Extrapyramidal Motor Signs in Degenerative Ataxias

Background  Extrapyramidal motor signs (EPS) are well-known symptoms of degenerative ataxia. However, little is known about frequency and appearance of EPS in subtypes of ataxia.

Methods  We characterized 311 patients with ataxia clinically and genetically. Course of the disease and EPS were investigated according to a standardized protocol. Diagnostic and prognostic impact of EPS in subtypes of ataxia was analyzed by Kaplan-Meier plots.

Results  Extrapyramidal motor signs occurred in all forms of ataxia, but frequency and type of EPS varied between genetically and clinically defined subtypes. Postural tremor in hereditary ataxias was typical for spinocerebellar ataxia type 2 (SCA2). Dystonia was generally rare in ataxias, but, if present, suggested SCA3. We observed a parkinsonian variant of SCA3 in which parkinsonism was present in the beginning of the disease and responded well to levodopa therapy, leading to diagnostic confusion. Parkinsonism in SCA3 was independent of CAG repeat length but ran in families, suggesting modifying genes. In idiopathic sporadic cerebellar ataxia (ISCA), EPS are more frequent in late-onset than in early-onset forms. In 50% of ISCA patients with parkinsonism, the diagnosis of multiple system atrophy remained questionable because of normal autonomic function.

Conclusions  Extrapyramidal motor signs can help to predict the genetic subtype of ataxia. Extrapyramidal motor signs were more frequent in genetic subtypes in which basal ganglia affection has been demonstrated by postmortem studies. However, no type of EPS was specific for an underlying mutation. In ISCA, EPS are an adverse prognostic factor. Parkinsonism is especially associated with a more rapid course of the disease.

CEREBELLAR ATAXIAS are a genetically, clinically, and pathologically heterogeneous group of neurodegenerative disorders in which extrapyramidal systems are involved in a highly variable manner. Extrapyramidal motor signs (EPS) vary even within a genetically well-defined subtype, such as Machado-Joseph disease.^1,2 On the other hand, EPS are constitutive for the clinical diagnosis of disease entities such as multiple system atrophy (MSA).³ Little is known about the diagnostic value of EPS in particular forms of ataxia, however, and data about the frequency of EPS in genetically defined subtypes of ataxia are scant.

Within the past few years, an increasing number of ataxia-causing mutations have been identified, leading to a new classification of hereditary ataxias in accordance with the molecular basis of these disorders.⁴ Molecular analyses of the causative mutations allow the confirmation or exclusion of several genetic subtypes of ataxia, namely, for the autosomal dominantly inherited spinocerebellar ataxias types 1, 2, 3, 6, and 7 (SCA1, SCA2, SCA3, SCA6, SCA7), and for Friedreich ataxia,^4,5 each for which direct genetic testing is available.

The phenotypes of some clinically or pathologically defined disorders such as SCA3 and Friedreich ataxia had to be revised according to genotyping. Genetic analyses revealed a broader clinical spectrum of these disorders than was believed before molecular genetic diagnosis became available.^2,6,7

To clarify the role of EPS in ataxia, we examined a cohort of 311 patients with ataxia for EPS. The prevalence of EPS was analyzed in 212 patients with well-defined mutations causing various types of hereditary ataxia and 99 patients with idiopathic ataxia were included to determine diagnostic and prognostic value of EPS in cerebellar ataxias.

A total of 311 patients with degenerative forms of ataxia were included in this study. Patients with ataxia secondary to alcohol abuse; inflammatory, neoplastic, paraneoplastic, vascular, or well-defined metabolic diseases such as vitamin B₁₂, vitamin E, or lysosomal enzyme deficiency, increased levels of phytanic acid or very long chain fatty acid, or thyroid dysfunction were excluded from this study. All but 7 patients were of German ancestry. Three patients were from Turkey, 2 were from Italy, 1 was from Portugal, and 1 was from Spain.

History was taken with special attention to family history, development of gait difficulties, and drug response. All patients were clinically examined with detailed tests for cerebellar symptoms and EPS. Since assessment of EPS is difficult in the presence of another movement disorder (ataxia), EPS were regarded as present only in cases of definite expression. Doubtful cases were regarded as missing data. However, in some instances, slowness of alternating movements, short-stepped gait, or impairment of postural reflexes in particular, may be caused by cerebellar rather than basal ganglia pathology. Despite this important pathogenetic consideration, this does not interfere with analyses of the prognostic value of clinical signs.

To study disease progression, we ascertained the year of onset of gait disturbances, the year when independent gait without walking aid or support became impossible, and the year when patients became permanently dependent on a wheelchair as suggested by Klockgether and coworkers.⁸ This was a prospective study in which all patients were seen by the same clinician (L.S.).

After informed consent was obtained, DNA was extracted from samples of peripheral blood leukocytes. All patients were genetically tested for the trinucleotide repeat expansions causing SCA1, SCA2, SCA3, SCA6, SCA7, and Friedreich ataxia using standard methods.^9-11

Based on genetic analyses and the patients' family history, we distinguished between patients with hereditary ataxias (212 patients from 138 families), including 12 patients (9 families) with SCA1, 11 patients (9 families) with SCA2, 80 patients with SCA3 (38 families), 32 patients (24 families) with SCA6, 35 patients (19 families) with undefined forms of autosomal dominant cerebellar ataxia, and 42 patients (39 families) with Friedreich ataxia. None of the patients with autosomal dominant cerebellar ataxia had a CAG expansion in the SCA7 gene. Ninety-nine patients were classified as having idiopathic ataxia based on negative family history and negative genetic testing.

In cases of idiopathic ataxia, we differentiated patients with probable MSA according to the diagnostic criteria of a recent consensus conference³ from patients with idiopathic sporadic cerebellar ataxia (ISCA) who did not fulfill the criteria of MSA. On the basis of age at onset, ISCA was subdivided into early-onset forms (age at onset <25 years) and late-onset forms (age at onset ≥25 years of age).

The comparisons concerning age at onset and duration of disease (time since onset) of patients with and without EPS and patients with and without signs of dysautonomia were performed by means of unpaired t tests. The influence of parkinsonism on age at onset and duration of disease was assessed by analyses of variance. Survival curves according to the method of Kaplan-Meier were used to investigate the impact of EPS, dysautonomia, time of onset (early or late), and parkinsonism on time until certain events, ie, loss of independent walking (walk with support) and time until progression to wheelchair dependence. The median time until these events, the 5- and 10-year event-free proportions, and the corresponding 95% confidence intervals were calculated. Log-rank tests were used to compare the event-free times between patients with and without EPS, dysautonomia, and parkinsonism, and between patients with early and late onset disease. P<.05 was considered statistically significant. However, because of the explorative nature of the analyses, the P values should be interpreted in a descriptive manner.

Extrapyramidal motor signs occur in all subtypes of ataxia, albeit with different frequencies (Table 1). Parkinsonian features such as facial impassivity, monotonous soft voice, rigidity, short-stepped gait, reduced arm swing, flexed posture, retropulsion, impaired postural reflexes, slowed alternating movements, or bradykinesia, are most frequent in MSA but also occur in late-onset cerebellar ataxia, SCA3, and genetically still undefined subtypes of autosomal dominant cerebellar ataxia. Resting tremor is rare even in patients with ataxia expressing other parkinsonian signs. Postural or action tremor is most frequently seen in SCA2 (50%), MSA (20%), early-onset cerebellar ataxia (25%), and late-onset cerebellar ataxia (27%), but is unusual in SCA1 and SCA6 (Table 1).

Dystonia is rare in all types of ataxia, but if present, dystonia strongly suggests the diagnosis of SCA3. Dystonia in SCA3 occurred mostly in early-onset patients, when the disease started before the age of 20 years. Myoclonus is infrequent in ataxias as a whole and is not restricted to a clinical or genetic subtype. Choreiform hyperkinesia is extremely rare in cerebellar ataxias. We observed generalized choreiform hyperkinesia in the end stage of 2 patients with SCA3 and 1 with SCA1. One patient with a genetically undefined form of autosomal dominant cerebellar ataxia presented with distally pronounced choreatic limb movements.

Bradykinesia and slow alternating movements are the most frequent EPS in all subtypes of ataxia and occur even in patients without any further sign of extrapyramidal disease. Especially in Friedreich ataxia and SCA1, other forms of EPS are restricted to single patients (Table 1).

In SCA3, EPS are more frequent than in other subtypes of autosomal dominant cerebellar ataxia. Five patients from 2 families of our cohort of 80 patients with SCA3 (38 families) presented with marked parkinsonism, including rigidity, reduced arm swing, slowed alternating movements, and bradykinesia. Some patients additionally experienced retropulsion, impaired postural reflexes, or resting tremor. These SCA3 patients had a substantial benefit from levodopa therapy of up to 15 years. On detailed examination, all patients had mild cerebellar oculomotor signs. The diagnosis of Parkinson disease had been made by neurologists, however, in all 5 patients for up to 15 years.

No genetic marker could be established that predisposed to the development of parkinsonism in patients with SCA3. Repeat length in the expanded allele was 71 CAG units in 3 patients, and 73 and 78 CAG in the 2 other patients, respectively. On the other hand, many other SCA3 patients with similar repeat lengths and similar disease duration did not develop parkinsonism.

Late-onset variants of ISCA were associated with a more rapid loss of independent walking than early-onset cerebellar ataxia. The median time from onset of symptoms to the use of a walking aid was 6 years in the late-onset group compared with 18 years in the early-onset group (95% confidence intervals: early-onset, 10-24 years; late onset, 5-16 years). Differences in progression to wheelchair dependence were not significant. The median for the early-onset cerebellar ataxia subgroup was not determinable because fewer than 50% of patients were wheelchair dependent.

Extrapyramidal motor signs are more frequent in late-onset than in early-onset cerebellar ataxias (Table 1). In late-onset cerebellar ataxia, parkinsonian features such as facial impassivity, monotonous soft voice, rigidity, short-stepped gait, reduced arm swing, bradykinesia, retropulsion, and impaired postural reflexes, in particular, were present in a substantial subset of cases, but were rare in patients with early-onset disease. Apart from slow alternating movements, only postural and/or action tremor were present in a substantial subset of patients with early-onset cerebellar ataxia.

The influence of EPS on the course of idiopathic ataxia has been assessed by survival time methods (Kaplan-Meier curves and log-rank tests) for hallmarks of progression (loss of independent walking and wheelchair dependence) as stated in the "Patients and Methods" section. Since Kaplan-Meier plots are an event-related approach, subgroup analyses (eg, excluding patients with probable MSA) were not possible because of critical low frequency of EPS. There was a general tendency toward later onset and a more rapid progression in patients with EPS (Table 2). Bradykinesia and impaired postural reflex had significant negative effects on the course of the disease concerning loss of the ability to walk unsupported and need for a wheelchair. Patients with short-stepped gait or slow alternating movements had a significantly increased risk to become dependent on a walking aid. Furthermore, dysarthria with a monotonous and soft voice and postural or action tremor increased the risk to become wheelchair dependent (Table 2). Statistical analyses for the influence of resting tremor, flexed posture, bradyphrenia, dystonia, myoclonus, and choreiform hyperkinesia on the course of the disease were not performed because of the small numbers of patients with ISCA presenting these signs.

To analyze how many patients with sporadic ataxia and parkinsonian features have MSA, we looked for signs of dysautonomia as an obligatory criterion for the diagnosis of probable MSA.³ Twenty patients with sporadic ataxia fulfilled the criteria of parkinsonism (bradykinesia plus rigidity, postural instability or tremor). Of these 20 MSA candidates, 10 were incontinent, and 4 of these patients had additional orthostatic hypotension, confirming the clinical diagnosis of probable MSA. None of our 20 MSA candidates had orthostatic hypotension (orthostatic fall in blood pressure by 30 mm Hg systolic or 15 mm Hg diastolic) without urinary incontinence. Ten patients with ataxia and parkinsonism had neither orthostatic hypotension nor incontinence; therefore, the diagnosis of MSA could not be made. Disease onset varied in the 10 patients with MSA, ranging between 45 and 62 years of age. Duration of the disease was 2 to 7 years in MSA patients. In the 10 patients without dysautonomia, disease onset was between 35 and 65 years of age, and duration of the disease was 1 to 9 years.

As MSA has a poor prognosis compared with other forms of idiopathic ataxia,⁸ we tested whether autonomic failure—as one of the main criteria of MSA—has a prognostic impact in association with ataxia if it is studied independent of parkinsonism.

Using Kaplan-Meier survival curves we determined median and event-free proportion for time to walking with support and time to wheelchair dependence in patients with sporadic ataxia with or without incontinence or orthostatic hypotension (Table 3). Progression to a walking aid as well as progression to wheelchair dependence was significantly more rapid in incontinent than in continent patients.

Symptomatic orthostatic hypotension or reduction of blood pressure in erecting from a recumbent position (systolic ≥30 mm Hg or diastolic ≥15 mm Hg) was observed in 7 patients, associated with incontinence in all but 1 patient. Because of the small numbers, the prognostic value of orthostatic hypotension could not be determined.

In this study, we analyzed EPS in a large cohort of patients with ataxia. Extrapyramidal motor signs occurred in all subtypes of degenerative ataxia, but expression and frequency varied. Extrapyramidal motor signs were more frequent in genetic subtypes of ataxia, such as SCA2 and SCA3, in which basal ganglia affection has been demonstrated by postmortem studies.^12-15 In contrast, EPS were rare in diseases in which basal ganglia are spared by the neurodegenerative process, such as in SCA6 and Friedreich ataxia.^16-18 In SCA1, clinical and pathological observations of extrapyramidal system involvement were variable.^13,19-21

In hereditary ataxias, some EPS were quite characteristic for the underlying mutation: more pronounced postural and/or action tremor was typical for SCA2, but may also occur in a mild form in Friedreich ataxia. Dystonia is generally rare in degenerative ataxia, but, if present, suggests the diagnosis of SCA3, with few exceptions.

We observed a parkinsonian phenotype comprising bradykinesia, rigidity, and mild resting tremor with only minor signs of cerebellar disease in some patients with SCA3. Parkinsonian features responded well to levodopa therapy, which widely relieved symptoms for up to 15 years in individual patients. Similar observations have been reported by other groups.^22,23 Due to dominating parkinsonian signs at the onset of disease, the diagnosis of SCA3 had frequently been overlooked for many years, even by experienced neurologists, before the more widespread nature of the underlying pathological process became more obvious. However, we found cerebellar oculomotor signs, such as impaired visual suppression of the vestibuloocular reflex or gaze-evoked nystagmus, in all patients; this observation may help to reliably distinguish between Parkinson disease and the parkinsonian variant of SCA3.

Some authors suggest the parkinsonian variant as a distinct subtype of SCA3, analogous to the dystonic-rigid, spastic, or neuropathy subtype of SCA3.^23,24 These 3 subtypes have been shown to be associated with a distinct course of the disease, and are widely determined by CAG repeat length. In contrast, the parkinsonian variant is not restricted to a distinct repeat length, but appears to run in families, suggesting a supporting genetic background, eg, a modifier gene.

Apart from SCA3, we observed levodopa-responsive parkinsonism as a prominent feature in patients with autosomal dominant cerebellar ataxia in whom the SCA1, SCA2, SCA3, SCA6, and SCA7 mutations had been excluded. In contrast to the SCA3 families where parkinsonian signs were present at the beginning of the disease, parkinsonism in these autosomal dominant cerebellar ataxia families developed many years after the onset of cerebellar deficits.

Prominent parkinsonism in a patient with sporadic ataxia is suggestive of MSA. However, 50% of patients with sporadic ataxia who fulfilled the criteria of parkinsonism, as stated by a consensus conference,³ had neither urinary incontinence nor orthostatic hypotension, and did not develop the full picture of MSA with obligatory autonomic failure despite a disease duration of up to 9 years. Classification of these ataxia patients with parkinsonism remains unclear. Follow-up studies are needed to observe whether these patients develop dysautonomia later in the course of the disease. Since MSA is a neuropathological diagnosis, evaluation by postmortem studies are mandatory to elucidate this problem. In the course of the disease, no differences could be established between patients with probable MSA and patients with sporadic ataxia and parkinsonism without autonomic dysfunction.

Our data suggest that parkinsonism is a prognostic factor in idiopathic ataxia even if autonomic function is intact and the diagnosis of MSA cannot be made (Table 4 and Figure 1). Specifically, bradykinesia, impaired postural reflex, parkinsonian dysarthria, short-stepped gait, postural or action tremor, and slow alternating movements are signs associated with a more rapid progression of the disease (Table 2). However, slow alternating movements are less specific for nigrostriatal degeneration, and may be caused by cerebellar problems as in SCA6 or pyramidal pathology as in Friedreich ataxia.

Harding²⁵ differentiated between early- and late-onset forms of ataxia. This concept has been questioned ever since genetically confirmed late-onset variants of Friedreich ataxia and early-onset variants of autosomal dominant cerebellar ataxias have been described. In our study, though, we found substantial differences in disease progression and clinical features such as EPS between late- and early-onset variants of ISCA. This substantiates onset age as an important prognostic factor, at least in presently diagnosed idiopathic forms of ataxia. Interestingly, progression was slower in early-onset forms of ISCA compared with late-onset ISCA, whereas the opposite is true for dominantly inherited ataxias.⁸ In the dominant spinocerebellar ataxias types 1, 2, 3 and 7, more extensive trinucleotide repeat expansions cause earlier onset and more rapid progression. Therefore, our data suggest that mechanisms different from trinucleotide repeat expansions are involved in the pathogenesis of ISCA, at least in the majority of patients.

In conclusion, our study demonstrated EPS to be a major factor in nosological and prognostic considerations in ataxias. Furthermore, EPS are important to recognize, since they respond better to symptomatic treatment than disabilities of cerebellar origin. This is especially true of the parkinsonian variant of SCA3 and presently undefined autosomal dominant cerebellar ataxia forms. Further studies are needed to solve the question whether the additive handicap of ataxia and EPS is responsible for the poorer prognosis, or whether EPS in ataxia signal a different and more deleterious pathomechanism.

Accepted for publication January 28, 2000.

We thank all of the patients who participated in this study. We gratefully acknowledge the excellent laboratory work of Ana Maria Menezes Vieira-Saecker.

Corresponding author: Ludger Schöls, MD, Neurologische Klinik der Ruhr-Universität, St Josef Hospital, Gudrunstrasse 56, D-44791 Bochum, Germany.