A functional evaluation of ansa cervicalis nerve transfer for unilateral vocal cord paralysis: Future directions for laryngeal reinnervation DAVID C. GREEN, MD, GERALD S. BERKE, MD, and MICHAEL C. GRAVES, MD, Los Angeles, California
There are a variety of methods for treating unilateral vocal cord paralysis. but to date there are few objective studies that evaluate the functional results of nerve transfer from the ansa cervlcalls. Six dogs underwent unilateral recurrent laryngeal nerve sec. tlon with Immediate reanastamosls to the sternothyroid branch of the ansa cervlcalls. After 5 to 6 months. measurements of vocal efficiency and acoustic parameters. vi. deolaryngoscopy. vldeostroboscopy. and evoked electromyography were performed. Identical measurements were made In eight control dogs during normal electrically Induced phonation and a simulated unilateral recurrent laryngeal nerve paralysis. Histologic analysis of both vocalls muscles. recurrent laryngeal nerves. ansa cervlcalls. and the ansa·recurrent laryngeal nerve anastamosls site was performed. Evidence of reinnervation was found In all of the animals that underwent nerve transfer. The vocal efficiency and acoustic quality after ansa cervlcalls nerve transfer were dependent on the degree of electrical stimUlation from the transferred nerve to the relnnervated cord during phonation. In the absence of electrical stimulation to the nerve transfer. physiologic vocal cord motion could not be elicited from the relnnervated cord. (OTOLARYNGOL HEAD NECK SURG 1991;104:453.)
T
here are a variety of methods for treating unilateral vocal cord paralysis. These include Teflon injection, I thyroplasty,' arytenoid adduction," and nerve" and nerve-muscle pedicle transfer.' A review of the literature reveals a lack of consensus regarding the optimal procedure for voice improvement. 6 Crumley et al. 4 have recommended nerve transfer from the ansa hypoglossi to the recurrent laryngeal nerve (RLN) as a treatment for unilateral vocal cord paralysis. This procedure was performed on five patients and the voice was thought to be superior to that achieved with Teflon injection because of the restoration of normal stiffness, mass, and symmetry of the cord. Spectral analysis was improved postoperatively," and stroboscopic examination revealed synchronous mucosal waves. This technique required an open procedure, but did not expose or manipulate the larynx. In addition, Crumley et al." postulated that the denervated
From the Division of Head and Neck Surgery, UCLA School of Medicine. Supported by VA Merit Review Research Fund. Received for publication July 3D, 1990; accepted Aug. 29, 1990. Reprint requests: David C. Green, MD. 10992 Ashton Ave., Apt. 304, Los Angeles, CA 90024.
23/1125135
sternothyroid muscle medialized the thyroid lamina and vocal cord. Tucker has also described a technique for laryngeal reinnervation that uses an ansa hypoglossi neuromuscular pedicle passed through a window in the thyroid cartilage. The literature on laryngeal reinnervation is extensive. I An excellent review of the subject was recently published by Rice." The idea goes back to 1909, when Horsley" first reported the repair of a recurrent laryngeal nerve after transection by a gunshot wound. He reported complete return of function in 3 months. In 1926, Blalock and Crowe" described some return of motion without atrophy of the cord after anastamosis of the distal segment of the RLN to either the ansa hypoglossi (endto-end), the phrenic, or into a slit in the opposite RLN. Histologic studies demonstrated regrowth of axons after anastarnosis in all three cases. Frazier and Mosser'! anastomosed the RLN to the ansa hypoglossi in 10 patients with paralysis present from 8 months to 11 years. They found improvement in 60% overall but little improvement in patients with paralysis for more than 6 years, In contrast to these early encouraging results, Hoover," in 1953, stated that attempts at restoration of recurrent laryngeal function by nerve anastamosis had been unsuccessful. This situation was clarified in 1963 453
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
OtoIaryngoJogyHead and Neck SUrgery
. . . GREENeta!.
by Siribodi et al.," who found a return of electromyographic activity after repair of the laryngeal nerve, but no functional movement. They proposed that aberrant regeneration of adductor and abductor fibers accounted for the lack of function. This finding is now termed laryngeal synkinesis." Although Doyle et al. IS reported almost complete return of function after anastamosis of the RLN in both dogs and human beings, most reports have been consistent with the findings of Siribodi et al." Gordon and McCabe l6 found that after section and immediate repair of the RLN, only adductor function returned. They attributed this to the fact that adductor muscles outnumbered abductor muscles. Boles and Fritzell 17 found polyphasic potentials by means of electromyography (EMG) 3 to 6 months after RLN section and anastamosis in dogs with little motion in the reinnervated cord. In contrast, in the crush injuries they studied, recovery took place in approximately 2 months with normal cord movement. They concluded that surgical repair of the RLN could possibly return tone and some adduction to the cord and improve the glottis as a phonator and a protective sphincter. Tashiro" found EMO evidence of recovery but no definite motion after repair of the RLN. Murakami and Kirchner" confirmed the presence of simultaneous activation of the adductors and abductors after repair of the RLN by EMO. This supported the idea of misdirection in axon regrowth. Satoh et al. 1O performed direct muscle stimulation and EMO studies and thought that impaired muscle function, in addition to axon misdirection, contributed to poor function of the reinnervated laryngeal muscles. Ded021 found a small degree of adduction in the reinnervated cord after RLN anastamosis. He believed the adduction resulted from the action of the cricothyroid muscle. As reinnervation of the larynx began to emerge as a potential treatment for unilateral vocal cord paralysis, controversy arose as to which nerve transfer would give the best results. Crumley et al." recommended the ansa cervicalis branch to the sternothyroid branch for several reasons. This branch to the sternothyroid is located very near the RLN. Although the pattern of sternothyroid activation is predominantly inspiratory, it fires throughout the respiratory cycle, giving a certain degree of continuous tone to the vocalis. Also, it is thought that the subsequent paralysis of the sternothyroid branch relaxes tension on the thyroid cartilage and further medializes the paralyzed cord. However, some investigators have argued against the use of the sternothyroid branch because the pattern of activation
from this nerve may result in cord adduction during inspiration rather than phonation." Marie et al. 23 found, after ansa-adductor branch transfer in the canine model, that during extended obstructive dyspnea, there was adduction of the cord, which was abolished after sectioning of the ansa. In this same canine model, the reinnervated cord failed to reach midline during whining, demonstrating a lack of ansa stimulation during phonation. Crumley" has advised against using the ansa's branch to the thyrohyoid despite its being the most active strap muscle during phonation, with peak activity during expiration. His reluctance stems from the requirement of a cable graft to reach the RLN stump and the resulting paralysis of the thyrohyoid, eliminating the adductor function of this muscle. The superior laryngeal nerve (SLN) has been proposed by Rice" to be ideal for nerve transfer to improve phonation after a unilateral vocal fold paralysis. This nerve is active during phonation and anatomically close to the RLN. Woodsen2S found that during normal respiration in human beings, the cricothyroid muscle is active predominantly during inspiration. This pattern of activation may be of concern because it is conjectured that inspiratory activation of the thyroarytenoid may cause airway obstruction. Rice" and Crumley" have performed a SLN-adductor branch anastamosis in dogs, which demonstrated spontaneous adduction of the cord with coughing or whining. However, Crumley" believes that use of the superior laryngeal nerve is "robbing Peter to pay Paul" in that this would eliminate the normal adductive effect of the cricothyroid muscle. Crumley" has also counseled against the use of the proximal RLN in RLN-RLN anastamosis. His concerns are again related to the possibility of airway obstruction during inspiration as a result of hypertrophy and medial bulging of the thyroarytenoid from RLN synkinesis. 13 Although Tashiro," Sirbodhi et al., 13 Murakami and Kirchner," Sato et al.,20 Veda et al.," and Iwamura" have shown activation of the adductors during inspiration, none of these studies has demonstrated respiratory obstruction with dyspnea from a unilateral RLNRLN anastamosis in the presence of a normal opposite cord. Iwamura" stated "such minor paradoxical adduction of the reinnervated vocal cord appears to be of no great importance, since the inspiratory adduction lasts for only a fraction of a second." As mentioned previously. the RLN·RLN anastamosis may not be the only "offender" in this regard. Marie et aI. 23 reported paradoxical inspiratory adduction of the reinnervated cord by means of the ansa branch to the sternothyroid.
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
Volume 104 Number 4
April 1991
Future directions for laryngeal reinnervation
A recent" clinical case report of a unilateral RLN-RLN anastamosis resulted in an excellent voice without airway obstruction. Another issue is the duration of time that a nerve transfer can still be used after vocal cord paralysis. Janecka," by means of the rat siatic nerve, found that severed axons will not regrow when the nerve segment has lost contact with the target muscle. How long denervated muscle can provide this trophic response is not known. Some authors obtain an EMG from any patient who has been paralyzed for more than 2 years. Patients with absence of muscle activity on a thyroarytenoid EMG are then treated with Teflon injection. Although the voice after nerve transfer has been reported as excellent, there are some disadvantages. The patient must undergo general anesthesia and an open surgical procedure. An intact distal stump of the RLN must be present. The patient must wait approximately 3 to 4 months before there is improvement from the procedure. Gelfoam injection into the paralyzed cord may improve glottal closure during the reinnervation period. Few reports have used objective voice measurements to study the results of nerve transfer for unilateral paralysis of the vocal folds. Vocal efficiency is an objective measure of the voice that was first studied by van den Berg" in 1956. He defined the efficiency of voice as the ratio of the acoustic power of the voice to the SUbglottic power. The subglottic air power can be estimated as the product of the mean glottic air flow rate and the subglottic pressure. Clinically, vocal efficiency has been shown to decrease with some forms of laryngeal disease, such as invasive carcinoma and vocal cord paralysis.F-" Ueda et al. 27 studied acoustic intensity. subglottic pressure, and flow rate during induced whining in dogs after a combination SLN-SLN and RLNRLN anastamosis. The vocal efficiency after reinnervation in this canine study was significantly less than normal. Although vocal efficiency is useful as a functional measure of voice, it may not correspond with vocal quality. The voice may be quite harsh with a normal vocal efficiency. Also. vocal efficiency does not indicate the degree of control the patient has over the glottis. 34 To measure voice quality, a number of acoustic measures have been used clinically. These include jitter. shimmer. and signal-to-noise ratio. Jitter is defined as the fluctuation in the time interval between successive peaks of the fundamental frequency. Shimmer is the cycle-to-cycle variation in the amplitudes of the peaks.
411
Signal-to-noise ratio is the ratio of the sound energy in the acoustic signal to the background noise. 35.36 Lieberman" was the first to report an increased jitter in pathologic phonation, and Zyski et al. 38 found increased jitter and shimmer in patients with laryngeal tumors and unilateral vocal cord paralysis. Efforts have been made to use these measures as screening devices for pathologic conditions of the larynx" and to document the results of laryngeal surgery." This study used an in vivo canine model to evaluate the functional results of nerve transfer from the ansa cervicalis to the RLN as a treatment for unilateral vocal cord paralysis. Measurements of vocal efficiency, acoustic analysis, videolaryngoscopy, videostroboscopy, and evoked EMG were performed after ansa cervicalis nerve transfer and compared with a control group of animals during normal phonation and a simulated RLN paralysis. Tissue samples were taken from both vocalis muscles, RLNs, ansa cervicalis, and the ansa-RLN anastamosis site for histologic analysis.
METHODS AND MATERIALS Experimental Design The experimental group consisted of six mongrel dogs (25 kg each) that underwent a unilateral RLN section with immediate reanastamosis to the sternothyroid branch of the ansa cervicalis under general anesthesia. One dog had a reaction to the anesthetic and died before the evaluation. After 5 to 6 months, measurements for vocal efficiency and acoustic analysis (jitter, shimmer, and signal-to-noise ratio) were made on the remaining five experimental dogs, in addition to videolaryngoscopy and videostroboscopy. In the experimental group, vocal efficiency and acoustic measures were made, with and without electrical stimulation of the ansa cervicalis nerve transfer. In addition, measurements of vocal efficiency and acoustic analysis (jitter, shimmer, and signal-to-noise ratio) were made on a group of eight control dogs during "normal" electrically stimulated phonation and a simulated RLN paralysis. In the experimental group, evoked EMG was performed to measure the nerve conduction velocities and the response amplitudes of the normal and reinnervated vocal folds. Tissue samples were taken of both vocalis muscles, RLNs, ansa cervicalis, and the ansaRLN anastamosis site for histologic comparison.
Surgical Technique In the Experimental Group Each animal was premedicated with acepromazine and then underwent general anesthesia with endotracheal intubation. The ventral neck was shaved and pre-
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
OtolaryngologyHead and Neck Surgery
.Q6 GREEN at oJ.
"'IICROPHONE SOUND LEVEL METER
"'IULTICHANNEl STORAGE OSCILLOSCOPE
DIREC T COMPUTER IZ ED DIGITIZ AT ION
.r-:
-J\-
PRESSURE TRANSDUCER
Vent ,lotor ~
Fig. 1. Schematic representation of experimental setup for In vivo canine model of phonatlon otter ansa cervlcalls nerve transfer.
pared with povidone iodine. The animal was placed in the supine position on a table in the operating room, and a midline neck incision was made from the midthyroid cartilage to 4 cm below the cricoid. The dissection was carried down to the trachea, and the RLN and the ansa cervicalis (ansa) were identified bilaterally. On the basis of the proximity of the ansa and the RLN, one side was chosen for reanastamosis . At approximately 5 cm from the cricothyroid joint, the RLN was sectioned sharply with a no. 15 blade. The proximal end of the RLN was ligated with 2-0 silk sutures. The distal end was then anastamosed end-toend to the ansa cervicalis. The anastamosis was performed with four 10-0 nylon sutures according to standard microsurgical technique . The site of the ansa-RLN anastomosis was marked with a loose 2-0 silk suture placed I em away from the anastamosis . The wound was closed in three layers with 3-0 VicryI. Six months postoperatively, the laryngeal phonatory characteristics were evaluated by means of the in vivo canine model of phonation ." Humane animal care was assured in compliance with The Principles of Laboratory Animal Care. formulated by the National Society for Medical Research, and the Guide for the Care and Use of Lab-
oratory Animals. prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23, revised 1978). In Vivo Canine Phonation Model
Each dog in the control and experimental groups was premedicated with intramuscular acepromazine. Intravenous thiopental was administered to a level of corneal anesthesia , and additional thiopental was used to maintain this level of anesthes ia throughout the experiment. The animal was placed in the supine position on the operating table, and a midline incision was made to expose the trachea from the hyoid to the sternal notch (Fig . I). In the animals undergoing experimental nerve transfer, the RLN on one side and the RLN-ansa anastamosis on the opposite side were identified and preserved. In the control group of animals, both RLNs were identified. For both control and experimental groups, the superior laryngeal nerves were identified along their course to the cricothyroid muscles . A low tracheotomy was performed at the level of the suprasternal notch, through which an endotracheal tube was passed to allow ventilator-assisted respiration. A second
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
Volume 104 Number 4 April 1991
Future directions for laryngeal reinnervation
tracheotomy was performed in a more superior location, through which a cuffed endotracheal tube was passed in a rostral direction and positioned with the tip 10 em below the vocal folds. The cuff was inflated to just seal the trachea. Humidified heated air was passed through this rostral endotracheal tube from a compressed air tank. Flow was controlled with a valve and measured with a Gilmont flowmeter (model FI500, Gilmont Instruments, Great Neck, N.Y.). The air flow was humidified and heated by bubbling it through 5 em of heated water so that the temperature of the air was 37° C when measured at the glottic outlet. A l-cm button was used to suspend the epiglottis from a fixed point to provide direct visualization of the larynx through the oral cavity. A l-cm segment of each superior laryngeal nerve (external branch) was isolated, and Harvard miniature electrodes (Harvard Apparatus Inc., Millis, Mass.) were applied around each nerve. Harvard electrodes were also applied to the RLN on one side and the ansa cervicalis on the opposite side in the animals undergoing nerve transfer or both RLNs in the control animals. Care was taken to place the electrode on the ansa cervicalis 2 em proximal to the anastamosis site in the animals undergoing nerve transfer, to avoid current spread to the distal RLN. The electrodes were then insulated from surrounding tissue. The cut proximal RLN on the side of the anastamosis in the animals undergoing nerve transfer was also dissected out and stimulated with a supramaximal current of 5 mA to verify an absence of vocal cord movement or EMG stimulation. Two Model S2LH constant-current nerve stimulators (WR Medical Electronics Co., St. Paul, Minn.) were used to stimulate the RLN, ansa cervicalis, and superior laryngeal nerves independently. These nerves were stimulated at 70 to 80 Hz stimulus frequency, with 0.5 to 2.0 mA intensity for 1.5-msec duration. Phonation was produced with an air flow of 318 to 523 cc/sec, applied through the larynx by the rostral endotracheal tube.
Acoustic Measures Acoustic measurements were obtained for the experimental nerve transfer group, both with and without stimulation of the ansa nerve transfer. Acoustic measurements for the control group were obtained during simulated normal phonation and RLN paralysis. Normal phonation required stimulation of all four laryngeal nerves. A unilateral RLN paralysis was simulated by not stimulating one RLN. The sound level measurements were made with a I-inch Quest condenser microphone placed 30 em from and level with the glottic outlet. The microphone was directed 90 degrees from
••7
the direction of the sound source. The sound level measurements were made in decibels, with a Quest sound level meter on the C scale. The acoustic signal was also digitized, after C scale filtering at 20 kHz, and stored on the hard disk of a personal computer. Subglottic pressure was measured with a Millar Mikro-Tip catheter pressure transducer (model No. SPC-330, size 3F, Millar Instruments, Inc., Houston, Texas) passed rostrally through the superior tracheotomy. It was placed 5 em below the glottis. This signal was low-pass filtered at 3 kHz, digitized at 20 kHz, and stored in a personal computer. Because of the variation in subglottic pressure during phonation, the peak pressures attained during the glottic cycle were used. These peaks were identified by means of a commercially available software package for the PC system ("C-Speech," Paul Milenkovic, University of Wisconsin, Madison, Wis.).36 The pressure transducer was calibrated before each experiment against a mercury manometer at 3r C. Vocal efficiency was calculated as the ratio of the acoustic power of the voice to the subglottic power. The total acoustic power was calculated according to the method of Koyama et al.," in which total sound power = 2 rPc 2/Poc. This formula applies for a sound power radiating with no known direction into a hemisphere of area (2 r'), a distance (r) away from the source. The product Po (the density of the medium) and c (the velocity of propagation) is the specific acoustic impedance of the medium, which is 41.1 dynes/sec/em' in air at 200 C. The term P, is the root mean square sound pressure in dynes/em' at the distance r from the sound source. The subglottic power was calculated as the product of the flow rate times the peak subglottic pressure. Acoustic analysis of the digitized acoustic signal was performed with a commercial software program ("CSpeech," Paul Milenkovic, University of Wisconsin, Madison, Wis.). Jitter, shimmer, and signal-to-noise ratio were calculated on approximately 0.3 second of stable phonation from each trial. The background noise in the experimental quarters was 35 dB lower than the values measured with the C scale. To normalize for varying fundamental frequency, jitter was calculated as a fraction of the period of the fundamental frequency. For each dog in either the experimental or the control group, the jitter, shimmer, signal-to-noise ratio, and vocal efficiency for that dog were calculated from the mean of five trials of phonation. A mean value for the entire experimental group, with and without ansa stimulation for jitter, shimmer, signal-to-noise ratio, and vocal efficiency, was calculated by taking the average across all five experimental dogs. A mean value for the
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
OtolaryngologyHead and Neck Surgery
.51 GREEN et 01.
control group during simulated normal phonation and RLN paralysis for jitter, shimmer, signal-to-noise ratio, and vocal efficiency was calculated by taking the average across all eight control dogs. VldeoltrobolcoPY and VldeolaryngolcoPY
Videolaryngoscopy was performed during single nerve stimulation of the RLN on the normal side or ansa cervicaJis on the nerve transfer side. The current stimulus was increased, starting from zero, to observe the effect of stimulation from each nerve separately. The image was detected by a Jed-Med CCD (chargecoupled device) video camera (Model 70-5110, JedMed, St. Louis, Mo.) and a Sony V-matic videocassette recorder (VO-5850. Park Ridge. N.J.). Videolaryngoscopy was also performed under a lighter plane of anesthesia to detect the degree of spontaneous motion of the cords during laryngeal stimulation with a cotton swab and obstructive respiratory dyspnea. Videostroboscopy was performed, both with and without stimulation of the ansa cervicalis, in the animals undergoing nerve transfer, and with and without stimulation of one RLN in the normal control animals, while the other three laryngeal nerves were being stimulated. For stroboscopic imaging of the larynx. a Bruel & Kjaer laryngostrobe unit (model 4914, Bruel & Kjaer, Orange. Calif.) was used. The stroboscope was connected to a Storz zero-degree telescope via a fluid-filled light cable. The video images were analyzed frame-by-frame with the videorecording unit. Evoked EMG
Conventional spontaneous EMG recordings were initially made during anesthesia. to detect fibrillation potentials in the thyroarytenoid muscle. Evoked EMG was then performed in the nerve transfer experimental animals by stimulating the ansa cervicalis nerve transfer or the normal RLN on the opposite side with a single 0.5 msec pulse of sufficient voltage through a Harvard miniature electrode. to obtain an adequate response. The response was detected with a bipolar needle placed transorally into the ipsilateral thyroarytenoid muscle. The signal was digitized at 20 kHz and stored in a personal computer. A conduction velocity for each nerve was calculated by measuring the latency of the response in milliseconds between two points along the nerve. The time difference in the latency between the stimulated points divided by the distance between these points gave the conduction velocity between these points. In the experimental animals, the conduction velocity was measured on the normal side between two
points along the distal RLN. On the nerve transfer side, conduction velocity was measured between a point proximal and distal to the ansa-Rl.N anastamosis. In addition, the amplitude of the evoked EMG was recorded in millivolts .as a measure of the amount of depolarizing muscle. Histologic AnaIY.'1
The larynx of each animal undergoing nerve transfer was removed. divided into right and left halves, and placed in 10% .formalin. A 3-mm thick cross-section of each vocalis muscle was taken midway between the anterior commissure and the tip of the vocalis process for staining with hematoxylin and eosin. A l-cm segment of the RLN was removed at a site I cm proximal to the cricothyroid joint bilaterally and at 5 cm proximal to the cricothyroid joint on the side not operated on. A l-cm segment of ansa cervicalis was taken I em proximal to the anastamosis on the side operated on and at a similar site on the ansa of the side not operated on. The site of the ansa-RLN anastamosis was also removed. All five nerve specimens were fixed in 4% paraformaldehyde-0.4% glutaraldehyde. The anastamosis site was embedded in parlodion and sectioned longitudinally. The remaining specimens were stained with osmium. embedded in Eponl Araldite (Electron Microscopy Sciences, Fort Washington, Pa.) and thin sections (0.5 "",m) were cut transversely. Statistical Ana'y." A paired t test was used to compare mean changes from ansa stimulation to nonstimulation within the experimental group (N = 5) for vocal efficiency, jitter, shimmer, and signal-to-noise ratio. Two group t tests were used to compare the mean difference in these measurements between the experimental group and the control group (N = 8). For the vocal efficiency measurement, the t test was computed on the log values. Between-dog comparisons by means of Mann-Whitney and Wilcoxon rank-sum criteria instead of the t test did not change any conclusions.
RESULTS Nerve and Vocal Cord Histologic Findings
There was little sign of atrophy in the reinnervated thyroarytenoid muscle when compared to the normal side. The average diameter of both the reinnervated and normal thyroarytenoid muscles was identical at 7.3 mm. Figure 2 shows the lack of atrophy in the reinnervated cord when compared to normal. Examination of the RLN on the side of the nerve transfer showed regenerated axons as seen in Fig. 3.
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
Volume 104 Number 4 April 1991
Future directions for laryngeal reinnervation 419
A
B Fig. 2. Cross section of A, normal cord and B, reinnervated cord.
VldeolaryngolcoPY of Groll Movement Increasing electrical stimulation of the normal RLN demonstrated increasing ipsilateral vocalis muscle bulging and arytenoid adduction in all of the dogs. The threshold for vocalis contraction on the normal side was between 0.05 and 0.19 rnA. With increasing stimulation of the ansa cervi calis on the side of the nerve transfer, increasing ipsilateral vocalis contraction was seen in all dogs. In four of the dogs, this vocalis contraction was associated with arytenoid adduction, but
in one animal there was distinct arytenoid abduction with vocalis contraction. The threshold for vocalis contraction from ansa stimulation was between 0.01 and 0.2 rnA. During obstructive respiratory dyspnea, under a light plane of anesthesia, after sectioning both superior laryngeal nerves, the normal cord was found to abduct with inspiratory efforts in all dogs. The reinnervated cord did not adduct or abduct to any degree during this time in any subject.
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
There are a variety of methods for treating unilateral vocal cord paralysis. but to date there are few objective studies that evaluate the functional results of nerve transfer from the ansa cervlcalls. Six dogs underwent unilateral recurrent laryngeal nerve sec. tlon with Immediate reanastamosls to the sternothyroid branch of the ansa cervlcalls. After 5 to 6 months. measurements of vocal efficiency and acoustic parameters. vi. deolaryngoscopy. vldeostroboscopy. and evoked electromyography were performed. Identical measurements were made In eight control dogs during normal electrically Induced phonation and a simulated unilateral recurrent laryngeal nerve paralysis. Histologic analysis of both vocalls muscles. recurrent laryngeal nerves. ansa cervlcalls. and the ansa·recurrent laryngeal nerve anastamosls site was performed. Evidence of reinnervation was found In all of the animals that underwent nerve transfer. The vocal efficiency and acoustic quality after ansa cervlcalls nerve transfer were dependent on the degree of electrical stimUlation from the transferred nerve to the relnnervated cord during phonation. In the absence of electrical stimulation to the nerve transfer. physiologic vocal cord motion could not be elicited from the relnnervated cord. (OTOLARYNGOL HEAD NECK SURG 1991;104:453.)
T
here are a variety of methods for treating unilateral vocal cord paralysis. These include Teflon injection, I thyroplasty,' arytenoid adduction," and nerve" and nerve-muscle pedicle transfer.' A review of the literature reveals a lack of consensus regarding the optimal procedure for voice improvement. 6 Crumley et al. 4 have recommended nerve transfer from the ansa hypoglossi to the recurrent laryngeal nerve (RLN) as a treatment for unilateral vocal cord paralysis. This procedure was performed on five patients and the voice was thought to be superior to that achieved with Teflon injection because of the restoration of normal stiffness, mass, and symmetry of the cord. Spectral analysis was improved postoperatively," and stroboscopic examination revealed synchronous mucosal waves. This technique required an open procedure, but did not expose or manipulate the larynx. In addition, Crumley et al." postulated that the denervated
From the Division of Head and Neck Surgery, UCLA School of Medicine. Supported by VA Merit Review Research Fund. Received for publication July 3D, 1990; accepted Aug. 29, 1990. Reprint requests: David C. Green, MD. 10992 Ashton Ave., Apt. 304, Los Angeles, CA 90024.
23/1125135
sternothyroid muscle medialized the thyroid lamina and vocal cord. Tucker has also described a technique for laryngeal reinnervation that uses an ansa hypoglossi neuromuscular pedicle passed through a window in the thyroid cartilage. The literature on laryngeal reinnervation is extensive. I An excellent review of the subject was recently published by Rice." The idea goes back to 1909, when Horsley" first reported the repair of a recurrent laryngeal nerve after transection by a gunshot wound. He reported complete return of function in 3 months. In 1926, Blalock and Crowe" described some return of motion without atrophy of the cord after anastamosis of the distal segment of the RLN to either the ansa hypoglossi (endto-end), the phrenic, or into a slit in the opposite RLN. Histologic studies demonstrated regrowth of axons after anastarnosis in all three cases. Frazier and Mosser'! anastomosed the RLN to the ansa hypoglossi in 10 patients with paralysis present from 8 months to 11 years. They found improvement in 60% overall but little improvement in patients with paralysis for more than 6 years, In contrast to these early encouraging results, Hoover," in 1953, stated that attempts at restoration of recurrent laryngeal function by nerve anastamosis had been unsuccessful. This situation was clarified in 1963 453
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
OtoIaryngoJogyHead and Neck SUrgery
. . . GREENeta!.
by Siribodi et al.," who found a return of electromyographic activity after repair of the laryngeal nerve, but no functional movement. They proposed that aberrant regeneration of adductor and abductor fibers accounted for the lack of function. This finding is now termed laryngeal synkinesis." Although Doyle et al. IS reported almost complete return of function after anastamosis of the RLN in both dogs and human beings, most reports have been consistent with the findings of Siribodi et al." Gordon and McCabe l6 found that after section and immediate repair of the RLN, only adductor function returned. They attributed this to the fact that adductor muscles outnumbered abductor muscles. Boles and Fritzell 17 found polyphasic potentials by means of electromyography (EMG) 3 to 6 months after RLN section and anastamosis in dogs with little motion in the reinnervated cord. In contrast, in the crush injuries they studied, recovery took place in approximately 2 months with normal cord movement. They concluded that surgical repair of the RLN could possibly return tone and some adduction to the cord and improve the glottis as a phonator and a protective sphincter. Tashiro" found EMO evidence of recovery but no definite motion after repair of the RLN. Murakami and Kirchner" confirmed the presence of simultaneous activation of the adductors and abductors after repair of the RLN by EMO. This supported the idea of misdirection in axon regrowth. Satoh et al. 1O performed direct muscle stimulation and EMO studies and thought that impaired muscle function, in addition to axon misdirection, contributed to poor function of the reinnervated laryngeal muscles. Ded021 found a small degree of adduction in the reinnervated cord after RLN anastamosis. He believed the adduction resulted from the action of the cricothyroid muscle. As reinnervation of the larynx began to emerge as a potential treatment for unilateral vocal cord paralysis, controversy arose as to which nerve transfer would give the best results. Crumley et al." recommended the ansa cervicalis branch to the sternothyroid branch for several reasons. This branch to the sternothyroid is located very near the RLN. Although the pattern of sternothyroid activation is predominantly inspiratory, it fires throughout the respiratory cycle, giving a certain degree of continuous tone to the vocalis. Also, it is thought that the subsequent paralysis of the sternothyroid branch relaxes tension on the thyroid cartilage and further medializes the paralyzed cord. However, some investigators have argued against the use of the sternothyroid branch because the pattern of activation
from this nerve may result in cord adduction during inspiration rather than phonation." Marie et al. 23 found, after ansa-adductor branch transfer in the canine model, that during extended obstructive dyspnea, there was adduction of the cord, which was abolished after sectioning of the ansa. In this same canine model, the reinnervated cord failed to reach midline during whining, demonstrating a lack of ansa stimulation during phonation. Crumley" has advised against using the ansa's branch to the thyrohyoid despite its being the most active strap muscle during phonation, with peak activity during expiration. His reluctance stems from the requirement of a cable graft to reach the RLN stump and the resulting paralysis of the thyrohyoid, eliminating the adductor function of this muscle. The superior laryngeal nerve (SLN) has been proposed by Rice" to be ideal for nerve transfer to improve phonation after a unilateral vocal fold paralysis. This nerve is active during phonation and anatomically close to the RLN. Woodsen2S found that during normal respiration in human beings, the cricothyroid muscle is active predominantly during inspiration. This pattern of activation may be of concern because it is conjectured that inspiratory activation of the thyroarytenoid may cause airway obstruction. Rice" and Crumley" have performed a SLN-adductor branch anastamosis in dogs, which demonstrated spontaneous adduction of the cord with coughing or whining. However, Crumley" believes that use of the superior laryngeal nerve is "robbing Peter to pay Paul" in that this would eliminate the normal adductive effect of the cricothyroid muscle. Crumley" has also counseled against the use of the proximal RLN in RLN-RLN anastamosis. His concerns are again related to the possibility of airway obstruction during inspiration as a result of hypertrophy and medial bulging of the thyroarytenoid from RLN synkinesis. 13 Although Tashiro," Sirbodhi et al., 13 Murakami and Kirchner," Sato et al.,20 Veda et al.," and Iwamura" have shown activation of the adductors during inspiration, none of these studies has demonstrated respiratory obstruction with dyspnea from a unilateral RLNRLN anastamosis in the presence of a normal opposite cord. Iwamura" stated "such minor paradoxical adduction of the reinnervated vocal cord appears to be of no great importance, since the inspiratory adduction lasts for only a fraction of a second." As mentioned previously. the RLN·RLN anastamosis may not be the only "offender" in this regard. Marie et aI. 23 reported paradoxical inspiratory adduction of the reinnervated cord by means of the ansa branch to the sternothyroid.
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
Volume 104 Number 4
April 1991
Future directions for laryngeal reinnervation
A recent" clinical case report of a unilateral RLN-RLN anastamosis resulted in an excellent voice without airway obstruction. Another issue is the duration of time that a nerve transfer can still be used after vocal cord paralysis. Janecka," by means of the rat siatic nerve, found that severed axons will not regrow when the nerve segment has lost contact with the target muscle. How long denervated muscle can provide this trophic response is not known. Some authors obtain an EMG from any patient who has been paralyzed for more than 2 years. Patients with absence of muscle activity on a thyroarytenoid EMG are then treated with Teflon injection. Although the voice after nerve transfer has been reported as excellent, there are some disadvantages. The patient must undergo general anesthesia and an open surgical procedure. An intact distal stump of the RLN must be present. The patient must wait approximately 3 to 4 months before there is improvement from the procedure. Gelfoam injection into the paralyzed cord may improve glottal closure during the reinnervation period. Few reports have used objective voice measurements to study the results of nerve transfer for unilateral paralysis of the vocal folds. Vocal efficiency is an objective measure of the voice that was first studied by van den Berg" in 1956. He defined the efficiency of voice as the ratio of the acoustic power of the voice to the SUbglottic power. The subglottic air power can be estimated as the product of the mean glottic air flow rate and the subglottic pressure. Clinically, vocal efficiency has been shown to decrease with some forms of laryngeal disease, such as invasive carcinoma and vocal cord paralysis.F-" Ueda et al. 27 studied acoustic intensity. subglottic pressure, and flow rate during induced whining in dogs after a combination SLN-SLN and RLNRLN anastamosis. The vocal efficiency after reinnervation in this canine study was significantly less than normal. Although vocal efficiency is useful as a functional measure of voice, it may not correspond with vocal quality. The voice may be quite harsh with a normal vocal efficiency. Also. vocal efficiency does not indicate the degree of control the patient has over the glottis. 34 To measure voice quality, a number of acoustic measures have been used clinically. These include jitter. shimmer. and signal-to-noise ratio. Jitter is defined as the fluctuation in the time interval between successive peaks of the fundamental frequency. Shimmer is the cycle-to-cycle variation in the amplitudes of the peaks.
411
Signal-to-noise ratio is the ratio of the sound energy in the acoustic signal to the background noise. 35.36 Lieberman" was the first to report an increased jitter in pathologic phonation, and Zyski et al. 38 found increased jitter and shimmer in patients with laryngeal tumors and unilateral vocal cord paralysis. Efforts have been made to use these measures as screening devices for pathologic conditions of the larynx" and to document the results of laryngeal surgery." This study used an in vivo canine model to evaluate the functional results of nerve transfer from the ansa cervicalis to the RLN as a treatment for unilateral vocal cord paralysis. Measurements of vocal efficiency, acoustic analysis, videolaryngoscopy, videostroboscopy, and evoked EMG were performed after ansa cervicalis nerve transfer and compared with a control group of animals during normal phonation and a simulated RLN paralysis. Tissue samples were taken from both vocalis muscles, RLNs, ansa cervicalis, and the ansa-RLN anastamosis site for histologic analysis.
METHODS AND MATERIALS Experimental Design The experimental group consisted of six mongrel dogs (25 kg each) that underwent a unilateral RLN section with immediate reanastamosis to the sternothyroid branch of the ansa cervicalis under general anesthesia. One dog had a reaction to the anesthetic and died before the evaluation. After 5 to 6 months, measurements for vocal efficiency and acoustic analysis (jitter, shimmer, and signal-to-noise ratio) were made on the remaining five experimental dogs, in addition to videolaryngoscopy and videostroboscopy. In the experimental group, vocal efficiency and acoustic measures were made, with and without electrical stimulation of the ansa cervicalis nerve transfer. In addition, measurements of vocal efficiency and acoustic analysis (jitter, shimmer, and signal-to-noise ratio) were made on a group of eight control dogs during "normal" electrically stimulated phonation and a simulated RLN paralysis. In the experimental group, evoked EMG was performed to measure the nerve conduction velocities and the response amplitudes of the normal and reinnervated vocal folds. Tissue samples were taken of both vocalis muscles, RLNs, ansa cervicalis, and the ansaRLN anastamosis site for histologic comparison.
Surgical Technique In the Experimental Group Each animal was premedicated with acepromazine and then underwent general anesthesia with endotracheal intubation. The ventral neck was shaved and pre-
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
OtolaryngologyHead and Neck Surgery
.Q6 GREEN at oJ.
"'IICROPHONE SOUND LEVEL METER
"'IULTICHANNEl STORAGE OSCILLOSCOPE
DIREC T COMPUTER IZ ED DIGITIZ AT ION
.r-:
-J\-
PRESSURE TRANSDUCER
Vent ,lotor ~
Fig. 1. Schematic representation of experimental setup for In vivo canine model of phonatlon otter ansa cervlcalls nerve transfer.
pared with povidone iodine. The animal was placed in the supine position on a table in the operating room, and a midline neck incision was made from the midthyroid cartilage to 4 cm below the cricoid. The dissection was carried down to the trachea, and the RLN and the ansa cervicalis (ansa) were identified bilaterally. On the basis of the proximity of the ansa and the RLN, one side was chosen for reanastamosis . At approximately 5 cm from the cricothyroid joint, the RLN was sectioned sharply with a no. 15 blade. The proximal end of the RLN was ligated with 2-0 silk sutures. The distal end was then anastamosed end-toend to the ansa cervicalis. The anastamosis was performed with four 10-0 nylon sutures according to standard microsurgical technique . The site of the ansa-RLN anastomosis was marked with a loose 2-0 silk suture placed I em away from the anastamosis . The wound was closed in three layers with 3-0 VicryI. Six months postoperatively, the laryngeal phonatory characteristics were evaluated by means of the in vivo canine model of phonation ." Humane animal care was assured in compliance with The Principles of Laboratory Animal Care. formulated by the National Society for Medical Research, and the Guide for the Care and Use of Lab-
oratory Animals. prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23, revised 1978). In Vivo Canine Phonation Model
Each dog in the control and experimental groups was premedicated with intramuscular acepromazine. Intravenous thiopental was administered to a level of corneal anesthesia , and additional thiopental was used to maintain this level of anesthes ia throughout the experiment. The animal was placed in the supine position on the operating table, and a midline incision was made to expose the trachea from the hyoid to the sternal notch (Fig . I). In the animals undergoing experimental nerve transfer, the RLN on one side and the RLN-ansa anastamosis on the opposite side were identified and preserved. In the control group of animals, both RLNs were identified. For both control and experimental groups, the superior laryngeal nerves were identified along their course to the cricothyroid muscles . A low tracheotomy was performed at the level of the suprasternal notch, through which an endotracheal tube was passed to allow ventilator-assisted respiration. A second
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
Volume 104 Number 4 April 1991
Future directions for laryngeal reinnervation
tracheotomy was performed in a more superior location, through which a cuffed endotracheal tube was passed in a rostral direction and positioned with the tip 10 em below the vocal folds. The cuff was inflated to just seal the trachea. Humidified heated air was passed through this rostral endotracheal tube from a compressed air tank. Flow was controlled with a valve and measured with a Gilmont flowmeter (model FI500, Gilmont Instruments, Great Neck, N.Y.). The air flow was humidified and heated by bubbling it through 5 em of heated water so that the temperature of the air was 37° C when measured at the glottic outlet. A l-cm button was used to suspend the epiglottis from a fixed point to provide direct visualization of the larynx through the oral cavity. A l-cm segment of each superior laryngeal nerve (external branch) was isolated, and Harvard miniature electrodes (Harvard Apparatus Inc., Millis, Mass.) were applied around each nerve. Harvard electrodes were also applied to the RLN on one side and the ansa cervicalis on the opposite side in the animals undergoing nerve transfer or both RLNs in the control animals. Care was taken to place the electrode on the ansa cervicalis 2 em proximal to the anastamosis site in the animals undergoing nerve transfer, to avoid current spread to the distal RLN. The electrodes were then insulated from surrounding tissue. The cut proximal RLN on the side of the anastamosis in the animals undergoing nerve transfer was also dissected out and stimulated with a supramaximal current of 5 mA to verify an absence of vocal cord movement or EMG stimulation. Two Model S2LH constant-current nerve stimulators (WR Medical Electronics Co., St. Paul, Minn.) were used to stimulate the RLN, ansa cervicalis, and superior laryngeal nerves independently. These nerves were stimulated at 70 to 80 Hz stimulus frequency, with 0.5 to 2.0 mA intensity for 1.5-msec duration. Phonation was produced with an air flow of 318 to 523 cc/sec, applied through the larynx by the rostral endotracheal tube.
Acoustic Measures Acoustic measurements were obtained for the experimental nerve transfer group, both with and without stimulation of the ansa nerve transfer. Acoustic measurements for the control group were obtained during simulated normal phonation and RLN paralysis. Normal phonation required stimulation of all four laryngeal nerves. A unilateral RLN paralysis was simulated by not stimulating one RLN. The sound level measurements were made with a I-inch Quest condenser microphone placed 30 em from and level with the glottic outlet. The microphone was directed 90 degrees from
••7
the direction of the sound source. The sound level measurements were made in decibels, with a Quest sound level meter on the C scale. The acoustic signal was also digitized, after C scale filtering at 20 kHz, and stored on the hard disk of a personal computer. Subglottic pressure was measured with a Millar Mikro-Tip catheter pressure transducer (model No. SPC-330, size 3F, Millar Instruments, Inc., Houston, Texas) passed rostrally through the superior tracheotomy. It was placed 5 em below the glottis. This signal was low-pass filtered at 3 kHz, digitized at 20 kHz, and stored in a personal computer. Because of the variation in subglottic pressure during phonation, the peak pressures attained during the glottic cycle were used. These peaks were identified by means of a commercially available software package for the PC system ("C-Speech," Paul Milenkovic, University of Wisconsin, Madison, Wis.).36 The pressure transducer was calibrated before each experiment against a mercury manometer at 3r C. Vocal efficiency was calculated as the ratio of the acoustic power of the voice to the subglottic power. The total acoustic power was calculated according to the method of Koyama et al.," in which total sound power = 2 rPc 2/Poc. This formula applies for a sound power radiating with no known direction into a hemisphere of area (2 r'), a distance (r) away from the source. The product Po (the density of the medium) and c (the velocity of propagation) is the specific acoustic impedance of the medium, which is 41.1 dynes/sec/em' in air at 200 C. The term P, is the root mean square sound pressure in dynes/em' at the distance r from the sound source. The subglottic power was calculated as the product of the flow rate times the peak subglottic pressure. Acoustic analysis of the digitized acoustic signal was performed with a commercial software program ("CSpeech," Paul Milenkovic, University of Wisconsin, Madison, Wis.). Jitter, shimmer, and signal-to-noise ratio were calculated on approximately 0.3 second of stable phonation from each trial. The background noise in the experimental quarters was 35 dB lower than the values measured with the C scale. To normalize for varying fundamental frequency, jitter was calculated as a fraction of the period of the fundamental frequency. For each dog in either the experimental or the control group, the jitter, shimmer, signal-to-noise ratio, and vocal efficiency for that dog were calculated from the mean of five trials of phonation. A mean value for the entire experimental group, with and without ansa stimulation for jitter, shimmer, signal-to-noise ratio, and vocal efficiency, was calculated by taking the average across all five experimental dogs. A mean value for the
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
OtolaryngologyHead and Neck Surgery
.51 GREEN et 01.
control group during simulated normal phonation and RLN paralysis for jitter, shimmer, signal-to-noise ratio, and vocal efficiency was calculated by taking the average across all eight control dogs. VldeoltrobolcoPY and VldeolaryngolcoPY
Videolaryngoscopy was performed during single nerve stimulation of the RLN on the normal side or ansa cervicaJis on the nerve transfer side. The current stimulus was increased, starting from zero, to observe the effect of stimulation from each nerve separately. The image was detected by a Jed-Med CCD (chargecoupled device) video camera (Model 70-5110, JedMed, St. Louis, Mo.) and a Sony V-matic videocassette recorder (VO-5850. Park Ridge. N.J.). Videolaryngoscopy was also performed under a lighter plane of anesthesia to detect the degree of spontaneous motion of the cords during laryngeal stimulation with a cotton swab and obstructive respiratory dyspnea. Videostroboscopy was performed, both with and without stimulation of the ansa cervicalis, in the animals undergoing nerve transfer, and with and without stimulation of one RLN in the normal control animals, while the other three laryngeal nerves were being stimulated. For stroboscopic imaging of the larynx. a Bruel & Kjaer laryngostrobe unit (model 4914, Bruel & Kjaer, Orange. Calif.) was used. The stroboscope was connected to a Storz zero-degree telescope via a fluid-filled light cable. The video images were analyzed frame-by-frame with the videorecording unit. Evoked EMG
Conventional spontaneous EMG recordings were initially made during anesthesia. to detect fibrillation potentials in the thyroarytenoid muscle. Evoked EMG was then performed in the nerve transfer experimental animals by stimulating the ansa cervicalis nerve transfer or the normal RLN on the opposite side with a single 0.5 msec pulse of sufficient voltage through a Harvard miniature electrode. to obtain an adequate response. The response was detected with a bipolar needle placed transorally into the ipsilateral thyroarytenoid muscle. The signal was digitized at 20 kHz and stored in a personal computer. A conduction velocity for each nerve was calculated by measuring the latency of the response in milliseconds between two points along the nerve. The time difference in the latency between the stimulated points divided by the distance between these points gave the conduction velocity between these points. In the experimental animals, the conduction velocity was measured on the normal side between two
points along the distal RLN. On the nerve transfer side, conduction velocity was measured between a point proximal and distal to the ansa-Rl.N anastamosis. In addition, the amplitude of the evoked EMG was recorded in millivolts .as a measure of the amount of depolarizing muscle. Histologic AnaIY.'1
The larynx of each animal undergoing nerve transfer was removed. divided into right and left halves, and placed in 10% .formalin. A 3-mm thick cross-section of each vocalis muscle was taken midway between the anterior commissure and the tip of the vocalis process for staining with hematoxylin and eosin. A l-cm segment of the RLN was removed at a site I cm proximal to the cricothyroid joint bilaterally and at 5 cm proximal to the cricothyroid joint on the side not operated on. A l-cm segment of ansa cervicalis was taken I em proximal to the anastamosis on the side operated on and at a similar site on the ansa of the side not operated on. The site of the ansa-RLN anastamosis was also removed. All five nerve specimens were fixed in 4% paraformaldehyde-0.4% glutaraldehyde. The anastamosis site was embedded in parlodion and sectioned longitudinally. The remaining specimens were stained with osmium. embedded in Eponl Araldite (Electron Microscopy Sciences, Fort Washington, Pa.) and thin sections (0.5 "",m) were cut transversely. Statistical Ana'y." A paired t test was used to compare mean changes from ansa stimulation to nonstimulation within the experimental group (N = 5) for vocal efficiency, jitter, shimmer, and signal-to-noise ratio. Two group t tests were used to compare the mean difference in these measurements between the experimental group and the control group (N = 8). For the vocal efficiency measurement, the t test was computed on the log values. Between-dog comparisons by means of Mann-Whitney and Wilcoxon rank-sum criteria instead of the t test did not change any conclusions.
RESULTS Nerve and Vocal Cord Histologic Findings
There was little sign of atrophy in the reinnervated thyroarytenoid muscle when compared to the normal side. The average diameter of both the reinnervated and normal thyroarytenoid muscles was identical at 7.3 mm. Figure 2 shows the lack of atrophy in the reinnervated cord when compared to normal. Examination of the RLN on the side of the nerve transfer showed regenerated axons as seen in Fig. 3.
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
Volume 104 Number 4 April 1991
Future directions for laryngeal reinnervation 419
A
B Fig. 2. Cross section of A, normal cord and B, reinnervated cord.
VldeolaryngolcoPY of Groll Movement Increasing electrical stimulation of the normal RLN demonstrated increasing ipsilateral vocalis muscle bulging and arytenoid adduction in all of the dogs. The threshold for vocalis contraction on the normal side was between 0.05 and 0.19 rnA. With increasing stimulation of the ansa cervi calis on the side of the nerve transfer, increasing ipsilateral vocalis contraction was seen in all dogs. In four of the dogs, this vocalis contraction was associated with arytenoid adduction, but
in one animal there was distinct arytenoid abduction with vocalis contraction. The threshold for vocalis contraction from ansa stimulation was between 0.01 and 0.2 rnA. During obstructive respiratory dyspnea, under a light plane of anesthesia, after sectioning both superior laryngeal nerves, the normal cord was found to abduct with inspiratory efforts in all dogs. The reinnervated cord did not adduct or abduct to any degree during this time in any subject.
Downloaded from oto.sagepub.com at Oakland University on June 5, 2016
