The Laryngoscope C 2014 The American Laryngological, V
Rhinological and Otological Society, Inc.
Nerve Branches to the Posterior Cricoarytenoid Muscle May Complicate the Laryngeal Reinnervation Procedure Wan-Fu Su, MD; Shao-Cheng Liu, MD; Shwun-De Wang, PhD; Wang-Yu Su, MD; Kuo-Hsing Ma, PhD; Tung-Tsun Huang, MD Objectives/Hypothesis: To better understand the reason for the low success rate of posterior cricoarytenoid (PCA) muscle reinnervation, we attempted to identify the communicating branches that may exist between the nerve branch to the PCA muscle and the other laryngeal adductors in addition to the interarytenoid (IA) muscle. Study Design: Quantitative histological assessment. Methods: Twenty human hemilarynges from patients with laryngeal or hypopharyngeal cancer were obtained after surgery and stained with Sihler’s stain, which rendered the muscle translucent while counterstaining the neuroanatomy of the recurrent laryngeal nerve (RLN) inside the larynges. Results: The nerve supply to the PCA muscle was separated into two main branches. One upper branch supplied the horizontal compartment, and the lower branch supplied the vertical/oblique compartment. In 14 specimens, two nerve branches to the PCA muscle arose separately from the RLN. In six specimens, one single main branch arose from the RLN and immediately ramified into two secondary branches. In all specimens except one, the nerve branch to the horizontal compartment was connected to the IA muscle. However, no communicating branches were observed between the nerve to the PCA muscle and the other laryngeal adductors. No anastomosis between nerve branches to the horizontal and vertical/ oblique compartments or other variations of nerve distribution were observed. Conclusions: The communicating branches between the nerve branches to the PCA muscle and the IA muscle may be the only branch, complicating the reinnervation procedure. More investigations are needed to identify variations in the terminal branch distribution from the RLN. Key Words: Posterior cricoarytenoid muscle, interarytenoid muscle, communicating branch, recurrent laryngeal nerve, lateral cricoarytenoid muscle. Level of Evidence: NA Laryngoscope, 125:419–423, 2015
INTRODUCTION Most patients with bilateral vocal cord palsy usually experience airway difficulties. Recurrent laryngeal nerve (RLN) injury from thyroid surgery represents the most frequent cause of bilateral vocal cord palsy, followed by intubation injury, cardiovascular accident, radiation therapy, and idiopathic causes.1 Many proce-
From the Department of Otolaryngology–Head and Neck Surgery (W.-F.S., W.-Y.S, T.-T.H.), Buddist Tzu Chi General Hospital, Taipei Branch School of Medicine, Tzu Chi University, Hualien; and the Department of Otolaryngology–Head and Neck Surgery (W.-F.S., S.-C.L.), and the Department of Biology and Anatomy (S.-D.W., K.-H.M.), Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan. Editor’s Note: This Manuscript was accepted for publication September 4, 2014. Wan-Fu Su, MD participated in the study design, writing, literature search, and manuscript review/editing. Shao-Cheng Liu, MD participated in data collection, picture compiling, and article submission. Shwun-De Wang, MD participated as a methodology director. Wang-Yu Su, MD participated in material collection. Kuo-Hsing Ma, PhD participated as a methodology director. Tung-Tsun Huang, MD participated in material collection The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Wan Fu Su, MD, Department of Otolaryngology-Head and Neck Surgery, Tri-Service General Hospital, 325, Sec 2, Chen-Kung Road, Taipei, 114 Taiwan. E-mail: [email protected] DOI: 10.1002/lary.24944
Laryngoscope 125: February 2015
dures have been proposed to resolve the respiratory problems statically or simultaneously, to preserve the phonatory and sphincteric functions dynamically. It is accepted that laser cordectomy, arytenoidectomy, or suture lateralization can resolve the respiratory problems but at the expense of phonatory quality.2 The laryngeal reinnervation of the posterior cricoarytenoid (PCA) muscle has received much attention because of its low success rate. Because a mobile vocal fold functions as a regulator of respiration and may enable maximal preservation of voice quality, many procedures have been investigated to ensure PCA muscle reanimation.3,4 Such investigations have included phrenic nerve implantation into the PCA muscle,5 phrenic nerve suture to the abductor branch of the RLN,6 and even artificial stimulator placement.7 However, all of these studies failed to offer a consistently mobile vocal fold. Several mechanisms have been proposed to explain the low success rate of PCA muscle reanimation. Before attempting PCA muscle animation, patients with cricoarytenoid joint fixation should be excluded.8 In addition to the other disadvantageous factors for PCA reanimation, such as the characteristics of the PCA muscle, time from RLN injury to the reinnervation procedure,9 anastomosis, and muscle fibrosis or atrophy of the Su et al.: Innervation of the PCA Muscle
419
PCA muscle, the communicating branch between the PCA muscle and interarytenoid (IA) muscle (CBPI) might play a major role via the synkinesis between the PCA and IA muscles.4 Attempts to reinnervate the PCA muscle usually ends up by reinnervating simultaneously some or all of the IA muscle. Consequently, the synkinesis between the PCA and IA muscles might eventually cause failure of laryngeal reinnervation. The IA muscle is not the most efficient adductor among the three adductors of the vocal folds, and the PCA muscle is the principal abductor of the vocal cords. Therefore, the strength of the IA muscle in adduction might be not sufficient to oppose the PCA muscle, but this does not adequately explain the unsatisfactory results after reinnervation both clinically and in animal experiments. To understand the reasons for unsatisfactory PCA muscle reanimation more thoroughly, we searched further for the other communicating branches between the PCA muscle and the other adductors (i.e., the lateral cricoarytenoid muscle and the thyroarytenoid muscle). We expected that these communicating branches would be so tiny and deep within the muscles they innervate that they might be found only with the use of special staining and with the aid of a dissecting microscope. We used the modified Sihler’s stain technique10 to analyze these communicating branches, and we describe the intralaryngeal neuroanatomy to explain the possible obstacles to PCA muscle reanimation.
MATERIALS AND METHODS Fifteen larynges were harvested from patients with laryngeal or hypopharyngeal cancer after total laryngectomy. All of these larynges had received concurrent radiochemotherapy with or without induction chemotherapy using a curative dosage of 7,000 cGy before surgery. Larynges showing disruption in the postcricoid region because of pathological processes or the cancer itself were excluded from this study. Nine of 15 patients had a tracheostomy before surgery, which had barely disrupted the postcricoid region. However, four of these nine patients were excluded from this study because of cancer extension or its subsequent pathological processes. Eleven larynges were immersed immediately after removal in 10% un-neutralized formalin. After the pathology department analyzed the cancerous tissue, the remnants of the laryngeal specimens were kept immersed in 10% un-neutralized formalin for another 2 weeks. Reagents were obtained and formulated in our laboratory. The whole larynx, including a portion of the trachea, was prepared using a modified Sihler’s staining technique in seven stages as described below.
Fixation The whole larynx was immersed immediately after removal in 10% un-neutralized formalin for 4 weeks.
Maceration and Depigmentation The fixed specimens were removed from the formalin and washed in distilled water for 30 minutes and then placed in a solution containing 3% aqueous potassium hydroxide with three drops of 3% hydrogen peroxide added to every 100 mL of solution. The specimens were incubated in this solution at room temperature. The solution was changed every second day ini-
Laryngoscope 125: February 2015
420
tially and at least once a week thereafter or whenever the solution became cloudy. Maceration continued for 3 weeks until the tissues became translucent.
Decalcification After washing in distilled water for 30 minutes, the laryngeal specimens were transferred into Sihler’s solution I for 3 weeks to decalcify the specimens. Sihler’s solution I was prepared as follows: one part glacial acetic acid, one part glycerin, and six parts 1% aqueous chloral hydrate. The specimens were transferred to fresh solution every 2 to 3 days.
Staining Following decalcification, each larynx was washed in distilled water for 30 minutes and incubated in Sihler’s solution II. Sihler’s solution II was prepared using one part Ehrlich’s hematoxylin (Lennox Laboratory Supplies, Dublin, Ireland), one part glycerin, and six parts 1% aqueous chloral hydrate. The solution was changed once per week. Each specimen was examined regularly under a dissecting microscope. The staining procedure was continued for at least 4 weeks or until the large nerves within the specimens turned dark purple and the terminal nerve branches were well stained.
Destaining Each larynx was washed in distilled water for 30 minutes and then immersed in Sihler’s solution I to remove excess stain. The destaining procedure was continued for at least 4 weeks.
Clearing Before clearing, specimens were washed in distilled water for 30 minutes. Clearing involved immersing the tissues in increasing concentrations of glycerin (40%, 60%, 80%, and 100%) in the dark. The specimens remained in 40% and 60% glycerin for 2 days, and in 80% and 100% glycerin for 1 day.
TRIMMING Each specimen was divided along the midline, and the hemilaryngeal preparations were dissected carefully to examine the intralaryngeal course of the RLN and its terminal branches under a dissecting microscope. All laryngeal muscles were dissected from their cartilaginous attachments and examined as whole-mount preparations.
RESULTS Twenty-two hemilaryngeal preparations were examined. Two preparations were not eligible for anatomical analysis because of inappropriate staining or disrupted anatomy caused by previous handling, but 20 were stained adequately. In the 20 eligible specimens, the RLN approached the larynx as a single trunk behind the cricothyroid joint. In six of the specimens, one single main branch arose from the RLN and ramified immediately into two secondary branches (Fig. 1A, Fig. 2). These nerves innervated their respective compartments completely without any overlap on one another. In 14 of the specimens, two nerve branches to the PCA muscle arose separately from the RLN (Fig. 1B, Fig. 3). In all Su et al.: Innervation of the PCA Muscle
Fig. 1. Schematic showing the distribution of secondary branching from the RLN. (A) One single main branch arose from the RLN and ramified immediately into two secondary branches in six of the 20 specimens. (B) Two nerve branches to the PCA muscle arose separately from the abductor division in 14 of the 20 specimens.
specimens except one, the nerve branches to the horizontal compartment (upper branch or superior secondary branch) were connected to the nerve to the IA muscle (Fig. 2, white arrow). However, no communicating branches were observed between the nerve branches to the PCA muscle and the nerves to the thyroarytenoid muscle or the lateral cricoarytenoid muscle. No anasto-
Fig. 3. Posterior view of a right human posterior cricoarytenoid muscle. The recurrent laryngeal nerve (R) is seen passing superiorly along the lateral edge of the muscle and the arytenoid cartilage (Ary). Note that the nerve branches to the vertical/oblique and horizontal compartments arise as separate branches (a, b) from the recurrent laryngeal nerve. No communicating nerves from these two separate branches to any other nerve branches were identified.
moses between nerve branches to the horizontal and vertical/oblique compartments or other variations of nerve distribution that were found in Sanders et al.’s study were observed.10 Specifically, the IA muscle received nerve branches from both the RLN and the nerve to the lateral cricoarytenoid muscle (Fig. 2, black arrowhead).
DISCUSSION
Fig. 2. Posterior view of a right human posterior cricoarytenoid muscle (PCA). A single branch (a) arises from the recurrent laryngeal nerve (R) and separates immediately into two secondary branches (b) and (c).The white arrow points to the connection between the nerve branch (c) and the nerve branch (d) to the interarytenoid muscle (IA). A communicating nerve branch (black arrowhead) is also seen between the nerve to the IA muscle and the branch (e) to the lateral cricoarytenoid muscle (LCA).
Laryngoscope 125: February 2015
Most bilateral vocal palsy reported in the past century has resulted from RLN injury during thyroid surgery. Consequently, laryngeal reinnervation has been the initial management following such injury.3 However, this procedure barely achieves consistent vocal abduction, and many modifications pertaining to the anastomosis between the phrenic nerve and the RLN or those abductors have been used.11,12 Several mechanisms have been proposed to explain these unfavorable outcomes. Among them, an inappropriate synkinesis between the IA muscle branch and the CBPI is considered to be one of the most important causes.4 Therefore, a number of methods have been used to investigate the innervation of the PCA muscle over the past decades. These have included cadaver dissection,13 Sihler’s whole-mount nerve staining,10 and electrophysiological methods such as nerve stimulation and electromyography.14 Each of these methods has advantages and disadvantages. Therefore, the Su et al.: Innervation of the PCA Muscle
421
neuroanatomy of the PCA muscle has differed somewhat between these methods, and different explanations have been suggested for the differences in laryngeal reinnervation outcomes. One clinically significant point in understanding the neuromuscular compartments of the PCA muscle and its functions relates to the reinnervation of the paralyzed PCA muscle. The division of the PCA muscle into horizontal and vertical/oblique compartments in humans and dogs is well documented and is based on a large body of evidence, including histochemical and functional evidence.15 The vertical/oblique compartment may be involved primarily in vocal cord abduction during inspiration.10 By contrast, the horizontal compartment may be responsible for stabilizing the arytenoid cartilage during phonation.10 The strong connection between the nerve to the PCA muscle and the nerve to the IA muscle might complicate the reinnervation of the PCA muscle. In some patients, any attempt at reinnervation of the PCA muscle may end up reanimating some or all of the IA muscle. To solve this problem, in patients with two separate branches to the PCA muscle (14 of the 20 in our study), reinnervation of the lower branch with simultaneous destruction of the upper branch may animate the vertical compartment. Similarly, in patients with a single nerve branch to the PCA muscle (six of the 20 in our study), the inferior secondary branch should be reinnervated precisely with the simultaneous destruction of the superior secondary branch or its IA muscle connection. Using the categories described in the necropsy study of Maranillo and Sanudo,13 those patients with separate branches to the PCA muscle (14 of the 20 in our study) would be categorized as displaying pattern III, those with a single branch to the PCA muscle with IA muscle connections (five of the 20 in our study) would be categorized as displaying pattern II, and those with single branch to the PCA muscle without any IA muscle connection (one of the 20 in our study) would be categorized as displaying pattern I. The findings in these two studies appear to be similar because patterns I (7.3%), II (42.7%), and III (34%) represented most (84%) of the 150 PCA muscles examined in the study by Maranillo and Sanudo.13 The secondary branches from the main abductor are usually too small to be traced in the clinic.16 In the PCA muscle reanimating procedure of Li et al.,4 the nerves supplying the PCA muscle were able to be traced only from the RLN. The terminal tracing of the upper and lower branches was not feasible in the study by Li et al. 4 The nerve anastomosis was performed feasibly at the RLN level on the right side and at the adductor level on the left side. In other words, the CBPI may have been intact although the IA muscle branch had been sectioned. Therefore, the synkinesis between the abductor and adductor muscles through the CBPI might have persisted, although they claimed that ligation of the IA muscle branch may have prevented the regrowth of the abductor branch into the proximal end of the sectioned IA muscle branch and consequently prevented synkinesis. Considering what is known about the CBPI, we decided to search further for other communicating Laryngoscope 125: February 2015
422
branches that may exist between the nerves to the PCA muscle and the nerves to the thyroarytenoid or lateral cricoarytenoid muscles. Fortunately, the absence of these communicating branches will not complicate the reinnervation procedure further based on our study. The absence of these communicating branches was also noted by Sanders et al.10 In their study, only one of the 15 PCA muscles showed a neural loop between the nerve branch to the horizontal compartment, the interarytenoid nerve, and the nerve branch to the lateral cricoarytenoid muscle. In addition, a necropsy study of 150 PCA muscles by Maranillo and Sanudo, which did not focus on searching for the communicating branches, also showed a lack of these communicating branches.13 This study and study by Sanders et al.10 using Sihler’s stain had small sample sizes, and further investigation is needed to confirm these findings. One of the variations found in our study is that the IA muscle simultaneously received nerve branches from the RLN and from the laryngeal adductor, but that these had no actual clinical implications for the PCA muscle reinnervation. By contrast, we observed no anastomoses between nerve branches to the horizontal and vertical/oblique compartments or other variations of nerve distribution, which were found in the study by Sanders et al. They found connections between the superior and inferior secondary branches off the RLN in six of 15 specimens. Even after the detailed neuroanatomy of the terminal branches of the RLN has been determined completely, several problems pertaining to PCA muscle reinnervation remain to be solved. First, regardless of whether the CBPI plays a major role in an unsatisfactory reinnervation, microdissection is usually not a reliable means of tracing the neural pathway from the extramuscular to intramuscular terminal branches in the clinic, because the latter are very small and difficult to distinguish from blood vessels and connective tissue. Second, the literature17 and our unpublished clinical data show that a simpler alternative procedure using direct implantation into the vertical/oblique compartments of the PCA muscle with a phrenic nerve usually fails to animate. This implies that the CBPI influences the result or that nerve implantation cannot produce a de novo motor endplate in the PCA muscle as observed for the thyroarytenoid muscle.18 Third, the ratio of fasttwitch to slow-twitch fibers differs between the PCA muscle and the thyroarytenoid muscle and between mammals.4 This difference might make PCA muscle reinnervation by nerve suture or nerve implantation not as desirable as laryngeal adductor reinnervation. If maximal PCA muscle reanimation is attempted, we recommend sectioning of both the adductors to improve the efficacy of the abductor and the upper branch or even the CBPI. In the study by Li et al.,4 the adductors were sectioned bilaterally, but the CBPI was preserved, although they claimed that synkinesis from the IA muscle was prevented by IA muscle branch ligation. Nevertheless, the abduction of the bilateral vocal cords was ultimately successful in their report. This suggests that synkinesis played a relatively minor role in the powerful regenerative abduction by the PCA muscle. Su et al.: Innervation of the PCA Muscle
CONCLUSION This study of the neuroanatomy of the terminal branch of the RLN revealed the lack of communicating branches between the PCA muscle branches and the adductor branches except the IA muscle and confirmed the IA muscle branch as the sole communicating branch responsible for synkinesis of the PCA muscle after PCA muscle reinnervation. Without the interference of other communicating branches, the CBPI still exhibited variability in the many disadvantageous factors for PCA muscle reanimation.
Acknowledgments The study was partially supported by a grant from the Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation TCRD-TPE-101-16.
BIBLIOGRAPHY 1. Jackson C. Ventriculocordectomy, a new operation for the cure of goitrous glottic stenosis. Arch Surg 1922;4:257–274. 2. Holinger L D, Holinger PC, Holinger PH. Etiology of bilateral abductor vocal cord paralysis: a review of 389 cases. Ann Otol Rhinol Laryngol 1976;85:428–436. 3. Crumley RL. Phrenic nerve graft for bilateral vocal cord paralysis. Laryngoscope 1983:93:425–428. 4. Li M, Chen S, Zheng H, et al. Reinnervation of bilateral posterior cricoarytenoid muscles using the left phrenic nerve inpatients with bilateral vocal fold paralysis. PLoS One 2013;8:e77233. 5. Brïndbo K, Hall C, Teig E, et al. Experimental laryngeal reinnervation by phrenic nerve implantation into the posterior cricoarytenoid muscle. Acta Otolaryngol 1987;103:339–344.
Laryngoscope 125: February 2015
6. Tixier C, Reyt E, Mezin P, et al. Selective laryngeal re-innervations in the dog by nervous microsutures of branches of the recurrent laryngeal nerve. Ann Otolaryngol Chir Cervicofac 1992;109:23–31. 7. F€orster G, Arnold D, Bischoff SJ, et al. Laryngeal pacing in minipigs: in vivo test of a new minimal invasive transcricoidal electrode insertion method for functional electrical stimulation of the PCA. Eur Arch Otorhinolaryngol 2013;270:225–231. 8. Yumoto E, Sanuki T, Toya Y, et al. Nerve-muscle pedicle flap implantation combined with arytenoid adduction. Arch Otolaryngol Head Neck Surg 2010;136:965–969. 9. Li M, Chen S, Wang W, et al. Effect of duration of denervation on outcomes of ansa-recurrent laryngeal nerve reinnervation. Laryngoscope 2014;124:1900–1905. 10. Sanders I, Wu BL, Mu L, et al. The innervation of the human posterior cricoarytenoid muscle: evidence for at least two neuromuscular compartments. Laryngoscope 1994;104:880–884. 11. van Lith-Bijl JT, Stolk RJ, Tonnaer JA, et al. Laryngeal abductor reinnervation with a phrenic nerve transfer after a 9-month delay. Arch Otolaryngol Head Neck Surg 1998;124:393–398. 12. Birchall M, Idowu B, Murison P, et al. Laryngeal abductor muscle reinnervation in a pig model. Acta Otolaryngol 2004;124:839–846. 13. Maranillo E, Sanudo JR. Variability of the nerve supply patterns of the human posterior cricoarytenoid muscle. Laryngoscope 2003;113: 602–606. 14. Zealear DL, Swelstad MR, Fortune S, et al. Evoked electromyographic technique for quantitative assessment of the innervation status of laryngeal muscles. Ann Otol Rhinol Laryngol 2005;114:563–572. 15. Asanau A, Timoshenko AP, Prades JM, et al. Posterior cricoarytenoid bellies: relationship between their function and histology. J Voice 2011;25: e67–e73. 16. Mu L, Sanders I. Sihler’s whole mount nerve staining technique: a review. Biotech Histochem 2010;85:19. 17. Tucker HM. Combined surgical medialization and nerve-muscle pedicle reinnervation for unilateral vocal fold paralysis: improved functional results and prevention of long-term deterioration of voice. J Voice 1997; 11:474–478. 18. Miyamaru S, Kumai Y, Ito T, et al. Nerve-muscle pedicle implantation facilitates re-innervation of long-term denervated thyroarytenoid muscle in rats. Acta Otolaryngol 2009;129:1486–1492.
Su et al.: Innervation of the PCA Muscle
423
Rhinological and Otological Society, Inc.
Nerve Branches to the Posterior Cricoarytenoid Muscle May Complicate the Laryngeal Reinnervation Procedure Wan-Fu Su, MD; Shao-Cheng Liu, MD; Shwun-De Wang, PhD; Wang-Yu Su, MD; Kuo-Hsing Ma, PhD; Tung-Tsun Huang, MD Objectives/Hypothesis: To better understand the reason for the low success rate of posterior cricoarytenoid (PCA) muscle reinnervation, we attempted to identify the communicating branches that may exist between the nerve branch to the PCA muscle and the other laryngeal adductors in addition to the interarytenoid (IA) muscle. Study Design: Quantitative histological assessment. Methods: Twenty human hemilarynges from patients with laryngeal or hypopharyngeal cancer were obtained after surgery and stained with Sihler’s stain, which rendered the muscle translucent while counterstaining the neuroanatomy of the recurrent laryngeal nerve (RLN) inside the larynges. Results: The nerve supply to the PCA muscle was separated into two main branches. One upper branch supplied the horizontal compartment, and the lower branch supplied the vertical/oblique compartment. In 14 specimens, two nerve branches to the PCA muscle arose separately from the RLN. In six specimens, one single main branch arose from the RLN and immediately ramified into two secondary branches. In all specimens except one, the nerve branch to the horizontal compartment was connected to the IA muscle. However, no communicating branches were observed between the nerve to the PCA muscle and the other laryngeal adductors. No anastomosis between nerve branches to the horizontal and vertical/ oblique compartments or other variations of nerve distribution were observed. Conclusions: The communicating branches between the nerve branches to the PCA muscle and the IA muscle may be the only branch, complicating the reinnervation procedure. More investigations are needed to identify variations in the terminal branch distribution from the RLN. Key Words: Posterior cricoarytenoid muscle, interarytenoid muscle, communicating branch, recurrent laryngeal nerve, lateral cricoarytenoid muscle. Level of Evidence: NA Laryngoscope, 125:419–423, 2015
INTRODUCTION Most patients with bilateral vocal cord palsy usually experience airway difficulties. Recurrent laryngeal nerve (RLN) injury from thyroid surgery represents the most frequent cause of bilateral vocal cord palsy, followed by intubation injury, cardiovascular accident, radiation therapy, and idiopathic causes.1 Many proce-
From the Department of Otolaryngology–Head and Neck Surgery (W.-F.S., W.-Y.S, T.-T.H.), Buddist Tzu Chi General Hospital, Taipei Branch School of Medicine, Tzu Chi University, Hualien; and the Department of Otolaryngology–Head and Neck Surgery (W.-F.S., S.-C.L.), and the Department of Biology and Anatomy (S.-D.W., K.-H.M.), Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan. Editor’s Note: This Manuscript was accepted for publication September 4, 2014. Wan-Fu Su, MD participated in the study design, writing, literature search, and manuscript review/editing. Shao-Cheng Liu, MD participated in data collection, picture compiling, and article submission. Shwun-De Wang, MD participated as a methodology director. Wang-Yu Su, MD participated in material collection. Kuo-Hsing Ma, PhD participated as a methodology director. Tung-Tsun Huang, MD participated in material collection The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Wan Fu Su, MD, Department of Otolaryngology-Head and Neck Surgery, Tri-Service General Hospital, 325, Sec 2, Chen-Kung Road, Taipei, 114 Taiwan. E-mail: [email protected] DOI: 10.1002/lary.24944
Laryngoscope 125: February 2015
dures have been proposed to resolve the respiratory problems statically or simultaneously, to preserve the phonatory and sphincteric functions dynamically. It is accepted that laser cordectomy, arytenoidectomy, or suture lateralization can resolve the respiratory problems but at the expense of phonatory quality.2 The laryngeal reinnervation of the posterior cricoarytenoid (PCA) muscle has received much attention because of its low success rate. Because a mobile vocal fold functions as a regulator of respiration and may enable maximal preservation of voice quality, many procedures have been investigated to ensure PCA muscle reanimation.3,4 Such investigations have included phrenic nerve implantation into the PCA muscle,5 phrenic nerve suture to the abductor branch of the RLN,6 and even artificial stimulator placement.7 However, all of these studies failed to offer a consistently mobile vocal fold. Several mechanisms have been proposed to explain the low success rate of PCA muscle reanimation. Before attempting PCA muscle animation, patients with cricoarytenoid joint fixation should be excluded.8 In addition to the other disadvantageous factors for PCA reanimation, such as the characteristics of the PCA muscle, time from RLN injury to the reinnervation procedure,9 anastomosis, and muscle fibrosis or atrophy of the Su et al.: Innervation of the PCA Muscle
419
PCA muscle, the communicating branch between the PCA muscle and interarytenoid (IA) muscle (CBPI) might play a major role via the synkinesis between the PCA and IA muscles.4 Attempts to reinnervate the PCA muscle usually ends up by reinnervating simultaneously some or all of the IA muscle. Consequently, the synkinesis between the PCA and IA muscles might eventually cause failure of laryngeal reinnervation. The IA muscle is not the most efficient adductor among the three adductors of the vocal folds, and the PCA muscle is the principal abductor of the vocal cords. Therefore, the strength of the IA muscle in adduction might be not sufficient to oppose the PCA muscle, but this does not adequately explain the unsatisfactory results after reinnervation both clinically and in animal experiments. To understand the reasons for unsatisfactory PCA muscle reanimation more thoroughly, we searched further for the other communicating branches between the PCA muscle and the other adductors (i.e., the lateral cricoarytenoid muscle and the thyroarytenoid muscle). We expected that these communicating branches would be so tiny and deep within the muscles they innervate that they might be found only with the use of special staining and with the aid of a dissecting microscope. We used the modified Sihler’s stain technique10 to analyze these communicating branches, and we describe the intralaryngeal neuroanatomy to explain the possible obstacles to PCA muscle reanimation.
MATERIALS AND METHODS Fifteen larynges were harvested from patients with laryngeal or hypopharyngeal cancer after total laryngectomy. All of these larynges had received concurrent radiochemotherapy with or without induction chemotherapy using a curative dosage of 7,000 cGy before surgery. Larynges showing disruption in the postcricoid region because of pathological processes or the cancer itself were excluded from this study. Nine of 15 patients had a tracheostomy before surgery, which had barely disrupted the postcricoid region. However, four of these nine patients were excluded from this study because of cancer extension or its subsequent pathological processes. Eleven larynges were immersed immediately after removal in 10% un-neutralized formalin. After the pathology department analyzed the cancerous tissue, the remnants of the laryngeal specimens were kept immersed in 10% un-neutralized formalin for another 2 weeks. Reagents were obtained and formulated in our laboratory. The whole larynx, including a portion of the trachea, was prepared using a modified Sihler’s staining technique in seven stages as described below.
Fixation The whole larynx was immersed immediately after removal in 10% un-neutralized formalin for 4 weeks.
Maceration and Depigmentation The fixed specimens were removed from the formalin and washed in distilled water for 30 minutes and then placed in a solution containing 3% aqueous potassium hydroxide with three drops of 3% hydrogen peroxide added to every 100 mL of solution. The specimens were incubated in this solution at room temperature. The solution was changed every second day ini-
Laryngoscope 125: February 2015
420
tially and at least once a week thereafter or whenever the solution became cloudy. Maceration continued for 3 weeks until the tissues became translucent.
Decalcification After washing in distilled water for 30 minutes, the laryngeal specimens were transferred into Sihler’s solution I for 3 weeks to decalcify the specimens. Sihler’s solution I was prepared as follows: one part glacial acetic acid, one part glycerin, and six parts 1% aqueous chloral hydrate. The specimens were transferred to fresh solution every 2 to 3 days.
Staining Following decalcification, each larynx was washed in distilled water for 30 minutes and incubated in Sihler’s solution II. Sihler’s solution II was prepared using one part Ehrlich’s hematoxylin (Lennox Laboratory Supplies, Dublin, Ireland), one part glycerin, and six parts 1% aqueous chloral hydrate. The solution was changed once per week. Each specimen was examined regularly under a dissecting microscope. The staining procedure was continued for at least 4 weeks or until the large nerves within the specimens turned dark purple and the terminal nerve branches were well stained.
Destaining Each larynx was washed in distilled water for 30 minutes and then immersed in Sihler’s solution I to remove excess stain. The destaining procedure was continued for at least 4 weeks.
Clearing Before clearing, specimens were washed in distilled water for 30 minutes. Clearing involved immersing the tissues in increasing concentrations of glycerin (40%, 60%, 80%, and 100%) in the dark. The specimens remained in 40% and 60% glycerin for 2 days, and in 80% and 100% glycerin for 1 day.
TRIMMING Each specimen was divided along the midline, and the hemilaryngeal preparations were dissected carefully to examine the intralaryngeal course of the RLN and its terminal branches under a dissecting microscope. All laryngeal muscles were dissected from their cartilaginous attachments and examined as whole-mount preparations.
RESULTS Twenty-two hemilaryngeal preparations were examined. Two preparations were not eligible for anatomical analysis because of inappropriate staining or disrupted anatomy caused by previous handling, but 20 were stained adequately. In the 20 eligible specimens, the RLN approached the larynx as a single trunk behind the cricothyroid joint. In six of the specimens, one single main branch arose from the RLN and ramified immediately into two secondary branches (Fig. 1A, Fig. 2). These nerves innervated their respective compartments completely without any overlap on one another. In 14 of the specimens, two nerve branches to the PCA muscle arose separately from the RLN (Fig. 1B, Fig. 3). In all Su et al.: Innervation of the PCA Muscle
Fig. 1. Schematic showing the distribution of secondary branching from the RLN. (A) One single main branch arose from the RLN and ramified immediately into two secondary branches in six of the 20 specimens. (B) Two nerve branches to the PCA muscle arose separately from the abductor division in 14 of the 20 specimens.
specimens except one, the nerve branches to the horizontal compartment (upper branch or superior secondary branch) were connected to the nerve to the IA muscle (Fig. 2, white arrow). However, no communicating branches were observed between the nerve branches to the PCA muscle and the nerves to the thyroarytenoid muscle or the lateral cricoarytenoid muscle. No anasto-
Fig. 3. Posterior view of a right human posterior cricoarytenoid muscle. The recurrent laryngeal nerve (R) is seen passing superiorly along the lateral edge of the muscle and the arytenoid cartilage (Ary). Note that the nerve branches to the vertical/oblique and horizontal compartments arise as separate branches (a, b) from the recurrent laryngeal nerve. No communicating nerves from these two separate branches to any other nerve branches were identified.
moses between nerve branches to the horizontal and vertical/oblique compartments or other variations of nerve distribution that were found in Sanders et al.’s study were observed.10 Specifically, the IA muscle received nerve branches from both the RLN and the nerve to the lateral cricoarytenoid muscle (Fig. 2, black arrowhead).
DISCUSSION
Fig. 2. Posterior view of a right human posterior cricoarytenoid muscle (PCA). A single branch (a) arises from the recurrent laryngeal nerve (R) and separates immediately into two secondary branches (b) and (c).The white arrow points to the connection between the nerve branch (c) and the nerve branch (d) to the interarytenoid muscle (IA). A communicating nerve branch (black arrowhead) is also seen between the nerve to the IA muscle and the branch (e) to the lateral cricoarytenoid muscle (LCA).
Laryngoscope 125: February 2015
Most bilateral vocal palsy reported in the past century has resulted from RLN injury during thyroid surgery. Consequently, laryngeal reinnervation has been the initial management following such injury.3 However, this procedure barely achieves consistent vocal abduction, and many modifications pertaining to the anastomosis between the phrenic nerve and the RLN or those abductors have been used.11,12 Several mechanisms have been proposed to explain these unfavorable outcomes. Among them, an inappropriate synkinesis between the IA muscle branch and the CBPI is considered to be one of the most important causes.4 Therefore, a number of methods have been used to investigate the innervation of the PCA muscle over the past decades. These have included cadaver dissection,13 Sihler’s whole-mount nerve staining,10 and electrophysiological methods such as nerve stimulation and electromyography.14 Each of these methods has advantages and disadvantages. Therefore, the Su et al.: Innervation of the PCA Muscle
421
neuroanatomy of the PCA muscle has differed somewhat between these methods, and different explanations have been suggested for the differences in laryngeal reinnervation outcomes. One clinically significant point in understanding the neuromuscular compartments of the PCA muscle and its functions relates to the reinnervation of the paralyzed PCA muscle. The division of the PCA muscle into horizontal and vertical/oblique compartments in humans and dogs is well documented and is based on a large body of evidence, including histochemical and functional evidence.15 The vertical/oblique compartment may be involved primarily in vocal cord abduction during inspiration.10 By contrast, the horizontal compartment may be responsible for stabilizing the arytenoid cartilage during phonation.10 The strong connection between the nerve to the PCA muscle and the nerve to the IA muscle might complicate the reinnervation of the PCA muscle. In some patients, any attempt at reinnervation of the PCA muscle may end up reanimating some or all of the IA muscle. To solve this problem, in patients with two separate branches to the PCA muscle (14 of the 20 in our study), reinnervation of the lower branch with simultaneous destruction of the upper branch may animate the vertical compartment. Similarly, in patients with a single nerve branch to the PCA muscle (six of the 20 in our study), the inferior secondary branch should be reinnervated precisely with the simultaneous destruction of the superior secondary branch or its IA muscle connection. Using the categories described in the necropsy study of Maranillo and Sanudo,13 those patients with separate branches to the PCA muscle (14 of the 20 in our study) would be categorized as displaying pattern III, those with a single branch to the PCA muscle with IA muscle connections (five of the 20 in our study) would be categorized as displaying pattern II, and those with single branch to the PCA muscle without any IA muscle connection (one of the 20 in our study) would be categorized as displaying pattern I. The findings in these two studies appear to be similar because patterns I (7.3%), II (42.7%), and III (34%) represented most (84%) of the 150 PCA muscles examined in the study by Maranillo and Sanudo.13 The secondary branches from the main abductor are usually too small to be traced in the clinic.16 In the PCA muscle reanimating procedure of Li et al.,4 the nerves supplying the PCA muscle were able to be traced only from the RLN. The terminal tracing of the upper and lower branches was not feasible in the study by Li et al. 4 The nerve anastomosis was performed feasibly at the RLN level on the right side and at the adductor level on the left side. In other words, the CBPI may have been intact although the IA muscle branch had been sectioned. Therefore, the synkinesis between the abductor and adductor muscles through the CBPI might have persisted, although they claimed that ligation of the IA muscle branch may have prevented the regrowth of the abductor branch into the proximal end of the sectioned IA muscle branch and consequently prevented synkinesis. Considering what is known about the CBPI, we decided to search further for other communicating Laryngoscope 125: February 2015
422
branches that may exist between the nerves to the PCA muscle and the nerves to the thyroarytenoid or lateral cricoarytenoid muscles. Fortunately, the absence of these communicating branches will not complicate the reinnervation procedure further based on our study. The absence of these communicating branches was also noted by Sanders et al.10 In their study, only one of the 15 PCA muscles showed a neural loop between the nerve branch to the horizontal compartment, the interarytenoid nerve, and the nerve branch to the lateral cricoarytenoid muscle. In addition, a necropsy study of 150 PCA muscles by Maranillo and Sanudo, which did not focus on searching for the communicating branches, also showed a lack of these communicating branches.13 This study and study by Sanders et al.10 using Sihler’s stain had small sample sizes, and further investigation is needed to confirm these findings. One of the variations found in our study is that the IA muscle simultaneously received nerve branches from the RLN and from the laryngeal adductor, but that these had no actual clinical implications for the PCA muscle reinnervation. By contrast, we observed no anastomoses between nerve branches to the horizontal and vertical/oblique compartments or other variations of nerve distribution, which were found in the study by Sanders et al. They found connections between the superior and inferior secondary branches off the RLN in six of 15 specimens. Even after the detailed neuroanatomy of the terminal branches of the RLN has been determined completely, several problems pertaining to PCA muscle reinnervation remain to be solved. First, regardless of whether the CBPI plays a major role in an unsatisfactory reinnervation, microdissection is usually not a reliable means of tracing the neural pathway from the extramuscular to intramuscular terminal branches in the clinic, because the latter are very small and difficult to distinguish from blood vessels and connective tissue. Second, the literature17 and our unpublished clinical data show that a simpler alternative procedure using direct implantation into the vertical/oblique compartments of the PCA muscle with a phrenic nerve usually fails to animate. This implies that the CBPI influences the result or that nerve implantation cannot produce a de novo motor endplate in the PCA muscle as observed for the thyroarytenoid muscle.18 Third, the ratio of fasttwitch to slow-twitch fibers differs between the PCA muscle and the thyroarytenoid muscle and between mammals.4 This difference might make PCA muscle reinnervation by nerve suture or nerve implantation not as desirable as laryngeal adductor reinnervation. If maximal PCA muscle reanimation is attempted, we recommend sectioning of both the adductors to improve the efficacy of the abductor and the upper branch or even the CBPI. In the study by Li et al.,4 the adductors were sectioned bilaterally, but the CBPI was preserved, although they claimed that synkinesis from the IA muscle was prevented by IA muscle branch ligation. Nevertheless, the abduction of the bilateral vocal cords was ultimately successful in their report. This suggests that synkinesis played a relatively minor role in the powerful regenerative abduction by the PCA muscle. Su et al.: Innervation of the PCA Muscle
CONCLUSION This study of the neuroanatomy of the terminal branch of the RLN revealed the lack of communicating branches between the PCA muscle branches and the adductor branches except the IA muscle and confirmed the IA muscle branch as the sole communicating branch responsible for synkinesis of the PCA muscle after PCA muscle reinnervation. Without the interference of other communicating branches, the CBPI still exhibited variability in the many disadvantageous factors for PCA muscle reanimation.
Acknowledgments The study was partially supported by a grant from the Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation TCRD-TPE-101-16.
BIBLIOGRAPHY 1. Jackson C. Ventriculocordectomy, a new operation for the cure of goitrous glottic stenosis. Arch Surg 1922;4:257–274. 2. Holinger L D, Holinger PC, Holinger PH. Etiology of bilateral abductor vocal cord paralysis: a review of 389 cases. Ann Otol Rhinol Laryngol 1976;85:428–436. 3. Crumley RL. Phrenic nerve graft for bilateral vocal cord paralysis. Laryngoscope 1983:93:425–428. 4. Li M, Chen S, Zheng H, et al. Reinnervation of bilateral posterior cricoarytenoid muscles using the left phrenic nerve inpatients with bilateral vocal fold paralysis. PLoS One 2013;8:e77233. 5. Brïndbo K, Hall C, Teig E, et al. Experimental laryngeal reinnervation by phrenic nerve implantation into the posterior cricoarytenoid muscle. Acta Otolaryngol 1987;103:339–344.
Laryngoscope 125: February 2015
6. Tixier C, Reyt E, Mezin P, et al. Selective laryngeal re-innervations in the dog by nervous microsutures of branches of the recurrent laryngeal nerve. Ann Otolaryngol Chir Cervicofac 1992;109:23–31. 7. F€orster G, Arnold D, Bischoff SJ, et al. Laryngeal pacing in minipigs: in vivo test of a new minimal invasive transcricoidal electrode insertion method for functional electrical stimulation of the PCA. Eur Arch Otorhinolaryngol 2013;270:225–231. 8. Yumoto E, Sanuki T, Toya Y, et al. Nerve-muscle pedicle flap implantation combined with arytenoid adduction. Arch Otolaryngol Head Neck Surg 2010;136:965–969. 9. Li M, Chen S, Wang W, et al. Effect of duration of denervation on outcomes of ansa-recurrent laryngeal nerve reinnervation. Laryngoscope 2014;124:1900–1905. 10. Sanders I, Wu BL, Mu L, et al. The innervation of the human posterior cricoarytenoid muscle: evidence for at least two neuromuscular compartments. Laryngoscope 1994;104:880–884. 11. van Lith-Bijl JT, Stolk RJ, Tonnaer JA, et al. Laryngeal abductor reinnervation with a phrenic nerve transfer after a 9-month delay. Arch Otolaryngol Head Neck Surg 1998;124:393–398. 12. Birchall M, Idowu B, Murison P, et al. Laryngeal abductor muscle reinnervation in a pig model. Acta Otolaryngol 2004;124:839–846. 13. Maranillo E, Sanudo JR. Variability of the nerve supply patterns of the human posterior cricoarytenoid muscle. Laryngoscope 2003;113: 602–606. 14. Zealear DL, Swelstad MR, Fortune S, et al. Evoked electromyographic technique for quantitative assessment of the innervation status of laryngeal muscles. Ann Otol Rhinol Laryngol 2005;114:563–572. 15. Asanau A, Timoshenko AP, Prades JM, et al. Posterior cricoarytenoid bellies: relationship between their function and histology. J Voice 2011;25: e67–e73. 16. Mu L, Sanders I. Sihler’s whole mount nerve staining technique: a review. Biotech Histochem 2010;85:19. 17. Tucker HM. Combined surgical medialization and nerve-muscle pedicle reinnervation for unilateral vocal fold paralysis: improved functional results and prevention of long-term deterioration of voice. J Voice 1997; 11:474–478. 18. Miyamaru S, Kumai Y, Ito T, et al. Nerve-muscle pedicle implantation facilitates re-innervation of long-term denervated thyroarytenoid muscle in rats. Acta Otolaryngol 2009;129:1486–1492.
Su et al.: Innervation of the PCA Muscle
423
