Introduction

Bilateral vocal fold paralysis (BVFP), which is caused by injury to the bilateral recurrent laryngeal nerves (RLNs), is a potentially fatal condition that usually presents as progressive airway compromise with or without obvious voice changes. The traditional treatments for BVFP include cordectomy, laser posterior cordectomy, arytenoidectomy and laterofixation of the vocal folds, tracheotomy, and tracheostomy. All are effective in improving airway compromise at the cost of possible deterioration of the voice and aspiration. Hence, an ideal treatment should aim at re-establishment of physiologic inspiratory abductive movement and phonatory adductive movement of the bilateral vocal folds with preservation of a normal or near-normal voice quality, which would theoretically be attained by selective reinnervation of the bilateral posterior cricoarytenoid (PCA) muscles.

The nerve muscle pedicle (NMP) technique which was popularized by Tucker [1] and used in the treatment of vocal fold paralysis in earlier studies, has been abandoned since other surgeons were unable to use the technique satisfactorily [2-4]. Electrophysiological, histological, and anatomical studies have shown that the phrenic nerve is the optimal transfer nerve for reinnervation of the PCA muscle in treating patients with BVFP [5-7]. Numerous animal experiments have proven the feasibility of reinnervation of the PCA muscle to reacquire vocal fold abduction using the phrenic nerve as a donor nerve [6,7], while reports on human beings in this area are relatively limited. This may be partly due to the variability and complexity of the nerve supply of the human laryngeal muscle [8,9]. Crumley first introduced a split-phrenic nerve graft procedure, and the results in animal experiments were encouraging [4,10]. However, he clinically applied phrenic nerve implantation technique in a series of five patients and failed to re-establish visible inspiratory vocal fold abduction [5]. In 2002, we reported a clinical series of six patients with BVFP, among whom one PCA muscle was reinnervated using one single phrenic nerve and the other using an ansa-NMP transfer. Five patients showed inspiratory abductive motion on the phrenic transfer side, while no patients showed motion on the NMP implantation side [11]. Marie et al. also reported a series of 12 patients with BVFP who underwent selective reinnervation of the bilateral PCA muscles with one right upper phrenic nerve root using an interposition-free nerve graft and the adductor muscles with thyrohyoid branches of the hypoglossal nerve. Three of six evaluable patients showed active arytenoid abduction [12].

In the present study, 44 patients underwent selective reinnervation of the bilateral PCA muscles with the left phrenic nerve. The feasibility, effectiveness, and safety of this procedure were evaluated.

Materials and Methods

Operation procedure

For patients who had not yet undergone tracheotomy, tracheotomy was first performed under local anesthesia. General anesthesia was then induced through the tracheotomy incision, direct laryngoscopy was performed, and the bilateral arytenoids were palpated to determine that they were passively mobile. If the arytenoids were found to be fixed, arytenoidectomy was adopted instead of phrenic nerve transfer because reinnervation could not be successful in such a case. A horizontal oblique incision was made at the original incision of the primary surgery or approximately 2 cm above the tracheotomy incision. The fascia between the strap muscles and the sternomastoid muscle was opened on the left side. The lower part of the carotid sheath was exposed, and the internal jugular vein was pulled laterally with a hook. The phrenic nerve was usually found deep under the transverse cervical artery. If the accessory phrenic nerve was found, it was kept intact. The phrenic nerve was cut off at the clavicle level. The cricopharyngeal muscle was incised on the left side, and the intralaryngeal segment of the RLN was dissected and exposed at the posterior-anterior part of the cricothyroid joint. The RLN was then traced proximally until the injured site of the RLN was found. The distal end of the RLN was prepared with resection of scarred tissue back to the nerve fascicles. The posteroinferior portion of the thyroid lamina was partially removed to allow for exposure of the intralaryngeal branches of the RLN (Figure 1A). All adductor branches of the RLN, including the thyroarytenoid branch, lateral cricoarytenoid branch, and interarytenoid branch, were sectioned, and injury to abductor branches was avoided (Figure 1B). The right RLN was exposed, and branches of the intralaryngeal segment were managed in the same way. The proximal end of the left phrenic nerve was transposed superiorly, to be anastomosed to the distal end of the left RLN using three to five epineurial 11-0 nylon sutures under a surgical microscope. The intralaryngeal adductor trunk of the left RLN was sectioned and sutured to a free nerve graft (a branch of the cervical plexus, usually the great auricular nerve or transverse nerve), 5 to 6 cm in length (Figure 1C), which was then passed contralaterally through a retrolaryngeal or retrotracheal tunnel and anastomosed to the distal end of the right RLN. The adductor branches of the right RLN were sectioned intralaryngeally, and their proximal ends were implanted to the right PCA muscle by suturing the adductor trunk epineurium with the PCA myolemma using a single 11-0 nylon suture (Figure 1D). Finally, the wound was closed in layers.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Intraoperative image of reinnervation of the bilateral PCA muscles using the left phrenic nerve.

The posteroinferior portion of the thyroid lamina was partially removed. All adductor branches of the RLN, including the thyroarytenoid branch, lateral cricoarytenoid branch, interarytenoid branch, and three abductor branches, were well exposed (A). All adductor branches of the RLN were sectioned, and three abductor branches were well protected (B). The distal stump of the left RLN was anastomosed to the left phrenic nerve. A free nerve graft was used as a bridge between the adductor trunk of the left intralaryngeal RLN and distal stump of the right RLN through a retrolaryngeal or retrotracheal tunnel (C). The proximal ends of the adductor branches of the right intralaryngeal RLN were implanted into the right PCA muscle belly by suturing the nerve epineurium with the PCA myolemma (D). a: Recurrent laryngeal nerve (RLN), b: Abductor branches; c: Interarytenoid branch; d: .Lateral cricoarytenoid branch; e: Thyroarytenoid branch; f: Posterior cricoarytenoid muscle; g: Lateral cricoarytenoid muscle; h: Thyroarytenoid muscle; i: Inferior horn of thyroid cartilage; j: Section of thyroid lamina; k: Phrenic nerve; l: Adductor trunk of the RLN; m: Free nerve graft; n: Trachea. The yellow arrow in Figure 1C: the anastomotic site between phrenic nerve and the distal end of left RLN trunk; The white arrow in Figure 1C: the anastomotic site between adductor trunk of left RLN and free nerve graft; The yellow arrow head in Figure 1D: the anastomotic site between free nerve graft and the distal end of right RLN trunk; The white arrow head in Figure 1D: the implantation site of the proximal ends of adductor branches of the right RLN.

https://doi.org/10.1371/journal.pone.0077233.g001

Results

All patients presented varying degrees of aspiration after surgery, which disappeared about 2 to 8 weeks postoperatively, and normal deglutition function was achieved. A local hematoma and chylous fistula developed in one patient each and were resolved with appropriate treatment within 1 to 2 weeks. Recovery of unilateral or bilateral vocal fold abduction was achieved in 41 cases, later deterioration was observed in 3 of them. The rest 38 cases (87% of this series) were successfully decannulated 3 to 8 months after surgery and could perform considerable physical activity. Those who were decannulated after subsequent arytenoidectomy were not included in calculating the success rate of decannulation. Evaluation results of the surgery effects are as follows.

Videostroboscopic findings

The preoperative and postoperative laryngeal appearance of the patients are shown in Table 1. Preoperative video recordings (see Video S1) showed that most vocal folds were fixed in the paramedian or midline position, with varying glottic chink during inspiration (Figure 2A), and with incomplete or complete glottal closure during phonation (Figure 2B). In addition, most vocal folds showed straight or mildly bowing edges with no significant difference between the left and right sides (P > 0.05). Two to 6 months after reinnervation surgery, video recordings (see Video S2) showed variable excursion of abductive movement during inspiration in the left vocal folds of 87% of patients and the right vocal folds of 72% of patients (Figure 2C). The excursion of abductive movement increased to varying degrees after physical activity, and the excursion of abductive movement on the left side was significantly larger than that on the right side (P < 0.05). The vocal folds were in the paramedian or midline position during phonation, with varying degrees of glottic chink (Figure 2D), but without a significant difference compared with the preoperative situation (P > 0.05) (Table 1). Three cases presented immobile vocal folds during inspiration and phonation, another 3 cases recovered evident vocal fold abduction during inspiration in early follow-up postoperatively, however, vocal fold adduction during inspiration (paradoxical movement) was observed in later follow-up, these 6 patients were defined as failure cases.

Rating Rank	No. of videostroboscopic findings		P value
	Preoperative	Postoperative
Glottal closure during phonation			0.9314
Complete	13	11
Slightly incomplete	11	13
Moderately incomplete	14	13
Severely incomplete	6	7
Functional abduction of left vocal fold			P<0.001
No abduction	44	6
Slight abduction	0	9
Moderate abduction	0	19
Severe abduction	0	10
Functional abduction of right vocal fold			P<0.001
No abduction	44	12
Slight abduction	0	19
Moderate abduction	0	9
Severe abduction	0	4
Edge of left vocal fold			0.7234
Straight	22	17
Mildly bowing	13	15
Moderately bowing	8	10
Severely bowing	1	2
Edge of right vocal fold			0.5607
Straight	23	18
Mildly bowing	15	19
Moderately bowing	6	6
Severely bowing	0	1

Table 1. Comparison of pre- and post-operative videostroboscopic findings.

CSV

Download CSV

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Preoperative and postoperative videostroboscopic findings.

Case No. 5 showed representative bilateral vocal fold paralysis. Preoperatively, the bilateral vocal folds were fixed in the paramedian position during inspiration (A). Severely incomplete glottic closure occurred during phonation (B). Six months postoperatively, the bilateral vocal folds abducted to an paramedian or intermediate position during inspiration (C) and adducted back to the near-midline position during phonation with moderate incomplete glottis closure (D).

https://doi.org/10.1371/journal.pone.0077233.g002

Discussion

Management of BVFP is a substantial challenge to laryngologists. Traditional treatments such as arytenoidectomy, which mainly widens the airway at the glottis level, may worsen the voice and lead to aspiration, greatly compromising the quality of life of patients. An optimal treatment should strike a balance between airway and voice, which could be theoretically attained by selective reinnervation of the PCA muscle [17].

Direct anastomosis of the two severed stumps of the RLN has been studied. This technique may reportedly achieve reinnervation of the laryngeal muscles; however, the vocal folds were fixed at the midline or paramedian position, or paradoxical movement of the vocal cords was observed due to misdirected regeneration of the adductor and abductor fibers, which is also called laryngeal synkinesis [18]. Later, a donor nerve such as the ansa cervicalis, vagus nerve, or phrenic nerve was anastomosed to the distal stump of the RLN for reinnervation of the PCA muscle. However, the results were disappointing because both the adductor and abductor muscles are reinnervated by the same nerve, so the vocal folds remain fixed in one position or show paradoxical motion. Many later researchers stated that selective reinnervation of the PCA muscles may help to avoid laryngeal synkinesis [5-7,11,17]. However, the abductor branches of the RLN are usually fairly small in diameter, and several branches which innervate the PCA muscle commonly originate from the RLN trunk. This makes direct neurorrhaphy technically challenging, even under surgical magnification. NMP implantation into the PCA muscles at one time seemed to have reached the goal of selective reinnervation of the PCA muscles. It has been used for treatment of BVFP in animal models [1,3,4,19] and has been clinically applied in humans [11,20]; however, the results were inconsistent and unpredictable. Similar results were also reported in direct implantation of the donor nerve into the PCA muscle [5,21].

The phrenic nerve is considered to be the ideal donor nerve for reinnervation of the PCA muscles from histological, anatomical, and physiologic viewpoints. Methods of reinnervating the PCA muscle using the phrenic nerve as reported in the literature include at least: 1) anastomosis of the phrenic nerve to the abductor branch of the RLN, 2) anastomosis of the phrenic nerve to the RLN with the adductor branches of the RLN severed, 3) implantation of the split phrenic nerve into the PCA muscle belly, 4) implantation of the phrenic nerve trunk into the PCA muscle belly, 5) anastomosis of the phrenic nerve to the RLN with anastomosis of the ipsilateral severed adductor branch of the RLN to the ansa cervicalis or other donor nerves, and 6) anastomosis of the phrenic nerve to one side of the RLN with anastomosis of the severed adductor trunk of the RLN to the contralateral RLN to achieve bilateral PCA muscles reinnervation. Most of the above studies were limited to animal experiments, in most of which restoration of evident vocal fold abduction was reported. In one of our previous animal studies, we anastomosed the upper branch of the phrenic nerve to the RLN, and at the same time, the adductor branches of the RLN were severed and implanted into the PCA muscle. We assumed that with this method, all regenerative phrenic nerve fibers would grow into the PCA muscle instead of wasting some in the ligated stump of the adductor trunk. Thus, two different sets of motor end plates would be established, which is termed “double reinnervation” [7]. This is in agreement with the point of view of Kwak, who stated that insertion of the proximal stump of the adductor branch of the RLN into the PCA muscle not only allows for twice the stimulation of the muscle to achieve abduction, but also prevents aberrant reinnervation of the adductor muscles by the abductor fibers [22]. Our results showed that all animals restored vocal fold abduction, and abduction excursion of some vocal folds even exceeded normal ones [7], indicating that this procedure is an effective method to reinnervate the PCA muscle. We treated six patients with BVFP using this surgical technique in clinical practice, and promising results were obtained in five of them. We modified the above-described technique by using a free nerve graft that was harvested from the cervical plexus for bridging between the adductor trunk of the left RLN and trunk of the right RLN. Our results in animal models using this modified technique showed satisfactory restoration of reinnervation of the bilateral PCA muscles by a single phrenic nerve (unpublished data), which is in accordance with the results reported by Baldissera et al [23]. Encouraged by these promising results, we applied this modified technique clinically in the present study.

Videostroboscopic, voice analysis and laryngeal electromyographic data confirmed the success of the modified technique. The evidence is as follows: 1) Inspiratory vocal fold abduction was restored on both sides, and abductive excursion increased after physical activity. 2) We successfully reinnervated the bilateral PCA muscles with a single phrenic nerve; the success rate of abductor reinnervation was 87%. 3) Voice function assessment showed that the voice of some patients improved after reinnervation surgery, although the results showed no significant difference. 4) Electromyographic examination showed typical inspiratory high-frequency discharge in the PCA muscles on both sides. 5) Resection of the left phrenic nerve was associated with no immediate clinical symptoms, such as airway compromise. Pulmonary function test results confirmed no long-term adverse effects on pulmonary function.

One phenomenon observed in this study is that abductive excursion of the left vocal fold was larger than that of the right side, and the recovery rate of vocal fold abduction on the left side was higher than that on the right side. We presume that this was partly due to the number of regenerative axons that pass through the anastomotic site. In the present study, we chose the left phrenic nerve as the donor nerve, which was transected and anastomosed to the distal stump of the left RLN trunk. The adductor branches were then cut, and their common trunk was anastomosed to a free nerve graft whose other end was carried contralaterally and sutured to the distal end of the right RLN trunk. The phrenic nerve fibers needed to pass through one anastomotic site to reach the left PCA muscle, but through three sites to reach the right PCA muscle. Some of the available regenerated axons were lost across each anastomotic site; thus, reinnervation procedures that involve an interposition nerve graft with two anastomoses are less effective [24]. However, our results showed that the right PCA could still gain adequate or partial reinnervation to restore abduction of the right vocal fold even when regenerative axons had to pass through three anastomotic sites; the recovery rate of vocal fold abduction was 70%. This indicates that satisfactory bilateral abduction movement can be achieved in humans by reinnervating the bilateral PCA muscles using a single phrenic nerve.

The voice quality of patients in this series did not worsen as did the voice quality of those who underwent arytenoidectomy. Interestingly, although we did not intentionally perform adductor reinnervation in the present study, we still observed evident vocal fold adduction during phonation, and some patients showed complete glottis closure. The reasons may be as follows: 1) The vocal folds passively adducted after abduction motion. 2) Laryngeal EMG showed that the motor unit potential was recorded in the TA muscles, which indicates subclinical reinnervation of the adductor muscles due to axonal regeneration across the transected distal ends of adductor branches of the RLN into the adductor muscles or neurotization that sprouted from the surrounding musculature, which may have arisen from the superior laryngeal nerve or from the nerve that innervates the cricopharyngeal muscle [9,18,25]. 3) Contraction of the uninjured cricothyroid muscle may also contribute to the adduction of the vocal folds [18].

Phrenic nerve transfer has been used to repair brachial plexus nerve injuries, and many researchers have studied pulmonary function following unilateral phrenic nerve transection. Fackler et al. reported 14 patients who underwent unilateral phrenic nerve transection; VC returned to the preoperative normal level in all patients 6 months after surgery [26]. Xu et al. reported a series of cases involving phrenic nerve neurotization and showed that the pulmonary functional parameters FVC and FEV1 were not affected by the procedure in the long term [27]. Luedemann et al. reported a series of 23 patients who underwent neurotization surgery with the phrenic nerve as the donor nerve and found that none of the 23 patients experienced pulmonary problems postoperatively; however, the patients incurred a higher risk of reduced pulmonary VC when using the right phrenic nerve than the left one. The authors suggested that if possible, the left phrenic nerve should be used [28]. Based on the above results and those of our previous report [11], we used the left phrenic nerve for transfer in all patients in the present study; none of them experienced pulmonary problems following the surgery. Pulmonary functional parameters indicated mild reduction in pulmonary function 3 months and 6 months postoperatively; however, all parameters—with the exception of PImax—recovered to normal reference value levels by 12 months postoperatively. A decrease in PImax is not an absolute indication of pulmonary function reduction because PImax is affected not only by the strength of the respiratory muscles, but also by the size of the glottic chink [29]. The absence of PImax recovery to the normal reference value in the present study may have been due to the glottic chink not recovering its normal size. However, the preoperative general condition of the patients is also very important. Pulmonary functional tests and chest X-rays must be performed to evaluate the preoperative diaphragm motion and exclude pulmonary diseases such as chest or lung trauma, emphysema, and chronic obstructive pulmonary disease.

There are several critical influencing factors which may affect the surgical outcome of phrenic nerve transfer surgery. The first important factor that may have influenced the outcome of surgery is the interval between the phrenic nerve transfer procedure and the onset of bilateral RLN injury. Theoretically, better results of reinnervation surgery are obtained if the interval between the injury and reinnervation is shorter, as is shown both in the literature and in our experimental study [6,30]. We performed the phrenic nerve transfer surgery for some BVFP patients with thyroid surgery-related RLN injury within 6 months after the injury, there are some reasons. Firstly, we performed exploratory surgery for BVFP patients whose denervation duration was less than 6 months only when the surgeons of primary thyroid surgery strongly suggested that bilateral RLNs were probably transected. Actually, bilateral RLNs were confirmed to be completely transected during exploratory surgery in 18 BVFP patients, and phrenic nerve transfer surgery was then performed immediately. However, when RLN exploration confirmed RLN continuity (not transected as suggested by the original thyroid surgeons), then nerve decompression surgery was performed (These patients were not included in the present study). Some of these patients did indeed benefit from the decompression surgery and their vocal folds regained physiologic vocal fold movement after decompression. This was consistent with some previous reports [31-34]. Secondly, progressive atrophy with denervation time may impair the regenerative capacity of skeletal muscles owing to irreversible pathologic alterations, such as muscle fibrosis and degeneration of muscle endplates [6]. If bilateral RLNs are completely transected, the degree of subclinical reinnervation is less, the atrophy of PCA muscles is more severe. Thirdly, as the only one abductor muscle of intrinsic laryngeal muscles, PCA muscle plays an important role in the process of respiration. Study on denervated laryngeal muscles of rats showed that denervation produced large shifts in the expression of the fast type IIX (ie, up-regulated) and IIB (ie, down-regulated) myosin heavy chain (MyHC) isoforms in the PCA muscle. However, reinnervation of the PCA muscle had little effect on reversing these shifts at the whole muscle level, in contrast, reinnervation of the TA muscle was highly effective in restoring a normal pattern of MyHC isoform expression at the whole-muscle level [35]. Consistent with these observations, Shiotani et al also reported that repair of the recurrent laryngeal nerve did not prevent denervation-like shifts in the MyHC isoform profile of the PCA muscle but was effective in minimizing shifts in the MyHC isoform profile in TA muscle [36]. Our research on denervated PCA muscle on humans also demonstrated that although type I MyHC isoforms and type II MyHC isoforms changed significantly with the denervation time increases, the prominent transitions in MyHC isoforms expression take place one year after RLN injury (unpublished data). Last but not least, in this series, three patients whose interval time between bilateral RLN injury and reinnervation surgery was >10 months did not show restoration of abduction in the follow-up, which strongly indicates that it is better to perform phrenic nerve transfer surgery for patients with BVFP whose dennervation course is <10 months. This is totally different from the situation in UVFP, in which reinnervation surgery could still be effective with denervation duration of several years. However, within 10 months after bilateral RLN injury, when is the better time range to perform phrenic nerve transfer surgery so as to achieve the most effectiveness needs further investigation, which will depend on further accumulation of cases and multivariate analysis on the influencing factors of the surgical procedure. In summary, the above findings indicate that early reinnervation surgery should be performed for BVFP patients if bilateral RLNs were confirmed to have been transected. However, the prognostic value of laryngeal PCA muscle EMG for BVFP is limited and sometimes not precise. So the most direct means to confirm the RLN transection (in our experience) is exploratory surgery in the first few months after RLN injury when surgeons of the primary thyroid surgery suggested that RLNs might have been transected. Our experience appears to show that RLN decompression can be helpful in patients in whom the RLNs have been ligated, but these data are not included in the present study. The authors concede that it is conceivable that we may have over-treated those patients by exploratory surgery during the early stage of RLN injury in a very small proportion of patients who might possibly have subsequently demonstrated spontaneous recovery of physiological vocal fold motion without any intervention. However, for most BVFP patients whose bilateral RLNs were indeed transected, they could benefit from early intervention because physiologic vocal fold abduction could be achieved if phrenic nerve transfer surgery were performed in early stage.

Patient age is another critical factor. All patients in this series were younger than 55 years with the exception of one, who was 59 years old at the time of the surgery and insisted on undergoing phrenic nerve transfer surgery. We informed the patient the possible ineffective outcome. The procedure was performed in this case after she signed the consent inform. Follow-up examinations showed no inspiratory vocal fold abduction. Considering the results of our previous report [11], we assume that the adverse surgical outcome was to some extent related to the patient’s older age. However, because of the small sample of this study, we don’t feel that it is conclusive to say that old age (>55 years old) leads to adverse outcomes of reinnervation surgery.

Based on our experience, the following surgical aspects of this procedure may help to improve the outcome. 1) It is critical to identify the interarytenoid branch of the RLN, thoroughly transect it and ligate its proximal end. The interarytenoid branch usually originates from the genu segment of the RLN just below the level of the cricoarytenoid joint. There are usually one to several abductor branches below the interarytenoid branch, which are also much smaller in diameter. Videostroboscopy showed immobile vocal folds in 3 cases, and abduction of vocal folds observed during early follow-up turned into paradoxical motion during later follow-up in another 3 patients. If the interarytenoid branch was not precisely identified and thoroughly cut or its proximal end was not ligated during the surgery, regenerative abductor fibers may regrow into the interarytenoid muscle through the distal end of the severed interarytenoid branch, or even sprout into other adductor muscles so as to result in laryngeal synkinesis, which most likely led to the failure of the surgery. 2) To expose and identify the interarytenoid branch, a small part of cartilage must be removed from the posteroinferior part of the thyroid lamina; however, the integrity of the superior and inferior process of the thyroid cartilage should be maintained. 3) The nerve supply pattern of the human PCA muscles is usually variable and complex and is present as two or more branches [1]. Care should be taken to avoid injuring the abductor muscle branches of the RLN. 4) The free nerve graft, which is usually harvested from one branch of the cervical plexus, should be comparable with the RLN in diameter and sufficiently long to achieve tension-free anastomosis. The length of the free nerve graft is not problematic if passed contralaterally through a retrolaryngeal or retrotracheal tunnel. 5) The nerve ends of the RLN distal stumps should be prepared by thorough resection of scarred or nonviable tissue to expose the normal nerve fascicles. 6) It is important to avoid injury of the cricoarytenoid joint and evaluate arytenoid mobility before and during surgery. 7) The left phrenic nerve should be used preferentially in order to avoid a higher risk of reduced pulmonary function.

Theoretically, the ideal treatment for BVFP is restoration of physiologic (bidirectional) motion of the bilateral vocal folds, including phasic abduction during inspiration, volitional closure for phonation, and ideally, a glottic closure reflex during swallowing or as a response to sensory stimuli [37]. Many researchers have studied combinations of adductor and abductor reinnervation techniques using animal models, and some reported restoration of physiologic motion of the vocal folds [38,39]. Marie clinically applied a combined reinnervation technique in 12 patients with BVFP, and only three of six evaluable patients achieved promising results. This success rate is not very satisfactory [12]. Other experimental studies reported good adduction, but no or poor abduction [40,41]. Crumley and Marie reported the results of animal experiments in which the abductor muscles were reinnervated by the phrenic nerve and the adductor muscles were reinnervated by the ansa cervicalis simultaneously. Coordinated vocal fold movements were observed in only a few of the animals. The reason may be that regenerative axons from the ansa cervicalis reach the abductor muscle before those from the phrenic nerve, so they may reinnervate the PCA muscle by sprouting into it from the ansa cervicalis anastomostic site [41]. This phenomenon was also evident in our previous animal experiments. Therefore, we preferentially studied the restoration of vocal fold abduction in the present study. The promising results of PCA muscle reinnervation in this pioneer study pave the path for our further work on combinations of reinnervation of both abductor and adductor muscles.

Conclusions

Reinnervation of the bilateral PCA muscles using the left phrenic nerve restored inspiratory vocal fold abduction to a satisfactory extent while preserving phonatory function at a preoperative level. None of our patients developed immediate dyspnea after surgery, and the pulmonary function parameters were recovered to normal reference value levels within 1 year, indicating the safety of this procedure.

Supporting Information