(1)
Department of Otlaryngology Head and Neck Surgery, Graduate School of Medical Sciences Kumamoto University, Kumamoto, Japan
Abstract
Vocal fold paralysis is a pathological finding of motion impairment of varying degrees caused by nervous system disorders. In this chapter, the detailed anatomy of the laryngeal nerves and their varying anastomotic patterns within the larynx are described. Regeneration of nerve fibers after damage varies among patients. Significant misdirected regeneration occurs in some, whereas others show little synkinesis, which is a major cause of considerable inconsistencies of the vibratory pattern of the vocal folds among patients with unilateral vocal fold paralysis. Second, the effects of changes in the physical properties of the affected vocal fold on the production of the mucosal wave are summarized. Third, the factors that must be considered before surgical treatment of paralytic dysphonia are highlighted. Finally, reacquisition of the thyroarytenoid muscle tonus by reinnervation is important in recovering the preinjury voice with a dynamic mucosal wave.
Keywords
Anastomosis of the laryngeal nervesMisdirected reinnervationStiffness of the vocal foldSynkinesis1.1 What Is “Vocal Fold Paralysis”?
Vocal fold immobility is caused by various disorders, including damage of the recurrent laryngeal nerve (RLN), vagus nerve, or motor neurons and central nervous system, fixation or subluxation of the cricoarytenoid joint, adhesion of the posterior part of the vocal folds, tumor invasion into the intrinsic laryngeal muscle(s), and severe laryngeal edema. This book deals with vocal fold immobility and motion impairment due to nervous system disorders, with an emphasis on diagnosis of the pathophysiology of vocal fold paralysis and treatment aimed at reinnervation of the intrinsic laryngeal muscles.
“Recurrent laryngeal nerve paralysis (RLNP)” has been often used in clinical settings to indicate immobile vocal fold due to nerve damage. However, the vocal fold on the affected side is not always “fixed” and usually presents with slight movement during phonation and inhalation, although whether such motion is active or passive remains to be clarified. The position of the affected vocal fold ranges from median or paramedian to intermediate. The position sometimes shifts from an intermediate to a median position as time elapses. As described later, electromyography (EMG) studies of most patients diagnosed with “RLNP” revealed the presence of laryngeal muscle activities, although they vary among individuals. The clinical diagnosis of “RLNP” includes varying degrees of nerve–muscle impairment: complete denervation of the superior and recurrent laryngeal nerves, resulting in total paralysis of the laryngeal muscles at one extreme and, as in most cases, partial denervation and reinnervation of the laryngeal muscles at the other extreme. Aberrant reinnervation may occur. Furthermore, the superior laryngeal nerve, including the internal and external branches and other lower cranial nerves, may be injured altogether (Fig. 1.1). Considering these issues, the term “vocal fold paralysis1 (VFP)” rather than “RLNP” is used in this book.
Fig. 1.1
Combinations of injured nerves in vocal fold paralysis
1.2 Anastomotic Patterns of the RLN and Its Regeneration After Damage
The recurrent laryngeal nerve (RLN) is a branch of the vagal nerve and contains both motor and sensory fibers. After branching out from the vagal nerve, the RLN runs below the aortic arch posteriorly on the left side and below the subclavian artery posteriorly on the right side to proceed cranially along the tracheoesophageal groove; finally, it enters the larynx through the space between the thyroid and cricoid cartilages. Because the RLN runs a long distance due to its departure from the skull base to the larynx, the nerve can be damaged by various types of disease at different levels. Sensory fibers innervate the mucosa of the trachea and esophagus. A sensory branch called “Galen’s branch” separates just before entering the larynx and makes an anastomosis with the internal branch of the superior laryngeal nerve (SLN) [1, 2]. Thus, the RLN after division of Galen’s branch contains motor fibers. These fibers innervate the adductor muscles (thyroarytenoid muscle (TA), lateral cricoarytenoid muscle (LCA), arytenoid muscle (AR), and the abductor muscle (posterior cricoarytenoid muscle (PCA). Figure 1.2 shows schematically the arrangement of the intrinsic laryngeal muscles, including the cricothyroid muscle (CT), which is innervated by the SLN. The medial bundle of the thyroarytenoid muscle is called the “vocalis muscle” because of its significant role in the production and fine-tuning of the voice.
Fig. 1.2
Schematic illustration of the location of the intrinsic laryngeal muscles. The medial bundle of the thyroarytenoid muscle is called the “vocalis muscle”
The SLN branches from the vagal nerve just below the inferior ganglion (i.e., the nodose ganglion), descends medially along the internal carotid artery, and then divides into internal and external branches. The internal branch (iSLN) contains mostly sensory fibers and enters the larynx through the thyrohyoid membrane or thyroid foramen (less often). The external branch (eSLN), which contains mainly motor fibers, descends along the connective tissue sheath containing the superior thyroid vessels and innervates the CT muscle.
The apparent position of the affected vocal fold varies among patients. As revealed by animal experiments [3–6] and clinical experiences [7–11], the laryngeal muscles almost always receive reinnervation after complete severance or even removal of a certain length of the RLN. Possible origins of regenerating nerve fibers have been reported not only from the RLN but also from branches of the vagal nerve to hypopharyngeal constrictor muscles, iSLN, eSLN, and autonomic nerves [2, 12–18]. Nasri et al. confirmed the presence of the motor nerve supply from eSLN to the TA muscle electromyographically in three of seven canine models [19]. Sanders et al. investigated the gross anatomy and interactions between SLN and RLN in human larynges [14]. They found connections between these nerves, with the exception of Galen’s anastomosis: (1) the RLN combines with the SLN in the neural plexus at the AR muscle; (2) the AR branches of the bilateral RLNs anastomose with each other; (3) the eSLN passes from the CT muscle and has a connection with the TA muscle branch of the RLN; and (4) the PCA branch of the RLN interacts with the AR branch. They indicated that there are significant neural connections between the RLN and SLN and that limited cross-innervation is seen from side to side in the area of the AR muscle. However, some connections were variable among the larynges, suggesting that the regeneration pattern after RLN injury would be determined individually considering these variations.
Semon reported a paralyzed vocal fold positioned at the median or paramedian position at the onset of VFP that gradually changed in position laterally. He explained this shift by assuming that the nerve fibers supplying the PCA muscle are more sensitive to injury than are the nerve fibers to the adductor muscles [20]. However, a time-dependent shift in a paralyzed vocal fold from a median to an intermediate position is rarely observed in a clinical setting as reported by Faaborg-Anderson [21]. The Wagner–Grossman theory2 has also been abandoned following recent experimental and clinical studies [6, 11, 23, 24].
Sanudo et al. scrutinized nerve connections in the human larynx and found four different anastomoses between the RLN and iSLN: (1) Galen’s anastomosis; (2) arytenoid plexus, including connections between the dorsal branch of the iSLN and ventral branch of the RLN in the mucosa and the AR muscle and connections to the contralateral side; (3) cricoid anastomosis between the arytenoid plexus and RLN branch before entering the larynx; and (4) thyroarytenoid anastomosis between the descending branch of the iSLN and ascending branch of the RLN. Anastomosis between the eSLN and RLN appears as a connecting branch throughout the CT muscle [2]. The various prevalences of the anastomotic patterns of this complex suggest functional differences in the regeneration of the sensory and motor fibers among individual subjects. Maranillo et al. also reported that some of the LCA and TA muscles receive a nerve supply from the iSLN or eSLN in addition to the RLN [25]. These anastomotic patterns vary among individuals and between the sides. When a motor fiber is transected, adjacent fibers sprout, extend into the remaining Schwann sheath, and reach a muscle to reinnervate before it falls into degeneration [26]. When the RLN is injured, nerve fiber regeneration occurs through the abovementioned anastomotic routes to supply motor fibers to each muscle. However, such reinnervation patterns differ among individuals. Regeneration is not limited to motor fibers but may also include autonomic systems. Thus, it seems there is no definite rule regarding what determines the position of the affected vocal fold.
Damrose et al. electrically stimulated the RLN when they performed arytenoid adduction on 15 patients and failed to detect any evoked potentials from the TA muscle. They concluded that little regeneration occurs through the RLN after its damage [27]. However, because surface electrodes on the skin were applied in 11 of 15 patients in their study, electrical activities of relatively small amplitude may have not been detected. It is generally accepted that the contraction strength and the timing of each laryngeal muscle influence the position of the “paralyzed” vocal fold. Another factor that may affect vocal fold position is the expiratory airflow. The expiratory airflow pushes the affected vocal fold upward and laterally when TA muscle contraction is insufficient.
Despite the accumulated evidence that the intrinsic laryngeal muscles receive a regenerated nerve supply in both animal models and patients, why physiological vocal fold movement fails to recover fully remains to be determined. Misdirected regeneration or synkinesis is considered to be the main cause [11, 16, 24, 28–37]. Many sprouting fibers grow from the central end of the nerve sheath after transection, and those that reach the myelin sheath on the peripheral side extend to make connections with target muscle fibers. Such growth of nerve fibers is very active, and their extension shows no target organ specificity [38]. The RLN innervates both adductor and abductor muscles, and fibers toward these muscles intermingle in the main nerve until it branches [26, 39–41]. Thus, the regenerated fibers have very little opportunity to reach and innervate the original target muscle. Further, a single nerve fiber may innervate both adductor and abductor muscle fibers (Fig. 1.3). The RLN contains not only motor fibers but also sensory and autonomic components [42], and regenerated fibers may connect with these non-motor fibers and vice versa. Second, only a few nerve fibers successfully reach the myelin sheath on the peripheral side, resulting in a decrease in the number of reinnervated muscle fibers [43]. Third, sufficient muscle contraction fails to recover even after reinnervation because laryngeal muscles degenerate during periods of denervation, accompanied by atrophic and fibrotic changes (Fig. 1.4) [44].
Fig. 1.3
Regenerated nerve fibers after denervation, showing possible misdirections. Regenerated fibers extend toward an original target muscle as well as its opponent muscle
Fig. 1.4
Possible causes of the failure of physiological vocal fold movement to recover after injury. Items 1–3 indicate factors related to misdirected regeneration of nerve fibers, and items 4 and 5 indicate those unrelated to misdirected regeneration
1.3 Symptoms
Unilateral vocal fold paralysis (UVFP) causes breathy hoarseness of varying degrees from a subtle change to a nearly aphonic state and swallowing difficulties. When the affected vocal fold is located off the midline, expectoration of food residue, sputum, and saliva in the supraglottic space and pyriform sinus becomes less effective because of insufficient elevation of subglottic pressure during coughing [45, 46]. Furthermore, latent penetration into the trachea may occur due to diminished sensation of laryngeal mucosa [47–49]. Unless other lower cranial nerve injuries are involved, dysphagia is usually mild and resolves in 10–20 days after onset. On the other hand, breathy dysphonia interferes with verbal communication, resulting in difficulty in social activities and physical symptoms such as fatigue, exhaustion, or numbness of the body due to hyperventilation after conversation. These symptoms definitely reduce the quality of life in patients with UVFP [50–53].
The symptoms of bilateral vocal fold paralysis are different from those of UVFP and depend mainly on the fixed position and residual motility of the vocal folds. When the glottis is wide (i.e., the vocal folds are fixed at an intermediate position), patients complain of breathy dysphonia without respiratory distress. When the glottis is narrow (i.e., the vocal folds are fixed in a median or paramedian position), inspiratory dyspnea is the main concern, although the voice is not hoarse. When the vocal fold positions are paramedian with slight vocal fold movement, the patients remain asymptomatic. Those with bilateral vocal fold paralysis who visit the ear, nose, and throat clinic are often indicated for tracheostomy or surgery to widen the glottis for relief of exertional dyspnea.
1.4 Vocal Fold Vibration
Considerable inconsistencies regarding the vibratory pattern of the vocal fold among patients with UVFP exist. Stroboscopic examination reveals that the closed phase during a vibratory cycle becomes very short or even absent. A mucosal wave may be irregular among cycles or asymmetric between the vocal folds due to the differences in the mass, tension, and stiffness of the vocal folds. Furthermore, stroboscopic illumination may not show vibratory images of the vocal folds when the acoustic signal of the voice does not contain periodic signals, specifically when the patient’s voice is extremely breathy. On aerodynamic examination, the maximum phonation time (MPT) decreases, the mean airflow rate (MFR) increases, and laryngeal efficiency such as the vocal efficiency index [54] is diminished. The auditory impression of the voice is characteristic in that breathiness is enhanced according to the GRBAS scale [55]. A normal vocal fold vibration is a result of vocal folds positioned midline, pliable vocal fold mucosa, and symmetrical physical properties such as shape, thickness, mass, stiffness, and tension. In patients with UVFP, in addition to the off-midline position of the vocal fold, lowered stiffness and tension, reduced thickness of the vocal fold, atrophic changes of the TA muscle, and synkinetic movement of the vocal fold on the affected side cause negative effects on vocal fold vibration.
1.4.1 Glottal Insufficiency
When normal vocal fold vibration begins in a soft attack mode, a slight prephonatory glottal gap exists; this closes following elapse of a few vibratory cycles [56]. However, the glottal gap remains even after several periods in patients with UVFP. Thus, glottal airflow is constantly leaked excessively through the glottis, which makes airflow turbulent rather than laminar, resulting in high-frequency noise (Fig. 1.5).
Fig. 1.5
Schematic illustration of the coronal section of the human vocal fold. Left: without contraction of the thyroarytenoid muscle; Right: with contraction of the thyroarytenoid muscle. Relative to actual thicknesses, the lamina propria is enlarged to facilitate understanding. Subsequent figures are drawn in a similar way
1.4.2 Atrophy of the Thyroarytenoid Muscle
Kobayashi et al. examined the effects of unilateral atrophy of the TA muscle on its vibratory mode in excised canine larynges [57]. The unilateral RLN was removed for a distance of 10 mm, and its proximal stump was sutured into the sternothyroid muscle to prevent reinnervation. The larynx was harvested 4 months after the procedure to allow the muscle to atrophy. Both vocal processes were sutured together to attain posterior glottal closure, and the larynx was fixed in a custom-made apparatus. The thyroid ala of the atrophied side was pressed medially to obtain sufficient glottal closure for steady phonation. Lateral and vertical displacements were measured simultaneously during phonation. Kobayashi et al. reported three main findings: (1) the vibration in the unilaterally atrophied larynges was periodical and symmetrical in phase, (2) the lateral amplitude was significantly greater than the vertical amplitude on both sides, and (3) the lateral and vertical amplitudes on the atrophied side were significantly greater than those on the normal side (Figs. 1.6 and 1.7) [57]. Thus, in the minimal prephonatory glottal gap, periodic vibration occurs in unilaterally atrophied larynges, and the amplitude is greater in the lateral and vertical directions than in the normal fold. The latter finding implies that phonosurgical procedures aimed at closure of the prephonatory glottal gap may have a beneficial effect on hoarseness in UVFP patients, although displacements of the vocal folds during vibration are not symmetrical.
Fig. 1.6
Lateral and vertical amplitudes in the four unilaterally atrophied larynges. The horizontal axis represents the airflow rate (liters per minute), and the vertical axis represents lateral (solid line) and vertical (dotted line) amplitudes. Circles represent the amplitude on the atrophied side; squares represent the amplitude on the normal side (Citation: Ref. [57])
Fig. 1.7
Lissajous trajectories of the vocal fold in a representative case when the larynx was blown at three different airflow rates. The horizontal and vertical axes represent lateral and vertical displacements of the vocal fold in millimeters, respectively (Citation: Ref. [57])
1.4.3 Lowered Stiffness and Tension of the Vocal Fold
The human vocal fold comprises the muscle layer (medial part of the TA muscle, vocalis muscle), vocal ligament, and mucosa (which consists of a very thin epithelial layer and a loose superficial layer of the lamina propria) (Fig. 1.5, left). Mechanical properties vary depending on the degree of contraction of the vocalis muscle. The vocal ligament consists of deep and intermediate layers of the lamina propria. The former is composed of dense fibrous tissue and is tightly connected with the muscle layer. The latter contains elastic and collagen fibers and functions as a transitional zone from the muscle layer to the superficial layer. The mucosa is connected loosely with the vocal ligament and can be moved in a different manner from that of the muscle layer. Hirano reported that the vocal fold could be regarded as a double-structured vibrator consisting of a body comprising the vocalis muscle and vocal ligament and a cover consisting of the epithelium and loose superficial layer of the lamina propria. A mucosal wave observed during phonation is elicited primarily by the cover [58].
When the CT muscle contracts, the vocal fold is elongated and becomes thinner, resulting in the elevation of the vocal fold tension. Such changes in tension produce a higher-pitch voice. On the other hand, contraction of the TA muscle causes bulging of the medial–inferior surface of the vocal fold and elevation of stiffness of the body of the vocal fold. The author and his colleagues investigated the effect of TA muscle contraction on the shape of the subglottic vault using a canine model [59]. The subglottic vault surrounded by the lower surface of the vocal fold formed a concave shape in the absence of TA muscle contraction and became convex during TA muscle contraction. The cover becomes relatively less stiff—i.e., more pliable—resulting in the more vigorous occurrence of a mucosal wave. In the human larynx, the cover is much thinner than the canine; thus, during moderate contraction of the CT muscle, TA muscle contraction plays a greater role in adjusting the stiffness of the whole vocal fold. During strong CT contraction, the cover is stretched, and its tension increases, diminishing the influence of TA muscle activity in the overall stiffness of the vibrating part of the vocal fold [60–63]. Sercars et al. reported a case of UVFP producing a voice with higher activities of the CT muscle [64]. They inferred that the strong CT muscle contraction elevated the tension of the cover and minimized the imbalance of stiffness between the vocal folds. As a result of such laryngeal adjustment, airflow escape during phonation decreases, and the voice becomes high pitched. Such compensatory phonation is termed “paralytic falsetto.”
Paralysis of the CT muscle produces asymmetry of vocal fold tension in the anterior–posterior direction. Based on animal studies, Tanabe et al. and Isshiki et al. found that, in such cases with a closed glottis, the vocal fold vibration is periodic with the vocal fold on the active side preceding the paralyzed side in the vibratory cycle, and the generated sound is not hoarse. With a wider glottal gap, glottal closure becomes irregular with less periodic vibration, and the voice is hoarse [65, 66]. Maunsell et al. reported similar results and theorized that a “coupling effect” occurs between the asymmetric vocal folds despite the presence of imbalance in tension, stiffness, and mass when disparities are insignificant [67]. They also reported that as airflow increases, the phase difference between the vocal folds decreases.
The author observed vocal fold vibration of the excised canine larynges from the tracheal side using a high-speed movie technique and found that the mucosal upheaval (MU), where the mucosal wave starts and propagates upward, appears between the anterior commissure and vocal process on the lower surface of the vocal fold when the vocal fold vibrates [68–71]. The pliability of the vocal fold mucosa was measured along the superior–inferior axis in the canine larynx. Pliability was defined as the maximal elevated distance in response to a constant focal negative pressure (−300 mmHg) applied through a 22-gauge needle with a flat tip (Fig. 1.8) [72]. After confirming contact of the needle tip with the mucosal surface under the operating microscope, the needle tip was slowly moved away using a micromanipulator. Thus, the mucosa was pulled off its original position by the needle tip until it was detached from the mucosa. These two critical moments when the mucosa and needle tip come into contact and become separated were carefully observed, and the distance between these two points was measured by the micromanipulator to an accuracy of 10 μm. It was revealed that the MU where the mucosal wave starts occurs around the point of minimal pliability [72]. Histological examination revealed that the MU occurred where the lamina propria became thinner and where the muscle layer neared the epithelial layer. When the vocal fold is lengthened, the MU actually shifts medially to the area where the lamina propria becomes thicker and where the free edge of the vocal fold is closer [69, 70].
Conversely, contraction of the TA muscle caused the lower surface of the vocal fold to bulge medially and to narrow the subglottic vault surrounded by the bilateral MU. The MU moved to the tracheal side during TA muscle contraction compared with its initial position when the TA muscle is relaxed. Therefore, TA muscle contraction causes expansion of the vibrating area toward the tracheal side and contributes to the production of a more dynamic mucosal wave in the vertical direction (Fig. 1.9) [59, 63, 71]. Although the cover of the human vocal fold is thinner than that of the canine, observation of the lower surface of the vocal folds through a tracheostomy confirmed the presence of MU in humans and that the mucosal wave starts at this point and propagates upward [73, 74]. Insufficient activities of the TA muscle due to UVFP may impair the development of MU, and the mucosal wave is not generated dynamically. Therefore, the patient’s voice becomes weak with little strength in addition to breathiness due to air escape through the glottis.
