Clinical manifestations

Presentation and course

The onset of spasmodic dysphonia is insidious in the majority of patients (84%), initially presenting as nonspecific hoarseness with gradual worsening of vocal quality over months to years (82; 06; 175). In a smaller proportion of patients, the onset is sudden (82). The most common inciting events identified by patients were stress, upper respiratory infection, and pregnancy and parturition (36; 04). Patients are often seen by multiple physicians prior to a definitive diagnosis.

Spasmodic dysphonia may be divided into three types, depending on the specific intrinsic laryngeal muscles that are most affected. If the adductor musculature is affected (thyroarytenoid, lateral cricoarytenoid, or interarytenoid) then the laryngeal dystonia is termed adductor spasmodic dysphonia, which results in the classic strained strangled speaking pattern (107). Adductor spasmodic dysphonia is the most common form and accounts for up to 82% of cases (22). If the abductor musculature is affected (posterior cricoarytenoid), the laryngeal dystonia is termed abductor spasmodic dysphonia, resulting in a breathy, effortful speaking pattern with poor generation of volume. Patients may have a mixed spasmodic dysphonia with both adductor and abductor musculature affected, which may result in speech with both phonatory breaks from adductor spasms and breathy breaks from adductor spasms. Abductor spasmodic dysphonia occurs in 17% of cases, and the mixed type is rare (23).

Spasmodic dysphonia most commonly affects normal speech, typically in the normal talking range of pitch, volume, and speed. Laughing, whispering, screaming, and yawning may normalize the voice. The fluency may also be improved by speaking in a higher pitch. These characteristics are likely related to the task-specific nature of dystonia. Similarly, some patients report that sensory tricks, such as touching the larynx, supporting the head, or lying down, may improve speech. Voice improvements also occur when the patient is taken by surprise or in sudden danger. Vacation often brings an improvement whereas stress can worsen vocal quality. Sometimes improvement is noted with sedative and alcohol use. Most patients report voice worsening with speaking in public or on the telephone (78).

Table 1. Factors Influencing Subjective Severity of Spasmodic Dysphonia

Spasmodic dysphonia has a female preponderance (77.6%), with an average age of onset at 51 years (134). Most patients present only with the vocal dysfunction without neurologic manifestations in other areas of the body (23). Spasmodic dysphonia less frequently occurs in the setting of segmental or generalized dystonia. In one study, patients with cough, dyscoordinate breathing, paroxysmal sneezing, and hiccups were found to have a higher incidence of extralaryngeal dystonia (137). The reported prevalence of extralaryngeal dystonia in patients with spasmodic dysphonia varies from 5% to 14% (62; 134). Those patients with spreading dystonia had a mean 3.3 body parts involved, mostly involving the lower face, jaw, tongue, or neck. The mean time before spreading was 7.3 years, with a median time of 4 years, and a range of 1 to 24 years (62). Tisch and colleagues report the occurrence of extralaryngeal dystonias (including blepharospasm and orofacial and cervical dystonia) preceding the onset of spasmodic dysphonia by a mean of 70.8 months (178). In one series that included 901 patients with vocal involvement, 82.5% had primary dystonia and 17.5% had secondary dystonia (23). In another series, 15.8% of patients with spasmodic dysphonia had spread of dystonia, all of whom developed cervical dystonia (17).

Reports of task-specific laryngeal dystonia occurring only with singing (singer’s laryngeal dystonia) have also been described (22). Respiratory laryngeal dystonia results in adductor spasms during respiration, resulting in dyspnea and inspiratory stridor. They do not usually lead to hypoxia.

Other associated neurologic conditions include dystonic head and neck tremors and postural and action tremors. Comorbid vocal tremor has also been reported (178; 189; 134).

Untreated spasmodic dysphonia may also be associated with increased psychological comorbidity. A study of 44 patients with adductor spasmodic dysphonia revealed substantial degrees of perceived handicap and low perceived control of the condition (86). Rates of anxiety and depression appear to be higher in patients with spasmodic dysphonia and other voice disorders compared to the general population (188). In an analysis of 142 patients receiving botulinum toxin injections for spasmodic dysphonia, anxiety was present in 13.4%, and depression was present in 2.8% (75).

Spasmodic dysphonia can functionally affect patients’ work productivity (78). Patients with spasmodic dysphonia report a 20% to 30% mean decrease in voice-related work productivity, which is ameliorated after treatment with botulinum toxin injections (123). The work productivity decreases are mainly in the form of presenteeism (122).

Prognosis and complications

Spasmodic dysphonia tends to emerge gradually in midlife and then reaches a plateau in terms of severity. Spontaneous remission has not been reported in spasmodic dysphonia as seen with cervical dystonia. Treatment with botulinum toxin injections has altered functional disability but repeated treatment is needed, possibly lifelong.

Clinical vignette

Case 1. The patient in his late 40s was working in law enforcement and doing a sting operation at a warehouse location. On-site, there was a lot of construction dust. He began to have vocal difficulties that he felt were due to laryngitis from irritant exposure, but his voice did not return even when he moved to a different project location. His voice continued to deteriorate and became more strained and effortful. About 10 years later, he saw a speech-language pathologist with some improvement in his voice. He then noticed that while on amusement-park rides with his grandchildren, his voice was reproducibly normal. His speech-language pathologist accompanied him to the amusement park to witness this change in his voice and, thus, diagnosed him with spasmodic dysphonia. He did not seek treatment until 2013 and was initially treated with 1 unit of botulinum toxin A to the bilateral thyroarytenoid muscles. He had significant breathiness and some difficulty with swallowing, and his dose was reduced. Currently, he receives an asymmetrical dose of 0.15 units of botulinum toxin A to one thyroarytenoid muscle and 0.25 units of botulinum toxin A to the contralateral thyroarytenoid muscle with vocal improvement from 40% to 50% normal function at baseline (when the effects of botulinum toxin have dissipated) to 80% to 90% of normal vocal function at the peak effect of the botulinum toxin injection with minimal breathiness or dysphagia.

Case 2. A 43-year-old female reported acute onset of voice changes in July 2016. She endorsed a strained effortful and “garbled” voice quality. She noted improvement in vocal quality with singing and with alcohol consumption. She was seen by her local ENT and referred to a speech-language pathologist and then further referred to the university for evaluation and management. As the vice president of her community college who frequently spoke to large groups, she had an extraordinary demand for voice use and quality in her vocational capacity. She was a nonsmoker and was otherwise healthy, with no other medical problems and no family history of neurologic disorders.

The patient was diagnosed with adductor spasmodic dysphonia and initially treated with LEMG-guided botulinum toxin injections with a good effect. But after 2 years of treatment, she struggled with her injections causing too much initial breathiness and a shorter-than-desired duration of peak voicing. She was changed to an endoscopic approach to her injections and currently receives botulinum toxin injections to the bilateral thyroarytenoid and to the interarytenoid musculature. Her total dose is 2.5 units every 10 weeks of which she typically experiences 1 week of initial breathiness and mild dysphagia to liquids and 9 weeks of effect. She reports 4 weeks of peak voicing (90% to 100% of normal function), with a gradual decline to her baseline of 15% to 20% of normal function.

Biological basis

Etiology and pathogenesis

The etiological and pathophysiological mechanisms underlying spasmodic dysphonia are not known. Spasmodic dysphonia may occur as an isolated dystonia, with Meige syndrome, or as part of generalized dystonia (116; 84; 66; 02; 115; 138).

Several causative genes have been identified in spasmodic dysphonia, often in the setting of a dystonic syndrome.

THAP1: Mutations in the DYT6 gene THAP1 typically present with dystonia affecting the cervical, cranial, and upper limb musculature, often with laryngeal involvement (21; 190). THAP1 mutations have also been identified in patients with early-onset generalized dystonia with spasmodic dysphonia (50).

TUBB4: Mutations in the TUBB4 gene on chromosome 19p13.3-p13.2, which encodes for a neuronally expressed tubulin, cause DYT4 dystonia, originally described in an Australian family (104). This form of dystonia is characterized by prominent spasmodic (“whispering”) dysphonia associated with craniocervical dystonia and a “hobby horse” type gait.

ANO3: Mutations in ANO3 (DYT24) typically present with craniocervical dystonia, including spasmodic dysphonia. Mild upper limb dystonia may also be present (172). Tremor is a characteristic feature of ANO3 mutations, differentiating it from the typical DYT6 phenotype.

GNAL: Mutations in GNAL (DYT25) are another cause of cervical dystonia and may be associated with head tremor and spasmodic dysphonia (11). Generalized dystonia occurs in approximately 10% of cases. Isolated laryngeal dystonia/spasmodic dysphonia has also been described (142).

KMT2B: This syndrome, which is caused by mutations in the KMT2B gene, results in a progressive childhood-onset dystonia, with prominent cervical, cranial, and laryngeal dystonia (121). It is associated with typical facial features of an elongated face and bulbous nasal tip. MR imaging reveals characteristic findings of subtle symmetrical globus pallidi hypointensity, with a hypointense lateral streak of bilateral globus pallidus externa.

A form of dysphonia similar to spasmodic dysphonia has also been described in dominantly inherited ataxia with dentate calcification, currently assigned the name spinocerebellar ataxia type 20 (94).

Careful clinical characterization of the dystonic syndrome allows accurate phenotype-genotype correlation and may assist in identifying an underlying genetic diagnosis. Genetic screening of patients with spasmodic dysphonia targeted at mutations in TOR1A, THAP1, and TUBB4 has a low diagnostic yield (63; 43).

Other reported associations include neuroleptic exposure, either immediately or as part of the tardive syndrome (186; 05), mitochondrial disease (139), valproic acid administration (with improvement after discontinuation) (132), central pontine myelinosis (163), amyotrophic lateral sclerosis (149), psychogenic dysphonia (157; 10), late-onset spasmodic dysphonia with low arylsulphatase A (117), essential tremor (105), palatal myoclonus (51), hereditary spastic paraplegia type 7 (64), multiple sclerosis (48), or trauma (56). Factors associated with an increased risk of spasmodic dysphonia, in small or isolated case-controlled studies, include a past history of mumps, blepharospasm, tremor, and intense occupational voice use (176), a personal history of cervical dystonia, sinus and throat illnesses, rubella, dust exposure (177), a family history of voice disorders and tremor (23), an immediate family history of vocal tremor and meningitis, and an extended family history of head and neck tremor, ocular disease, and meningitis (177).

There is evidence to suggest that spasmodic dysphonia is caused by abnormalities of large-scale brain networks, rather than due to pathology limited to the basal ganglia (60). Neurophysiological studies reveal that patients with adductor spasmodic dysphonia have a shortened cortical silent period compared with healthy controls. Similar findings are also observed in patients with focal hand dystonia, suggestive of widespread cortical excitability in both of these conditions affecting the motor cortex representing asymptomatic regions of the body (152; 153). Measurements of excitability over the dominant primary motor cortex during “linguistic” and “non-linguistic” tasks following transcranial magnetic stimulation have been shown to differ in patients with spasmodic dysphonia compared with healthy controls, with restoration of these changes following treatment with botulinum toxin (174). A study utilizing EEG found larger gamma band coherence in late vocalization between somatosensory and premotor cortical areas in patients with spasmodic dysphonia, indicating excessively large synchronization between these two areas (89).

The spasmodic dysphonia patient has abnormal blink reflex recovery, pointing to a loss of inhibitory control (37). The difference in R2 amplitude attenuation to electrical and mechanical stimulation suggests that the dystonia involves not only the larynx but also other anatomical structures (37). Loss of inhibitory control is also seen in many other focal dystonias (133) and the underlying defect may be impaired inhibition at the cortical level leading to a decrease in the requirement for activation with voluntary motor commands (34).

Sharbrough and colleagues described abnormal auditory brainstem responses in 7 of 18 patients with spasmodic dysphonia, indicating slower brainstem conduction along the auditory pathway (164). In another study of six patients with adductor spasmodic dysphonia using the auditory brainstem response, five of six patients had a compromised capacity of the auditory brainstem to conduct impulses (57). In a study of 12 spasmodic dysphonia patients using three different auditory brainstem response parameters, 75% were abnormal. Three of the 12 had prolonged wave I-V interpeak latency. Seven had pathologic wave V latency shifts at a high stimulus rate (158). These findings were not confirmed by Middleton in a study of 14 spasmodic dysphonia patients with normal hearing (124). Postmortem brainstem examination in two patients with spasmodic dysphonia revealed several neuropathological changes compared to controls (169). This included small clusters of inflammation in the reticular formation surrounding the solitary tract and spinal trigeminal nuclei and in the pyramids, in addition to neuronal degeneration and depigmentation in the substantia nigra and locus coeruleus.

Using quantitative topographic electrophysiologic mapping and SPECT, Devous and colleagues suspected that dysfunction of cortical perfusion, cortical electrophysiology, or both, occurred in spasmodic dysphonia (45). Studying a 59-year-old male patient with adductor type spasmodic dysphonia during phonation with PET, Hirano and colleagues noted remarkable activities during phonation in the left motor cortex, Broca area, cerebellum, and the auditory cortices, whereas the supplementary motor area was not activated (70). In normal subjects, significant activities were observed during vocalization in the motor area, Broca area, the supplementary motor area, and the cerebellum, whereas the auditory association area was not activated, even though the subjects heard their own voice (68). Compared with normal activities there were two apparent differences: one was the lack of activity in the supplementary motor area, and the other was activation in the auditory association area. The auditory association area was not activated during normal vocalization, but came to be activated when the speaker’s own voice was distorted (68; 69), as is the case with the spasmodic dysphonia strained voice. As the supplementary motor area is known to function for motor planning, programming (145) is usually activated in normal phonation and damage of the supplementary motor area causes a severe disturbance of voluntary vocalization. Hirano and colleagues suggested that the functional deficit of the supplementary motor area might be related to spasmodic dysphonia (70). In a PET study using [11C] raclopride (RAC) to assess striatal dopaminergic neurotransmission, patients with spasmodic dysphonia demonstrated decreased RAC displacement during symptomatic speech production compared with controls, indicating decreased dopaminergic transmission (167). RAC displacement was increased during the unaffected task of asymptomatic tapping, possibly representing a compensatory adaptation of the nigrostriatal dopaminergic system.

When fMRI measurements were performed during vocal motor tasks in patients with laryngeal dystonia and compared with healthy volunteers, reduced activation of primary sensorimotor as well as premotor and sensory association cortices during vocalization in patients with laryngeal dystonia were noted (65). Another study utilizing fMRI found reduced functional connectivity between the left inferior parietal cortex, putamen, and bilateral premotor cortex (14). Gray matter volume, cortical thickness, and brain activation on fMRI were increased in the laryngeal sensorimotor cortex, inferior frontal gyrus, superior or middle temporal and supramarginal gyri, and cerebellum in a study of 40 spasmodic dysphonia patients compared with controls (170). Phenotype-specific abnormalities have been reported in adductor and abductor forms of spasmodic dysphonia on high-resolution MRI and diffusion-weighted imaging. Differences in cortical thickness and white matter fractional anisotropy were observed in the left sensorimotor cortex and angular gyrus and the white matter bundle of the right superior corona radiate (18). In addition, genotype-specific abnormalities were seen between sporadic and familial cases in the left superior temporal gyrus, the supplementary motor area, and the arcuate portion of the left superior longitudinal fasciculus.

Lenticular nucleus hyperechogenicity has been observed on transcranial sonography in patients with spasmodic dysphonia (185). A study examining a combination of independent component analysis and linear discriminant analysis of resting-state functional magnetic resonance imaging data found abnormal functional connectivity within sensorimotor and frontoparietal networks in patients with spasmodic dysphonia compared with control subjects (13). In addition, differences in abnormal functional connectivity appeared to distinguish between the different clinical phenotypes of spasmodic dysphonia, as well as between the genetic and sporadic forms. Differences in cortical surface area have also been described in subjects with spasmodic dysphonia compared with control subjects in regions associated with sensorimotor integration, motor preparation, and motor execution, as well as in areas of processing of auditory and visual information (98).

Histologic analysis of the recurrent laryngeal nerves in two patients revealed no apparent signs of either destruction or degeneration. The percentage of thin nerve fibers, the diameter of which may range from 5 to 10 microns, however, was higher than in normal controls (97). In another study, recurrent laryngeal nerve removed from patients with spasmodic dysphonia at the time of surgery, using light and electron microscopy, were compared with control recurrent laryngeal nerves (33). Slight morphometric differences were found between the two groups, but these cannot explain causation of spasmodic dysphonia (33).

Epidemiology

The prevalence of primary laryngeal dystonia is estimated to be 5.9 per 100,000 (09). Spasmodic dysphonia occurs more often in women than in men (23; 171; 01). In one series, women made up to 79.3% of the population with spasmodic dysphonia (01). The overall ratio ranges between 1.4 to 3.8 females to 1 male (23; 171). Broken down into subgroups, the female-to-male ratio was 4.1:1 for the adductors and 2.2:1 for the abductors (162). Izdebski and colleagues report in a series of 200 patients with spasmodic dysphonia, the age of onset at 41 years (+13.25 SD) with a range of 6 to 65 years for males. For females, the mean age at onset was similar, at 45.4 years (+13.3 SD), with a range of 7 to 78 years (82).

Prevention

There are no known methods to prevent spasmodic dysphonia.

Differential diagnosis

Confusing conditions

The differential diagnosis of spasmodic dysphonia is broad and includes both organic and functional disorders. Most of these conditions can be excluded by careful clinical examination; however, in some cases the diagnosis may be challenging. Essential voice tremor and muscle tension dysphonia can cause voice breaks and can form the most important entities of the differential diagnosis. The movement disorder in essential tremor is rhythmic rather than spasmodic, and it often involves pharyngeal and strap muscles. Lundy and colleagues, by applying acoustic and motor speech parameters to the problem, found that unlike spasmodic dysphonia, tremor is more often marked by fluctuations in frequency rather than just in intensity (110). Finally, it should be noted that tremor may coexist with dystonia in as many as one third of patients (95). Laryngeal endoscopy helps in the differential between spasmodic dysphonia, vocal tremor, and muscle tension dysphonia.

Muscle tension dysphonia may also mimic the strained voice quality of abductor spasmodic dysphonia (74). Evidence of consistent sound-specific phonatory breaks should raise suspicion of adductor spasmodic over muscle tension dysphonia. The hyperadduction of muscle tension dysphonia is generally sustained and is unlikely to be spasmodic. Neither tremor nor muscle tension dysphonia demonstrate task specificity, although the voice may deteriorate with stress in both conditions. Although differences between muscle tension dysphonia and adductor spasmodic dysphonia have been described on fiberoptic laryngoscopy, phonatory airflow measurement, and acoustic analysis, there is currently no single diagnostic test to differentiate these two disorders (150). The diagnosis is made clinically based on perceptual voice evaluation. In psychogenic dysphonia, there are often a number of atypical characteristics, including loss of normal shouting, yawning, and laughing. Psychogenic dysphonia also does not present with tremor (103). In psychogenic dysphonia, the vocal symptoms are more often invariant across phonetic and most phonatory variables and not associated with sound prolongations or voice arrests (103). In one study, psychogenic speech and voice disorders were found to occur in 16.5% of 182 patients with psychogenic movement disorders. Among these patients, stuttering was the most common speech abnormality (n = 16, 53.3%), followed by speech arrests (n = 4, 13.3%), foreign accent syndrome (n = 2, 6.6%), hypophonia (n = 2, 6.6%), and dysphonia (n = 2, 6.6%) (10).

Vocal cord polyps or other mass lesions can cause dysphonia characterized by strain and vocal breaks. These lesions would be visualized and diagnosed on laryngeal endoscopy. There are also reports of stridor resulting from laryngeal dystonia, and in some cases, this may be misdiagnosed as asthma (178). Table 2 lists some differential diagnoses of spasmodic dysphonia.

Even among experts, there can be notable variation when diagnosing patients with spasmodic dysphonia, vocal tremor, or muscle tension dysphonia. In one study, international experts were asked to classify the vocal diagnosis of 178 patients across four different sites into 11 categories. The experts reviewed a portion of these patients’ video-laryngeal endoscopy and speech recordings, but due to a lack of diagnostic guidelines, there was poor interrater agreement regarding patient diagnoses (108).

Table 2. Differential Diagnosis for Spasmodic Dysphonia

Associated or underlying disorders

Most commonly, spasmodic dysphonia occurs sporadically, and there are no associated or underlying disorders.

Diagnostic workup

Understanding of the nature of the symptoms, including onset, aggravating factors, alleviating factors, associated symptoms, duration of symptoms, and the character of symptoms is the focus of the history and helps lead to a correct diagnosis.

Table 3. Voice History and Symptoms of Spasmodic Dysphonia

The diagnosis is based primarily on auditory-perceptual features of the speaking voice (106). The primary modality to assess spasmodic dysphonia is by listening to the patient speak during conversational speech and during elicited speech tasks with either vowel or voiceless consonant predominant sentences (87). One should assess the fluidity of the voice, the quality of articulation, and the quality of the vocal signal. A normal voice should be smooth, without voice breaks or spasm.

Listening to the patient reading sentences that are designed to elicit either adductor or abductor spasms is also helpful.

Adductor spasmodic dysphonia phrases or sentences would include words with vowels or voiced sounds such as:

Abductor spasmodic dysphonia phrases or sentences would include words with vowels following a voiceless consonant (such as d, f, h, p, s or t).

Some clinicians have patients read the phonetically balanced “rainbow passage”:

Because vocal tremor, muscle tension dysphonia, psychogenic dysphonia, and, occasionally, other laryngeal disorders can cause speech characteristics similar to spasmodic dysphonia, a physical examination is essential for the evaluation. Attention should be paid to a complete neurologic examination. Spasmodic dysphonia usually presents sporadically and without other neurologic findings, so any tremor, weakness, or cranial nerve neuropathy should prompt further neurologic evaluation. Furthermore, laryngeal endoscopic imaging will help to rule out other types of dysphonia diagnoses including vocal fold lesions, paralysis or paresis, or scarring. The addition of a laryngeal stroboscopic assessment to the endoscopic imaging will allow a detailed evaluation of the vibratory function and architecture of the vocal folds. Small vocal fold cysts and polyps may result in alteration of the voice, and stroboscopic examination may help determine the lesion’s contribution to the patient’s vocal complaints.

History, physical examination, and laryngeal stroboscopy are all critical components of the diagnostic algorithm, but the most important component remains the auditory perceptual evaluation of voice (120).

Researchers have tried to identify modalities to diagnose spasmodic dysphonia and differentiate it from muscle tension dysphonia, but these have not gained wide clinical use. The long-term average spectrum (LTAS) assesses the average amplitude spectrum across a selective frequency range and provides information on the spectral distribution of the speech signal over a period of time (74). One study suggested that the long-term average spectrum may identify spectral noise differences between muscle tension dysphonia and adductor spasmodic dysphonia in women. The use of neural network and support vector machine-based methods, in combination with a pattern recognition algorithm, has also been studied (160). Fine kinemetric analysis from high-speed digital imaging may assist in the clinical differentiation of adductor spasmodic dysphonia and muscle tension dysphonia, although further studies are required (136). High-speed videoendoscopy has been able to detect a rapid sustained adduction shortly after onset of phonation in a small sample of patients with adductor spasmodic dysphonia compared to muscle tension dysphonia (35).

Other measures of voice quality, which may provide further characterization of spasmodic dysphonia, include the IINFVo (Impression, Intelligibility, Noise, Fluency, and Voicing) perceptual rating scale and the AMPEX (Auditory Model Based Pitch Extractor) acoustical analysis (165). Measurements in perceptual, acoustic, and self-assessment dimensions all demonstrate significant improvement in symptoms following botulinum toxin therapy. However, these three parameters were found to have poor intrinsic correlation; thus, a tridimensional approach may be preferable (44). Another study found that the latency between the initiation of thyroarytenoid electrical activity and the onset of phonation was significantly related to the severity of adductor spasmodic dysphonia (38). The technique of airflow interruption may also provide additional quantitative information regarding laryngeal function in spasmodic dysphonia (71). One study examining the perceptual structure of adductor spasmodic dysphonia and the acoustic correlates of underlying perceptual factors identified a two-factor model of adductor spasmodic dysphonia, characterized by hyperadduction and hypoadduction (32). Perceived ratings of roughness appeared to be most representative of the hyperadduction factor, whereas breathiness was most strongly associated with the hypoadduction factor. Onset of relative fundamental frequency measures were found to be related to perceived vocal effort in patients with adductor spasmodic dysphonia in another study (52). Cepstral analysis and machine-learning algorithms have been able to distinguish between spasmodic dysphonia patients and controls and between pre- and post-botulinum toxin treated patients (173).

In the setting of associated clinical features, such as segmental dystonia, tremor, or gait disturbance, careful clinical characterization of the phenotype may provide clues to an underlying genetic dystonic syndrome (see etiology and pathogenesis).

Management

Botulinum toxin. Botulinum toxin injections are the mainstay of treatment for spasmodic dysphonia. The American Academy of Otolaryngology-Head and Neck Surgery as well as the American Academy of Neurology endorses botulinum toxin as primary therapy for spasmodic dysphonia. Blitzer and colleagues performed the first botulinum toxin treatment on a spasmodic dysphonia patient (24), and their work was confirmed by a subsequent double-blind study (181). Meta-analysis of 30, mostly single-blind studies, indicated moderate overall improvement as a result of botulinum toxin treatments (28). The Cochrane Collaboration Review noted that although only one study (181) was double-blind and fulfilled their inclusion criteria out of the nearly 77 reports, most of the studies had similarly positive effects related to the length of treatment, degree of improvement, patient satisfaction, and observed side effects (187).

Treatment of adductor spasmodic dysphonia results in an average onset of action of 2.4 days. The average peak benefit is 9 days, and this lasts for approximately 15.1 weeks (26). Patients report an improvement to 91.2% of normal (22). Another study noted average subjective phonation improvement of 33% over a similar duration (113). Lower functional gain was noted in patients with dystonia in other body regions (113). In another study examining outcomes of onabotulinumtoxinA, 88.1% of injection cycles for 328 patients with spasmodic dysphonia were noted to be maximally beneficial (135). Treatment with botulinum toxin leads to a reduction in spasmodic contractions observed on video laryngoscopy and reduces the voice handicap index score (53). A metaanalysis showed significant improvement in several quality of life measures, including the Voice Handicap Index and Voice-Related QoL scales (54). Fine wire electromyography has revealed that both the thyroarytenoid and the lateral cricoarytenoid muscles might be affected in adductor spasmodic dysphonia, even if the thyroarytenoid is more predominant (93). In contrast, the thyroarytenoid and lateral cricoarytenoid muscles are equally involved in spasmodic dysphonia with tremor.

Different protocols have been proposed for injecting botulinum toxin into the thyroarytenoid muscle, either unilaterally or bilaterally. Initial dosages of 1.25 units bilaterally of onabotulinumtoxinA resulted in a statistically significant shorter duration of breathiness without significant differences in clinical effectiveness or voice outcome in one study, compared with an initial dosage of 2.5 units bilaterally (147). Unilateral injection may result in fewer adverse events such as breathiness or hoarseness (96; 19). Unilateral injections improve the performance ratio of strong voice interval divided by breathy voice interval (96). Some studies have shown that the duration of benefit in men is significantly longer and with less swallowing difficulty after bilateral injections (112). Upile and colleagues found that low-dose unilateral injections resulted in no significant difference in patient outcome in terms of voice duration, voice score, and complication rate when compared with bilateral injections (183). However, another study found that bilateral injections had a significantly higher proportion of patients than unilateral injections (36% vs. 33%, p=0.0228), with both greater than 3-month duration of effect and less than 2-week duration of side effects (47). The dose of botulinum toxin required for the treatment of adductor spasmodic dysphonia tends to remain stable over time (148). There appears to be a lower rate of immunoresistance with botulinum toxin in spasmodic dysphonia compared with other dystonias, and this may be related to the low antigen challenge associated with the low doses used (126). Higher doses were required for symptom control in female patients with adductor spasmodic dysphonia compared to males in one retrospective chart review, although the reasons for this observation are unclear (102).

In the authors’ experience, for a chemodenervation naïve patient, we start with botulinum toxin A at a dose of 0.5 to 1 unit in the bilateral thyroarytenoid muscle for the adductor spasmodic dysphonia. This starting dose is based on Blitzer's original work. Patients are warned about potential coughing with liquids and breathy vocal quality, which typically last no more than a week. They are then instructed to return when their voice breaks start to worsen. At that time, they are asked to judge their best voice on a scale of 0% to 100%, with 100% being a completely normal voice and 0% being no voice at all. They are then asked about any initial side effects and how long those symptoms lasted. At that point, the decision is made to increase or decrease the botulinum toxin dose or change the injection interval.

There are different techniques of laryngeal botulinum toxin injection. The most common is the EMG-guided percutaneous approach (125). Injections can also be performed using external landmarks, termed the point-touch technique (127). Injections under direct visualization, either by endoscopy or mirror examination, include a transoral approach (58), a percutaneous approach, and a transnasal approach (143).

To perform the EMG-guided percutaneous approach to the thyroarytenoid muscle, a Teflon-coated injection needle is connected to the EMG machine. The needle is inserted just off midline through the space between the cricoid and thyroid cartilages, pointing towards the ipsilateral thyroarytenoid muscle. The localization of the needle is verified by crisp motor unit potentials heard on EMG when the patient performs a long “/i/” (125; 29). Others additionally use laryngoscopic control (181).

Patients with adductor spasmodic dysphonia and associated vocal tremor may benefit from combined interarytenoid and thyroarytenoid muscle botulinum toxin injections (88). Botulinum toxin injections into the interarytenoid muscle have a 50% response rate in patients previously unresponsive to traditional injection sites (87).

Possible complications from botulinum toxin injections into the thyroarytenoid musculature include breathiness, throat pain, and dysphagia. Initial functional decline following botulinum toxin injection was observed in 28.5% of cases (131). In a series of 1300 patients over 24 years of age, adductor injections were associated with a 25% risk of mild, transiently breathy voice and a 10% risk of transient coughing with drinking fluids. Breathy dysphonia has been reported as high as 50.9%, lasting on average 20 days (49). Local pain, bruising, and itch were reported in less than 1% of individuals (22). Rarely, bilateral abductor paralysis has been reported and is likely a result of toxin diffusion around the muscular process of the arytenoid to the posterior cricoarytenoid muscles (184). In one series of 131 patients, 67 of 574 injections (12%) were categorized as failures: failure was associated with injection by a new physician, the patient being a professional voice user, and a lack of breathiness in the voice after injection (192).

Treatment of abductor spasmodic dysphonia is less satisfactory, with patients reporting an improvement to 70.3% of normal (22). In addition to a less successful outcome, the injection of botulinum toxin into the posterior cricoarytenoid muscle is more challenging. The average onset of effect is 4.1 days, with a peak effect at 10 days. The mean duration of benefit is 10.5 weeks (26). Successful outcomes have been reported using an initial unilateral posterior cricoarytenoid muscle injection, followed by a contralateral injection after 2 weeks if necessary (22). Simultaneous bilateral posterior cricoarytenoid injections have also shown success in improving vocal outcome measures (91). Injection techniques usually include EMG guidance and laryngeal rotation (25), endoscopic guidance (146), and transcricoid rostrum (118). A major concern for bilateral injections is bilateral vocal fold abductor weakness and the consequent narrowing of the airway for adequate respiration causing dyspnea, exertional wheezing, stridor, and dysphagia.

Botulinum toxin doses. Five different botulinum toxin products are available on the market. Four of these toxins are botulinum toxin type A and include Botox® (Allergan, Inc., Irvine, California, United States), Dysport® (Ipsen Ltd., Slough, Berkshire, United Kingdom), Xeomin® (Merz Pharmaceuticals, United States), and CBTXA (China). Myobloc®/Neurobloc® (Solstice Neurosciences, Inc., South San Francisco, California, United States) is botulinum toxin type B. In the United States, the available preparations include Botox®, Xeomin®, and Myobloc®/NeuroBloc®.

Doses of botulinum toxin used for the treatment of spasmodic dysphonia vary depending on the practitioner and the particular brand of toxin used. In the early literature, the doses of botulinum toxin (Botox®) used for adductor spasmodic dysphonia ranged from 3.75 to 7.5 U for bilateral injections (31; 29; 30; 181) to 15 U for unilateral injections (125; 109). Up to 50 U per vocal cord has been used (85). Later literature and common practice have recommended the use of lower doses (27). In one retrospective study of 126 patients, the mean dosage of onabotulinumtoxinA was 1.54 +/- 0.35 units per side (59). Injections can be bilateral, unilateral, staggered, or with asymmetrical doses (118; 180; 183).

The optimal dose of botulinum toxin varies between patients, and does not appear to correlate with severity of spasmodic dysphonia, age, or gender. In one study, BMI and overall health were correlated with a higher effective dose (191). In one retrospective study, patients with adductor spasmodic dysphonia who received long-term injections appeared to have an initial reduction followed by stabilization in their dosage (129). Those who have a fluctuating dosing trend tended to have a shorter interval between injections.

Surgery. After recognition of the organic etiology of spasmodic dysphonia, the development of recurrent nerve resection confirmed the neurologic basis of the disease process and provided initial remarkable results. Dedo reported the results of sectioning of the recurrent laryngeal nerve in 34 patients (40). The first patient was a woman with a 17-year history of spasmodic dysphonia who had been treated by more than 29 physicians from different specialties without improvement. Blocking of the recurrent nerve with lidocaine improved her spasm. After five repeated reproducible results with lidocaine, the patient underwent sectioning of the right recurrent laryngeal nerve with essentially normal voice within 1 week after the procedure, with the help of speech therapy. Subsequently, 33 additional patients underwent similar procedures from a group of 41 patients who responded to recurrent nerve block. Dedo also performed placebo injections, which did not improve the patient’s voice (40). Short- and mid-term results were good, with a recurrent rate of 10% to 15%, respectively (83; 41; 42). Long-term results showed gradual decline in the benefit from the procedure.

Another less invasive procedure involved crushing the recurrent nerve, which appears initially equally effective. This procedure too suffered from widespread recurrence of symptoms with a success rate of 13% at 3 years (20). Gross recurrent laryngeal nerve destruction is no longer a contemporary management for spasmodic dysphonia although this procedure is of great historical importance regarding the recognition that this disease has a neurologic and not psychiatric pathophysiology.

Selective laryngeal adductor denervation-reinnervation (SLAD-R), first described in 1999 (16), involves selective denervation of the adductor branch of the recurrent laryngeal nerve. The distal nerve stumps are reinnervated with a nonlaryngeal nerve, generally the sternohyoid branches of the ansa cervicalis nerve (119). The initial report presented favorable results in 19 of 21 patients. Only one patient underwent further treatment with Botox postoperatively (16). Similar results were also reported in a smaller series in which expert and untrained judges undertook perceptual evaluation of voice quality (03). Additional studies have also reported favorable results after a mean follow-up time of 7.5 years (119). Functional reinnervation of the vocal cord adductors by ansa cervicalis has been observed 10 years after successful SLAD-R surgery (39).

Isshiki and colleagues offered an apparently mechanical solution to the problem of hyperadduction in adductor spasmodic dysphonia with laryngeal framework surgery (81; 79). In this procedure, a lateralization laryngoplasty is performed by means of a midline division and expansion of the thyroid cartilage. They reported success in a small series with informal judgments regarding voice quality made by both patients and physicians. A further small study using laryngeal framework surgery, or type II thyroplasty, also resulted in significant acoustic and aerodynamic improvement in adductor spasmodic dysphonia (156). Follow-up after 2 to 5 years revealed unsatisfactory outcomes in 7 of 90 cases and were mainly attributed to inadequate technique and in one case, an incorrect indication (154). Surgical mechanical faults appear to be the main cause of unsuccessful outcomes (80). Another long-term follow-up study (average of 41.3 months) following type II thyroplasty also demonstrated sustained improvement in voice symptoms and quality of life (155).

Bilateral thyroarytenoid muscle myectomy is also used in the treatment of adductor spasmodic dysphonia. One study found no difference in the postoperative voice handicap index 10 score between patients who underwent bilateral thyroarytenoid muscle myectomy compared to those who received type II thyroplasty (130). Bilateral thyroarytenoid muscle myectomy appeared to improve strangulation, interruption, and tremor and worsen breathiness when compared to type II thyroplasty. Successful treatment of adductor spasmodic dysphonia has also been reported in a small series of patients using selective lateral laser thyroarytenoid myotomy (76; 61). Sustained improvements in voice handicap index and GRBAS scale have been observed following endoscopic laser thyroarytenoid myoneurectomy (182; 61).

Surgical treatment for abductor spasmodic dysphonia is more nuanced and has been less described in the literature. Bilateral vocal fold medialization has been used by some. In one case series, six patients underwent vocal fold medialization via type 1 thyroplasty and demonstrated significant benefit on Voice Handicap Index (VHI) and Voice-Related Quality of Life (VR-QOL) measures. The four patients who had undergone the procedure at least 1 year prior to their survey reported a sustained benefit to their symptoms (46). One patient with staged bilateral posterior cricoarytenoid partial myoneurectomy demonstrated sustained benefit up to 8 years postoperatively (15).

Deep brain stimulation (DBS) for adductor spasmodic dysphonia is being increasingly described for the treatment of spasmodic dysphonia and vocal tremor. Deep brain stimulation of the globus pallidus internus (GPi) resulted in improvement of segmental dystonia, including spasmodic dysphonia in one patient, with maintenance of symptom control at 10-year follow up (73). Some patients with primary dystonia with a component of spasmodic dysphonia had no benefit to their spasmodic dysphonia with GPi stimulation (77). One patient with essential tremor and adductor spasmodic dysphonia had sustained benefit to her spasmodic dysphonia with bilateral ventral intermediate (Vim) nucleus of the thalamus stimulation (111). Another patient with essential tremor and coincident adductor spasmodic dysphonia treated with unilateral thalamic deep brain stimulation experienced significant improvement in vocal dysfunction with isolated Vim and combined Vim and ventral oralis anterior (Voa) nucleus of the thalamus (141). Honey and Kruger have reported promising results with bilateral thalamic deep brain stimulation (99; 72).

Other therapeutic modalities. Voice therapy has been reported to be helpful in conjunction with botulinum toxin treatment to retrain a long-term compensatory behavior superimposed on spasmodic dysphonia (128), although this has not been consistent. Another study did not demonstrate additional benefits in patients receiving voice therapy compared to those receiving botulinum toxin alone or to those who received botulinum toxin with sham voice therapy. Further studies are required to confirm these results (166). Significant improvement with speech therapy may suggest an alternate diagnosis of muscle tension dysphonia (12). In a systematic review of 21 studies of voice therapy in patients with spasmodic dysphonia, the adherence rate was quite low (about 60%) (114). Laryngeal vibrotactile stimulation temporarily improved symptoms in 9 of 13 patients (90). This improvement was associated with a suppression of theta band power in the left somatosensory-motor cortex, similar to what is seen during successful sensory tricks.

Although to a lesser degree than in the past, many spasmodic dysphonia patients are still subjected to psychiatric treatment. Psychotherapy is useful to help coping with social stresses, which can be considerable; the patient, however, needs to be made aware first that his or her condition is an organic disorder and to be treated accordingly either with botulinum toxin or surgery. Acupuncture has not been proven effective by expert raters, although Lee and colleagues found a majority of patients reported subjective improvements in voice production (101). Electrical stimulation of the thyroarytenoid muscle was reported to improve symptoms in one small case series (140). An open-label study of sodium oxybate demonstrated reduced voice symptoms in alcohol-responsive spasmodic dysphonia in those with and without coexistent voice tremor (151); unfortunately, the medication is expensive and only has a duration of up to 3 or 4 hours, which has limited more common clinical use (168).

Special considerations

Pregnancy

Pregnancy has been associated with the sudden onset of adductor spasmodic dysphonia in small studies (04). Botulinum toxin injections are not given to women who are pregnant.

Anesthesia

Laryngeal spasms associated with spasmodic dysphonia will relax under sedation or anesthetic; thus, there are no additional considerations for anesthesia in abductor spasmodic dysphonia or adductor spasmodic dysphonia.