Clinical features, pathophysiology, treatment, and controversies of tremor in dystonia

Abstract

Dystonia and tremor frequently co-occur. In some cases, they have shared biological mechanisms, while in others dystonia and tremor are two comorbid conditions. The term “dystonic tremor” is used to describe tremor in those who have dystonia. Two mutually exclusive definitions of “dystonic tremor” were proposed. According to one definition, dystonic tremor is the tremor in the dystonic body part. An alternate definition of dystonic tremor entails irregular and jerky oscillations that have saw tooth appearance with or without overt dystonia. This paper outlines the differences in two definitions of dystonic tremor and identifies their limitations. Given the diverse views defining “dystonic tremor”, this paper will use the term “tremor in dystonia”. In addition, we will outline different ways to separate the subtypes of tremor in dystonia. Then we will discuss pathophysiological mechanisms derived from the objective measures and single neuron physiology analyses of tremor in dystonia.

1. Introduction

Dystonia is defined as a disorder characterized by excessive muscle contractions and involuntary postures, often with repetitive movements or jerky oscillations [1]. Tremor is defined as involuntary, rhythmic oscillations of the body parts [2,3]. Tremor has some resemblance to dystonia; for example, dystonia and tremor are chronic and involuntary; dystonia can be rhythmic like tremor. Many patients diagnosed with dystonia also have tremor, or vice versa. The prevalence of clinically apparent tremor in those with dystonia ranges from 10 to 90% [4,5]. On the contrary, those diagnosed with tremor can also have subtle dystonia; the co-prevalence is up to 21%. Latter condition is called essential tremor plus by 2018 consensus definition of tremor [6]. The extensive variations in co-prevalence are the result of the varied definitions of tremor that is seen along with dystonia, i.e., dystonic tremor. The movement disorders society tremor consensus group had defined dystonic tremor as a tremor of the dystonic body part [6,7]. The same group also defined a different entity, the tremor associated with dystonia (TAWD), where tremor affects a body part that is not dystonic [7]. The shortcoming of the consensus definition is that dystonia can be subtle and it can be missed particularly if there is robust tremor. Particularly the challenge is that if rhythmic and regular oscillations are robust and they are combined with very subtle dystonia then such movement be called TAWD or dystonic tremor. In order to address such controversy the recent consensus statement suggested the use of term essential tremor plus [8]. The classic definition of dystonic tremor relies on the phenomenology of the movement. Accordingly, the dystonic tremor has jerky, coarse, and irregular oscillations with “saw-tooth” appearance [9] (Video 1). Small and regular oscillations, appearing like ET are also described with dystonia. Such oscillations can be present in conjunction with irregular and jerky oscillations [10] (Video 2). The limitation of the classic definition is that categorizing the oscillations into jerky or irregular can be subjective leading to interrater differences. This review will focus on the clinical perspectives on dystonic tremor, how to diagnose them, their differential diagnoses, treatment and pathophysiology. Since many studies use different criteria to define dystonic tremor (classic versus consensus definitions) the review will inclusively discuss both, and for the purpose of this paper we will inclusively use term “tremor in dystonia”.

2. Clinical perspectives on tremor in dystonia: identification, controversy, and differential diagnosis

Clinically apparent tremor has been reported in 14–87% of the patients diagnosed with dystonia [11], with an overall average of 53% [4]. The prevalence of tremor increases with older age and longer duration of dystonia or those who have segmental and multifocal dystonia. Tremor in dystonia is mostly seen during postural holding and reaching tasks; but resting tremor can be seen elderly patients with dystonia [12,13]. The body regions affected by tremor and the temporal relationship between tremor and dystonia onset are variable, with an exception of head tremor that is very common in cervical dystonia [4]. Truong and Hermanowicz pointed out that careful observation may reveal that movement to one side is slower than to the other. Removal of visual cues, such as asking the patient to close eyes or darkening the room, may cause the neck to turn to one side. The side that the neck turns to is the direction of the slow movement. The rapid movement is the correcting effort. Moving the head to each side will help to identify the side involved: toward the ipsilateral side, the head shaking will disappear; toward the opposite side, the head shaking will worsen. They also suggested that treatment with botulinum toxin should therefore be aimed at the muscles precipitating the slow movements [5].

The peak frequency of tremor in dystonia varies from 4 to 10 Hz; the range similar to ET, ET-plus, parkinsonian tremor, and functional tremor [14]. Nevertheless, irregular, jerky, asymmetric, and directional oscillations with or without concomitant twisting or turning of the body part suggests dystonia rather than ET or ET-plus. In tremulous patients it is clinically challenging to identify dystonia; especially if the latter is mild [4]. Looking for overflow, mirror dystonia on the contralateral side to dystonia, sensory trick, and abnormal posturing can be helpful in identifying dystonia [1], but the sensitivity and specificity of these signs vary with distribution of dystonia.

There is a substantial inter-rater disagreement in the diagnosis of tremor in dystonia [15]. Small, regular tremors of the body region affected by dystonia or TAWD are indistinguishable from ET suggesting a possible overlap between dystonia and ET. Tremor in dystonia can occur at rest, mimicking parkinsonian tremor [13]. However, irregular tremor with thumb extension is more suggestive of dystonic subtype [16] (Video 3). Predominant rest tremor with mild postural and kinetic tremor, suppression of tremor with voluntary movements, and re-emergence also support the diagnosis of parkinsonian tremor [17]. Additional dystonic postures and parkinsonian signs are essential clues for differentiation between these two syndromes. Tremor in dystonia can be misdiagnosed as functional or psychogenic tremor because directional, position-specific and irregularity of tremor in dystonia can be clinically misinterpreted as inconsistency of functional tremor. Distractibility and entrainment should be carefully observed for distinguishing the tremor in dystonia from the functional tremor [18].

3. Objective tremor assessment

Wearable sensor technology and surface electromyography (EMG) can be helpful to characterize tremor in dystonia and differentiate it with other syndromes [19]. Archimedes spirals with multidirectional axis were commonly observed for tremor in dystonia than ET but had poor sensitivity and specificity [20]. Signs of dystonia including co-contraction between the agonist-antagonist pairs and overflow as well as irregular tremor and influence of sensory tricks on tremor, were the proposed criteria for tremor in dystonia [21]. The surface EMG tracing of the agonist and the antagonist muscles is useful to visualize the irregularity of tremor and co-contraction (Fig. 1). Mirror dystonia with tremor is present in the contralateral limb and can be diagnosed with EMG recording. The effects of sensory trick on reduction of the tremor amplitude in dystonic patients is a critical differentiating factor between dystonia and ET [22]. The irregular quality of tremor in dystonia can be quantified by the large half-power bandwidth of peak spectral analysis of the accelerometer or EMG and the high tremor stability index. This index is a parameter that can be calculated from the signal transducers to measure the range of changes in the frequency over time [23]. A greater tremor stability index means increased variation of tremor frequency. A higher index has been demonstrated for tremor in dystonia compared to ET [8]. These objective features of tremor in dystonia still can not rule out co-existing ET or ET-plus, as it is possible to have both at the same time.

Fig. 1.

Surface EMG recordings in a patient with craniocervical dystonia with head tremor. The tracing demonstrated irregular and jerky tremor with co-contraction of the antagonist pairs of neck muscles bilaterally. SC: splenius capitis, SCM: sternocleidomastoid.

Certain electrophysiological tests that underpin the physiology of dystonia can assist for distinguishing tremor in dystonia from that of ET. These tests would be appropriate for tremulous patients with subtle dystonia or TAWD subtypes that look like ET. The temporal discrimination threshold is the shortest interval to discriminate between two tactile stimuli. The tremor discrimination threshold was increased in patients with TAWD while it was normal in ET [24]. Loss of the blink reflex recovery curve was found in those with tremor with dystonia, but not in ET [25]. Loss of presynaptic reciprocal inhibition was reported in CD with hand tremor, i.e., TAWD, but not ET. Moreover, increased co-contraction of the antagonist pairs during ballistic movements of hands was also demonstrated in this subgroup [26].

Objective analyses of oscillation trajectories define tremor characteristics providing additional phenomenological insight. We used magnetic search-coil, providing information on three dimensional angular head orientation to study head oscillations in 14 randomly selected CD participants [10,27]. The study participants aimed their head toward a target that was projected in front of them, three meters away. We measured spontaneous changes in angular head position as the participants aimed at the straight-ahead target. In order to address the effects of head-on-trunk position on head oscillations, the angular positions were measured when the participants aimed their gaze (orientation of eye and head) to the target projected 10°, 20°, and 30° to the right or to the left. A laser projector mounted on the head provided visual feedback to ensure accuracy of head position. It eliminated the possibility of compensation by the eye movement for the imprecise head position. Data were digitized and processed to compute angular head positions in three dimensions [10,27]. Subsequently, we separated sinusoidal oscillations and jerky head movements in three distinct groups (Fig. 2). One group had prominent dystonia with minimal head tremor; another had prominent head tremor with minimal dystonia. In contrast, the third group had an equal combination of both types of oscillations. These results showed that the head oscillations were evident in the majority of subjects, even when they were not clinically apparent. These observations confirm prior suspicions that clinical methods for detecting head oscillations lack sensitivity.

Fig. 2.

Two distinct types of head movements. Non-sinusoidal (green) and sinusoidal (red) traces are decomposed from parent (blue) waveform. Examples of three subjects are shown, one subject had dystonic movements of the head, another had essential tremor of the head, and the third was mixed. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Quantitative analyses of head oscillations disclosed two subtypes with distinct measures of frequency, amplitude, and sinusoidality (Fig. 3). The sinusoidal type had 4–6 Hz frequency; these oscillations resembled ET. A lower frequency (<2 Hz) subtype had alternating slow and fast movements giving it a “saw tooth” appearance resembling Fahn’s (1984) definition of “dystonic tremor”. These observations reveal that clinical methods for detecting head oscillations lack specificity, because the head tremor (sinusoidal) and dystonic (jerky) subtypes could not be discriminated by clinical examination. We then measured the energy of both types of head movements; the cervical dystonia subject had 96% of total energy distributed in jerky waveform (Fig. 2). The subject who had head oscillations resembling ET had 79% energy distributed to sinusoids (Fig. 2). The subject with mixed tremor had 85% energy distributed to jerky movements and 15% energy in sinusoids (Fig. 2). We also found that the distinct head oscillations subtypes were often combined in the same participant. One-third of the participants had only the jerky head oscillations consistent with CD, while two-thirds had both subtypes of oscillations. These combinations could not be detected by clinical examination alone, and they provide strong evidence against widely held clinical views that these head oscillations should be considered mutually exclusive diagnoses.

Fig. 3.

Two types of head oscillations in cervical dystonia – sinusoidal (like ET) and dystonic (jerky waveforms). In these studies of 14 subjects, there was some overlap in amplitude, but no overlap in frequency or sinusoidality.

Adapted with permission from [27].

4. Pathophysiology of tremor in dystonia

The recent studies in electrophysiology, functional imaging, and the long-term efficacy of deep brain stimulation at different targets have increased our understanding of the pathophysiology of tremor in dystonia and point to a central mechanism. The involvement of both basal ganglia and cerebello-thalamocortical circuits has been proposed as the underlying pathophysiology of tremor in dystonia [28–30]. However, it is still unclear whether the basal ganglia or the cerebellar networks, or a combination of both systems, drive the tremor pathophysiology. Plenty of evidence has supported that jerky oscillations in dystonia shares a similar neuroanatomy and physiology as dystonia without tremor including loss of inhibition at spinal, brainstem, and cortical level [24–26,31,32], suggesting that jerky oscillations in dystonia and “pure” dystonia (without any oscillations) might have overlapping mechanisms.

4.1. Evidence of cerebello-thalamo-cortical involvement

CD patients with tremor of the head showed higher clinical scores of cerebellar dysfunctions than those without tremor [33]. A structural imaging study revealed cerebellar atrophy with or without cerebellar signs or lesions in 14% (26 out of 188) of patients with cervical and segmental dystonia with concomitant tremor [34]. Moreover, a functional MRI (fMRI) study in laryngeal dystonia with voice tremor demonstrated functional abnormalities in the cerebellum and its connection to the medial frontal gyrus, indicating the involvement of the cerebello-thalamocortical network [35]. A recent functional imaging study in dystonia showed that cerebellar connectivity during grip-force induced tremor in those with baseline tremor was not different from ET, supporting the idea of similar cerebellar mechanisms of tremor genesis in both [28,36]. Finally, a longitudinal follow-up study of tremor in dystonia demonstrated the long-lasting beneficial effect of deep brain stimulation at the ventrointermediate nucleus of thalamus, the nucleus that receives afferents from the cerebellum, supporting the involvement of the cerebello-thalamo-cortical loops [37].

4.2. Evidence of basal ganglia-thalamo-cortical involvement

Data from a resting-state fMRI study focusing on the sensorimotor network showed widespread involvement of the cortex and the basal ganglia in laryngeal dystonia with and without tremor. This indicated that jerky oscillations in dystonia and pure dystonia share similar neuroanatomy including the basal ganglia-thalamo-cortical networks. Interestingly, loss of inhibition in the right premotor cortex, left parietal cortex and the left putamen was greater in tremor in dystonia compared to those without tremor [38]. This finding supports that the basal ganglia circuits are also significantly related to jerky oscillations in dystonia.

4.3. Evidence of both basal ganglia-thalamo-cortical and cerebello-thalamo-cortical networks

Recent evidence suggested that both basal ganglia and cerebello-thalamo-cortical networks likely contribute to pathophysiology of jerky oscillations in dystonia. A study of head oscillation in CD proposed that jerky head movements in CD patients could be analogous to nystagmus, and it could be due to an impairment in a head neural integrator secondary to abnormal cerebellar and basal ganglia connections [10,39]. Functional imaging studies showed that there was greater reduction of functional connectivity in the cortical-basal ganglia-cerebellar pathway in those who have tremor in dystonia compared to ET. These studies support the presence of the widespread networks involving both the cerebellar and the basal ganglia [28,36]. A recent study of the optimal deep brain stimulation targets using a combination of volumes of tissue activated estimation and functional and structural connectivity in tremor in dystonia and ET revealed that the best target for the former was the ventralis oralis posterior nucleus of thalamus that receives the inputs from the basal ganglia loop while the appropriate target for ET was the ventrointermediate nucleus of thalamus. Tremor improvement was significantly correlated with structural connectivity in the dentatothalamic tract and pallidothalamic tracts that project to the thalamus [29]. Taken together, a combination of both cerebello-thalamo-cortical and basal ganglia-thalamo-cortical networks are likely involved in driving pathophysiology of tremor in dystonia (Fig. 4).

Fig. 4.

The proposed mechanisms of tremor genesis in dystonia. Both basal ganglia-thalamo-cortical and cerebello-thalamo-cortical circuits are likely involved in pathophysiology of dystonic tremor syndrome. A combination of the basal ganglia-thalamo-cortical and cerebello-thalamo-cortical pathways may contribute to tremor in dystonic tremor (DT) subtype. In contrast, tremor associated with dystonia (TAWD) primarily involves the cerebellar networks similar to ET. The basal ganglia connect to premotor cortex and supplementary motor area via ventralis oralis posterior (VOp) nucleus of thalamus. The cerebellum connects to the motor cortex via the ventrointermediate nucleus of thalamus (VIM).

4.4. The pathophysiology of subtypes of tremor in dystonia

Two studies demonstrated that the pathophysiology of two subtypes of oscillations in dystonia might be different. Pallidal single neuron recording during deep brain stimulation in CD patients showed that physiology of pallidal neurons in those who had dystonia with jerky oscillations was similar to those with pure dystonia without any oscillations; it was not the case in those with dystonia and sinusoidal tremor [40]. The single-unit physiology revealed the fundamental differences between neck tremor that is sinusoidal (i.e., “dystonic tremor” according to the MDS 1998 and 2018 definitions) and neck oscillations that were jerky (i.e., “dystonic tremor” according to Fahn 1984 definition) [40]. The differences were then compared to those with no neck oscillations but pure dystonia. The study measured the single-unit activity of 727 pallidal single neurons during deep brain stimulation surgery in patients with CD. Cluster analyses of spike-train recordings allowed classification of the pallidal activity into burst, pause, and tonic. Burst neurons were more common, and the number of spikes within spike and inter-burst intervals was shorter in pure dystonia and jerky oscillation groups (i. e., dystonic tremor according to Fahn 1984 definition) compared to the sinusoidal tremor group (i.e., dystonic tremor according to MDS 1998 and 2018 definitions) [40]. Pause neurons were more common and irregular in the pure tremor group compared to pure dystonia and jerky oscillation groups (Fig. 5). There was bi-hemispheric asymmetry in spontaneous firing discharge in pure dystonia and jerky oscillation groups, but not in the sinusoidal oscillation group. These results demonstrated that the physiology of pallidal neurons in patients with pure CD is similar to those who have CD with jerky oscillations. The pallidal physiology was different from those who have CD combined with sinusoidal oscillations. These results further implied distinct mechanistic underpinnings for different types of head movements in CD [40].

Fig. 5.

(A-C) Examples of three cervical dystonia patients, one with sinusoidal head oscillations (CD-s) (A), jerky head oscillations (CD-J) (B) and no head oscillations (pure CD) (C). Angular head orientation normalized to maximal excursion are plotted on y-axis while corresponding timestamp is plotted on x-axis. Distribution of pause, burst and tonic neurons in globus pallidum interna (GPi) (D) and externa (GPe) (E). Number of cells in each category is plotted on y-axis, the x-axis depicts the patient category, color of bar plot depicts cell type.

Adapted with permission from [40].

The functional influence of the cerebello-thalamo-cortical pathway by assessing the cerebellar inhibition was studied in tremor in dystonia, TAWD and ET. The tremor with dystonia exhibited decreased cerebello-thalamo-cortical inhibition compared to TAWD and ET. It was further hypothesized that the cerebello-thalamo-cortical integrity would be weaker in tremor with dystonia allowing the influence from the basal ganglia on these networks while the cerebello-thalamo-cortical functional connection is stronger in TAWD similar to ET [8]. Thus, tremor in dystonic and TAWD could be distinct entities with different pathophysiology. Tremor in dystonia probably involves both the basal ganglia and the cerebellar circuits, while TAWD prominently involves the cerebellar networks resembling ET (Fig. 4).

5. Treatment

Focal and segmental dystonia with tremor of the same organ have favorable responses to botulinum toxin injections. Indeed it is known that the tremor in generalized dystonia may respond to anti-dystonic medications including anticholinergics and clonazepam [41]. In contrast, the hand tremor with craniocervical dystonia or laryngeal dystonia (TAWD subtype) may have better benefit with anti-tremor medications. Focused ultrasound is a noninvasive technique to improve quality of life in patients with dystonia. The efficacy of this modality had been tested in few early studies. Magnetic resonance guided focused ultrasound thalamotomy had shown improvement in dystonia in 10 patients with focal hand dystonia [42]. The focus ultrasound of pallidothalamic tract (Forel’s field H1-otomy) was effective in treatment of dystonia [43]. On the contrary, focused ultrasounded guided treatment of tremor is known for a side effect of making coexisting dystonia worse, or lead to new onset of dystonia [44]. For medically refractory cases, deep brain stimulation at the ventrointermediate nucleus of thalamus or the border of ventrointermediate nucleus/ventralis oralis posterior nucleus of thalamus should be considered [37]. In addition, subsequent globus pallidus interna deep brain stimulation might be suggested if the tremor improvement is not sustained after thalamic deep brain stimulation [45]. Such disparities tell us that some tremor in dystonia is treatable with anti-dystonia medications, while others with anti-tremor medications. Likewise, some tremor are treated with thalamic cerebellar receiving region deep brain stimulation; while others at the pallidal or thalamic pallidal receiving region deep brain stimulation. This means that the tremor in dystonia is not just phenomenologically different, but they may have different etiology and pathophysiology.

6. Conclusions

The clinical spectrum of tremor in dystonia are heterogeneous but can be categorized into subtypes. Given difficulties in the clinical diagnosis of such tremor subtypes in dystonia, electrophysiological tests for dystonia are helpful. Recent evidence show that both basal ganglia and the cerebello-thalamo-cortical networks might contribute to tremor genesis in dystonia. The beneficial effects of treatment are fair and may depend on the subtypes of tremor in those who have dystonia.