Dear Sir:
Cerebral small vessel disease (SVD) encompasses a spectrum of neurological manifestations, including stroke, cognitive decline, and progressive gait difficulty [
1]. The associated gait abnormalities are commonly characterized by slowed and shuffling gait, resembling vascular parkinsonism. Lesions associated with parkinsonism generally affect the striatal and related motor pathways, while SVD neuroimaging marker burden can predict the risk of parkinsonism [
2,
3]. Genetic forms of SVD, such as cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), also commonly present with parkinsonism [
4]. Gait difficulties are common in advanced CADASIL; however, the prevalence and characteristics of parkinsonism remain underexplored [
5].
The present study aimed to comprehensively evaluate the motor features of parkinsonism in patients with CADASIL, assess their prevalence, and identify potential clinical and neuroimaging markers. We hypothesized that the pathophysiology of parkinsonism is primarily vascular, distinct from the α-synucleinopathy seen in idiopathic Parkinson’s disease (PD).
Patients from the Taiwan CADASIL Registry and patients with idiopathic PD from the same hospital were evaluated and compared (
Supplementary Figure 1). In patients with CADASIL, parkinsonism was diagnosed based on gait impairment, bradykinesia, or rigidity, excluding stroke-related paresis [
3,
6]. The severity of parkinsonism motor symptoms was assessed by two movement disorder specialists using the Unified Parkinson’s Disease Rating Scale part III (UPDRS-III), and items in the UPDRS-III were further categorized into four domains: tremor, rigidity, bradykinesia, and postural instability. Overall disability was determined using the Hoehn-Yahr (H-Y) stage. Gait speed was examined using the 4-meter walking test. Magnetic resonance imaging (MRI) features were analyzed visually and quantitatively. Plasma α-synuclein levels were measured using an immunomagnetic reduction method. The first part of the analysis identified factors associated with parkinsonism in CADASIL, and the second part compared the phenotypes and severity of Parkinsonism between patients with CADASIL and those with PD. Detailed methods are provided in the
Supplementary Methods.
Among the 75 CADASIL patients enrolled (median age 64 years, interquartile range [IQR] 57-72 years, 52% male), 55 (73%) exhibited parkinsonism features, with the prevalence increasing with age (
Supplementary Figure 2). The median UPDRS-III score was 9 (IQR, 3-23), median H-Y stage was 1 (1-2), and median gait speed was 0.83 m/s (0.57-1.09 m/s). The neuroimaging markers are shown in
Table 1.
Of the 55 patients with parkinsonism, 19 reported subjective motor symptoms, while others were identified during examination. The median age at parkinsonism onset was 66 years (IQR, 62-70 years). Twenty-two patients (40%) received levodopa at a median levodopa-equivalent daily dose of 225 mg (100-400 mg). Among the 32 patients who underwent 99mTc-TRODAT-1 SPECT (single-photon emission computed tomography), 31 (97%) showed decreased dopamine uptake, which was graded as mild (75%), moderate (16%), or severe (6%). A decrease in asymmetrical uptake was observed in 24 (77%) patients.
CADASIL patients with parkinsonism were older (median age 67 vs. 57) than those without parkinsonism, though no significant differences were noted in vascular risk factors, stroke history, or plasma α-synuclein levels (
Table 1). Neuroimaging findings revealed that CADASIL patients with parkinsonism had thinner cortices (median 2.35 vs. 2.39 mm), larger white matter hyperintensity (WMH) volumes (47.3 vs. 20.2 mL), and a higher prevalence of basal ganglia lacunes (60% vs. 30%). Patients with higher H-Y stages also had larger WMH volumes, thinner cortices, and a trend toward more lacunes and microbleeds (
Supplementary Table 1).
Figure 1A shows examples of brain MRI scans in patients with CADASIL at different H-Y stages.
In the age- and sex-adjusted logistic regression models, age, WMH volume, Fazekas scale score, and basal ganglia lacunes were associated with parkinsonism (
Table 1). In multivariate analysis, only the WMH volume remained statistically significant (odds ratio [OR], 1.04; 95% confidence interval [CI], 1.003-1.07). Among patients with parkinsonism, thinner cortical thickness was associated with higher UPDRS-III scores (adjusted beta -5.48, 95% CI -9.98 to -0.99) and advanced H-Y stages (OR 0.30, 95% CI 0.07-1.00) (
Supplementary Table 2).
Compared to 158 patients with idiopathic PD (median age, 70 years; 51% male), CADASIL patients with parkinsonism had lower UPDRS-III scores (median, 12 vs. 18) but similar H-Y stages (2 vs. 2) (
Table 2). Patients with CADASIL had significantly lower tremor (0 vs. 0) and rigidity scores (1 vs. 4), with borderline lower bradykinesia (7 vs. 9), but no differences in postural instability (4 vs. 4) (
Figure 1B). Plasma α-synuclein levels were significantly lower in CADASIL patients with parkinsonism (0.08 vs. 0.95 pg/mL) compared with patients with PD. Plasma α-synuclein levels correlated with UPDRS-III scores in PD patients (beta 2.13, 95% CI 0.83-3.43), but not in CADASIL patients (beta 5.27, 95% CI -2.21-12.75) (
Supplementary Table 3 and
Figure 1C). Neuroimaging comparisons revealed that CADASIL patients with parkinsonism had thinner cortices (median 2.35 vs. 2.44 mm), higher Fazekas scores (6 vs. 2), and more lacunes (7 vs. 0) and microbleeds (21 vs. 0) than PD patients.
In summary, we found that over 70% of symptomatic patients with CADASIL exhibited parkinsonism, a higher prevalence than previously reported. Prior studies have reported parkinsonism in 11% of 45 Italian patients and 33% of 157 Korean patients [
4,
5]. These discrepancies may be due to the older age of our cohort, differences in
NOTCH3 variants, and more thorough motor examinations. Motor symptoms in CADASIL, characterized by postural instability and bradykinesia with less rigidity and tremor, resemble “lower-body” type of vascular parkinsonism.
Our study suggests that WMH volume contributes to the development of parkinsonism in CADASIL, while cortical thinning correlates with motor symptom severity. These findings are consistent with existing evidence showing that the severity of SVD features, such as WMH volume and cortical thinning, may serve as surrogate markers of parkinsonism. For example, a cross-sectional study in sporadic SVD patients found that the total SVD burden was associated with mild parkinsonism features [
7]. Similarly, in sporadic SVD patients without parkinsonism at baseline, a high WMH volume and number of lacunes were associated with the incident parkinsonism over five years [
2]. Furthermore, in PD patients, WMH was associated with worsening axial motor performance, independent of nigrostriatal dopaminergic denervation [
8], while cortical thinning was linked to advanced motor stage [
9]. In CADASIL, a diffuse SVD, the cumulative WMH burden could potentially damage the cortico-basal ganglia-thalamus interconnecting fibers, and secondary neurodegeneration, such as cortical thinning, may occur.
We speculate that the association of cortical thinning, but not WMH, with motor severity could be linked to the ceiling effect of large WMH volumes in patients with CADASIL; however, secondary cortical thinning, reflecting brain atrophy, is more closely related to the severity of clinical symptoms. This is consistent with findings that brain atrophy is strongly associated with cognitive score, disability scales, and predicting clinical deterioration in CADASIL [
10]. The significantly lower plasma α-synuclein levels in CADASIL patients, which did not correlate with motor symptoms, indicate that α-synucleinopathy is not the underlying pathology of parkinsonism in CADASIL.
While parkinsonism can occur in sporadic SVD or in individuals with brain frailty, CADASIL is a genetically driven subtype with a more severe vascular pathology, particularly in the cortico-basal ganglia-thalamic circuits. Some of our CADASIL cases were initially diagnosed with vascular parkinsonism, only later receiving a genetic diagnosis due to neuroimaging features such as disproportionately severe WMH, multiple lacunes, and microbleeds. We propose that clinicians review brain MRI scans carefully in cases of vascular parkinsonism and be aware that genetic SVD can also result in such a phenotype.
Our study has several limitations. First, a selection bias may exist because severe cases were excluded, and minor cases might have been missing. Second, the study enrolled CADASIL patients with the predominant NOTCH3 p.R544C variant; therefore, the results might not be generalizable to populations with different NOTCH3 variants. Third, NOTCH3 was not tested in patients with PD. Finally, we did not compare the differences between CADASIL patients with parkinsonism and those with non-genetic vascular parkinsonism.
In conclusion, we found that parkinsonism is common in patients with CADASIL. The burden of cerebral SVD, particularly WMH volume and cortical thinning, may contribute to its development and severity.