Clinical manifestations

Presentation and course

Patients with cervical dystonia have involuntary head and neck movements resulting in abnormal postures, spasmodic motions, or tremor. The most prominent feature is usually sustained rotational deviation of the head to one side.

Terms such as "laterocollis," "anterocollis," and "retrocollis" describe the direction of head movement tilted laterally, forwards, and backward, respectively. Anterocollis may be subdivided into conceptual anterocollis (when it involves exclusively the position of the cervical spine), anterocaput (when it affects exclusively the head position), and anterior or posterior sagittal shift (when it affects head and cervical spine positions) (52). Similarly, lateral tilt (laterocollis) can be distinguished from lateral shift of the head. A combination of these distortions of posture is usually present. Associated asymmetric hypertrophy of involved neck muscles is expected. Superimposed on the sustained abnormal posture may be fast components in the form of spastic jerks or head oscillation (tremor). Head oscillations are present in 28% to 68% of patients and can have a regular or irregular component (114). In fact, isolated head tremor can be a presenting feature. One longitudinal study followed 20 patients with an isolated head tremor and no other neurologic symptoms (51). After 5 years, 15 of those patients had developed dystonic posturing in the neck. The abnormal head and neck movements disappear when the patient is sleeping. There should be no contractures in most cases, but in patients with a prolonged history of cervical dystonia, there may be fixed deformities in the neck.

Patients may describe actions that improve the symptoms of cervical dystonia. This phenomenon, previously known as a “sensory trick” or “geste antagoniste,” has been renamed “alleviating maneuver” (101; 18). Alleviating maneuvers are helpful in establishing the diagnosis of cervical dystonia. In a study involving 154 subjects with cervical dystonia, 138 (89.6%) reported an alleviating maneuver (102). There was marked improvement of cervical dystonia in 39.8% of subjects and partial improvement in 43.4%. The most common alleviating maneuver was a light touch to the lower face or neck. Other alleviating maneuvers that have been described include resting the occiput against a wall, abducting the ipsilateral shoulder, placing objects on the head, or placing a hand above the head. The pathophysiological mechanisms of alleviating maneuvers are not well understood, but using paired-pulse transcranial magnetic stimulation in patients with cervical dystonia, these maneuvers were thought to improve dystonia through an inhibitory effect on motor cortex excitability (04).

Neck pain is a common feature of cervical dystonia. This and many other features of cervical dystonia were explored in the multicenter study, CD-PROBE, involving 1046 patients with cervical dystonia assessed after three consecutive treatments with botulinum toxin (64). In this study, neck pain was reported in over 95% of participants, and the rate of moderate to severe pain was 70.7% (24). In one multicenter, cross-sectional study, substance abuse was found in 11% of patients with cervical dystonia (85). Risk factors identified were younger age, male gender, and comorbid mood disorder.

Patients with cervical dystonia have been found to have an increased risk of secondary degenerative changes of the upper cervical spine, particularly C2/C3 and C3/4 on the side ipsilateral to the head tilt (27). Patients with degenerative changes had more restriction in head mobility and longer time without adequate treatment. Cervical arthritis can contribute to pain, limitation of head movement, and poor response to botulinum toxin and surgical therapy.

Headache attributed to craniocervical dystonia is included in the classification of the International Headache Society as a secondary headache (12). The pain is in the location of the muscles affected by the dystonia, and it follows a temporal progression with the cervical dystonia. However, a study that evaluated the frequency of headache in patients with and without cervical dystonia showed that the frequency of headache in both groups was similar (around 50%), and the most common headache type was migraine (06).

Swallowing function may be abnormal even before treatment, especially in patients with extreme retrocollis. In a study of 43 patients, over 50% of patients had objective findings of swallowing abnormalities under videofluoroscopic examination in the form of delayed swallowing reflex and vallecular residue (107). In another study with 25 patients with cervical dystonia, 72% had dysphagia based on electrophysiological evaluation (49).

Extra cervical involvement is not uncommon in cervical dystonia. In an analysis of 1477 patients with cervical dystonia, 60.7% did not develop any other type of abnormal movements (97). On the other hand, 17.9% who initially had cervical dystonia, developed dystonia in other areas, most commonly hands, face, and head. One study demonstrated that 28% of patients diagnosed with focal cervical dystonia also had one other body part affected in addition to the neck. The most common site was the shoulder, followed by hand, face, and larynx. Movement disorder specialists often diagnosed focal cervical dystonia as dystonia isolated to the neck or the neck and a shoulder, whereas dystonia affecting the neck and the arm, face, jaw, or larynx was diagnosed as a segmental dystonia (71). Postural hand tremor or head tremor are more common associations. Tremor was noted in 63% of patients with isolated focal or segmental cervical dystonia, and the spread of dystonia commonly occurs in areas affected by the tremor (97). In fact, patients presenting with head tremor as the primary manifestation of cervical dystonia may have a tremor-dominant subtype of cervical dystonia with evidence of cerebellar dysfunction as demonstrated by greater cerebellar disability scores on the Scale for the Assessment and Rating of Ataxia (SARA), increased gait variability, and reduced cerebellar vermian volume (86). Hand tremor is seen in at least 30% of patients and usually manifests with symmetrical postural and action tremor (22). Whether the hand tremor in cervical dystonia is a manifestation of dystonia or coexisting essential tremor is a matter of debate (112). Notably, the co-occurrence of dystonia, especially cervical dystonia, in families with essential tremor was reported (83).

The Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS), a comprehensive cervical dystonia rating scale has been found to be internally consistent and potentially useful as an outcome measure for therapeutic trials of cervical dystonia (35). In the CD-PROBE study, the minimal clinically important change (MCIC) on TWSTRS for improvement was noted to be ≥8 points based on mean TWSTRS total scores in patients with cervical dystonia when compared against the Patient Global Impression of Change (41).

Prognosis and complications

Spontaneous remission occurs in about 10% to 20% of patients (61; 22), usually taking place within the first year of symptoms and with decreasing frequency as the illness becomes more chronic. In general, younger patients are more likely to have spontaneous remission, and the occurrence of remission falls off exponentially after 5 years.

The natural course of cervical dystonia is not well described, though most experts agree that symptoms tend to worsen in the first 5 to 10 years after onset after which it usually remains stable (02).

Cervical dystonia can worsen quality of life and may be associated with depression and other mood changes. Psychiatric features and level of pain seem to be the main determinants of disability (127). One study showed that even with effective pain relief in cervical dystonia with botulinum toxin use, mood symptoms, such as depression and anxiety, remained unchanged, suggesting that cervical dystonia is more than just a motor disorder (92). An international survey of patients with cervical dystonia reported that although the majority were satisfied with their relationship with their physician, they often had high expectations from treatment, hoping to be free from spasms and pain and return to their normal routine (32). Educating patients about the natural history of cervical dystonia and reasonable expectations from therapy is an important part of the treatment plan.

Given that the onset of cervical dystonia in most patients is in the 4th or 5th decade, it is not surprising that employment and work productivity have been found to be negatively affected even in treated patients (91). Pain has been found to be a primary driver of these adverse impacts, and preliminary evidence suggests that initiating treatment with onabotulinumtoxinA may reduce absenteeism and presenteeism (reduced productivity).

Clinical vignette

A 48-year-old woman presented with a seven-year history of involuntary head rotation to the right, which was persistent throughout the day and was associated with superimposed bouts of forceful spasms that would throw her head backward. The involuntary movements were worse when she was nervous or in front of a crowd. She felt that using her hand to touch her chin would partially relieve her symptoms, allowing her to maintain her head straight for a short period of time. She had severe pain in her neck on the right side. She was a telephone operator and had stopped working for 3 years because of this condition. She admitted that she was depressed.

On examination, she looked anxious. Her head posture was abnormal, being held to the right. There were episodic quick jerky movements of her head being pulled backward. Her left sternocleidomastoid muscle appeared hypertrophic. She was able to partially correct the head posture by touching her chin with her right hand. There was a fine hand tremor of her outstretched hands. The rest of her examination was normal.

Routine blood tests were normal. X-rays of her cervical spine revealed abnormal head and neck posturing and early changes of cervical spondylosis. CT scan of her head was normal.

A diagnosis of cervical dystonia was made. She was subsequently treated with botulinum toxin injections given intramuscularly to her neck muscles. Her left sternocleidomastoid, right splenius capitis, and right levator scapulae were treated. She reported improvement in her head posturing and a reduction in neck pain after 3 days, and clinical response lasted 10 weeks. She was treated with repeated injections every 3 to 4 months and continued to respond favorably to the treatment.

Biological basis

Etiology and pathogenesis

The etiology of cervical dystonia remains unknown but appears to be heterogeneous. Genetic causes and a relationship to trauma have been studied. In terms of underlying neurophysiology, dystonia has been considered to be a network disorder and involves basal ganglia, cerebellum, pedunculopontine nucleus, thalamus, and cortex (68; 122; 89).

Genetics. The role of inheritance is poorly defined. Although the majority of cases are sporadic, a family history of dystonia may be found in some patients. In a review of 1000 patients with idiopathic cervical dystonia, 14% of patients reported a family history (76). A handful of variant genes have been identified in patients with cervical dystonia, but they are disease-causing only in a small percentage of patients.

DYT7 gene locus on chromosome 18p has been associated with families from Germany and Central Europe with cervical dystonia (77), but the significance of this association has been questioned (131). The DYT13 gene locus has been described in idiopathic torsion dystonia, presenting with predominantly craniocervical and upper limb dystonia in an Italian family (126). Polymorphism in the dopamine D5 receptor gene has been reported to link with the development of cervical dystonia, suggesting a candidate gene at the DRD5 locus (15). The contribution of polymorphism Val66Met in the gene encoding brain-derived neurotropic factors (BDNF) for the pathogenesis of dystonia was debated, but a meta-analysis that pooled 784 patients with dystonia and 1020 controls supported a true association (109).

More convincing and potentially monogenetic causes of the cervical dystonia phenotype are also now supported in the literature. Over 50 families have been described with mutations in DYT6 (encoding for THAP1, a proapoptotic protein), with dystonia primarily affecting the cervical, cranial, and upper limb muscles (14). THAP1 mutations decrease transcription regulation and DNA binding (106). It also interferes the myelination within the oligodendrocyte lineage, which may implicate abnormal timing of myelination with dystonia. A mutation was found in exon 7 of CIZ1 (c.790A> G, p.S264G) in a large Caucasian pedigree with adult-onset cervical dystonia using whole-exome sequencing (134). CIZ1 encodes CIZ1 (Cip1-interacting zinc finger protein 1) involved in DNA synthesis and cell cycle control. A study of 12 Chinese probands with familial cervical dystonia failed to identify a mutation in CIZ1 (84). Another study of 45 Chinese patients with isolated cervical dystonia found mutations in the COL6A3 gene, suggesting an autosomal recessive cause of early-onset isolated dystonia (133). This gene encodes the collagen alpha-3 chain, which is one of the three subunits of collagen type IV. Mutations in GNAL have been found to be a cause of primary torsion dystonia in several families, with the neck as a common site of onset (55; 74; 21). GNAL encodes the protein Gαolf, which is important for dopamine D1 receptor function. However, a study with 192 British probands failed to identify GNAL mutations as the causative gene for cervical dystonia (25). Mutations in the anoctamin3 gene (ANO3) have been associated with autosomal dominant craniocervical dystonia (26). The gene encodes a calcium-gated chloride channel, which is highly expressed in the striatum. The most common site of onset was the neck. Head and hand tremors were frequently present, but myoclonus was only occasionally seen. DYT-TUBB4A, formerly known as DYT4 of “whispering dysphonia,” involves mutations to TUBB4A, with the most consistent feature being laryngeal involvement though cervical dystonia was seen in 60% of the reported cases (05).

In a multicentric cohort involving 1000 patients with cervical dystonia, the contribution of known genes to cause dystonia was small. Pathogenic or likely pathogenic sequence variants of GNAL, TAHP1, and TOR1A were found in 0.8% of the participants (76). Synonymous and noncoding sequence variants in THAP1 and GNAL were found in 4% of the patients. In the largest genome-wide association study (GWAS) of cervical dystonia to date, which included 919 patients and 1491 controls, no variants were found that could be validated in a replication study (121). Based on these studies, a very small percentage of patients with cervical dystonia have monogenic disease. The majority of patients are likely to have a combination of multiple risk-modifying sequence variants and exposure to environmental contributors.

Metabolomics. Metabolomics may offer additional insight into the pathophysiology of cervical dystonia. The metabolome reflects the combined systemic effects of genetics, lifestyle, environmental exposure, and secondary biological responses. One study identified nine biological processes associated with cervical dystonia, including oxidized lipid 2-hydroxydecanoate, which may be abnormal in cervical dystonia (80).

Trauma. The role of trauma in the pathogenesis of cervical dystonia is unclear. Clinically, torticollis occurring shortly after neck injury differs from typical idiopathic cervical dystonia in that usually no improvement appears during or after sleep, and no help is provided by "sensory tricks" (125). Patients who develop cervical dystonia immediately after a relatively mild neck injury have been described (57), but the significance remains unclear. In another report, however, 12% of patients developed symptoms similar to classical cervical dystonia shortly after head or neck injury (110). These patients had an earlier age of onset with a higher prevalence of neck pain and might reflect a triggering effect of trauma on patients who were predisposed to the development of cervical dystonia. Consistently, a subgroup of patients exists: those who present with symptoms quickly after trauma (with severe pain being a major symptom), have fixed torticollis with a greater tendency for laterocollis, and are resistant to treatment (108; 98). It has been proposed that the disorder should be called “posttraumatic painful torticollis” to differentiate these patients from those with the usual form of the disorder (108).

Pathophysiology. Despite the advances in the last 20 years in identifying genes that are present in patients with cervical dystonia, there is an important gap in knowledge of how those changes translate to abnormal function of higher-order motor control that generates dystonic symptoms (16). Mutant mice for TOR1A have impaired formation of afferent synapsis in a set of interneurons called GABApre (135). Those interneurons are located in the dorsal spinal cord and inhibit proprioceptive sensory afferent terminals, and they negatively regulate sensory-motor signaling. Changes in brain circuits are also implicated in the genesis of cervical dystonia. A meta-analysis of studies with voxel-based morphometry of gray matter volume in individuals with primary focal dystonia showed changes in the post-central gyrus, primary motor cortex, thalamus, and putamen, which implicates structural abnormalities in the sensorimotor network (136). Another study of functional MRI and cervical dystonia implicated other areas of the brain, including dorsal cingulate gyrus, cerebellum, and different areas of the frontal lobe (79). A postmortem neuropathological study in six cervical dystonia brains revealed a significantly lower density of Purkinje cell density in the cerebellum compared to controls, suggesting possible cerebellar dysfunction in cervical dystonia (104). Another study with 188 patients with cervical dystonia showed that 14% had clinical or radiological evidence of cerebellar involvement (07). Finally, a study of 32 patients with cervical dystonia showed significant reductions in diffusion tensor imaging (DTI)-based tractography, suggesting microstructural abnormalities within dentatorubrothalamic tracts consistent with decreased axonal integrity (116). This is suggestive of abnormal functional connectivity between the cerebellum and somatosensory cortex.

The presence of alleviating maneuvers also indicates a dysfunction of sensorimotor integration. Data from a primate model of dystonia show that there are enlarged and overlapped sensory receptive fields (117). Positron emission tomography studies on the alleviating maneuvers reported on ipsilateral (to direction of head turn) superior and inferior parietal lobe as well as bilateral occipital cortex activation associated with reduction of activities in the contralateral supplementary motor areas and primary sensorimotor cortex (13). Evidence expands on this important role of dysfunction in sensorimotor cortex, including defective inhibitory plasticity (48), abnormal sensory gating (38), and impaired cortical modulation of brainstem circuits (100).

Epidemiology

The incidence of cervical dystonia was evaluated in a study in Rochester, Minnesota between 1960 and 1979. The overall incidence rate was 1.2 per 100,000 (29). Based on a 2-million subject commercial database in the United States, a study estimated the prevalence in the United States was estimated to be 390 per 100,000 (66). A systematic literature review on the epidemiology of cervical dystonia found a wide range in the reported values, with an estimated prevalence of 28 to 183 cases per million and an incidence of 8 to 12 cases per million person-years (43). This is consistent with the reported incidence estimate for cervical dystonia of 1.07 per 100,000 person-years (95% CI: 0.86-1.32) (118).

The peak age at onset is in the fourth to fifth decade with a mean age of 42 years (43). A 2:1 female-to-male ratio has been consistently reported. Occasionally, it starts at a much younger age and has been reported as early as 3 years of age in a series of young-onset cases. In a report of 76 patients with cervical dystonia with disease onset before the age of 28, the male-to-female ratio was 1.24:1 and the mean onset age was 21 (3 to 28) years (73). A family history of a movement disorder was present in 35.7% of patients in a data set of focal dystonia in Ireland, whereas a positive history of dystonia was present in 16.5% of the cases (130).

Also see the article titled Epidemiology of movement disorders.

Differential diagnosis

Confusing conditions

There is often a delay in the diagnosis of cervical dystonia. In one study, a total of 108 patients saw a mean of 3.5 providers over a mean period of 44 months from symptom onset to diagnosis (123). Cervical dystonia should be distinguished from an isolated head tremor that can be seen in essential tremor. In contrast to essential tremor, cervical dystonia is associated with an abnormal posture, tends to be irregular, and is associated with a null point (a particular head position that attenuates the tremor).

Secondary causes of torticollis should be excluded before reaching a diagnosis of idiopathic cervical dystonia (120; 75). Secondary causes include structural lesions, drug exposures, neurodegenerative diseases such as Parkinson disease, multiple system atrophy, and progressive supranuclear palsy. Congenital or traumatic bony abnormalities in the cervical spine may be detected by x-rays of the region. Septic arthritis of C1 to C2 lateral facet joint, pharyngeal abscess, cervical epidural abscess, and osteomyelitis have been reported to present with torticollis.

A variety of tumors, including cerebellar cavernous angioma, frontal meningioma, posterior fossa tumor, and spinal astrocytoma or ependymoma, have all been reported to present as torticollis.

In children, one should be aware of the condition known as Sandifer syndrome, in which tilting of the head may relieve pain associated with a hiatal hernia. Congenital shortening and fibrosis of the sternocleidomastoid muscle may present in infancy as torticollis. This may sometimes be associated with congenital vertebral anomalies.

Ocular causes may produce head tilting as a compensatory act for diplopia. Trochlear nerve palsy is an example, but head tilting can occur in other ocular muscle palsies. Similarly, patients with vestibular diseases may tilt their heads to find the position with least vertigo.

Rare causes of torticollis include basal ganglia infarction, systemic lupus erythematosus, psoriatic spondylitis, midbrain lesion in multiple sclerosis, and ataxia-telangiectasia.

Cervical dystonia can be caused by exposure to dopamine receptor-blocking agents such as neuroleptics and antiemetics (56). Tardive cervical dystonia should be suspected in the setting of such medication exposure. In one cross-sectional study, clinical features such as predominant retrocollis and extra cervical involvement were associated with tardive cervical dystonia whereas predominant torticollis, head tremor, sensory trick, and family history were associated with idiopathic cervical dystonia (90). There have also been isolated reports of cervical dystonia caused by cholinesterase inhibitors such as donepezil (60).

Diagnostic workup

The diagnosis is usually made clinically, and no confirmatory test for cervical dystonia exists. Excluding secondary causes is most important. A detailed drug history should be taken to exclude drug-induced dystonia. In patients with onset of illness younger than the age of 40 years, Wilson disease should be ruled out.

CT scan and magnetic resonance imaging may be performed in selected patients with cervical dystonia. However, these investigations provide little diagnostic information, and usually no pertinent pathology can be found.

One study proposed an algorithm to help diagnose cervical dystonia and better exclude mimickers (42). In a patient with patterned, repetitive muscle spasms inducing stereotyped abnormal movements/postures in the head/neck (tremor may or may not be present), one should first ascertain if a sensory trick is present. If so, cervical dystonia is a likely diagnosis. If not, one must actively exclude the following to rule out mimickers: fixed involuntary head and neck postures, focal neck weakness, and diplopia induced by correction of the abnormal head posture. This study compared 43 patients with idiopathic cervical dystonia and 41 control subjects. This algorithm achieved a 96% sensitivity and 81% specificity in diagnosis of cervical dystonia in these patients (42).

Management

Botulinum toxin. Botulinum toxin injected intramuscularly into the dystonic neck muscles is helpful in relieving symptoms of cervical dystonia and is considered a first-line treatment for this disorder (01; 62; 103). Type A (onabotulinumtoxinA, abobotulinumtoxinA, incobotulinumtoxinA) and type B (rimabotulinumtoxinB) formulations of botulinum toxin are approved for the treatment of cervical dystonia (115). This toxin prevents the release of acetylcholine at nerve terminals and induces temporary, graded weakness in the dystonic muscles with improvement in head posture, involuntary movements, and pain. Botulinum toxin has also been shown to improve depression and anxiety in patients with cervical dystonia (39). The available formulations of botulinum toxin have been extensively evaluated in clinical trials, demonstrating significant improvement in cervical dystonia compared to placebo (78; 124; 34; 23).

Anterocollis is often excluded from clinical trials of botulinum toxin because it is presumed that this treatment is not effective; however, some patients clearly improve with injections, although the risk of dysphagia is higher than with other forms of cervical dystonia (63). However, one study found that targeting muscles such as longus colli, longus capiti, suprahyoid muscles (predominantly the anterior heads of the digastric), and the sternocleidomastoid muscles could be targeted in anterocollis, with benefit reported by patients (88). Injection of these muscles remains technically challenging. This study utilized ultrasound combined with EMG guidance for injections into the longus colli muscles and aid from an ENT using an endoscopic approach via the nose to inject the longus capiti muscles. The side effects of botulinum toxin when treating cervical dystonia include dysphagia, flu-like symptoms, and neck weakness. Based on pivotal trials, the general recommendation is that botulinum toxin be administered no sooner than every 3 months; however, the optimal dosing frequency is still being investigated.

One posthoc analysis evaluated flexible dosing of incobotulinumtoxinA at intervals ranging from six to 20 weeks in patients with cervical dystonia, based on patient request and clinical examination (50). Intervals between injections were less than 12 weeks in 369 of 821 (44.9%) treatment cycles for cervical dystonia. The severity and frequency of adverse events were similar in subjects receiving injections at intervals less than 12 weeks compared 12 weeks or more. The most common side effects were dysphagia, muscular weakness, neck pain, and injection-site pain. This study evaluated up to six treatment sessions and, thus, long-term studies are needed to assess the development of immunoresistance with chronic therapy at shorter intervals. A study utilizing incobotulinumtoxinA, showed that patients with cervical dystonia who experience early waning of incobotulinumtoxinA effect benefit from shorter injection cycles without an increased risk of side effects or loss of benefit over time (33).

With the potential advantage of a longer duration of response, injectable DaxibotulinumtoxinA is in development as an alternative to currently available botulinum toxins. In an open-label, dose-ranging, phase 2 study, the median duration of response in 33 subjects was nearly 6 months (65).

There is no clear consensus on the optimal injection technique for botulinum toxin; however, an injection plan is generally based on head position, location of pain, and presence of hypertrophy in muscles likely to contribute to the observed head position. Despite differences in technique among practitioners, there is general agreement that it is best to use the lowest effective dose at each treatment session and only repeat injections as often as necessary.

Muscles are targeted for injection based on a variety of factors including the direction of head position. The classic head positions along with most common muscles injected include torticollis (splenius, levator scapulae, semispinalis, contralateral sternocleidomastoideus), laterocollis (levator scapulae, semispinalis, scalenus medius), retrocollis (bilateral semispinalis), and anterocollis (bilateral sternocleidomastoideus, bilateral levator scapulae) (69). Injection of more anterior muscles can cause dysphagia when doses are too high, so great caution should be exercised. To further help with targeting muscles, the injector will often rely on EMG activity as well as muscle palpation, feeling for tenderness and muscle hypertrophy. Whether automated kinematic assessments can improve on injection plans based on traditional examination methods is a subject of investigation (111). Jost and colleagues proposed a system (Col-Cap concept) that further subdivides head positions by differentiating caput from collis. Collis involves neck movement above C2, whereas with caput, the neck remains fixed above C2. For example, torticollis can be subdivided as torticaput (no upper neck rotation) or torticollis (includes upper neck rotation) (69). They further propose muscle injection patterns according to these subdivisions. At this time, it is unclear if this approach will improve outcomes.

Oral medications. Various oral agents are commonly used in clinical practice in off-label fashion. However, there are limited data to support the use of oral medications for the treatment of cervical dystonia (103). Anticholinergic drugs such as trihexyphenidyl have been described to be effective in a small number of patients. Baclofen, a GABA B receptor agonist, may be helpful in some patients as well. Benzodiazepines, such as lorazepam and clonazepam, can also be helpful. Rarely, levodopa has been found to ameliorate cervical dystonia in patients with adult-onset dopa-responsive dystonia (113). These agents are often limited by side effects but may be helpful as an adjunct treatment with botulinum toxin.

Other agents have been tried in botulinum toxin–refractory cervical dystonia, such as riluzole, a medicine aimed at antagonizing glutamatergic action in the basal ganglia. A six-patient, open-label study showed clinical improvement in half the patients, but these results have not been replicated in controlled clinical trials (93). Perampanel, an antiseizure agent that antagonizes AMPA receptors, was evaluated in an open-label phase 2a study (54). At doses typical of its use in epilepsy, perampanel was poorly tolerated and showed no motor benefit for motor symptoms of cervical dystonia.

Deep brain stimulation. Deep brain stimulation can be helpful for patients with cervical dystonia who are refractory to other treatments. A prospective study of bilateral GPi deep brain stimulation in 10 patients with cervical dystonia found a significant clinical improvement based on blinded assessments after a mean of 7.7 years (129). All 10 patients had 5 years of open-label follow-up with significant improvement. A study of GPi deep brain stimulation in cervical dystonia involving 28 patients found that neither advanced age nor severity of symptoms was associated with a poor response (132). However, those patients with lateral shift had less improvement. A multicenter, randomized, sham-controlled study evaluated the efficacy of GPi deep brain stimulation for the treatment of cervical dystonia (128). This trial involved 62 patients, and the primary endpoint was the change in the TWSTRS severity score from baseline to 3 months, assessed by two blinded dystonia experts. At 3 months, there was significant improvement in severity for the deep brain stimulation group compared to sham stimulation (-5.1 points vs. -1.3, p=0.0024). There were also significant improvements in the TWSTRS disability score (41% vs. 11%) and Bain Tremor Score (61% vs. 16%) in the stimulation group compared to the sham group. There were no significant differences between groups in improvement of measures of pain and quality of life. A retrospective study of 30 patients with primary cervical dystonia treated with GPi deep brain stimulation demonstrated that those with a phasic type had greater response than those with a tonic-type of dystonia (28). More specifically, they propose that the range of voluntary neck motility with respect to midline is an objective factor that is useful in predicting the prognosis of patients with cervical dystonia who undergo deep brain stimulation surgery. Patients who could move their heads past midline presurgery showed an average TWSTRS improvement of 73.9% compared to 55.3% for the group that could not move their heads past midline (59). There is also evidence of benefit from GPi deep brain stimulation in patients with cervical dystonia who previously underwent peripheral denervation surgery (20). A small study with 13 patients evaluated cognition after 12 months of continuous stimulation (46). There was minimal impact in the number of produced words. The other cognitive domains were intact, supporting that deep brain stimulation seems a safe treatment modality.

Another small study used functional MRI to evaluate the amplitude of low frequency fluctuations (ALFF) in patients with bilateral GPi deep brain stimulation at a priori clinically optimal settings, nonoptimal settings, and with the device off. ALFF has been shown to reflect the intensity of spontaneous neural activity. All patients showed significantly decreased ALFF in the primary motor cortex and somatosensory cortex comprising leg, trunk, and arm regions with optimal settings compared to the off state. Decreased ALFF in the sensorimotor cortex and increased ALFF in the pons were seen in optimal settings compared to nonoptimal settings. This suggests that optimized GPi deep brain stimulation normalizes aberrant sensorimotor cortex hyperactivity in dystonia (82).

A separate investigation observed patients who underwent bilateral GPi deep brain stimulation (70). It showed improvement in the TWSTRS scores for disease severity, disability, and pain by 72%, 59%, and 46%, respectively. These benefits persisted up to 5 years after the initial surgery, though there has been little other long-term data published to confirm these findings.

Subthalamic nucleus deep brain stimulation has been explored as an alternative target because of observations of parkinsonian signs, such as bradykinesia in patients with cervical dystonia treated with pallidal stimulation (87). In a series of nine patients assessed with blinded rating scales (TWSTRS), bilateral subthalamic stimulation appeared to be an effective alternative (99). Side effects observed were depression and weight gain, but there was no bradykinesia observed.

Experience with the recording of oscillatory activity in neurons during deep brain stimulation surgery has led to the identification of increased beta synchronization as a physiomarker in Parkinson disease, and current efforts have focused on taking advantage of this to develop responsive closed-loop deep brain stimulation systems to treat this disorder. Similarly, an investigation in cervical dystonia may have identified abnormal theta activity in the internal pallidum as a useful physiomarker that could be of use in future development of responsive deep brain stimulation treatment of cervical dystonia (95).

Other surgical procedures. Selective peripheral denervation has been performed in limited numbers of patients with cervical dystonia and appears to be modestly beneficial. This surgical procedure involves sectioning of the nerve twigs supplying the dystonic neck muscles (11). This operation may be considered for patients who are resistant to all other forms of therapy. In a series of 168 patients, 77% had a moderate to excellent improvement in head position, and pain improved in 81% (30). These results were sustained for a follow-up period of 3.4 years in 70% of those who improved. A review of selective peripheral denervation in 53 patients with cervical dystonia evaluated clinical outcomes postoperatively, at 6 months, and at a mean of 42 months (10). All except three patients were secondary nonresponders to botulinum toxin. Follow-up was available for 55 procedures at 6 months and 34 procedures at late follow-up. Statistically significant improvements were seen at both time points in the Tsui scale score and in a Visual Analogue Scale. The major postoperative issues included recurrence of dystonia with a change in muscles involved (51%) and reinnervation of muscles (43%). Clinical symptoms of major concern were seen after 29% of procedures, and 13% of patients had repeat denervations. Thus, although selective peripheral denervation can result in improvement of symptoms and quality of life, the risk of return of symptoms needs to be taken into consideration. A retrospective study evaluating the long-term outcome of selective peripheral denervation in comparison to bilateral GPi deep brain stimulation for cervical dystonia included 35 patients who had at least one year of follow-up (37). Eleven patients who underwent selective peripheral denervation needed additional surgery due to insufficient benefit, whereas only one patient with deep brain stimulation needed a second surgery for repositioning of electrodes. The combined global surgical outcome score for selective peripheral denervation versus deep brain stimulation was “bad” in 65 and 13.3%, “fair to good” in 30 and 26.7%, and marked improvement in five and 60% (p < 0.001). Side effects were reported in 75% of selective peripheral denervation patients and 60% of deep brain stimulation patients and were not significantly different. The most common side effects were sensory disturbance in the selective denervation patients and hardware complications and speech changes in the deep brain stimulation patients. One deep brain stimulation patient had an intraoperative hemorrhage.

Noninvasive neuromodulation. Some noninvasive neuromodulation methods have been described and may offer therapeutic techniques to correct abnormal excitatory circuits in cervical dystonia. These include repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS) (81). It has been shown that the cerebellum and its circuits seem to play a role in the pathogenesis of cervical dystonia. One study used cerebellar continuous theta burst stimulation (cTBS), a novel form of repetitive transcranial magnetic stimulation, in an attempt to activate synaptic plasticity in the cerebellar-thalamo-cortical circuits. Evaluation after 2 weeks of cerebellar stimulation demonstrated improvement for patients with cervical dystonia in the TWSTRS scale. The effect of transcranial direct current stimulation in patients with cervical dystonia is still being studied (72).

Physical therapy. Although physical therapy may be helpful as adjunctive therapy for cervical dystonia, the studies evaluating its efficacy have been limited. A systematic literature review of physical therapy as a treatment for cervical dystonia included 16 studies (45). Seven were clinical trials, and the remaining were case reports or case series. Most studies were of low methodological quality, and the limitations included a small sample size and a lack of randomization and blinding. Reported treatments include EMG biofeedback training, transcutaneous electrical nerve stimulation, muscle elongation, massage, and relaxation training. Although there are reports of improved head position, range of motion, pain, and quality of life, further high-quality studies are needed to determine the efficacy of physical therapy for cervical dystonia. One randomized trial showed no statistically or clinically significant benefit from specialized physiotherapy compared to standard neck physiotherapy (40).

Outcomes

A retrospective, longitudinal descriptive study evaluated the long-term benefit and side effect profile of botulinum toxin for 89 patients with dystonia, including 51 patients with cervical dystonia (105). Long-term therapeutic benefit was maintained with a mean follow-up of 18.5 years. This retrospective analysis was limited to only those patients who had been receiving botulinum toxin injections regularly for at least 10 years at a single movement disorder center. Comparing the benefit after last injection with first injection, there was an improvement in global response effect (3.57 vs. 3.18, p < 0.0001), greater peak effect on dystonia (3.75 vs. 3.45, p < 0.0005), shorter latency to effect (3.3 days vs. 5.5 days, p < 0.0001), longer duration of maximal (15.78 vs. 13.45, p = 0.0084), and total benefit (19.42 vs. 16.33, p = 0.0037). Adverse events were reported in 10% of patients with cervical dystonia; the most common side effects were dysphagia and neck weakness. Overall, this report provides support for the long-term efficacy and safety of botulinum toxin for the treatment of dystonia.

Another prospective, observational study demonstrated a reduction in objective cervical dystonia severity in patients receiving onabotulinumtoxinA injections that was sustained over the duration of the study (36 months). TWSTRS scores continued to improve during this period, from 31.59 at baseline, to 27.52 at 12 months, 25.79 at 24 months, and 24.49 at 36 months. Furthermore, the study demonstrated patient satisfaction scores between 88% to 90% throughout the duration of the study (31). The study shows high rates of both objective improvement and subjective satisfaction. Of note, subjective improvement remained stable despite objective improvement over time, suggesting that there is some degree of dissociation between the two measures.

The development of neutralizing antibodies and loss of clinical response to botulinum toxin is a potential long-term problem. This makes sense given the repeated injection of a large protein over time akin to a vaccine, but the risk is felt to be offset by the minute amount of toxin being injected (nanograms). Assessing immunity through antibody testing is problematic. Antibodies detected by ELISA are not necessarily neutralizing, and the more specific mouse protection assay is not readily available. In addition, multiple factors can lead to secondary treatment failure other than immunoresistance. Looking at this issue, a review of 39 patients identified the following causes of unsatisfactory response: suboptimal doses and muscle targeting, intolerable side effects, complex movement patterns, discordant perceptions, and incorrect diagnoses (67). Only one patient had true immunoresistance

Regarding the true incidence of neutralizing antibody-mediated toxin resistance, historical estimates have been quite low. One early study found four of 326 (1.2%) patients with cervical dystonia had true neutralizing antibodies (17). A meta-analysis looking across different indications at over 2000 patients treated with onabotulinumtoxinA exclusively found a similar low overall frequency of conversion to antibody positivity in 0.49% and in cervical dystonia specifically of 1.28% (94). However, a much higher prevalence than expected was found in a current study looking at 596 patients with cervical dystonia as well as other dystonia subtypes treated at least four times with botulinum toxin type A (03). Overall, 13.9% had neutralizing antibodies by mouse protection assay. Factors that were associated with a higher risk of developing antibodies were higher individual and cumulative dose of toxin; botulinum toxin formulation with incobotulinumtoxinA was associated with the lowest risk. Unfortunately, this study suffers from several flaws (08). For example, the majority of the 596 patients were treated with abobotulinumtoxinA, which has a higher protein content than the other types of botulinum neurotoxin and, hence, may account for the greater antigenicity. The authors fail to mention other studies showing a much lower frequency of neutralizing antibodies (17). They provide no data on the correlation between the presence of antibodies detected by hemidiaphragm assay and clinical response, although they state that their patients were “still responding.” In contrast, none of our patients with positive titers for blocking antibodies tested by mouse protection assay had any clinical response, indicative of true immunoresistance. Mouse protection assay is more clinically meaningful as it correlates well with clinical response, which may explain the differences in the reports of secondary unresponsiveness among various studies using different assays and botulinum neurotoxin products (58). There is clearly a need to develop more sensitive and specific assays for neutralizing antibodies and to provide correlative analyses of the titers and lack of clinical response (09).

In practice, patients who have developed immunoresistance to botulinum toxin A can be switched to botulinum toxin type B with re-establishment of a clinical response. One report found that eight subjects who had developed immunoresistance to abobotulinumtoxinA or onabotulinumtoxinA were able to re-establish clinical responsivity to incobotulinumtoxinA after a one-year hiatus from treatment (47). This response was sustained over 2.5 years without an elevation in antibody titers, and it was proposed that this was possible due to the “low antigenicity” of incobotulinumtoxinA.

Special considerations

Pregnancy

Pregnancy is not known to affect the symptoms of cervical dystonia.

Botulinum toxin preparations were previously labeled Food and Drug Administration category C. The Food and Drug Administration eliminated this letter category labeling system in 2015 in favor of a more comprehensive narrative labeling system to avoid misinterpretation, confusion, and misuse (19). The following is the current label from the Food and Drug Administration:

Anesthesia

Cervical dystonia has been described to occur during general anesthesia (119) and postoperatively (44). In the latter case, thiopentone, fentanyl, enflurane, and nitrous oxide were the agents used.