Clinical and Demographic Characteristics of Upper Limb Dystonia

Abstract

Background:

Knowledge of characteristics in upper limb dystonia remains limited, derived primarily from small, single-site studies.

Objective:

The objective of this study was to characterize demographic and clinical characteristics of upper limb dystonia from the Dystonia Coalition data set, a large, international, multicenter resource.

Methods:

We evaluated clinical and demographic characteristics of 367 participants with upper limb dystonia from onset, comparing across subcategories of focal (with and without dystonia spread) versus nonfocal onset.

Results:

Focal onset occurred in 80%; 67% remained focal without spread. Task specificity was most frequent in this subgroup, most often writer’s cramp and affecting the dominant limb (83%). Focal onset with spread was more frequent in young onset (<21 years). Focal onset occurred equally in women and men; nonfocal onset affected women disproportionately.

Conclusions:

Upper limb dystonia distribution, focality, and task specificity relate to onset age and likelihood of regional spread. Observations may inform clinical counseling and design, execution, and interpretation of future studies.

Adult-onset isolated limb dystonia more commonly affects the upper rather than lower limbs and can be focal or part of segmental, multifocal, hemi-, or generalized dystonia.¹ Most studies of adult-onset isolated upper limb dystonia (ULD) focus on subtypes of task-specific focal dystonia such as writer’s cramp,^2–5 typist’s cramp,⁶ musician’s cramp,^7–11 or sports-related dystonia.¹² In contrast, the demographic and clinical features of ULD are primarily derived from relatively small populations of predominantly young-onset, generalized dystonia.^13,14 Prior studies reported that body region of onset, positive family history, and a younger age at onset may increase the risk of spread of dystonia.^13,15–20 In fact, prospective analysis in a relatively small adult-onset component of the current data set demonstrated a relationship between initial body region affected (including hand) and risk of spread.²¹ Thus, clinical and demographic features are critical for understanding important subtypes of ULD. Such clinical characteristics are essential for design and implementation of future studies that aim to test symptomatic or disease-modifying interventions in isolated limb dystonia.^21,22 The primary aim of this report is to characterize clinical features and demographics of isolated ULD where the upper limb was involved at onset. Data are derived from the Dystonia Coalition database, the largest standardized, multicenter cohort ever assembled.

Methods

Data were obtained from the Dystonia Coalition database (https://www.rarediseasesnetwork.org/cms/dystonia/). A detailed description of the study population and dystonia classification is provided in the Supporting Information. All participants gave written informed consent at the recruiting site according to the Declaration of Helsinki and The Common Rule. Analysis of deidentified aggregate data was approved by the Emory University Human Subjects Review Board and the Human Research Protection Office at Washington University. We only included data from the first visit if a participant had multiple visits.

Eligibility for this study required a diagnosis of isolated dystonia and an age of at least 18 years. A clinician-administered standardized questionnaire and video were collected for each participant according to the standard protocol at the time of enrollment.²³ All participants were evaluated by movement disorders specialists either in person or by a standardized video to note involvement of each affected body region (upper face, lower face, tongue, larynx, neck, shoulder, upper arm, hand, trunk, pelvis, upper leg, foot), the affected side of the body, and Burke-Fahn-Marsden dystonia rating scores. Spread was determined by any change in the participants’ recall of onset regions and the body regions affected at study intake. In addition, video recordings included each participant’s demonstration of a sensory trick.²³

We performed comparisons of clinical and demographic variables across subgroups that included focal versus nonfocal onset, focal onset without spread versus focal onset with spread, and young (<21 years old) versus adult (≥21 years old) onset according to established criteria.²⁴ The following variables were compared across subgroups: gender, mean age of onset, duration of dystonia (excluding focal vs. nonfocal onset subgroups), frequency of task specificity and sensory tricks, and Burke-Fahn-Marsden score. We applied independent 2-tailed Student’s t tests for comparisons of continuous data and Pearson χ² statistics to test the main effect of associations between nominal clinical characteristic variables across each dystonia subgroup with the null hypothesis of no association. All statistics were considered significant with P < 0.05, adjusted using Bonferroni correction to control for family-wise error rate in independent subgroup analyses. We performed all statistics with IBM (Armonk, NY) SPSS Statistics for Windows, version 25.0.

Results

Clinical Characteristics

A total of 367 participants reported ULD at onset (demographics in Supporting Information Table S1). Focal onset occurred in 292/367 (80%), of which 195 (67%) remained focal without spread. Adult onset was reported in 286/367 (78%). Limb onset affected the dominant upper limb in 303/367 (83%), preferentially involving the hand with (85%) or without arm (21%) at onset. Table 1 displays clinical characteristics for all participants across subgroups. The dominant limb was affected more frequently (χ² = 4.5, P = 0.034) in those with (212/246, 85%) versus without (92/121, 76%) task-specific dystonia. Gene mutations thought to be associated with the dystonia were reported in 13/367 participants, including 12 with onset before age 21 years (Supporting Information Table S2).

TABLE 1.

Clinical characteristics of focal limb dystonia

	All	Focal Onset	Versus	Nonfocal Onset	χ2 or t	P
Number of participants	367	292 (80%)		75 (20%)
Female	210 (57%)	157 (54%)a		53 (71%)a	7.0	0.008
Mean age of onset, y	37.5 ± 18.1	35.8 ± 17.5a		44.4 ± 18.4a	−3.6	0.001
Task specificity	246 (67%)	205 (70%)		41 (55%)	6.5	0.011
Sensory trick	153 (42%)	107 (37%)a		46 (61%)a	15.0	<0.001
Onset, dominant limb	303 (83%)	239 (82%)		64 (85%)	0.5	0.478
BFM total	10.2 ± 11.2	9.6 ± 11.6		12.2 ± 9	−2.3	0.023
		No Spread	Versus	Spread	χ2 or t	P
	Number of participants	195		97
	Female	105 (54%)b		52 (54%)	0.001	0.969
	Mean age of onset, y	38.9 ± 15.4b		29.5 ± 19.6b	4.1	<0.001
	Dystonia duration, y	14.5 ± 12.9b		24.1 ± 16b	−5.1	<0.001
	Sensory trick	57 (29%)b		50 (52%)b	13.9	<0.001
	Onset dominant limb	164 (84%)		75 (77%)b	2.0	0.157
	BFM total	4.9 ± 4.2b		18.6 ± 15.6b	−8.4	<0.001

^aSignificant differences across focal versus nonfocal upper limb onset following focal upper limb onset.

^bSignificant differences across no spread versus spread following focal upper limb onset.

All participants reporting upper limb involvement at onset, then divided into (top) focal onset versus nonfocal (limb plus another body region) onset and (bottom) without and with spread to other body regions. Pearson χ² or t statistic are reported for nominal or continuous group comparisons, respectively. Significant results appear in bold font. Data are n (%) or mean ± standard deviation.

Abbreviation: BFM, Burke-Fahn-Marsden scale.

Young Onset Versus Adult Onset

Table 2 compares participants reporting young-onset (age of onset <21) and adult-onset (≥21 years) ULD. The proportion of those with focal onset was similar across young onset versus adult onset, but the young-onset group had more frequent regional spread (Supporting Information Figure S1).

TABLE 2.

Clinical characteristics based on onset age

	Young Onset	Versus	Adult Onset	χ2 or t	P
Number of participants	81		286
Female	44 (54%)		166 (58%)	0.3	0.566
Mean age of onset, y	11.6 ± 5.3a		44.9 ± 12.9a	−34.4	<0.001
Dystonia duration, y	31.6 ± 17.2a		13.3 ± 10.6*	9.1	<0.001
Task specificity	37 (45%)a		208 (73%)a	21.2	<0.001
Sensory trick	34 (42%)		119 (42%)	0.001	0.972
BFM total	15.8 ± 14.8a		8.6 ± 9.3a	4.1	<0.001
Focal onset	71 (86%)		221 (76%)	4.2	0.040
Spread after onset	41 (49%)a		56 (19%)a	25.2	<0.001
Onset, dominant limb	64 (79%)		238 (84%)	0.9	0.347
Focal	27 (33%)		160 (56%)
Segmental	22 (27%)		92 (32%)
Multifocal	12 (15%)		18 (6%)
Generalized	19 (23%)		14 (5%)

^asignificant differences across onset age subgroups.

Young (age <21 years) onset and adult (age ≥21). Pearson χ² or t statistic are reported for nominal or continuous group comparisons, respectively. Percentage focal onset and spread after onset reflect a percentage of upper limb onset, whereas all other percentages reflect the percentage of participants in young versus adult onset. Significant results appear in bold font. Data are n (%) or mean ± standard deviation.

BFM, Burke-Fahn-Marsden scale.

Task Specificity

Participants with task specificity (n = 246) reported writer’s cramp (n = 194), musician’s cramp (n = 30), typist cramp (n = 25), and sport-related cramp (n = 4). Participants may have had more than 1 type (n = 7), including writer’s cramp and typist cramp (n = 3), musician’s cramp and writer’s cramp (n = 1), and musician’s cramp and typist cramp (n = 3). Musician’s cramp was more commonly associated with focal onset without spread (27/30 instances) than focal onset with spread (3/30). Speech-related (laryngeal) dystonia was present in those with nonfocal ULD onset (10/13 instances) more frequently than focal onset with spread (3/13 instances).

Sensory Tricks

Of participants reporting a sensory trick, 57/57 (100%) with focal onset and no spread, and 32/50 (64%) with focal onset and spread reported sensory tricks isolated to the limb. The remaining 18 participants with spread reported sensory tricks relieved dystonia in other body parts, but not limb dystonia.

Discussion

Demographic and clinical characteristics from the largest multicenter, international cohort of patients with standardized evaluations of ULD suggest the existence of distinct subgroups based on focal versus nonfocal onset and focal onset with and without spread. Specifically, age of onset, sex, and frequency of task specificity and sensory tricks differed across subgroups. Consideration of such features may be necessary for accurate patient prognosis and in the design and interpretation of treatments with symptomatic or disease-modifying properties.

ULD onset in our entire cohort occurs across the age spectrum, with peak frequency in the fifth decade. Subgroup analysis indicates increased prevalence of focal onset and spread among young onset, consistent with prior observations following lower limb onset.^13,14,25 In fact, 12/13 participants in our cohort with an identified gene recall onset in youth, 8 with TOR1A-associated dystonia. Although this supports the fact that TOR1Aassociated and other genetic dystonia commonly contribute to ULD onset in youth, we advise caution given likely clinical selection bias in testing this population. A prospective assessment of 47 adults with hand dystonia (included in the current study) indicated a mean span of 3.3 years from onset until spread to another body region.²¹ This relatively short time to spread and a plateau of the Kaplan-Meier curve in conjunction with our current data support the notion that ULD spread is not solely related to dystonia duration but also age of onset. One must consider inherent study bias when comparing duration of dystonia in our study to others^15,19,20 because we only recruited those at least 18 years of age and asked for recall regarding onset region(s).

Our data indicate that a higher proportion of women to men are affected by ULD, but this ratio is nearly 1:1 in the focal-onset group, independent of spread. A recent meta-review suggests a predominance of women with isolated focal dystonia, except for ULD.²⁶ Others have previously observed gender differences across various forms of task-specific limb dystonia. For instance, a higher prevalence of men may be affected by musician’s cramp.^7,27 This contrasts with a report that 62% of 105 consecutive patients with writer’s cramp were women.⁴ Inconsistences in the gender ratios across various forms of task-specific dystonia may represent true differences between men and women, but may also reflect population differences in the number of men versus women who are musicians, for instance. Indeed, 77% of participants with musician’s dystonia were men in our study, whereas 38% and 48% were men among those with writer’s cramp and typist cramp, respectively. Thus, future studies should account for onset site and spread as well as task specificity when attempting to assess the potential contribution of sex differences in dystonia.

Task specificity is most frequent in focal-onset ULD without spread and least frequent with nonfocal onset. When we analyzed 324 participants in an independent Dystonia Coalition cohort without limb involvement at onset that later developed ULD, only 21% reported task specificity. Thus, task-specific ULD appears to favor those with focal onset without regional spread. Across all subgroups analyzed, the frequency of writer’s cramp outnumbered all additional forms of task specificity. This probably reflects the high prevalence of writer’s cramp in the general population, although there are no well-designed epidemiological studies estimating its prevalence. The observation that ULD affects the dominant limb in 85% supports this concept and may support prior reports that relatively intensive use of a body region contributes to task specificity.^4,28,29 Our findings indicate that studies of task specificity should focus on focal ULD without regional spread to better differentiate the pathophysiological basis of altered sensory input,^30,31 maladaptive plasticity,³² and disruption of large-scale neural connectivity.³³

There is relative paucity of upper arm involvement that influences classification of body distribution (eg, segmental vs. multifocal). Whether scarceness of upper arm dystonia results from discrepancies in the sensitivity of examination or anatomical or pathophysiological mechanisms produce such differences remains uncertain. Muscles controlling the upper arm extend across shoulder girdles, thus complicating clear demarcation of “segmental” dystonia and application of currently available scales for dystonia.³⁴ Future efforts should focus on improving diagnostic criteria for detecting upper segment limb dystonia.

Study limitations include retrospective recall of the body region affected at dystonia onset. In addition, those with mild disease that are either underrecognized or not seeking treatment are less likely to attend one of our recruiting sites. The Dystonia Coalition does not require neuroimaging of the neural axis to exclude potentially overlooked secondary structural lesions that may contribute to a small percentage of those categorized as isolated dystonia for study entry.^35,36 Furthermore, all examinations were performed at least 2 months after botulinum toxin injection, but lingering effects could still alter dystonia manifestations.³⁷ Nevertheless, these data on ULD focality, task specificity, and likelihood of regional spread may inform clinical counseling and future efforts in the design, execution, and interpretation of clinical trials.