate’ and provides the basis for clinical evaluations of diagnostic tools and therapeutic interventions. Keywords Vocal fold paralysis Electromyography Dysphonia Paramedian
Introduction There is still controversy surrounding the origin of paralysed vocal fold positions. In the 19th century, ‘Semon’s law’ postulated a vocal fold lateralisation during the course of laryngeal paralyses, based on the fact that predominant activity of adductors in early stages decreases during persistent paralyses [1]. This ‘law’ was replaced by the ‘Wagner–Grossmann theory’, which described the cricothyroid muscle (CT) as the sole adductor of paralysed vocal folds, because of its independent innervation (independent of the recurrent laryngeal nerve; RLN) via the external branch of the superior laryngeal nerve (extSLN) [2, 3]. Several authors rejected this ‘theory’ after experimental studies on animals and clinical studies failed to prove this causal connection [4, 5]. In our department, laryngeal electromyography (EMG) and objective voice analyses are part of routine diagnostics in the case of laryngeal paralyses. The aim of this study was to evaluate the influence of laryngeal vocal fold position and EMG activity on the clinical outcome in unilateral vocal fold paralyses (uVFP). To improve the ability to classify the vocal fold position beyond the terms ‘paramedian’ and ‘intermediate’, an area measurement of the glottic plane was introduced.
123
Eur Arch Otorhinolaryngol
Patients and methods All analyses were done after routine clinical diagnostics in outpatients who had given their consent. Patients with tumour-related paralyses (e.g., lung cancer) were not considered in this study. To detect and describe vocal fold mobility impairments, a rigid endoscope (90°) (Wolf, Knittlingen, Germany) was used for laryngoscopy and videostroboscopy. A camera system was connected to the endoscope and all films were stored to a hard disc recording system (Rheder&Partner, Hamburg, Germany). From these films, two pictures were taken for further analyses: one picture from the opened glottis during respiration and one picture during phonatory closure. All films of the 56 patients and the respective pictures were presented to two laryngologists (specialised MDs). In a blinded design (both raters were unaware of each others’ analyses, as well as of patient´s personal, anamnestic and clinical data) both had to classify the position of the paralysed vocal fold into ‘paramedian’ and ‘intermediate’ types. Using a graphic feature of the recording system (Rheder&Partner, Hamburg, Germany), they also had to mark the glottic midline in the laryngeal picture with maximal vocal fold abduction (during respiration) on the video screen. Now the area between this midline and the rim of vocal folds was determined bilaterally in number of pixels, using the above-mentioned graphic feature (Fig. 1). From the two areas, an area quotient (AQ) was calculated: AQ = Ap/Am (‘‘Ap’’ represents the glottic area of the paralysed and ‘‘Am’’ the glottic area of the mobile vocal fold). Thus AQ varied continuously between 0 and 1: A low AQ described a more medial and a high AQ a more lateral vocal fold (VF) position. To demonstrate that paralysis was the cause of vocal fold mobility impairments in all patients, an EMG of the impaired thyroarytenoid muscle (TA) was performed. Hooked wire electrodes were placed transorally using a special bowed applicator (Inomed, Teningen, Germany). An EMG of both CTs was also recorded using bipolar needle electrodes (Inomed, Teningen, Germany). EMG measures were done consecutively in one outpatient treatment. The EMG response was classified on a 4-point scale (0 to 3): ‘0’ represented silence, possible fibrillation, possible single muscle unit action potentials (MUAPs); ‘1’ represented severe and ‘2’ a moderate rarefied response; ‘3’ represented a finding of interfering MUAPs during phonation. The duration of paralysis (in months) was registered for each patient and the clinical outcome considering recovery of vocal fold mobility was evaluated after 12 or more months after onset of paralysis. Voice qualities were determined perceptually using the RBH scale, which classifies roughness (R), breathiness (B) and hoarseness (H, corresponding to G in the
123
Fig. 1 Vocal fold paralysis of the right side (small glottal area). The indirect endoscopic view into the larynx is shown. The glottal midline and both glottal areas (left and right) are marked and quantified in pixels. The area quotient (AQ) was defined as AQ = Ap/Am (‘Ap’ represents the glottic area of the paralysed and ‘Am’ the glottic area of the mobile vocal fold)
international GRBAS scale) on a scale from ‘0’ to ‘3’ (‘0’ is normal, ‘3’ severely impaired) [6]. An additional objective computerised measurement was performed using the Go¨ttingen hoarseness diagram (GHD), which allows independent measures of roughness (signal irregularity) and breathiness (additive noise) [7]. Statistics Inter-rater reliability in determining the VF position (paramedian, intermediate) was assessed using Cohens Kappa. Correlation between AQ values from rater 1 and rater 2 was estimated using Pearson’s correlation coefficient and tested to be significantly unequal to 0. Differences between EMG findings between left and right side of paralysis were tested using the Wilcoxon signed-rank test. In the case of low inter-rater reliability, differences in the AQ according to the classified VF position were tested using the Wilcoxon rank-sum test for both rater 1 and rater 2 separately. Correlation between AQ measurements and voice quality values was tested for using Spearman’s correlation coefficient. All correlations were tested separately for rater 1 and rater 2. The analyses performed were conducted using the statistics software R version 3.1.1 (2014-07-10).
Results Demographic and clinical data Unilateral VFPs were diagnosed in 56 patients (33 women and 23 men). Their mean age was 52 years, with a range
Eur Arch Otorhinolaryngol
from 16 to 79 years. In 24 cases the uVFP followed thyroid, in 3 cases endothoracal, in 2 cases carotid and in 2 cases neck bone surgery. In 3 cases it was caused by neurosurgical brainstem treatments and in 3 cases intubations were the only interventions in the region of the neck and chest. Three patients were suffering from a polytrauma of the neck and chest and 1 patient from a brainstem insult. No potential origin was found in 15 (27 %) patients and they were classified as ‘idiopathic paralyses’ (Table 1). Thyroid surgery was the main cause of paralyses in women (61 %), but not in men (17 %). Idiopathic lesions were found more often in males (39 %) than in females (18 %). Considering all surgical treatments (thyroid, endothoracal, carotid, neck bone, brainstem surgery and intubation), we saw 37 iatrogenic lesions (66 %). The mean latency between onset of paralyses and EMG measures was 30 months and ranged between 7 days and 312 months. Clinical outcome regarding recovery of vocal fold mobility could be evaluated of 32 (57 %) patients after more than 12 months. No recovery was observed in 24/32 (75 %) patients. Recovery within the first 3 months was seen in 1/32 (3 %) patients, between 3 and 6 months in 5/32 (16 %) patients, between 6 and 9 months in 2/32 (6 %) patients, and between 9 and 12 months in no patients. Males and females did not differ significantly in AQ measurements of both raters, rater 1 (p = 0.205) and rater 2 (p = 0.4635).
Kappa = 0.6316, 95 % CI (0.3897, 0.8118)], differences in the AQ according to the classified VF position were tested using the Wilcoxon rank-sum test separately for both raters. The distributions of the AQ differed significantly for vocal folds with either ‘paramedian’ or ‘intermediate’ position for both raters, rater 1 (p = 0.0193) and rater 2 (p \ 0.0001). In ‘paramedian’ positions, we saw low and in ‘intermediate’ positions high AQ values (Fig. 2). AQ measurements of rater 1 and rater 2 were significantly correlated (r = 0.52, p \ 0.0001) and VF classifications were shown to be similar. There was no significant correlation between AQ and duration of paralyses. Functional data AQ only had an impact on perceptual voice quality values. Lower AQ (more medial position of the paralysed vocal fold) correlated with superior voice qualities. This result was significant considering ‘Hoarseness’ values from rater 1 (r = 0.44, p = 0.001) and showed a tendency taking rater 2 into account separately (r = 0.26, p = 0.06) (Table 2). Regarding objective measures obtained from the GHD, there was no significant correlation with AQ even though the correlation of breathiness measured with the GHD and RBH scale was significant (r = 0.56, p \ 0.001)
Endoscopic data Due to the low inter-rater reliability in determining the VF position (‘paramedian’, ‘intermediate’) [Cohens Kappa: Table 1 Demographic and clinical data Variable
Females (n = 33)
Males (n = 23)
Total (n = 56)
Age (years) Mean ± SD
54.0 ± 12.4
48.4 ± 16.9
51.7 ± 14.6
Range
25–79
16–75
16–79
Origin of paralysis, number (%) Thyroid surgery
20 (61 %)
4 (17 %)
24 (43 %)
Endothoracal surgery
0 (0 %)
3 (13 %)
3 (5 %)
Carotid surgery
2 (6 %)
0 (0 %)
2 (4 %)
Neck bone surgery
0 (0 %)
2 (9 %)
2 (4 %)
Brainstem surgery
3 (9 %)
0 (0 %)
3 (5 %)
Intubation
1 (3 %)
2 (9 %)
3 (5 %)
Polytrauma
1 (3 %)
2 (9 %)
3 (5 %)
Brainstem insult
0 (0 %)
1 (4 %)
1 (2 %)
unknown
6 (18 %)
9 (39 %)
15 (27 %)
Fig. 2 Area quotient (AQ) values from rater 1 and rater 2 and the corresponding assignment to ‘paramedian’ or ‘intermediate’ VF position. If VF position assignment of rater 1 and rater 2 differs, the VF position is referred to as ‘undecided’. Results of all patients (n = 56) are shown. The AQ describes the vocal fold position. Higher values represent a more lateral vocal fold position. The figure illustrates that higher AQ values tend to be classified as ‘intermediate’, while lower AQ values tend to be classified as ‘paramedian’
123
Eur Arch Otorhinolaryngol Table 2 Correlation between AQ measurements and voice quality values, separately for rater 1 and rater 2 Method
Variable
Rater 1
Rater 2
GHD
Irregularity
r = 0.1496, p = 0.3910
r = 0.09015, p = 0.6065
Noise component
r = 0.2460, p = 0.1544
r = 0.2770, p = 0.1072
Roughness
r = 0.2893, p = 0.03946
r = 0.1633, p = 0.2522
Breathiness
r = 0.3743, p = 0.0068
r = 0.2646, p = 0.0606
Hoarseness
r = 0.4423, p = 0.001
r = 0.2641, p = 0.0611
RBH scale
Bold values indicate statistical significance p \ 0.05
and the correlation of roughness showed a tendency (r = 0.32, p = 0.06). Electromyographic data Paralyses were confirmed by EMG of the affected TA muscle. There were no significant differences in EMG findings between left and right side paralyses (Wilcoxon signed-rank test p = 0.8537). We saw no significant correlation of CT-EMG values with AQ, not even if the synergistic TA-EMG activity was taken into account. Regarding recovery of paralyses, no significant correlation with EMG activities (CT or TA) could be found. A comparison of iatrogenic versus idiopathic lesions in terms of paralysis recovery failed due to small sample sizes.
Discussion Demographic and clinical data The fact that only 32/56 (57 %) cases could be followed up after 12 months might be the reason for the high percentage of permanent paralyses (75 %). In particular patients who had idiopathic lesions missed the follow-up diagnostics at our clinic. Many of them were free of complaints and not willing to undergo a long journey for a follow-up investigation. The majority of recoveries were evident during the first 6 months. Later recoveries were observed in some cases up to 10 months. This result is consistent with findings of other studies and supports the recommendation of permanent phonosurgical interventions after this period [8]. The high percentage of iatrogenic lesions (66 %) emphasises the importance of pre- and postoperative endoscopic evaluations of VF mobilities to evaluate the risk of various surgical treatments (thyroid, neck bone,
123
carotid, thoracic, cervical, brainstem). The majority of iatrogenic lesions are the direct consequence of thyroid surgery (24/37, 65 %). Because thyroid surgery is performed more often in women, they are affected disproportionately [9]. Endoscopic data To achieve continuous data about positions of paralysed vocal folds, we decided to introduce the AQ. Our computerised videostroboscopic system includes a function for measuring areas in the number of pixels (Rheder&Partner, Hamburg, Germany). Other authors described additional measurements of the anterior commissure´s angle for the same purpose [10]. Both methods allow the evaluation of vocal fold positions without being dependent on absolute measurements of laryngeal dimensions. Because the AQ represents a ratio of left and right glottic areas individual and gender-dependent impacts (morphologies) were neutralised. In our study, the AQ values differed significantly between the traditional classification into ‘paramedian’ and ‘intermediate’ for rater 1 (p = 0.01925) and rater 2 (p \ 0.0001). Correlation analyses of AQ and duration of uVFP could not confirm the (historical) observation of Semon, who described the change from ‘paramedian’ to ‘intermediate’ vocal fold positions over the course of time in permanent VFP [1]. The low inter-rater reliability illustrates the doubtful relevance of classified vocal fold positions in unilateral paralyses. The AQ measurements as well as the dichotomous classification into ‘paramedian’ and ‘intermediate’ depend on the identification of midlines in paralysed larynges. Our study revealed an uncertainty in identifying midlines in paralysed larynges. Alternative and more useful clinical criteria might be: Tension of the paralysed VF, ability to adduct the paralysed VF, videostroboscopic findings regarding glottis closure, mucosal wave and amplitude. This assumption should be tested in further studies. Functional data A more medial vocal fold position had a positive impact on voice qualities evaluated perceptually by the researchers. This result could not be confirmed in the objective measures of roughness and breathiness. The perceptual evaluation of voice might be influenced by the impression of the previously classified videostroboscopic findings. This finding underlines the need for objective measures of voice qualities. The AQ did not correlate with objectified functional deficits (voice quality) or clinical outcome (prognosis, recovery) or site and cause of lesion evaluated by
Eur Arch Otorhinolaryngol
anamneses, endoscopy and EMG. This raised the question whether classifications of paralysed VF positions are useful. Even for indications of phonosurgical treatments after a period of 10 months, videostroboscopic findings might contain more valuable information than documentations of VF positions. This aspect was not resolved by the present study. Electromyographic data Regarding the findings of the CT-EMG and the vocal fold position, we could not detect a significant effect of CT activity on the AQ. This might be interpreted as a contradiction of the ‘Wagner–Grossmann theory’, which stated a more medial vocal fold position accompanied by high CT activity [2, 3]. Kirchner ascribed this aspect of ‘flaccid’ paralyses of vocal folds in an intermediate position (Kirchner: ‘cadaveric position’) to ‘vagal paralyses’. He assumed an inhibited, non-paralysed CT caused by the interruption of afferent vagal stimuli. At that time he regretted the lack of CT-EMG diagnostics [11, 12]. A similar idea of an ‘Inaktivita¨tsatrophie’ (translation: atrophy caused by inactivity) was first published by Grossmann [13]. In our study, we saw increased CT activities as well as decreased CT activities without significant effect on vocal fold position (AQ). Increased CT activity might be compensatory to adduct the vocal fold. This ‘‘medializing’’ effect of the CT should work in non-paralysed vocal folds that have a normal muscle tension. If paralysed vocal folds are bowed and fluffy, even increased CT activities might not be able to cause a medialisation of the paralysed vocal fold. Additionally, there was no significant impact of EMG activities either on clinical outcome regarding recovery of vocal fold mobility or on voice quality. The group of patients who were diagnosed early (e.g., within the first 3 months) was unfortunately too small to enable sound analyses. Later diagnostics are influenced by the ‘synkinesis’ (Crumley) or ‘autoparalytic syndrome’ (Stennert) that appears in permanent paralyses beyond the 6th month [14, 15]. In our study, EMG did not provide useful prognostic data in uVFP even under consideration of continuous data for paralysed vocal fold positions. This observation is consistent with previous data of earlier investigations [16, 17]. The fact that we found iatrogenic paralyses in 66 % of cases emphasises the relevance of the laryngeal EMG for medicolegal reasons. In 27 % of cases, we saw idiopathic paralyses and used the EMG within the scope of detailed diagnostics. In the case of posttraumatic lesions (7 %), laryngeal EMGs were often required for medical reports. In all cases the laryngeal EMG was a valuable diagnostic tool to prove laryngeal paralyses. It should be performed early, within 3 months after the onset of paralyses.
Neither CT- or TA-EMG activity nor duration of uVFP had a significant influence on the position of the paralysed vocal fold (AQ) or on voice quality. EMG remains the gold standard diagnostic tool for demonstrating and classifying vocal fold paralyses. However, EMG results are not suitable for predicting their clinical outcome [18]. The classification of paralysed vocal folds into ‘paramedian’ and ‘intermediate’ positions may not be helpful. More diagnostically conclusive criteria like videostroboscopic data should be preferred instead. The AQ provides sound continuous data that can be used to resolve this issue in further studies. Conflict of interest flicts of interest.
All the authors declare that there are no con-
References 1. Clerf LH, Baltzell WH (1953) Re-evaluation of Semon’s hypothesis. Laryngoscope 63(8):693–699 2. Grossmann M (1897) Experimentelle Beitra¨ge zur Lehre von der ,,Posticusla¨hmung‘‘. Arch LaryngRhinol 6:282–360 3. Dedo HH (1970) The paralyzed larynx: an electromyographic study in dogs and humans. Laryngoscope 80(10):1455–1517 4. Woodson GE (1993) Configuration of the glottis in laryngeal paralysis. I: Clinical study. Laryngoscope 103(11):1227–1234 5. Koufman JA, Walker FO, Joharji GM (1995) The cricothyroid muscle does not influence vocal fold position in laryngeal paralysis. Laryngoscope 105(4):368–372 6. Nawka T, Anders LC, Wendler J (1994) Die auditive Beurteilung heiserer Stimmen nach dem RBH-system. Sprache-Stimme-Geho¨r 18:130–133 7. Fro¨hlich M, Michaelis D, Strube HW, Kruse E (2000) Acoustic voice analysis by means of the hoarseness diagram. J Speech Lang Hear Res 43:706–720 8. Friedman AD, Burns JA, Heaton JT, Zeitels SM (2010) Early versus late injection medialization for unilateral vocal cord paralysis. Laryngoscope 120:2042–2046 9. Sanabria A, Silver CE, Sua´rez C et al (2013) Neuromonitoring of the laryngeal nerves in thyroid surgery: a critical appraisal of the literature. Eur Arch Otorhinolaryngol 270:2383–2395 10. Inagi K, Khidr AA, Ford CN, Bless DM, Heisey DM (1997) Correlation between vocal functions and glottal measurements in patients with unilateral vocal fold paralysis. Laryngoscope 107(6):782–791 11. Kirchner JA (1966) Atrophy of laryngeal muscles in vagal paralysis. Laryngoscope 76(11):1753–1765 12. Kirchner JA (1977) Intrathoracic injury to the motor nerve supply of the larynx. Acta Otolaryngol 83(1–2):163–169 13. Grossmann M (1906) Beitrag zur Lehre von der wechselseitigen funktionellen Beziehung der Kehlkopfmuskeln untereinander. Arch Laryng Rhinol 18:463–474 14. Crumley RL (1989) Laryngeal synkinesis: its significance to the laryngologist. Ann Otol Rhinol Laryngol 98(2):87–92 15. Stennert E (1979) I. Hypoglossal facial anastomosis: its significance for modern facial surgery. II. Combined approach in extratemporal facial nerve reconstruction. Clin Plast Surg 6(3): 471–486 16. Laskawi R, Drobik C, Baaske C (1994) Prognostic value of electrodiagnosis of Bell’s palsy. Laryngorhinootologie 73:338–341
123
Eur Arch Otorhinolaryngol 17. Sataloff RT, Mandel S, Mann EA, Ludlow CL (2004) Practice parameter: laryngeal electromyography (an evidence-based review). Otolaryngol Head Neck Surg 130:770–779
123
18. Blitzer A, Crumley RL, Dailey SH et al (2009) Recommendations of the Neurolaryngology Study Group on laryngeal electromyography. Otolaryngol Head Neck Surg 140:782–793
