The Pharyngeal Plexus–Mediated Glottic Closure Response and Associated Neural Connections of the Plexus

Importance  There continues to be a paucity of data regarding the pharyngeal plexus (PP) and its interconnectivity with the laryngeal nerves and function.

Objective  To identify the specific neural pathways involved in the glottic closure reflex (GCR)-like pathway of the PP and other pathways to the thyroarytenoid (TA) muscle in the porcine model.

Design, Setting, and Animal Subjects  Animal experimental study from September 2013 to June 2014 conducted in a tertiary academic medical center on male Yorkshire pigs.

Interventions  Contraction of the TA was detected with electromyography (EMG) during electrical stimulation of the PP in 7 porcine necks. Subsequently, the external branch of the superior laryngeal nerve (eSLN), communicating nerve of Galen (NG), and the recurrent laryngeal nerve (RLN) were sequentially transected to help elucidate the path of neural conduction.

Main Outcomes and Measures  Confirmation of TA muscle contraction by EMG.

Results  Stimulation of the PP evoked a response from the TA muscle in 6 of 7 subject animals. In 3 of 7 subjects, a long latency response (mean, 14.62 milliseconds) was identified, which was eliminated only after transection of the RLN. In 3 of 7 subjects, a short latency response (mean 3.05 milliseconds) was identified, which disappeared in 1 subject each by eSLN, RLN, and NG transection.

Conclusions and Relevance  We identified the specific neural pathway involved in the PP’s GCR-like pathway. We also noted a variable direct pattern of innervation to the TA.

Laryngeal physiology has come a long way since Galen in the 2nd century AD noted the “reversivi” or recurrent laryngeal nerves and performed his famous experiments in front of the Roman Elders.¹ He showed in likely the first laryngeal experiment in the porcine model that transection of the nerve would cease the animal’s squealing. Almost 2000 years later, our laboratory continues to search for various innervation patterns to the larynx with a focus on helping to augment the glottic closure reflex (GCR). The importance of the GCR cannot be overstated owing to its critical role in airway protection.

The GCR is composed of the internal branch of the superior laryngeal nerve (iSLN) and the recurrent laryngeal nerve (RLN), providing the afferent and efferent pathways, respectively. This classic pathway involves ascending the iSLN to the vagus to the nucleus tractus solitaries to the nucleus ambiguus and then descending down the vagus to the RLN. In addition to the GCR arc, the iSLN and RLN have variable interconnecting neural pathways that bypass the brainstem. Anatomical studies in human cadaveric larynges have shown the presence of the classic nerve of Galen (NG) anastomosis between the iSLN and RLN to be present 100% of the time, in addition to other anastomotic patterns in the larynx between the RLN, iSLN, and external SLN (eSLN).² For simplicity, we will refer to the communicating nerve of Galen and all other possible communicating nerves between branches of the vagus simply as communicating nerves, since their identification and isolation intraoperatively is variable.

Cephalic to the RLN and iSLN off of the vagus nerve lies the pharyngeal plexus, which provides variable sensory and motor innervation to the pharynx with contributions from the glossopharyngeal nerve. In our group’s previous porcine pharyngeal plexus experiments,³ we identified 3 separately distinguishable latency-based innervation patterns leading to vocal cord adduction. Our findings supported previous laboratory observations that the electrical and mechanical stimulation of the plexus leads to glottic closure.^4-6 Our experiments showed variable latencies to thyroarytenoid (TA) muscle contraction. The long latency responses in our previous experiments³ were likely evoked GCR-like responses. The short latency m-wave responses were likely direct patterns of innervation to the TA muscle. To help understand the variable latency responses, we established a protocol in which we could trace the neural pathway involved directly to the larynx and confirm our previous findings in search of a possible target for rehabilitation.

Experiments were carried out in 7 male Yorkshire pig necks. Procedures followed the principles laid out in Guide for the Care and Use of Laboratory Animals⁷ and the Animal Welfare Act (7 USC § 2131 et seq, 1966), and the Yale institutional animal care and use committee (IACUC) approved all experiments.

To help optimize efficiency, we performed each experiment independently on each side. The animals were approximately 3 months old, and their weights ranged from 20 to 30 kg. The porcine model was selected because of its anatomic and neurophysiologic resemblance to the human.^8,9 Ketamine hydrochloride (30 mg/kg), xylazine hydrochloride (2.2 mg/kg), and telazol (4.4 mg/kg) were injected intramuscularly as induction agents. General anesthesia was then provided with isoflurane inhalation via a nose cone. Subsequently, a tracheotomy was performed through a midline neck incision, and a 6-mm (inner diameter) endotracheal tube was inserted as the route of inhalational anesthesia. The endotracheal tube was then secured to the skin using a stay suture. The depth of anesthesia was maintained at a minimum alveolar concentration of less than 1.0 during the procedure without use of pharmacologic muscle relaxation. Body temperature was controlled at 38°C using an external heating pad.

Under aseptic technique, a midline skin incision was extended from the hyoid bone to the sternal notch. The iSLN was surgically exposed at the thyrohyoid membrane. The eSLN was subsequently identified on its course to the cricothyroid muscle. Branches of the ipsilateral pharyngeal plexus were carefully dissected at the level of middle and inferior pharyngeal constrictor muscles. Attention was paid during the pharyngeal plexus dissection to preserve any small branches. The ipsilateral RLN and the intermittently identified nerve of Galen anastomosis and other possible communicating nerves were also identified and preserved. The thyroid cartilage was then partially removed to allow both vocal cords to be visualized and to allow for a transmucosal superior insertion of recording electrodes.

The XLTEK Neuromax electromyograph (Natus Medical Inc) was used to stimulate the target nerves and to record electromyographic waveforms. A monopolar recording electrode was inserted into the mid-portion of the ipsilateral TA muscle. A reference electrode and ground electrode were positioned in the subcutaneous tissue at the skin incision at least 5 cm from the larynx.

All subject animals underwent initial stimulation of the exposed branches of the pharyngeal plexus. A bipolar dual-needle electrode was placed into the exposed branches of the pharyngeal plexus. To avoid ground circuit and artifact, each nerve floated freely on mineral oil–soaked pledgets. Stimulus presentation to the branches of the pharyngeal plexus began in early inspiration at 0.1 mA as a single pulse and continued up to 20 mA. The applied stimulus intensity was increased until contraction of the TA muscle was identified by electromyography. Once TA muscle contraction was identified during stimulation of a specific branch of the pharyngeal plexus, the latency, duration, amplitude, and voltage needed was recorded. The branch was then selectively transected with a No. 15 blade. We subsequently stimulated the central (proximal) or distal end of the nerve to identify which pathway led to TA muscle contraction. Following isolation of either the central or distal end, we proceeded in a staged manner as shown in Figure 1.

The eSLN was transected first, followed by repeat stimulation of the branch of the pharyngeal plexus, which was previously isolated and shown to cause TA muscle contraction. If TA muscle contraction continued, anastomosis of the NG or any communicating nerve that appeared to be from the pharyngeal plexus (PP) down to the RLN was transected. Once again, if TA muscle contraction continued, we proceeded to the last step in which the RLN was transected at the cricothyroid joint. If at any point TA muscle contraction ceased after transection, we continued to stimulate any remaining distal branches to help trace the pathway. The latency, duration, amplitude, and voltage needed were recorded throughout the procedure, and each recording was duplicated 6 times to ensure reproducibility. In addition to electromyographically confirmed responses, contraction of the TA muscle with its accompanying vocal cord adduction was directly visualized.

We were able to elicit a TA muscle response from PP stimulation in 6 of the 7 experimental necks. A summary of the positive responses is provided in the Table. The PP stimulation leading to TA muscle contraction had a long latency response in 3 necks, numbers 3, 4, and 6, and this persisted until RLN transection. The average latency for these 3 necks was 14.62 milliseconds; duration, 2.34 milliseconds; and amplitude, 0.56 mV, with an average of 3.3 mA needed to initiate a response.

In experimental neck 2, stimulation of the distal branch of the pharyngeal plexus elicited TA muscle contraction with an average short latency of 3.3 milliseconds; duration, 2.2milliseconds; and amplitude, 0.5 mV, with an average of 3.8 mA needed to elicit the response. The pathway was noted to be through the eSLN and persisted through the distal transected branch of the eSLN after RLN transection.

Experimental neck 1 showed a short latency response during central branch stimulation, which persisted after eSLN transection but terminated after a communicating branch to the RLN was isolated and transected. This central branch led to TA muscle contraction with an average latency of 2.95 milliseconds; duration, 2.38 milliseconds; and amplitude, 0.2 mV, with an average of 3.8 mA needed to initiate a response. This communicating branch is shown in Figure 2. The transected distal branch stimulated until the RLN was transected, at which point it ceased.

Finally, experimental neck 5 showed a short latency response to stimulation of the central branch that persisted until RLN transection. The average latency to TA contraction was 2.92 milliseconds; duration, 1.8 milliseconds; and amplitude of 0.2 mV, with 3.2 mA needed to initiate a response. Of note, stimulation of the pharyngeal plexus in 1 of the 7 experimental necks failed to elicit TA muscle contraction. We attribute this to the inadvertent premature transection of a PP branch resulting from the delicate nature of the dissections and the procedural difficulty of isolating the various PP branches.

The variable pattern of PP innervation that leads to TA muscle contraction depicts the complex neurophysiologic nature of our experiments. The PP involvement in glottic closure is supported by our data, albeit there are multiple pathways involved. We summarize our findings herein and detail the pathways that we have previously suspected.

To better understand the innervation patterns, separating the short and long latency responses is essential. Simply put, the latency to contraction is the amount of the time it takes the neural impulse to travel after the nerve is stimulated to its end target. The long latency responses were all noted only during stimulation of the central end of the PP branch, with an average latency to TA muscle contraction of 14.62 milliseconds. This occurred in a 20- to 30-kg animal with a GCR reflex arc length of approximately 27 ± 3 cm when measured intraoperatively with the help of the veterinarian. The porcine neural impulse rate is approximately 5 cm/milliseconds, and 3 to 5 synapses are involved in the classic GCR, with an average of 1.5 milliseconds per synapse.¹⁰ Therefore, a GCR latency in the range of 10 to 15 milliseconds is supported, given the variable number of synapses and size of the animal. In our long latency subset of experimental necks, our average latency time was well supported and suggested a GCR-like evoked brainstem response. The latency time was too long for a direct m-wave pattern to the TA muscle and could only be explained by a multi-synaptic pathway through nucleus tractus solitarius and nucleus ambiguus of the brainstem through RLN to its target TA muscle. This is further reinforced by the finding that the response was eliminated only after transection of the RLN in these experiments. Finally, in our group’s previous experiments,³ we had shown that simultaneous simulation of the long latency branch of the PP and the iSLN required lower individual voltage thresholds, indicating that there was a neural impulse that likely summated at the level of the brainstem. A collection of verifiable data now supports that the pharyngeal plexus has afferent branches that play a critical role in the GCR-like pathway, with the RLN acting as the classic terminal efferent route, as shown in Figure 3.

The short latency to TA muscle contraction responses can be separated into 2 groups. The first group includes only 1 experimental neck in which the short latency response to TA muscle contraction occurred during stimulation of the distal branch of the PP. This response was eliminated after eSLN transection. Interestingly, stimulation of the distal branch of the eSLN in this animal led to TA muscle contraction even after subsequent RLN transection, indicating a direct pattern of innervation from the PP through the eSLN directly to the TA muscle bypassing the RLN, as shown in Figure 4. This finding was not totally unexpected, since Nasri et al¹¹ showed the cross-innervation of the TA muscle by the eSLN in the canine model.

The second group contained 2 experimental necks that showed a short latency response to TA muscle contraction that occurred during stimulation of the central branch of the PP. Knowing that the larynx had been shown in multiple human studies to have a significant anastomotic network of neural structures,^2,12 we were able to identify in 1 of the necks a communicating nerve that was responsible for TA muscle contraction. The second neck that showed a short latency response on central end stimulation only ceased conduction after transection of the RLN. This neck might also have had a communicating nerve from the PP to the RLN, but this was not identified intraoperatively. The findings in these 2 experimental necks support the presence of variable communicating nerves between the PP and the RLN, as depicted in Figure 5.

Three distinct pathways were identified. First, an evoked, long-latency GCR-like pathway was identified, with sensory fibers of the PP acting as the afferent route projecting to the brainstem NTS and NA, and the RLN as the efferent route. Second, we identified a short-latency connection between the PP and the eSLN leading to direct TA muscle contraction. Finally , we noted a likely communicating branch from the PP directly to the RLN and ultimately to TA muscle contraction. These pathways all influence glottic closure. Identification and clarification of these pathways in the human larynx could offer surgical targets for rehabilitation while avoiding direct manipulation of the RLN itself, with a focus on patients with chronic aspiration due to incomplete or weakened glottic closure. In future studies, we will explore possible pacing strategies that would augment TA muscle contraction during swallowing.

Submitted for Publication: June 12, 2014; final revision received August 5, 2014; accepted August 22, 2014.

Corresponding Author: Boris Paskhover, MD, Section of Otolaryngology, Department of Surgery, Yale School of Medicine, 333 Cedar St, PO Box 208041, New Haven, CT 06520-8041 (Boris.Paskhover@yale.edu).

Published Online: October 16, 2014. doi:10.1001/jamaoto.2014.2440.

Author Contributions: Dr Paskhover had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Paskhover, Sasaki.

Acquisition, analysis, or interpretation of data: Paskhover, Wadie, Sasaki.

Drafting of the manuscript: Paskhover, Wadie.

Critical revision of the manuscript for important intellectual content: Paskhover, Sasaki.

Statistical analysis: Wadie.

Obtained funding: Paskhover, Sasaki.

Administrative, technical, or material support: Paskhover, Wadie.

Study supervision: Paskhover, Sasaki.

Conflict of Interest Disclosures: None reported.

Funding/Support: This research was supported by the Charles W. Ohse Endowment and the Virginia Alden Wright Fund.

Role of the Funder/Sponsor: The supporting institutions had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.