Vocal Cord Palsy in Hypoxic Brain Injury: A Qualitative Systematic Review

Abstract

This study aimed to consolidate evidence linking vocal cord palsy (VCP) to hypoxic injury, and its pathophysiology, and explore related topographical representations along laryngeal innervation. PubMed, Embase, and Scopus were the databases used. This review adheres to PRISMA guidelines. We included case series or case reports published before December 6, 2023. These studies must document VCP, diagnosed via direct or indirect laryngoscopic evaluation, resulting from hypoxic injury with imaging documentation specifying the level of injury. Screening, review, quality assessment, and extraction were done using Covidence. Our search strategy yielded 380 articles, of which 11 papers met the inclusion criteria for final review. A total of 12 patients were included across the selected studies, evenly split between genders. The causes of hypoxic injury were stroke in 11 cases and perinatal asphyxia in one. The hypoxic injury affected the medulla, pons, basal ganglia, thalamus, internal capsule, cortex, and cerebellum. The distribution of hypoxic injuries was: left side (6), bilateral (2), right side (2), and not mentioned (2). Vocal cord involvement was unilateral in eight cases and bilateral in four cases. In five cases, the involvement was ipsilateral to the hypoxic injury, bilateral when the injury was bilateral, and contralateral in two cases. Our study provides insights into how hypoxic brain injury can cause VCP and correlates the level of lesions along the innervation pathway with the clinical presentation. VCP can be induced by hypoxic injuries to the neurons extending from the laryngeal motor cortex to the laryngeal motor neurons in the medulla.

Introduction

The vocal cords play a crucial role in both sound production (phonation) and airway protection [1]. The intricate movements of abduction and adduction of the vocal folds are controlled by the intrinsic laryngeal muscles, primarily innervated by the recurrent laryngeal nerve [2, 3]. However, the cricothyroid muscle receives innervation from the superior laryngeal nerve. Both the recurrent and superior laryngeal nerves stem from branches of the vagus nerve [2, 3].

Vocal cord palsy (VCP) presents a significant challenge for airway surgeons and laryngologists, bearing considerable clinical importance. A variety of causes have been identified for unilateral or bilateral VCP, with trauma to the recurrent laryngeal nerve being the most common factor in both groups [4]. Among well-documented etiologies such as trauma, neoplastic infiltration, and inflammation [2, 5, 6], there remains an unexplored possibility: hypoxic injury to the central nervous system (CNS). Hypoxia induces neuronal injury through several mechanisms, including impaired Na+-K + ATPase pump function leading to increased excitatory amino acids in the extracellular fluid, generation of nitric oxide and free radicals, mitochondrial dysfunction, inflammatory and immune responses, and hydrolysis of matrix metalloproteinases [7].

Our interest in investigating the role of hypoxia in VCP is spurred by evidence from various case reports documenting VCP following hypoxic injury. For instance, Fujiki et al. reported a case of permanent left vocal cord paralysis following a cerebral infarct in the territory of the left middle cerebral artery [8], while Saito et al. presented a case of neonatal asphyxia resulting in bilateral vocal cord paresis alongside lesions in the pontomedullary tegmentum, bilateral putamen, and thalamus [9]. Furthermore, laryngeal innervation is complex. While motor innervation is received via the Vagus nerve, the Vagal nuclei receive corticobulbar innervation via the laryngeal motor cortex in the cerebrum [10, 11]. This suggests that injuries beyond the Vagus nerve or its branches, specifically at the supranuclear level of the vagal nucleus, could potentially lead to laryngeal denervation.

This systematic review aims to consolidate evidence linking VCP to hypoxic injury and develop a deeper understanding of its pathophysiology. We also seek to explore the topographical representations of anatomical levels that may affect laryngeal innervation when subjected to hypoxia, believing this will enhance our comprehension of laryngeal innervation as a whole.

Methods

The reporting of this review strictly adheres to PRISMA (Preferred Reporting Items for Systemic Review and Meta-Analysis) guidelines [12].

Inclusion Criteria

We included any prospective or retrospective case series or case reports involving the human population, without gender or age restrictions, published in the English language before December 6, 2023. These studies must document VCP (unilateral or bilateral/ partial or complete), diagnosed via direct or indirect laryngoscopic evaluation, resulting from hypoxic injury with imaging documentation (MRI or CT scan) specifying the level of injury.

Exclusion Criteria

We excluded animal studies, studies not in English, studies lacking laryngoscopic documentation of VCP, studies where cases had premorbid conditions or inciting factors other than hypoxic injury that could have led to the VCP, and studies lacking imaging documentation of hypoxic brain injury or mention of anatomical levels of brain injury.

To prevent confusion and acknowledge paresis as a result of partial denervation, we utilized the terms “partial palsy” and “complete palsy” for paresis and completed paralysis, respectively.

Literature Search, Study Selection, and Data Extraction

The following databases were reviewed for published studies before December 6, 2023: PubMed, Embase, and Scopus. Boolean logic was used for conducting database search and Boolean search operators “AND” and “OR” were used to link search terms. The following search strategy was adopted:

(“hypoxic ischemic injury” OR “hypoxic ischemic injuries” OR “hypoxic brain injury” OR"hypoxic brain injuries” OR “hypoxic ischaemic encephalopathy” OR “hypoxic ischaemic encephalopathies” OR “HIE” OR “perinatal hypoxia” OR “perinatal asphyxia” OR “Ischemic Stroke” OR “Hemorrhagic Stroke” OR “Stroke” OR “cerebrovascular accident” OR “cerebrovascular accidents” OR “hypoxic injury” OR “hypoxic injuries” OR “hypoxia, brain“[MeSH Terms] OR “hypoxia ischemia, brain“[MeSH Terms] OR “Asphyxia Neonatorum“[MeSH Terms] OR “Asphyxia“[MeSH Terms] OR “Stroke“[MeSH Terms] OR “Hemorrhagic Stroke“[MeSH Terms] OR “Ischemic Stroke“[MeSH Terms] OR “Hypoxia“[MeSH Terms] OR “Fetal Hypoxia“[MeSH Terms]) AND (“vocal cord palsy” OR “vocal cord palsies” OR “vocal fold palsy” OR “vocal fold palsies” OR “vocal fold paresis” OR “vocal cord paresis” OR “Vocal Cord Paralysis“[MeSH Terms]) AND (‘case report’ OR ‘Case series’ OR ‘case study’ OR ‘cross sectional study’ OR ‘observational study’ OR ‘prospective study’).

Our search strategy yielded 380 articles, all of which were imported into Covidence Software. Ninety-six duplicates were identified and removed: 93 by the software itself and three manually. Initial screening of the remaining 284 articles was conducted by AKM and SB, resulting in 43 articles selected for full-text screening. AKM and SB performed the full-text screening, with any discrepancies resolved by consulting a third author, BRG. Ultimately, only 11 studies met the inclusion criteria for the final review. The article screening process is depicted in Fig. 1.

Fig. 1.

The PRISMA flow diagram generated by covidence

AKM and BRG extracted the following data using a data extraction sheet in Covidence software: Author, year and place of publication, total number of cases, age and gender, cause of hypoxic brain injury, imaging modality with anatomical level of injury, vocal cord involvement including the side, type of involvement (complete or partial palsy), symptoms, and outcomes of management.

Quality Assessment

Quality assessment of the included studies was done using the Quality Assessment Tool for Case Series Studies developed by the National, Heart, Lung, and Blood Institute [13] by authors AKM and BRG.

Statistical Analysis

We didn’t use any statistical analysis in this study. The descriptive data were manually entered and extracted using the data extraction sheet of Convivence software.

Results

Our literature search identified a total of 380 articles. Following the removal of duplicates and articles not meeting inclusion criteria, 11 papers were deemed suitable for our systematic review. Figure 1 illustrates the results of our literature search and selection process, while Table 1 provides a summary of the characteristics of each included study. The selected papers comprised seven case reports, three case series studies, and one letter to the editor. Of all the studies, three were from Japan, two from the USA, and one each from Spain, the UK, India, Portugal, Canada, and Switzerland. Quality assessment, conducted using the Quality Assessment Tool for Case Series Studies (NIH), categorized nine studies as good quality and two as fair quality.

Table 1.

Data extracted from the selected studies

Study ID	Country	Type of study	Sex	Age	Cause of hypoxic brain injury	Imaging	Level of Injury (CNS)	Side of injury	Vocal cord involvement	Side of involvement	Type of involvement	Symptoms	Outcome after management	Assessed Quality
Fujiki 1999 [8]	Japan	CR	M	69 y	Ischemic Stroke (PAN)	CT	Cortex	Left	UL	Left	Paralysis	Hoarseness	Permanent paralysis	Good
Takase 2011 [22]	Japan	CR	M	68 y	Ischemic stroke (CVA)	MRI	NM (Probably cortex)	Left	UL	Left	Paresis	Dysarthria	Recovered	Good
Bathini 2014 [23]	Canada	CS	F	8 y	Stroke (Type NM)	MRI	NM (Probably cortex)	Left	BL	BL	NM	Dysphonia	Recovered	Fair
Sawalha 2021 [25]	USA	CR	F	60s	Ischemic stroke	MRI & CT	Basal Ganglia	Left	UL	Left	Paralysis	Hoarseness	Recovered	Good
Bogousslavsky 1990 [26]	Switzerland	CS	M	63 y	Stroke (Type NM)	CT	Internal capsule genu	Left	UL	Right	Paresis	Dysarthria	Permanent paralysis	Good
Bathini 2014 [23]	Canada	CS	F	4 m	Stroke (Type NM)	MRI	Thalamus	NM	BL	BL	Paralysis	Dysphonia	Permanent paralysis	Fair
Saito 2006 [9]	Japan	CS	F	7 m	Neonatal asphyxia	MRI/FLAIR	Striatum, thalamus, Medulla, Pons,	BL	BL	BL	Paresis	Stridor	NM	Good
Allam 2015 [27]	USA	CR	F	47 y	Ischemic stroke	MRI	Medulla, Pons	BL	BL	BL	Paresis	Stridor	Permanent paralysis	Good
Shetty 2016 [30]	UK	CR	M	59 y	Ischemic stroke	CT	Cerebellum, Medulla	Right	UL	Left	Paralysis	Dysarthria	NM	Fair
Diaz-Perez 2023 [31]	Spain	CR	F	88 y	Ischemic stroke	CT	cerebellum, bulbar region	Left	UL	Left	Paralysis	Stridor	Permanent paralysis	Good
Darshan 2022 [32]	India	CR	M	40 y	Ischemic stroke	MRI	Medulla	Right	UL	Right	Paralysis	Hoarseness	Permanent paralysis	Good
Valente 2023 [33]	Portugal	LTE	M	69 y	Stroke (Type NM)	MRI	Medulla	NM	UL	Right	Paresis	Dysarthria	Recovered	Good

CR: Case report, CS: Case series, LTE: Letter to the editor, y: years, m: months, F: Female, M: Male, CT: Computed Tomography, MRI: Magnetic Resonance Imaging, NM: Not Mentioned, BL: Bilateral, UL: Unilateral

A total of 12 patients were included across the selected studies, evenly split between genders. Among them, three were pediatric cases, and the remaining nine were adults. Ischemic stroke was identified as the cause of hypoxic injury in seven cases, with an additional four cases with the type of stroke not mentioned. Perinatal asphyxia was noted as the cause in one pediatric case.

MRI was the imaging modality used in five cases, while CT was utilized in four cases. One case underwent both MRI and CT, and another underwent MRI with FLAIR sequence. The hypoxic injury resulting in vocal cord palsy (VCP) primarily affected the brainstem, particularly the medulla and pons. Other affected regions included the basal ganglia, thalamus, internal capsule, and cortex (MCA territory which involves lateral and dorsal aspects of the temporal, parietal, and frontal lobes, as well as deeper structures including the caudate, internal capsule, and thalamus) [14], and cerebellum. In two cases, the level of injury was not specified.

The distribution of hypoxic injuries was as follows: six cases on the left side, two bilateral cases, two on the right side, and two cases where the side was not mentioned. Vocal cord involvement was unilateral in eight cases and bilateral in four cases. In five cases, the involvement was ipsilateral to the hypoxic injury (cortex, basal ganglia, cerebellum, and medulla), bilateral (Striatum, thalamus, medulla, pons) when the injury was bilateral and contralateral (internal capsule genu, cerebellum, medulla) in two cases. This relationship could not be determined in three cases due to insufficient details. The topographic representation of hypoxic injuries along the laryngeal innervation found in our study is shown in Fig. 2.

Fig. 2.

MRI Brain sagittal section showing the areas involved by hypoxic injury in different cases included in our study.*As the laryngeal motor cortex, basal ganglia, and internal capsule cannot be well appreciated on the sagittal section, their approximate locations are presented here

Laryngoscopic examination revealed total immobility (paralysis) in six cases and decreased mobility (paresis) in five cases, with one case lacking documentation. Presenting symptoms included dysarthria (four cases), hoarseness (three cases), stridor (three cases), and dysphonia (two cases). Among the cases with paralysis, four did not recover, while two cases of paresis progressed to paralysis. Two cases of paresis and one of paralysis showed improvement, and the case with undocumented laryngoscopy also improved. Outcome details were unavailable for two cases.

Discussion

Laryngeal innervation is intricately structured. The signals for the motor control of voice begins in the association cortices composed of prefrontal, cingulate, inferior parietal, precuneal, and middle temporal areas, which are sent to Laryngeal Motor Cortex (LMC) [15, 16]. LMC is located in the primary motor cortex (Brodmann area 4) of the frontal lobe [17]. The vocal motor commands from LMC are modulated via the subcortical loops of basal ganglia (putamen, globus pallidus), pontine gray matter, and cerebellum. The modified program is relayed back to LMC via the ventrolateral thalamus [15]. LMC comprises ventral LMC (vLMC) and dorsal LMC (dLMC) leading to two separate pathways: the LMC pathway and the limbic vocal control pathway, respectively [10, 11]. The projection fibers of pyramidal cells (first-order neurons) of dLMC travel in the corticobulbar tract passing through the corona radiata, genu of internal capsule, middle third of crus cerebri, and pons to medulla [18]. Here, half of the fibers travel ipsilaterally to establish monosynaptic synapses with the laryngeal motor neurons (LMNs) of ipsilateral nucleus ambiguus, and half of the fibers cross to the contralateral side to synapse with LMNs contralaterally [15]. From LMNs, the efferent fibers (second-order neurons) exit the medulla via the vagus nerve to supply the laryngeal muscles [19]. This pathway is referred to as the LMC pathway [10, 15]. It is concerned with complex and learned voluntary laryngeal movements required during speaking and singing [10, 11]. Unlike the fibers from dLMC, the projection fibers (first-order neurons) from the vLMC reach the anterior cingulate cortex (ACC) in the limbic lobe to synapse with second-order neurons [10, 17]. The fibers from ACC travel to the periaqueductal gray (PAG) to supply third-order neurons, which in turn sends innervation to fourth-order neurons in the pontine reticular formation (PRF) in the pons [15, 17]. From PRF, the nerve fibers descend to the nucleus ambiguous on each side such that LMNs receive equal innervations from bilateral LMCs [15]. This pathway forms the limbic vocal control pathway, which controls innate nonverbal and emotional vocalizations, such as laughter and cry [10].

Vocal cord palsy can result from insult to any of the structures along this innervation pathway. A wide variety of causes have been described for VCP, however, hypoxic neuronal injury as a cause of VCP has been rarely discussed. Hypoxic brain injury can occur in three patterns either as a watershed, basal ganglia-thalamus, or total brain injury, commonly affecting the periventricular white matter, basal ganglia, thalamus, posterior limb of the internal capsule, and medial temporal lobe [20, 21]. Hypoxia can damage the innervation pathway to the vocal cord from LMC to laryngeal motor neurons (LMNs) in the nucleus ambiguus [15]. Hypoxic injury drives the neurons into very less efficient anaerobic metabolism resulting in rapid depletion of ATP stores [22]. This results in the failure of energy-dependent active transporters of ions across the cell membrane (e.g. Na⁺-K⁺ ATPase pump). It causes a rapid influx of sodium and calcium ions into cells culminating in cellular swelling, membrane damage, increased concentration of excitatory amino acids in ECF, generation of nitrous oxide and free radicals, mitochondrial dysfunction, and an inflammatory response. This cascade of events results in neuronal death, thus, denervation [7, 22, 23].

The lesions in the cases included in our study involved hypoxic injuries ranging from the cerebral cortex to LMNs in the medulla. A patient had a cerebral infarct in the left Middle Cerebral Artery (MCA) territory (specific site not mentioned) resulting in permanent left vocal cord paralysis [8]. MCA has four segments M1, M2, M3, and M4. M1 gives of lenticulostriate arteries which supply deeper structures such as basal ganglia, internal capsule, and thalamus, while others are concerned with supplying the majority of dorsolateral cortex [24] including LMC and association areas [11]. This infarct probably involved one or more of these areas. However, the finding is in contrast to the fact that unilateral lesions rarely result in vocal fold paralysis as each vocal cord receives equal innervation from each LMC [15]. Likewise, in another patient, an MR angiography revealed severe stenosis in the second portion of left MCA resulting in ipsilateral paresis which recovered [25]. Here, the ischemia would have involved the dorsolateral cortex only as stenosis was in M2, more so involving only the neurons related to ipsilateral vocal cord innervation. Moreover, the third case also had left MCA territory infarction resulting in bilateral vocal cord involvement which also recovered [26]. Here, the area of involvement in the brain would be similar to the first case, with ischemic injury involving the neurons related to vocal cord innervation bilaterally. The recovery in both cases could be due to resolving mild hypoxic injury or probable regeneration [27], and the intact innervation from right LMC [11].

A patient with an infarct in the left basal ganglia had left vocal cord paralysis which recovered over time [28]. Basal ganglia are involved in the modulation of vocal motor commands of LMC along with other subcortical structures [15]. Thus, the infarct could have involved the neurons in the left basal ganglia concerned with innervation to the left vocal cord only, as the right vocal cord was intact. The return of left cord function can be attributed to the probable regeneration of the neurons [27] or bilateral innervation of each vocal cord wherein the intact innervation from the right side led to the recovery [11].

A small infarct in the genu of the left internal capsule resulted in the paresis of the right vocal cord which later evolved into permanent paralysis [29]. The paresis can be due to the fact the corticobulbar fibers pass via the genu of the internal capsule [18]. Maybe only the fibers that cross to the contralateral side (left to right) were injured resulting in paresis of the right cord. But, the evolution into permanent paralysis cannot be explained as the ipsilateral supply from the right was still intact [15].

A case with thalamic stroke had bilateral permanent VCP [26]. The side of the lesion was not mentioned but, probably involved bilateral thalami. As the ventrolateral thalamus relays the modified motor program for vocal control to LMC which is executed [15], the stroke in thalami led to interruption of this pathway bilaterally leading to bilateral vocal cord paralysis.

Left paramedian pons and right medial medullary infarction in a case resulted in permanent bilateral VCP [30]. The corticobulbar tract carrying first-order neurons of the LMC pathway passes via pons [18] and PRF is involved in the limbic vocal control pathway [10]. All LMNs concerned with innervating ipsilateral laryngeal muscles lie in the medulla and are usually involved in lateral medullary infarction [31]. Here, the left paramedian pons infarct could have led to bilateral vocal cord paresis as far as equal bilateral innervation is concerned [15]. And, medial medullary infarct usually spares vocal pathways [32]. So, bilateral vocal cord paralysis is unexplainable based on these lesions. Thus, the infarction might have involved a larger area than outlined in the imaging also involving lateral medulla or pons on the right side.

A left cerebellar and medullary infarct in a patient caused ipsilateral VCP the outcome of which was not mentioned [33]. Similarly, an infarct established in the left hemi-cerebellum and the lateral bulbar region, congruent with the territory of the left inferior posterior cerebellar artery in another patient resulted in ipsilateral permanent paralysis [34]. The LMNs in the left lateral medulla [11]might have been damaged irreversibly in both cases. In addition, the co-ordinating pathways to and from the cerebellum [17] might have been disrupted.

A right lateral medulla infarct in a patient caused permanent ipsilateral paralysis [35] while a small medullar stroke (side not mentioned) in another patient resulted in right vocal cord paresis which recovered [36]. Here, the infarct involved almost all LMNs in the right lateral medulla resulting in permanent paralysis in the former, while only a few LMNs later resulted in paresis which recovered in the latter.

Apart from the causes of hypoxic brain injury discussed in the above cases, neonatal asphyxia has also been associated with vocal cord involvement. In our study, a patient who suffered neonatal asphyxia at birth, had bilateral pontomedullary tegmentum, bilateral putamen, and thalamus lesions [9]. These areas are usually involved in an acute hypoxic-ischemic insult to neonatal brains due to their high susceptibility attributed to high metabolic demands and level of myelination [37, 38]. The brainstem tegmentum being a watershed area in the vertebrobasilar vascular system is further vulnerable to hypoxia which forms part of Dorsal Brainstem Syndrome [37, 38]. She suffered bilateral vocal cord paresis, the outcome of which was not mentioned [9]. As discussed in previous cases, the corticobulbar tract and PRF lie in the pons, LMNs in the medulla, and putamen and thalamus are involved in the modulation of vocal motor program. The lesions involving these areas bilaterally in this patient led to bilateral VCP.

Conclusions

Our study was based on 12 cases grouped from different case reports and series. There are no large sample studies published addressing the topic in question. Further, for some of the studies that met the eligibility criteria, the full text could not be retrieved even via the Research4Life database and direct mailing to the corresponding authors, and hence excluded. Among the studies included, few lacked details about the site and type of hypoxic injury, vocal cord involvement, and outcomes. While acknowledging the limitations of our study, this study sheds light on the complexities of the laryngeal innervation pathway and the vulnerability of neurons to hypoxic insults. Given the scant literature available and the substantial morbidity associated with VCP, there’s a pressing need for further large-scale investigations to assess the effects of hypoxic injury on laryngeal innervation.

In conclusion, our study provides insights into how hypoxic brain injury can cause VCP and correlates the level of lesions along the innervation pathway with the clinical presentation. VCP can be induced by hypoxic injuries like ischemic stroke and neonatal asphyxia to the neurons in the nuclear and supra-nuclear levels of the vagus nerve extending from the LMC to the LMNs in the medulla. Hypoxia damages neurons causing failure of energy-dependent active transporters of ions across the cell membrane. Findings in some cases could not be justified with available knowledge on innervation from LMC to LMN which calls for further comprehensive research in this area.

Funding

This study was not funded from any source.

Declarations

Ethical Approval

Not applicable.

Informed Consent

Not applicable.

Compliance with Ethical Standards

This study is done adhering to applicable ethical standards.

Conflict of interest

The authors have no conflicts to disclose.