- Email: [email protected]
Visual observations of glottal configuration and vocal outcomes in arytenoid adduction
Visual Observations of Glottal Configuration and Vocal Outcomes in Arytenoid Adduction Katsuhide Inagi, MD,* Nadine P. Connor, PhD,† Tastutoshi Suzuki, MD,† Diane M. Bless, PhD,† and Takahiro Kamijo, MD* Arytenoid adduction procedures involve approximation of the arytenoid cartilages with the goal of reducing posterior glottal gap size and improving voice. However, voice outcomes after arytenoid adduction are not always optimal and may be improved by precise use of suture placements, forces, and direction angles. The development of intraoperative methods of assessing optimal suture direction appears critical for achieving the best voice outcome. The goal of this study was to examine the relationship of visual classification of glottal configuration, digital measures of the glottis, acoustic and aerodynamic measures, and voice outcome. Our results suggested that visual classification of glottal configuration was not useful in distinguishing voice outcome, except for cases in which there was a large posterior glottal gap. In contrast, acoustic and aerodynamic measures were related to digitized glottal measures and may be developed into a useful method of intraoperative monitoring. (Am J Otolaryngol 2003;24:290-296. © 2003 Elsevier Inc. All rights reserved.)
Arytenoid adduction is a phonosurgical procedure in which the arytenoids are approximated to reduce posterior glottal gap size and improve voice.1,2 This procedure is considered technically challenging by many practitioners,3,4 and as such, appears to be used less frequently5 than other medialization operations, such as thyroplasty type I. This is unfortunate because arytenoid adduction can be very effective. Failure to improve voice after arytenoid adduction may be a function of arytenoid cartilage overrotation,3,6 shortening of the vocal fold,3 and/or inadequate correction of vertical plane differences between the vocal folds.7 However, there have been few studies systematically investigating these issues,7-12 resulting in only a somewhat im-
From the *Department of Otolaryngology–Head and Neck Surgery, Kitasato Institute Medical Center Hospital, Saitama, Japan; and †Department of Surgery, Division of Otolaryngology–Head and Neck Surgery, University of Wisconsin Medical School, Madison, WI. Supported by the National Center for Voice and Speech through grant number P6DC000976 from the National Institute for Deafness and Other Communicative Disorders and research grant 2001 from the Kitasato Institute. Address correspondence to: Nadine P. Connor, PhD, University of Wisconsin Clinical Science Center, Room K4/711, 600 Highland Avenue, Madison, WI 537927375. E-mail: [email protected]. © 2003 Elsevier Inc. All rights reserved. 0196-0709/03/2405-0000$30.00/0 doi:10.1016/S0196-0709(03)00054-1 290
proved understanding of the impact of induced alterations in glottal configuration on laryngeal biomechanics and voice. More research is needed in this area to assist with the development of methods for maximizing positive voice outcomes after surgery. Human excised larynx paradigms allow investigation of 3-dimensional changes in glottal configuration after adjustments in arytenoid position, along with associated changes in phonatory functioning within a single larynx.7 In a previous study of excised human larynges,7 we found that changes in suture direction and force of pull affected glottal configuration in both the horizontal and vertical planes. Increased force of pull on the muscular process resulted in increased adduction of the vocal process for all suture directions. Changes in suture direction and force of pull also affected acoustic and aerodynamic measures of induced voice. This information may be clinically useful in allowing more precise arytenoid positioning and therefore more consistent and optimal vocal outcomes. Because suture direction appears critical in achieving optimal glottal configuration and positive voice outcomes, it would be clinically useful to devise intraoperative methods for deciding when optimal glottal configuration has been achieved. Proposed methods could include visual classification of glottal
American Journal of Otolaryngology, Vol 24, No 5 (September-October), 2003: pp 290-296
GLOTTAL CONFIGURATION
291
TABLE 1. Subjects Larynx No.
Age
Gender
Cause of Death
Intubation
1 2 3 4
31 42 53 82
M F F M
Aneurysm of aorta Pulmonary hypertension Myocardial infarction Bronco pneumonia
Yes Yes No No
configurations, digital measures of glottal parameters (width and area), and aerodynamic, acoustic measures. The purpose of this study was to evaluate the relationships among these measures of glottal configuration and vocal outcomes after arytenoid adduction procedures in a human excised laryngeal model with simulated unilateral vocal fold paralysis to determine if a particular method may hold promise for future development as a method of intraoperative monitoring. MATERIAL AND METHODS Subjects Four excised cadaveric human larynges (2 men and 2 women) were obtained from the Department of Pathology, University of Wisconsin, as listed in Table 1. These 4 larynges were selected because they did not have any vocal fold mucosal lesions and minimal laryngotracheal frame damage. Cause of death for each donor and the presence of an endotracheal tube at autopsy are shown in Table 1.
2. A medialization thyroplasty (type I) was performed on the left side to allow optimal tension of the left vocal fold (Fig 1). 3. An arytenoid adduction procedure was performed on the left side, using 100 g of applied weight to simulate a standard force of suture pull for all experiments. Sutures were directed to 1 of 6 positions on the thyroid cartilage (Fig 1). Before the each trial, the position of the left vocal fold was returned to the neutral reference position using the CCD camera for confirmation. 4. The larynx was mounted in the excised larynx study apparatus; phonation was induced; and video, acoustic, and aerodynamic recordings were made (Fig 2). 5. Suture direction was altered to a new position and recordings were repeated. Alteration in suture direction induced changes in glottal configuration and allowed study of the relationship between glottal configuration and voice outcome. Laryngeal preparation. Extralaryngeal tissue and muscle were dissected away to increase stabil-
Experimental Procedures The same experimental sequence was used for each of the 4 larynges. The general sequence is listed and then described in further detail. 1. The larynx was prepared for study. The right vocal fold was fixed at the same level across the 4 larynges using a needle and suture technique. To accomplish this, we positioned the right vocal fold at the level that yielded the best assessed phonatory outcome when the left vocal fold was adducted with sutures and 150-g weights to both the superior notch (SN) and inferior notch (IN) suture directions. A fine needle was used for simulating the thyroarytenoid and lateral cricoarytenoid muscles. In addition, a suture was placed into the right muscular processes for simulating the interarytenoid muscle. Under these conditions, a unilateral (left) vocal fold paralysis model was prepared (Fig 1).
Fig 1. A schematic view of unilateral paralysis model is shown. The right vocal fold was medialized and fixed using a suture that simulated the intra-arytenoid muscle and a needle that simulated the thyroarytenoid and lateral cricoarytenoid muscles. Thyroplasty type I was performed. Sutures attached at muscular process were directed toward 6 different directions, including the SN, anterior fourth of superior border of thyroid cartilage, middle of superior border of thyroid cartilage (MS), inferior notch (IN), anterior fourth of inferior border of thyroid cartilage (FI), and middle of inferior border of thyroid cartilage (MS).
292
INAGI ET AL
Fig 2. A schematic view of the excised laryngeal experimental apparatus. Compressed air was directed through a humidifier and artificial lung and then released. Subglottal air pressure and airflow were transduced at a position between the artificial lung and the larynx. A microphone was placed at 30 cm above the vocal folds. A CCD camera and a stroboscopy unit positioned above the larynx were used to record a superior view for analysis of vocal fold vibratory characteristics. Aerodynamic and acoustic data were acquired on personal computer using available computer software.
ity of the larynx. The epiglottis and superior part of aryepiglottic folds were extracted from each larynx, without any effect on arytenoid displacement, to allow improved visualization of the vocal folds and phonation. Arytenoid adduction procedure. Arytenoid adduction was performed on the left side of the larynx in a manner similar to the description of Isshiki et al.2 The muscular process of the left arytenoid cartilage was exposed and confirmed to be mobile without intrusion at the cricoarytenoid joint. A single 3-0 Tevdek suture was tied through the exposed muscular process. As shown in Figure 1, the sutures were passed through the larynx and emerged at 6 separate locations on the thyroid cartilage: (1) the SN, (2) the junction of the superior edge and the anterior fourth of the superior border, (3) the junction of the superior edge and the middle of the superior border, (4) the inferior notch (IN), (5) the junction of the inferior edge and the anterior fourth of inferior border, and (6) the junction of the inferior edge and the middle of the inferior border. Six separate experiments were performed within each larynx, in a serial fashion with 1 experiment per suture direction, to allow study of vocal param-
eters in a variety of glottal configurations. It was necessary to control the position of the right vocal fold across larynges to allow comparison of phonatory outcomes. Therefore, the right vocal fold was fixed at the same level across the 4 larynges, based on phonatory outcomes when the left vocal fold was adducted via medialization thyroplasty, and with sutures and 150-g weights to both the SN and IN suture directions. While monitoring voice production on the computer monitor, the position of the right vocal fold was adjusted to a position that met acoustic criteria for best phonation. The subsequent phonatory outcomes were based on changes in the left adducted vocal fold created by 100-g weights maintaining different suture directions. Medialization thyroplasty. Unilateral (left) vocal fold paralysis was simulated for this arytenoid adduction study and was treated using a left medialization thyroplasty (Fig 1). Because arytenoid adduction procedures do not correct the tension of vocal fold, medialization thyroplasty was necessary to produce suitable tension of the vocal fold and to adduct the membranous vocal fold on each larynx. Because laryngeal size and
GLOTTAL CONFIGURATION
293
TABLE 2. Shim Sizes and Locations Larynx
Shim Size
Shim Location
No. of Larynx
Shim Depth (mm)
From anterior border of window (mm)
From superior border of window (mm)
1 2 3 4
6 3 5 5
2 0 3 2
0 1 1 1
length of vocal fold on obtained larynges were different, different shim sizes and locations were need. The thyroplasty window was 5 mm ⫻ 10 mm with the anterior border located 5 mm or 7 mm from anterior edge of left thyroid cartilage for female or male larynges, respectively, and the superior border located at midportion between superior and inferior notch. Shim sizes and locations are listed in Table 2.
4. Open 2: both vocal processes met at the midline. However, there was a posterior gap because the body of the arytenoids did not meet at the midline (Fig 3B). 5. Open 3: both vocal processes and arytenoids do not meet at the midline, resulting in a large posterior gap (Fig 4). In addition to the glottal configuration classifications, 4 measurements were made from each digitized laryngeal view during the closed phase of vocal fold vibration under stroboscopy: vocal gap area, posterior gap area, vocal gap width, and posterior gap width with NIH Image 1.62 software (USDHSS-National Institutes of Health, Bethesda, MD) using an Apple PowerBook G4 (Cupertino, CA) (Fig 4). Total gap area was calculated as the following: total gap area ⫽ vocal gap area ⫹ posterior gap area. Acoustic and aerodynamic properties. Measurements included airflow, subglottal pressure, vocal efficiency index (AC flow/DC flow), jitter, percent shimmer, and signal-to-noise ratio (SNR).
Excised larynx apparatus. Each of the 4 larynges was mounted on scaffolding in the excised laryngeal research laboratory at the University of Wisconsin. The cricoid cartilage was stabilized on the horizontal plane, and the thyroid cartilage was tilted forward between 10° and 15° from the vertical plane. For glottal configuration recordings, a video camera was mounted above the larynx for recording a superior laryngeal view. Video images of the superior laryngeal view were used in the analysis and included videostroboscopic images recorded at 85 dB SPL during the closed phase of the vocal vibratory cycle. For phonatory recordings, compressed air was directed through a humidifier and artificial lung and then released through the larynx. Subglottal air pressure and airflow were transduced subglottally (Fig 2). A microphone was placed 30 cm above the vocal folds. Aerodynamic and acoustic data were acquired on a dedicated personal computer using available software (C-Speech, Madison, WI).
Measures Glottal configuration. The superior laryngeal video images were classified during induced phonation based on glottal configuration into 1 of 5 groups, as described here. 1. Overrotation: the adducted vocal process moved beyond the center of the larynx and over-or underlapped the contralateral vocal process (Fig 3A). 2. Complete closure: both arytenoids met at the midline without any apparent glottal gap. 3. Open 1: both vocal processes and arytenoids met at the midline with a small posterior gap.
Fig 3. The superior laryngeal video images were classified during phonation based on glottal configuration. (A) Overrotation: the adducted vocal process moved beyond the center of the larynx and over-or underlapped the contralateral vocal process. (B) Open 2: both vocal processes met at the midline. However, there was a posterior gap because the body of the arytenoids did not meet at the midline.
294
INAGI ET AL
vocal processes and arytenoids did not meet at center (open 3), and this measure was negatively correlated with total gap area, posterior gap area, and posterior gap width. No significant difference in vocal efficiency index was observed among the overrotation, open 1, and open 2 configurations. Acoustic Measurements
Fig 4. Four measurements were made from each digitized laryngeal view at the closed phase of vocal vibration during the induced phonation under stroboscopy: vocal gap area, posterior gap area, vocal gap width, and posterior gap width. Also, this glottal configuration was classified as open 3 (both vocal processes and arytenoids do not meet at the midline, resulting in a large posterior gap) in visual observation classification.
Statistical Analysis Each phonatory characteristic was compared across glottal configuration groups with KruskalWallis nonparametric analysis of variance. All testing was done at the nominal 5% level. Spearman rank correlation analysis was used to examine the relationship between 4 digital measurements of glottal configuration and each phonatory characteristic. The Spearman correlation coefficient (RS) was considered significant at P values of less than .05.
No significant differences in the acoustic measures were found among glottal configuration classification groups (Fig 6). However, all acoustic measures were significantly correlated with the digital measures of glottal configuration. Jitter was positively correlated with total gap area and vocal gap area, whereas SNR was negatively correlated (Table 3). Percent shimmer was positively correlated with all configuration measurements except vocal gap width (Table 3). DISCUSSION The arytenoid adduction technique has been proposed as a useful procedure for patients presenting with a large posterior glottal chink12 or with vocal folds that are positioned off the vertical plane. The arytenoid adduction procedure can be done alone or can be combined with other medialization procedures such as thyroplasty. Although techni-
RESULTS Aerodynamic Measurements Subglottal pressure required to produce phonation at 85 dB SPL in each glottal configuration group was between 37 cmH2O and 41 cmH2O (Fig 5). No significant differences in subglottal pressure were found among the glottal configuration groups, and no significant relationships among the glottal configuration measurements and subglottal pressure were found. Airflow was significantly greater for the open 3 configuration, in which a large posterior gap was present, in comparison with the other configurations. Furthermore, airflow was positively related to total gap area, posterior gap area, and posterior gap width (P ⬍ .05, Table 3). The lowest vocal efficiency index (0.78 ⫾ 0.05) was observed when both the
Fig 5. Aerodynamic measurements were compared among visual observation classification. No significant difference in subglottal pressure was found among the glottal configuration groups. However, airflow was significantly greater for the open 3 configuration in comparison with the other configurations. The lowest vocal efficiency index (0.78 ⴞ 0.05) was observed when both the vocal processes and arytenoids did not meet at center (open 3). No significant difference in vocal efficiency index was observed among the overrotation, open 1, and open 2 configurations.
GLOTTAL CONFIGURATION
295
TABLE 3. Correlation (Rs) Between Digitized Laryngeal Configuration Measurements and Vocal Outcomes Aerodynamic
Total gap area Vocal gap area Posterior gap area Vocal gap width Posterior gap width
Acoustic
Airflow
Subglottal Pressure
Vocal Efficiency Index
0.75† 0.318 0.715† 0.17 0.673†
0.215 0.335 0.133 0.064 0.089
0.449* 0.352 0.381* 0.142 0.399*
Jitter 0.44* 0.602† 0.297 0.152 0.301
% Shimmer
SNR
0.585† 0.406* 0.512† 0.149 0.446*
0.44* 0.504† 0.325 0.204 0.307
*P ⬍ .05. † P ⬍ .01.
cally difficult, arytenoid adduction may have some advantage over other procedures because surgical intervention does not cause trauma to the vocal fold mucosa or postoperative vocal fold scaring. In this vein, the arytenoid adduction technique might have an important role in obtaining optimal or better vocal results than those achieved via phonosurgical procedures that potentially cause trauma such as to the vocal fold, such as those involving injectates. In cases of glottal incompetence, baseline arytenoid positions can be highly variable13-15 in both the horizontal and vertical planes. In a previous study,7 we found that a fine-grained arytenoid positioning may be useful for achieving improved vocal quality after arytenoid adduction. By changing the direction of sutures attached to the muscular process, arytenoid position was adjusted in both the horizontal and vertical planes and phonatory outcomes were dramatically improved. To obtain optimal vocal outcomes after arytenoid adduction, it is necessary to assess the voice during the operation and to alter glottal con-
Fig 6. Acoustic measurements were compared among visual observation classification. No significant differences in the acoustic measures were found among glottal configurations.
figuration via manipulation of suture directions accordingly. Ostensibly this could be determined by assessing the voice or viewing the apparent position of the arytenoids. The purpose of this study was to examine possible methods of intraoperative voice or laryngeal configuration assessment that could serve as these intraoperative indicators of optimal voice during arytenoid adduction procedures. Our results suggest that complete closure of both vocal folds and arytenoids is necessary to obtain optimal voice. However, we found no significant differences between vocal outcome and visual classification of laryngeal configuration by using a superior laryngeal view. Interestingly, vocal outcomes in the case of arytenoid overrotation showed very similar results to cases with a large posterior gap. It appears that judging laryngeal configuration and assigning a configuration to a particular classification is not an effective means of intraoperative monitoring or prediction of voice outcome. During an arytenoid adduction procedure in an operating room, surgeons usually use a flexible nasopharyngeal fiberscope for visualizing the glottis during surgery. The small, 2-dimensional image obtained by this type of intraoperative imaging protocol may be even more difficult than the method used in the current excised larynx study. As such, it might be particularly difficult for surgeons to accurately judge complete closure of both vocal folds and arytenoids under nasopharyngeal fiberscopic observation. In contrast to the results of laryngeal configuration classification, data derived from digitized image analysis of glottal configuration were significantly correlated with vocal outcomes. Airflow, vocal efficiency index, and percent shimmer were significantly cor-
296
related with total gap area, vocal gap area, total gap area, and posterior gap width. Jitter and SNR were significantly correlated with vocal gap area and total gap area. For obtaining better results from arytenoid adduction, intraoperative measurement of glottal configuration during the closed phase in vocal vibratory pattern via stroboscopy or high speed imaging could be potentially useful for achieving optimal voice outcomes. However, methods for performing these measured during a surgical situation must be developed so that they are feasible and not cumbersome for the surgeon and staff. This study was conducted on excised larynges when it was possible to carefully control and measure the variables impacting vocal outcomes. Whether exercising the same control in an actual surgical situation is possible has not been determined. Possible limitations with the use of an excised laryngeal model include the following issues: (1) clinical laryngeal paralysis does not typically manifest as a complete denervation of the larynx, hence there may be forces exerted by residual or reinnervated muscle fibers16; (2) extrinsic laryngeal muscles and contralateral intrinsic laryngeal muscles may affect the position of the arytenoid on the paralyzed side11; and (3) the reduced support of structures such as traction of the trachea and pharynx may effect the results.8 A model, in the best of circumstances, can only serve as an approximation to the human clinical situation it is designed to reflect. As such, findings of this work should be interpreted with appropriate caution. Technological development of appropriate image analysis tools that increase feasibility of use in the operating room are needed to fully understand complete glottal closure based on the dynamic movements of both the arytenoid and vocal fold during arytenoid adduction procedures. Surgeons must be aware of the implication that glottal configuration, which is achieved via manipulation of suture directions, has a significant impact on glottal clo-
INAGI ET AL
sure and resultant voice outcome after arytenoid adduction. ACKNOWLEDGMENT The authors would like to thank Ms Joan Kozel for illustrations and Ms Jessica Adsit for editorial assistance. REFERENCES 1. Morrison LF: Reverse King operation. Ann Otol Rhinol Laryngol 57:945-956, 1948 2. Isshiki N, Tanabe M, Sawada M: Arytenoid adduction for unilateral vocal fold paralysis. Arch Otolaryngol 104:555-558, 1978 3. Slavit DH, Maragos NE: Physiologic assessment of arytenoid adduction. Ann Otol Rhinol Laryngol 101:321327, 1992 4. Isshiki N, Kojima H, Sawada M: Special instruments for laryngeal framework surgery. Ann Otol Rhinol Laryngol 100:728-730, 1991 5. Rosen CA: Complications of phonosurgery: Results of national survey. Laryngoscope 108:1697-1703, 1998 6. Isshiki N: Recent modifications in laryngeal framework surgery. Jibi Inkoka Rinsho 83:1-8, 1990 7. Inagi K, Connor PN, Suzuki T, et al: Glottal configuration, acoustic, and aerodynamic changes induced by variation in suture direction in arytenoid adduction procedures. Ann Otol Rhinol Laryngol 111:861-870, 2002 8. Noordzu JP, Perrault DF, Woo P: Biomechanics of arytenoid adduction surgery in an ex vivo canine model. Ann Otol Rhinol Laryngol 107:454-461, 1998 9. Haji T, Mori K, Omori K, et al: Mechanical properties of the vocal fold. Stress-strain studies. Acta Otolaryngol (Stockh) 112:559-565, 1992 10. Tran QT, Berke GS, Gerratt BR, et al: Measurement of Young’s modulus in the in vivo human vocal folds. Ann Otol Rhinol Laryngol 102:584-591, 1993 11. Woodson GE, Hengesteg A, Rosen CA, et al: Changes in length and spatial orientation of the vocal fold with arytenoid adduction in cadaver larynges. Ann Otol Rhinol Laryngol 106:552-555, 1997 12. Kraus DH, Orlikoff RF, Rizk SS, et al: Arytenoid adduction as an adjunct to type I thyroplasty for unilateral vocal cord paralysis. Head Neck 21:52-59, 1999 13. Sellars IE: A re-appraisal of intrinstic laryngeal muscle action. J Otolaryngol 7:450-456, 1978 14. Nasri S, Sercarz JA, Azizzadeh B, et al: Measurement of adductory force of individual laryngeal muscles in an in vivo canine model. Laryngoscope 104:1213-1218, 1994 15. Hirano M, Kiyokawa K, Kurita S: Laryngeal muscles and glottic shaping, in Fujimura O (ed): Vocal Physiology: Voice Production, Mechanisms, and Functions. New York, NY, Raven Press, 1988, 49-65 16. Hiroto I, Hirano M, Tomita H: Electromyographic investigation of human vocal cord paralysis. Ann Otol Rhinol Laryngol 77:296-304, 1968
Arytenoid adduction is a phonosurgical procedure in which the arytenoids are approximated to reduce posterior glottal gap size and improve voice.1,2 This procedure is considered technically challenging by many practitioners,3,4 and as such, appears to be used less frequently5 than other medialization operations, such as thyroplasty type I. This is unfortunate because arytenoid adduction can be very effective. Failure to improve voice after arytenoid adduction may be a function of arytenoid cartilage overrotation,3,6 shortening of the vocal fold,3 and/or inadequate correction of vertical plane differences between the vocal folds.7 However, there have been few studies systematically investigating these issues,7-12 resulting in only a somewhat im-
From the *Department of Otolaryngology–Head and Neck Surgery, Kitasato Institute Medical Center Hospital, Saitama, Japan; and †Department of Surgery, Division of Otolaryngology–Head and Neck Surgery, University of Wisconsin Medical School, Madison, WI. Supported by the National Center for Voice and Speech through grant number P6DC000976 from the National Institute for Deafness and Other Communicative Disorders and research grant 2001 from the Kitasato Institute. Address correspondence to: Nadine P. Connor, PhD, University of Wisconsin Clinical Science Center, Room K4/711, 600 Highland Avenue, Madison, WI 537927375. E-mail: [email protected]. © 2003 Elsevier Inc. All rights reserved. 0196-0709/03/2405-0000$30.00/0 doi:10.1016/S0196-0709(03)00054-1 290
proved understanding of the impact of induced alterations in glottal configuration on laryngeal biomechanics and voice. More research is needed in this area to assist with the development of methods for maximizing positive voice outcomes after surgery. Human excised larynx paradigms allow investigation of 3-dimensional changes in glottal configuration after adjustments in arytenoid position, along with associated changes in phonatory functioning within a single larynx.7 In a previous study of excised human larynges,7 we found that changes in suture direction and force of pull affected glottal configuration in both the horizontal and vertical planes. Increased force of pull on the muscular process resulted in increased adduction of the vocal process for all suture directions. Changes in suture direction and force of pull also affected acoustic and aerodynamic measures of induced voice. This information may be clinically useful in allowing more precise arytenoid positioning and therefore more consistent and optimal vocal outcomes. Because suture direction appears critical in achieving optimal glottal configuration and positive voice outcomes, it would be clinically useful to devise intraoperative methods for deciding when optimal glottal configuration has been achieved. Proposed methods could include visual classification of glottal
American Journal of Otolaryngology, Vol 24, No 5 (September-October), 2003: pp 290-296
GLOTTAL CONFIGURATION
291
TABLE 1. Subjects Larynx No.
Age
Gender
Cause of Death
Intubation
1 2 3 4
31 42 53 82
M F F M
Aneurysm of aorta Pulmonary hypertension Myocardial infarction Bronco pneumonia
Yes Yes No No
configurations, digital measures of glottal parameters (width and area), and aerodynamic, acoustic measures. The purpose of this study was to evaluate the relationships among these measures of glottal configuration and vocal outcomes after arytenoid adduction procedures in a human excised laryngeal model with simulated unilateral vocal fold paralysis to determine if a particular method may hold promise for future development as a method of intraoperative monitoring. MATERIAL AND METHODS Subjects Four excised cadaveric human larynges (2 men and 2 women) were obtained from the Department of Pathology, University of Wisconsin, as listed in Table 1. These 4 larynges were selected because they did not have any vocal fold mucosal lesions and minimal laryngotracheal frame damage. Cause of death for each donor and the presence of an endotracheal tube at autopsy are shown in Table 1.
2. A medialization thyroplasty (type I) was performed on the left side to allow optimal tension of the left vocal fold (Fig 1). 3. An arytenoid adduction procedure was performed on the left side, using 100 g of applied weight to simulate a standard force of suture pull for all experiments. Sutures were directed to 1 of 6 positions on the thyroid cartilage (Fig 1). Before the each trial, the position of the left vocal fold was returned to the neutral reference position using the CCD camera for confirmation. 4. The larynx was mounted in the excised larynx study apparatus; phonation was induced; and video, acoustic, and aerodynamic recordings were made (Fig 2). 5. Suture direction was altered to a new position and recordings were repeated. Alteration in suture direction induced changes in glottal configuration and allowed study of the relationship between glottal configuration and voice outcome. Laryngeal preparation. Extralaryngeal tissue and muscle were dissected away to increase stabil-
Experimental Procedures The same experimental sequence was used for each of the 4 larynges. The general sequence is listed and then described in further detail. 1. The larynx was prepared for study. The right vocal fold was fixed at the same level across the 4 larynges using a needle and suture technique. To accomplish this, we positioned the right vocal fold at the level that yielded the best assessed phonatory outcome when the left vocal fold was adducted with sutures and 150-g weights to both the superior notch (SN) and inferior notch (IN) suture directions. A fine needle was used for simulating the thyroarytenoid and lateral cricoarytenoid muscles. In addition, a suture was placed into the right muscular processes for simulating the interarytenoid muscle. Under these conditions, a unilateral (left) vocal fold paralysis model was prepared (Fig 1).
Fig 1. A schematic view of unilateral paralysis model is shown. The right vocal fold was medialized and fixed using a suture that simulated the intra-arytenoid muscle and a needle that simulated the thyroarytenoid and lateral cricoarytenoid muscles. Thyroplasty type I was performed. Sutures attached at muscular process were directed toward 6 different directions, including the SN, anterior fourth of superior border of thyroid cartilage, middle of superior border of thyroid cartilage (MS), inferior notch (IN), anterior fourth of inferior border of thyroid cartilage (FI), and middle of inferior border of thyroid cartilage (MS).
292
INAGI ET AL
Fig 2. A schematic view of the excised laryngeal experimental apparatus. Compressed air was directed through a humidifier and artificial lung and then released. Subglottal air pressure and airflow were transduced at a position between the artificial lung and the larynx. A microphone was placed at 30 cm above the vocal folds. A CCD camera and a stroboscopy unit positioned above the larynx were used to record a superior view for analysis of vocal fold vibratory characteristics. Aerodynamic and acoustic data were acquired on personal computer using available computer software.
ity of the larynx. The epiglottis and superior part of aryepiglottic folds were extracted from each larynx, without any effect on arytenoid displacement, to allow improved visualization of the vocal folds and phonation. Arytenoid adduction procedure. Arytenoid adduction was performed on the left side of the larynx in a manner similar to the description of Isshiki et al.2 The muscular process of the left arytenoid cartilage was exposed and confirmed to be mobile without intrusion at the cricoarytenoid joint. A single 3-0 Tevdek suture was tied through the exposed muscular process. As shown in Figure 1, the sutures were passed through the larynx and emerged at 6 separate locations on the thyroid cartilage: (1) the SN, (2) the junction of the superior edge and the anterior fourth of the superior border, (3) the junction of the superior edge and the middle of the superior border, (4) the inferior notch (IN), (5) the junction of the inferior edge and the anterior fourth of inferior border, and (6) the junction of the inferior edge and the middle of the inferior border. Six separate experiments were performed within each larynx, in a serial fashion with 1 experiment per suture direction, to allow study of vocal param-
eters in a variety of glottal configurations. It was necessary to control the position of the right vocal fold across larynges to allow comparison of phonatory outcomes. Therefore, the right vocal fold was fixed at the same level across the 4 larynges, based on phonatory outcomes when the left vocal fold was adducted via medialization thyroplasty, and with sutures and 150-g weights to both the SN and IN suture directions. While monitoring voice production on the computer monitor, the position of the right vocal fold was adjusted to a position that met acoustic criteria for best phonation. The subsequent phonatory outcomes were based on changes in the left adducted vocal fold created by 100-g weights maintaining different suture directions. Medialization thyroplasty. Unilateral (left) vocal fold paralysis was simulated for this arytenoid adduction study and was treated using a left medialization thyroplasty (Fig 1). Because arytenoid adduction procedures do not correct the tension of vocal fold, medialization thyroplasty was necessary to produce suitable tension of the vocal fold and to adduct the membranous vocal fold on each larynx. Because laryngeal size and
GLOTTAL CONFIGURATION
293
TABLE 2. Shim Sizes and Locations Larynx
Shim Size
Shim Location
No. of Larynx
Shim Depth (mm)
From anterior border of window (mm)
From superior border of window (mm)
1 2 3 4
6 3 5 5
2 0 3 2
0 1 1 1
length of vocal fold on obtained larynges were different, different shim sizes and locations were need. The thyroplasty window was 5 mm ⫻ 10 mm with the anterior border located 5 mm or 7 mm from anterior edge of left thyroid cartilage for female or male larynges, respectively, and the superior border located at midportion between superior and inferior notch. Shim sizes and locations are listed in Table 2.
4. Open 2: both vocal processes met at the midline. However, there was a posterior gap because the body of the arytenoids did not meet at the midline (Fig 3B). 5. Open 3: both vocal processes and arytenoids do not meet at the midline, resulting in a large posterior gap (Fig 4). In addition to the glottal configuration classifications, 4 measurements were made from each digitized laryngeal view during the closed phase of vocal fold vibration under stroboscopy: vocal gap area, posterior gap area, vocal gap width, and posterior gap width with NIH Image 1.62 software (USDHSS-National Institutes of Health, Bethesda, MD) using an Apple PowerBook G4 (Cupertino, CA) (Fig 4). Total gap area was calculated as the following: total gap area ⫽ vocal gap area ⫹ posterior gap area. Acoustic and aerodynamic properties. Measurements included airflow, subglottal pressure, vocal efficiency index (AC flow/DC flow), jitter, percent shimmer, and signal-to-noise ratio (SNR).
Excised larynx apparatus. Each of the 4 larynges was mounted on scaffolding in the excised laryngeal research laboratory at the University of Wisconsin. The cricoid cartilage was stabilized on the horizontal plane, and the thyroid cartilage was tilted forward between 10° and 15° from the vertical plane. For glottal configuration recordings, a video camera was mounted above the larynx for recording a superior laryngeal view. Video images of the superior laryngeal view were used in the analysis and included videostroboscopic images recorded at 85 dB SPL during the closed phase of the vocal vibratory cycle. For phonatory recordings, compressed air was directed through a humidifier and artificial lung and then released through the larynx. Subglottal air pressure and airflow were transduced subglottally (Fig 2). A microphone was placed 30 cm above the vocal folds. Aerodynamic and acoustic data were acquired on a dedicated personal computer using available software (C-Speech, Madison, WI).
Measures Glottal configuration. The superior laryngeal video images were classified during induced phonation based on glottal configuration into 1 of 5 groups, as described here. 1. Overrotation: the adducted vocal process moved beyond the center of the larynx and over-or underlapped the contralateral vocal process (Fig 3A). 2. Complete closure: both arytenoids met at the midline without any apparent glottal gap. 3. Open 1: both vocal processes and arytenoids met at the midline with a small posterior gap.
Fig 3. The superior laryngeal video images were classified during phonation based on glottal configuration. (A) Overrotation: the adducted vocal process moved beyond the center of the larynx and over-or underlapped the contralateral vocal process. (B) Open 2: both vocal processes met at the midline. However, there was a posterior gap because the body of the arytenoids did not meet at the midline.
294
INAGI ET AL
vocal processes and arytenoids did not meet at center (open 3), and this measure was negatively correlated with total gap area, posterior gap area, and posterior gap width. No significant difference in vocal efficiency index was observed among the overrotation, open 1, and open 2 configurations. Acoustic Measurements
Fig 4. Four measurements were made from each digitized laryngeal view at the closed phase of vocal vibration during the induced phonation under stroboscopy: vocal gap area, posterior gap area, vocal gap width, and posterior gap width. Also, this glottal configuration was classified as open 3 (both vocal processes and arytenoids do not meet at the midline, resulting in a large posterior gap) in visual observation classification.
Statistical Analysis Each phonatory characteristic was compared across glottal configuration groups with KruskalWallis nonparametric analysis of variance. All testing was done at the nominal 5% level. Spearman rank correlation analysis was used to examine the relationship between 4 digital measurements of glottal configuration and each phonatory characteristic. The Spearman correlation coefficient (RS) was considered significant at P values of less than .05.
No significant differences in the acoustic measures were found among glottal configuration classification groups (Fig 6). However, all acoustic measures were significantly correlated with the digital measures of glottal configuration. Jitter was positively correlated with total gap area and vocal gap area, whereas SNR was negatively correlated (Table 3). Percent shimmer was positively correlated with all configuration measurements except vocal gap width (Table 3). DISCUSSION The arytenoid adduction technique has been proposed as a useful procedure for patients presenting with a large posterior glottal chink12 or with vocal folds that are positioned off the vertical plane. The arytenoid adduction procedure can be done alone or can be combined with other medialization procedures such as thyroplasty. Although techni-
RESULTS Aerodynamic Measurements Subglottal pressure required to produce phonation at 85 dB SPL in each glottal configuration group was between 37 cmH2O and 41 cmH2O (Fig 5). No significant differences in subglottal pressure were found among the glottal configuration groups, and no significant relationships among the glottal configuration measurements and subglottal pressure were found. Airflow was significantly greater for the open 3 configuration, in which a large posterior gap was present, in comparison with the other configurations. Furthermore, airflow was positively related to total gap area, posterior gap area, and posterior gap width (P ⬍ .05, Table 3). The lowest vocal efficiency index (0.78 ⫾ 0.05) was observed when both the
Fig 5. Aerodynamic measurements were compared among visual observation classification. No significant difference in subglottal pressure was found among the glottal configuration groups. However, airflow was significantly greater for the open 3 configuration in comparison with the other configurations. The lowest vocal efficiency index (0.78 ⴞ 0.05) was observed when both the vocal processes and arytenoids did not meet at center (open 3). No significant difference in vocal efficiency index was observed among the overrotation, open 1, and open 2 configurations.
GLOTTAL CONFIGURATION
295
TABLE 3. Correlation (Rs) Between Digitized Laryngeal Configuration Measurements and Vocal Outcomes Aerodynamic
Total gap area Vocal gap area Posterior gap area Vocal gap width Posterior gap width
Acoustic
Airflow
Subglottal Pressure
Vocal Efficiency Index
0.75† 0.318 0.715† 0.17 0.673†
0.215 0.335 0.133 0.064 0.089
0.449* 0.352 0.381* 0.142 0.399*
Jitter 0.44* 0.602† 0.297 0.152 0.301
% Shimmer
SNR
0.585† 0.406* 0.512† 0.149 0.446*
0.44* 0.504† 0.325 0.204 0.307
*P ⬍ .05. † P ⬍ .01.
cally difficult, arytenoid adduction may have some advantage over other procedures because surgical intervention does not cause trauma to the vocal fold mucosa or postoperative vocal fold scaring. In this vein, the arytenoid adduction technique might have an important role in obtaining optimal or better vocal results than those achieved via phonosurgical procedures that potentially cause trauma such as to the vocal fold, such as those involving injectates. In cases of glottal incompetence, baseline arytenoid positions can be highly variable13-15 in both the horizontal and vertical planes. In a previous study,7 we found that a fine-grained arytenoid positioning may be useful for achieving improved vocal quality after arytenoid adduction. By changing the direction of sutures attached to the muscular process, arytenoid position was adjusted in both the horizontal and vertical planes and phonatory outcomes were dramatically improved. To obtain optimal vocal outcomes after arytenoid adduction, it is necessary to assess the voice during the operation and to alter glottal con-
Fig 6. Acoustic measurements were compared among visual observation classification. No significant differences in the acoustic measures were found among glottal configurations.
figuration via manipulation of suture directions accordingly. Ostensibly this could be determined by assessing the voice or viewing the apparent position of the arytenoids. The purpose of this study was to examine possible methods of intraoperative voice or laryngeal configuration assessment that could serve as these intraoperative indicators of optimal voice during arytenoid adduction procedures. Our results suggest that complete closure of both vocal folds and arytenoids is necessary to obtain optimal voice. However, we found no significant differences between vocal outcome and visual classification of laryngeal configuration by using a superior laryngeal view. Interestingly, vocal outcomes in the case of arytenoid overrotation showed very similar results to cases with a large posterior gap. It appears that judging laryngeal configuration and assigning a configuration to a particular classification is not an effective means of intraoperative monitoring or prediction of voice outcome. During an arytenoid adduction procedure in an operating room, surgeons usually use a flexible nasopharyngeal fiberscope for visualizing the glottis during surgery. The small, 2-dimensional image obtained by this type of intraoperative imaging protocol may be even more difficult than the method used in the current excised larynx study. As such, it might be particularly difficult for surgeons to accurately judge complete closure of both vocal folds and arytenoids under nasopharyngeal fiberscopic observation. In contrast to the results of laryngeal configuration classification, data derived from digitized image analysis of glottal configuration were significantly correlated with vocal outcomes. Airflow, vocal efficiency index, and percent shimmer were significantly cor-
296
related with total gap area, vocal gap area, total gap area, and posterior gap width. Jitter and SNR were significantly correlated with vocal gap area and total gap area. For obtaining better results from arytenoid adduction, intraoperative measurement of glottal configuration during the closed phase in vocal vibratory pattern via stroboscopy or high speed imaging could be potentially useful for achieving optimal voice outcomes. However, methods for performing these measured during a surgical situation must be developed so that they are feasible and not cumbersome for the surgeon and staff. This study was conducted on excised larynges when it was possible to carefully control and measure the variables impacting vocal outcomes. Whether exercising the same control in an actual surgical situation is possible has not been determined. Possible limitations with the use of an excised laryngeal model include the following issues: (1) clinical laryngeal paralysis does not typically manifest as a complete denervation of the larynx, hence there may be forces exerted by residual or reinnervated muscle fibers16; (2) extrinsic laryngeal muscles and contralateral intrinsic laryngeal muscles may affect the position of the arytenoid on the paralyzed side11; and (3) the reduced support of structures such as traction of the trachea and pharynx may effect the results.8 A model, in the best of circumstances, can only serve as an approximation to the human clinical situation it is designed to reflect. As such, findings of this work should be interpreted with appropriate caution. Technological development of appropriate image analysis tools that increase feasibility of use in the operating room are needed to fully understand complete glottal closure based on the dynamic movements of both the arytenoid and vocal fold during arytenoid adduction procedures. Surgeons must be aware of the implication that glottal configuration, which is achieved via manipulation of suture directions, has a significant impact on glottal clo-
INAGI ET AL
sure and resultant voice outcome after arytenoid adduction. ACKNOWLEDGMENT The authors would like to thank Ms Joan Kozel for illustrations and Ms Jessica Adsit for editorial assistance. REFERENCES 1. Morrison LF: Reverse King operation. Ann Otol Rhinol Laryngol 57:945-956, 1948 2. Isshiki N, Tanabe M, Sawada M: Arytenoid adduction for unilateral vocal fold paralysis. Arch Otolaryngol 104:555-558, 1978 3. Slavit DH, Maragos NE: Physiologic assessment of arytenoid adduction. Ann Otol Rhinol Laryngol 101:321327, 1992 4. Isshiki N, Kojima H, Sawada M: Special instruments for laryngeal framework surgery. Ann Otol Rhinol Laryngol 100:728-730, 1991 5. Rosen CA: Complications of phonosurgery: Results of national survey. Laryngoscope 108:1697-1703, 1998 6. Isshiki N: Recent modifications in laryngeal framework surgery. Jibi Inkoka Rinsho 83:1-8, 1990 7. Inagi K, Connor PN, Suzuki T, et al: Glottal configuration, acoustic, and aerodynamic changes induced by variation in suture direction in arytenoid adduction procedures. Ann Otol Rhinol Laryngol 111:861-870, 2002 8. Noordzu JP, Perrault DF, Woo P: Biomechanics of arytenoid adduction surgery in an ex vivo canine model. Ann Otol Rhinol Laryngol 107:454-461, 1998 9. Haji T, Mori K, Omori K, et al: Mechanical properties of the vocal fold. Stress-strain studies. Acta Otolaryngol (Stockh) 112:559-565, 1992 10. Tran QT, Berke GS, Gerratt BR, et al: Measurement of Young’s modulus in the in vivo human vocal folds. Ann Otol Rhinol Laryngol 102:584-591, 1993 11. Woodson GE, Hengesteg A, Rosen CA, et al: Changes in length and spatial orientation of the vocal fold with arytenoid adduction in cadaver larynges. Ann Otol Rhinol Laryngol 106:552-555, 1997 12. Kraus DH, Orlikoff RF, Rizk SS, et al: Arytenoid adduction as an adjunct to type I thyroplasty for unilateral vocal cord paralysis. Head Neck 21:52-59, 1999 13. Sellars IE: A re-appraisal of intrinstic laryngeal muscle action. J Otolaryngol 7:450-456, 1978 14. Nasri S, Sercarz JA, Azizzadeh B, et al: Measurement of adductory force of individual laryngeal muscles in an in vivo canine model. Laryngoscope 104:1213-1218, 1994 15. Hirano M, Kiyokawa K, Kurita S: Laryngeal muscles and glottic shaping, in Fujimura O (ed): Vocal Physiology: Voice Production, Mechanisms, and Functions. New York, NY, Raven Press, 1988, 49-65 16. Hiroto I, Hirano M, Tomita H: Electromyographic investigation of human vocal cord paralysis. Ann Otol Rhinol Laryngol 77:296-304, 1968