Topography of cortical thinning associated with white matter hyperintensities in Parkinson's disease

Parkinsonism and Related Disorders 21 (2015) 372e377

Contents lists available at ScienceDirect

Parkinsonism and Related Disorders journal homepage: www.elsevier.com/locate/parkreldis

Topography of cortical thinning associated with white matter hyperintensities in Parkinson's disease Jee Hyun Ham a, Hyuk Jin Yun b, Mun-Kyung Sunwoo c, Jin Yong Hong d, Jong-Min Lee b, Young H. Sohn a, Phil Hyu Lee a, e, * a

Department of Neurology, Yonsei University College of Medicine, Seoul, South Korea Department of Biomedical Engineering, Hanyang University, Seoul, South Korea Department of Neurology, Bundang Jesaeng General Hospital, Seongnam, South Korea d Department of Neurology, Yonsei University Wonju College of Medicine, Wonju, South Korea e Severance Biomedical Science Institute, Seoul, South Korea b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 September 2014 Received in revised form 22 December 2014 Accepted 28 January 2015

Background: Although white matter hyperintensities (WMHs) are associated with cognitive impairments in Parkinson's disease (PD), the relationships between WMHs and cortical atrophy in regard to cognitive impairments are unknown. Here, we investigated the topography of cortical thinning related to deep (DWMHs) and periventricular WMHs (PWMHs) and their differential impacts on cognitive performance in PD. Methods: We enrolled 87 patients with non-demented PD and evaluated WMH scores using a semiquantitative visual rating system. The patients were divided into low-, moderate-, and high-grade groups based on WMH severity for total WMHs (TWMHs), DWMHs, and PWMHs, and cortical thickness was measured using a surface-based method according to the WMHs severity. Additionally, the correlations between WMH-associated cortical thinning and neuropsychological performance were analyzed. Results: The detailed neuropsychological test demonstrated that PD patients with high-grade WMHs showed poorer performance on frontal lobe-based cognitive tasks compared with those with low-grade DWMHs. The areas of cortical thinning were more extensive in patients with DWMHs, involving the entire frontal areas and restricted temporoparietal areas, whereas in patients with PWMHs, cortical thinning was localized in the small frontal areas. A multiple regression analysis of the relationships between WMH-associated cortical thickness and cognition revealed that DWMH-associated frontal thickness had an independent effect on frontal lobe-based cognition, while frontal thickness related to PWMHs did not have a significant correlation with cognitive tasks. Conclusions: These data suggest that in patients with PD, DWMHs are closely coupled with decreased cortical thickness in the frontal areas and may lead to declines in executive function. © 2015 Published by Elsevier Ltd.

Keywords: Parkinson's disease White matter hyperintensities Cortical thinning

Cognitive impairments are one of the most disabling non-motor symptoms associated with Parkinson's disease (PD). Patients with PD have a 3- to 6-fold higher risk of developing dementia compared with controls [1], and one-fifth of patients with untreated early PD exhibit mild cognitive impairments (MCI) [2]. A number of possible pathological mechanisms underlying PD-related cognitive dysfunction have been suggested. Of these, a-synuclein is a key * Corresponding author. Department of Neurology, Yonsei University College of Medicine, 250 Seongsanno, Seodaemun-gu, Seoul 120-752, South Korea. Tel.: þ82 2 2228 1608; fax: þ82 2 393 0705. E-mail address: [email protected] (P.H. Lee). http://dx.doi.org/10.1016/j.parkreldis.2015.01.015 1353-8020/© 2015 Published by Elsevier Ltd.

pathological contributor to the development of dementia in PD [3]. Moreover, co-existing Alzheimer's disease (AD) pathologies may influence the onset and progression rate of cognitive decline in patients with PD [4]. Some studies have suggested that a silent vascular pathology known as white matter hyperintensities (WMHs) may be associated with cognitive dysfunction in PD. In fact, WMHs are associated with cognitive decline in normal aging [4] and have been identified as a risk factor for the transition to AD in patients with MCI [5]. WMHs, revealed by T2-weighted magnetic resonance imaging (MRI) scans, are typically categorized as deep white matter hyperintensities (DWMHs), patchy areas of WMHs in subcortical

J.H. Ham et al. / Parkinsonism and Related Disorders 21 (2015) 372e377

white matter, or periventricular white matter hyperintensities (PWMHs), which are adjacent to the cerebral ventricles [6]. Even though DWMHs and PWMHs are commonly observed on the MRI scans of elderly persons, the differential effects of specific WMHs on neuropsychological function remain controversial. Both DWMHs and PWMHs are significantly associated with executive dysfunction in patients with MCI [7], but only the progression of PWMH volume in non-demented elderly individuals was associated with a decline in cognitive processing speed [8]. Meanwhile, other evidence suggests that DWMHs are primarily related to executive dysfunction in older adults with MCI [9] and in middle-age individuals [10], indicating that DWMHs and PWMHs may have differential effects on cognitive dysfunction. Similarly, several studies have found that WMHs in patients with PD are associated with cognitive impairments [11,12]; however, these results are inconsistent among studies [13]. In terms of association between WMHs and cortical atrophy, unlike AD [14] or vascular cognitive impairments [15], the relationships between different types of WMHs and cortical atrophy in PD, as well as their association with cognitive function, are not yet fully understood. Thus, we evaluated the patterns of cortical thinning related to DWMHs and PWMHs and analyzed the correlations between WMH-associated cortical thinning and neuropsychological performance in patients with non-demented PD. 1. Patients and methods

373

grade WMH groups. WMH scores were rated blindly (by H.J.H. and S.M.K.), and the intra- and inter-scorer reliability (expressed as correlation coefficients) were 0.93 and 0.86, respectively. Additionally, a high-resolution T1-weighted MRI was obtained using a three-dimensional T1-TFE sequence configured with the following acquisition parameters: axial acquisition with a 224  224 matrix; 256  256 reconstructed matrix; 220  220 mm field of view; 0.86  0.86  1.0 mm voxels; echo time, 4.6 ms; repetition time, 9.6 ms; flip angle, 8 ; and slice gap, 0 mm. 1.3. Image acquisition and processing T1-weighted images were registered in the ICBM-152 average template using a linear transformation and were corrected for intensity non-uniformity artifacts [19]. The images were then classified as white matter (WM), gray matter (GM), cerebrospinal fluid (CSF), or background using an advanced neural net classifier. Hemispheric cortical surfaces were automatically extracted from each T1-weighted image using the Constrained Laplacian-based Automated Segmentation with Proximities (CLASP) algorithm, which reconstructs the inner cortical surface by deforming a spherical mesh onto the WM/GM boundary and then expanding the deformable model to the GM/CSF boundary [20]. The reconstructed hemispheric cortical surfaces consisted of 40,962 vertices, each forming high-resolution meshes. The inner and outer cortical surfaces had the same number of vertices, and there was a close correspondence between the counterpart vertices of the inner and outer cortical surfaces. Cortical thickness was defined using the t-link method, which captures the Euclidean distance between these linked vertices (Supplementary Fig. 1) [20]. Diffusion smoothing with a full-width half maximum of 20 mm was used to blur each map of cortical thickness, which increased the signal-to-noise ratio and statistical power. Each smoothed individual thickness map was then transformed to a surface group template using a 2-dimensional (2D) surface-based registration that aligns variable sulcal folding patterns through sphere-to-sphere warping [21]. For lobar regional analysis, a lobe-parcellated surface group template was used; the definition of the lobar regions has been described in detail previously [22].

1.1. Subjects This cross-sectional study enrolled 87 patients with non-demented PD from a university hospital between January 2008 and January 2013. PD was diagnosed according to the clinical diagnostic criteria of the United Kingdom PD Society Brain Bank [16]. The study subjects were divided into total WMH (TWMH), DWMH, and PWMH groups based on the locations of WMHs and each subgroup was further classified as having high-, moderate- or low-grade WHMs. The Seoul Neuropsychological Screening Battery (SNSB) [17] was employed to determine impairments in specific cognitive subsets. The SNSB measures attention, language, visuoconstructive function, verbal and visual memory, and frontal/executive function. The quantifiable tests consisted of a digit span task, the Korean version of the Boston Naming Test, the Rey Complex Figure Test, the Seoul Verbal Learning Test, the Controlled Oral Word Association Test, a go/no-go test and contrasting programming, and the Stroop Test. Each of these quantifiable cognitive tests has age-, sex-, and education-specific norms available that are based on 447 normal subjects, and the scores on the tests in the current study were classified as abnormal if they were below the 16th percentile for matched normal subjects. Parkinsonian motor symptoms were assessed using the Unified PD Rating Scale Part III. The basic demographic data for gender, age, and histories of hypertension, diabetes mellitus, or cerebrovascular accidents were also analyzed. Patients with a pharmacological history of drugs that induce parkinsonism were excluded. Additionally, exclusion criteria consisted of evidence of focal brain lesions on MRI scans or the presence of other neurodegenerative diseases that might account for cognitive dysfunctions. An [18F] FP-CIT positron emission tomography scan was performed on all subjects, all of whom exhibited decreased dopamine transporter uptake in the posterior putamen. Informed consent was obtained from all patients and control subjects. This study was approved by the Institutional Review Board of Yonsei University Severance Hospital. 1.2. Brain MRI All scans of patients were acquired using a 3.0-T system (Intera or Achieva, Philips Medical System; Best, The Netherlands). WMHs were determined using fluid attenuated inversion recovery sequence images (TR/TE/TI, 8502/132/2100 ms, 5-mm section thickness) and the WMH scores were rated using a semi-quantitative visual rating system [18]. PWMHs were identified as continuous, confluent areas of high signal intensity adjacent to the anterior or posterior horns of the lateral ventricles (“cap”) and along the lateral ventricles (“bands”). Absence of lesions was a score of 0, 5 mm lesions was a score of 1, and a score of 2 was given for lesions >5 mm. DWMHs were more than 10 mm from the lateral ventricle, and were located in the deep or subcortical white matter. These rating criteria were also applied to regions of the basal ganglia (caudate, putamen, globus pallidus, thalamus, and internal capsule) and infratentorial regions (cerebellum, midbrain, pons, and medulla). This rating scale provides four sum scores in a semi-quantitative manner: PWMHs (0e6), DWMHs (0e24), basal ganglia WMHs (0e30), and infratentorial WMHs (0e24). That all combined to make TWMHs score (0e84). Based on the distribution of WMH scores, patients with PD were divided into tertiles of the low-, moderate- and high-

1.4. Statistical analysis The statistical analyses were performed using the Statistical Package for the Social Sciences, version 20.0 (SPSS, Inc.; Chicago, IL, USA). Differences in the baseline demographic characteristics between low-, moderate-, and high-grade WMH groups were evaluated using an analysis of variance test for continuous variables or the chi-Square test for categorical variables. An analysis of covariance (ANCOVA) was used to compare differences between the low-, moderate- and high-grade WMH groups for the neuropsychological testing, and it was adjusted for age, gender, education level, and Mini-Mental State Examination (MMSE) scores. The global difference and corrected t-statistical maps of cortical thickness were analyzed between the low- and high-grade WMH groups, after adjusting for age, sex, years of education, disease duration, and intracranial volume (ICV) as covariates. Statistical analyses were performed using the SurfStat toolbox (http://www.math.mcgill.ca/keith/ surfstat/) for Matlab (R2010b; MathWorks; Natick, MA, USA). The group differences in cortical thickness were considered to be significant at a random-field theorycorrected value of P < 0.05. To investigate the correlation between WMH-associated cortical thickness and cognitive performance, we obtained the mean cortical thickness of cortical thinning regions in all study subjects where cortical thickness in high-grade WMHs was significantly decreased relative to low-grade WMHs. Additionally, the cortical thickness of atrophic region in individual subject where cortical thinning was more severe in the high-grade WMH group than in the low-grade WMH group was parcellated into frontal, parietal, temporal, and occipital lobes, and we used frontal thickness in following correlation and regression analyses. A partial correlation analysis was performed between WMH-associated cortical thickness and WMH severity after controlling age, gender, education, MMSE and ICV. In Model 1 of a multiple regression analysis, we used cognitive subsets, age, gender, education level, MMSE, and ICV as independent variables and the WMHassociated mean cortical thickness (depending on each WMH from the ANCOVA test) as a dependent variable. In Model 2 to explore independent effects of WMHassociated cortical thickness on cognitive performance, the WMH type was added to predictors of Model 1. Next, a multiple regression analysis was performed using cognitive subsets, age, gender, education level, MMSE, and ICV as independent variables and WMH-associated cortical thickness in frontal area as a dependent variable (Model 1). Finally, the WMH type was added to the predictors for each multiple regression model (Model 2). For all tests, a p value <0.05 was considered to be statistically significant.

2. Results 2.1. Demographic characteristics and neuropsychological data Based on the severity of the WMHs, 29 patients exhibited highgrade TWMHs (>15), while 28 had low-grade TWMHs (0e5); 26 patients had high-grade DWMHs (>10), while 32 had low-grade

374

J.H. Ham et al. / Parkinsonism and Related Disorders 21 (2015) 372e377

DWMHs (0e2); 22 patients had high-grade PWMHs (5e6), while 26 had low-grade PWMHs (0e1). The demographic characteristics for all subjects are provided in Table 1. No significant differences were found between low- and high-grade patients regarding age, gender, education level, MMSE, or vascular risk factors for DWMHs, PWMHs, or TWMHs. However, patients with high-grade TWMHs exhibited a greater frequency of hypertension and higher UPDRS III scores compared with low-grade TWMH patients. The detailed neuropsychological test results are shown in Table 2 and Supplementary Tables 1 and 2 After adjusting for age, sex, years of education, and MMSE, patients with high-grade DWMH demonstrated lower scores in the word Stoop (111.6 versus 99.8, p ¼ 0.007) tasks compared with those with low-grade DWMHs (Table 2). However, patients with TWMH (Supplementary Table 1) and PWMHs (Supplementary Table 2) did not show a significance difference. 2.2. Differential effects of TWMHs, DWMHs, and PWMHs on cortical thinning Patients with high-grade TWMHs exhibited cortical thinning in the dorsolateral prefrontal cortex, orbitofrontal cortex, anterior

cingulate cortex, left anterior temporal area, and the right superior and medial temporal gyri compared with patients with low-grade TWMHs (Fig. 1A). In patients with DWMHs, the high-grade group showed decreased cortical thickness in the dorsolateral frontal cortex, right anterior cingulate cortex, left orbitofrontal cortex, primary motor cortex, basal forebrain, inferior temporal gyrus, and right inferior parietal regions compared with the low-grade group (Fig. 1B). Finally, in patients with PWMHs, the high-grade group displayed decreased cortical thickness in the dorsolateral prefrontal cortex relative to the low-grade group, but this was small and localized (Fig. 1C). There were no areas in which patients with lowgrade TWMHs, DWMHs, or PWMHs exhibited decreased cortical thickness relative to patients with matched high-grade WMHs. 2.3. Correlation between WMH and cortical thickness After adjusting age, gender, education, MMSE and ICV, the score of TWMH showed a significant negative correlation with mean cortical thickness (r ¼ 0.546, p < 0.001) and frontal thickness (r ¼ 0.461, p < 0.001). Similarly, the score of DWMH showed a significant negative correlation with mean cortical thickness (r ¼ 0.507, p < 0.001) and frontal thickness (r ¼ 0.492,

Table 1 Demographic characteristics in Parkinson's disease according to the severity of white matter hyperintensities (WMH). Total WMH

Age Gender (men) hypertension Diabetes mellitus Total cholesterol CVA smoking Disease duration (m) UPDRS III(0e108) MMSE(0e30) Education(yr)

Low (n ¼ 28)

Moderate (n ¼ 30)

High (n ¼ 29)

a

b

b

67.7 ± 6.1 15 (53.6%) 8 (28.6%) 4 (13.8%) 185.8 ± 37.6 0 (0.0%) 0 (0.0%) 31.1 ± 37.6 18 ± 10(2e37) 27 ± 2(23e30) 9.7 ± 4.4

68.4 ± 7.5 12 (44.4%) 12 (44.4%) 5 (18.5%) 183.8 ± 36.1 1 (1.2%) 0 (0.0%) 51.2 ± 54.2 25 ± 11(5e48) 27 ± 2(22e30) 9.5 ± 4.4

70.3 ± 5.8 12 (41.4%) 16 (55.2%) 3 (10.3%) 181.6 ± 32.2 0 (0.0%) 0 (0.0%) 36.4 ± 34.5 26 ± 14(2e48) 27 ± 2(22e30) 10.0 ± 4.3

0.252 0.633 0.125 0.683 0.901 0.350

0.309 0.293 0.033 0.955 NS

NS 0.498 0.221 0.671 NS 0.313

0.889 0.817 0.422 0.382 NS 0.296

0.213 0.020 0.614 0.872

NS 0.035 NS NS

0.274 0.074 0.975 NS

0.606 NS NS NS

Low (n ¼ 32)

Moderate (n ¼ 29)

High (n ¼ 26)

a

b

b

71.1 ± 4.7 16 (50.0%) 11 (34.4%) 3 (9.4%) 176.7 ± 33.8 0 (0 0%) 1 (2.4%) 33.2 ± 44.6 20 ± 11(2e39) 27 ± 2(22e30) 9.5 ± 4.5

67.7 ± 6.7 15 (51.7%) 12 (41.4%) 3 (10.3%) 181.3 ± 30.5 1 (3.4%) 0 (0.0%) 39.1 ± 38.0 23 ± 11(9e45) 27 ± 2(21e30) 9.6 ± 5.1

71.5 ± 6.2 13 (50.0%) 16 (61.5%) 2 (7.7) 173.4 ± 30.6 0 (0.0%) 0 (0.0%) 38.3 ± 37.5 23 ± 14(8e44) 26 ± 3(22e30) 10.0 ± 4.3

0.092 0.989 0.106 0.943 0.726 0.370 0.408 0.826 0.581 0.347 0.936

NS NS 0.039 0.820 NS

0.085 0.498 0.221 0.671 NS 0.313

0.056 0.898 0.135 0.733 NS 0.339

0.356 NS NS 0.444 NS

NS NS NS NS

NS NS NS NS

Low (n ¼ 26)

Moderate (n ¼ 39)

High (n ¼ 22)

a

b

b

b

69.8 ± 4.8 15 (60.0%) 12 (48.0%) 4 (16.0%) 191.0 ± 41 8 1 (4.0%) 1 (4.0%) 33.8 ± 45.5 21 ± 10(4e39) 27 ± 1(24e30) 9.5 ± 4.7

69.9 ± 5.8 14 (51.9%) 10 (37.0%) 4 (14.8%) 184.6 ± 34.1 1 (3.7%) 1 (3.7%) 36.4 ± 49.8 24 ± 10(8e47) 26 ± 3(24e30) 8.6 ± 4.7

72.3 ± 6.5 12 (54.5%) 13 (59.1%) 2 (9.1%) 184.0 ± 32.9 0 (0.0%) 1 (4.5%) 25.0 ± 29.3 25 ± 9(5e44) 27 ± 2(22e30) 10.1 ± 3.6

0.251 0.836 0.305 0.764 0.802 0.646 0.989 0.637 0.411 0.328 0.471

0.433 0.706 0.447 0.479 NS 0.343 0.926 NS 0.665 NS NS

NS 0.554 0.424 0.906 NS 0.956 0.956 NS 0.911 0.691 NS

p

p1

p2

b

p3

Deep WMH

Age Gender (mean) hypertension Diabetes mellitus Total cholesterol CVA smoking Disease duration (m) UPDRS III(0e108) MMSE(0e30) Education(yr)

p

p1

p2

b

p3

Periventricular WMH

Age Gender (men, %) hypertension Diabetes mellitus Total cholesterol CVA smoking Disease duration (m) UPDRS III(0e108) MMSE(0e30) Education(yr)

p

p1

p2

Values are expressed as number of subjects (%) or mean ± SD. p1: low-grade group vs high-grade group, p2: low-grade group vs moderate-grade group, p3: moderate-grade group vs high-grade group. NS: not significant, CVA: Cerebrovascular Accident, UPDRSIII: Unified Parkinson's Disease Rating Scale part III, MMSE: Mini-Mental State Examination. a p-value from analysis of variance test or Chi-square test (or Fisher's exact test). b Bonferroni corrected p-values of the post-hoc pairwise comparison tests.

p3

0.437 0.851 0.124 0.543 NS 0.362 0.882 NS NS 0.539 0.678

J.H. Ham et al. / Parkinsonism and Related Disorders 21 (2015) 372e377

375

Table 2 Neuropsychological data in patients with Parkinson's disease based on the severity of deep white matter hyperintensities. Deep white matter hyperintensities Low (n ¼ 32) Attention Digit span (forward) Digit span (backward) Language and related function K-BNT Repetition Visuospatial function RCFT Verbal memory function (SVLT) Immediate recall Delayed recall Recognition Visual memory function (RCFT) Immediate recall Delayed recall Recognition Frontal executive function Contrasting program Go-no-go test Phonemic fluency COWAT (Animal) COWAT (Supermarket) Word stroop test Color stroop test

Moderate (n ¼ 25)

High (n ¼ 26)

a

p

b

b

b

p1

p2

p3

6.4 ± 1.4 3.5 ± 1.0

6.5 ± 1.5 3.9 ± 1.6

5.8 ± 1.8 3.1 ± 1.1

NS NS

NS NS

NS NS

NS NS

44.6 ± 8.8 14.4 ± 1.0

42.9 ± 11.4 14.6 ± 0.9

39.3 ± 9.7 14.4 ± 1.1

NS NS

NS NS

NS NS

NS NS

31.6 ± 5.4

32 1 ± 6.9

30.9 ± 7.3

NS

NS

NS

NS

16.0 ± 5.2 4.6 ± 2.6 9.1 ± 2.3

16.7 ± 5.3 4.6 ± 2.9 8.8 ± 2.9

15.5 ± 4.2 3.8 ± 2.8 8.6 ± 2.8

NS NS NS

NS NS NS

NS NS NS

NS NS NS

12.1 ± 6.6 11.8 ± 6.7 9.4 ± 2.0

12.4 ± 6.7 11.8 ± 6.6 9.1 ± 2.4

9.4 ± 6.6 11.4 ± 6.8 9.0 ± 2.2

NS NS NS

NS NS NS

NS NS NS

NS NS NS

NS NS NS NS NS 0.007 NS

NS NS NS NS NS 0.007 NS

NS NS NS NS NS NS NS

NS NS NS NS NS NS NS

19.8 18.2 19.4 13.5 15.3 111.6 66.8

± ± ± ± ± ± ±

0.7 3.5 9.7 3.6 5.2 1.0 27.4

18.6 18.3 20.7 14.1 16.8 109.2 76.8

± ± ± ± ± ± ±

3.5 3.3 12.3 4.1 4.5 11.0 27.4

18.8 17.3 16.6 11.3 12.6 99.8 57.7

± ± ± ± ± ± ±

3.6 4.3 0.9 3.4 5.4 18.1 28.7

Values are expressed as mean ± SD. p1: low-grade group vs high-grade group, p2: low-grade group vs moderate-grade group, p3: moderate-grade group vs high-grade group. Bold: statistically significant (p < 0.05), NS: not significant, K-BNT: Korean version of Boston Naming Test, RCFT: Rey Complex Figure Test, SVLT: Seoul Verbal Learning Test, COWAT: Controlled Oral Word Association Test. a p-value adjusted for age, gender, years of education and MMSE. b Bonferroni corrected p-values of the post-hoc pairwise comparison tests and multiple tests. (raw p-value x the number of groups x the number of neuropsychological tests in each domain).

p < 0.001). The score of PWMH was also negatively correlated with mean cortical thickness (r ¼ 0.603, p < 0.001) and frontal thickness (r ¼ -0.516, p < 0.001; Supplementary Fig. 2).

2.4. Correlations between WMH-associated cortical thickness and cognition The results of correlations between WMH-associated mean cortical thickness and cognition are shown in Supplementary Table 3. Analysis of the relationships between mean cortical thickness and scores on the cognitive subsets revealed that TWMHrelated mean cortical thickness was significantly correlated with the word Stroop task (p ¼ 0.006) and that DWMH-associated mean cortical thickness had a significant correlation with the word Stroop (p ¼ 0.009) and semantic COWAT (p ¼ 0.015) tasks. On the other hand, mean cortical thickness related to PWMHs did not have a significant correlation with any of the cognitive tasks. When each type of WMH was added as a covariate, the significant effects of WMH-associated mean cortical thickness on clinical scores from the neuropsychological tests were no longer observed. Because WMH-associated cortical thinning occurred primarily in the frontal region, we used frontal thickness as a predictor in the regression analyses. In patients with TWMHs, frontal thinning area was significantly associated with the word Stroop task (p ¼ 0.005) and frontal thickness in patients with DWMH was significantly correlated with the word Stroop task (p ¼ 0.001) and semantic fluency (p ¼ 0.039). However, frontal thickness in patients with PWMHs did not. When each type of WMH was added as a covariate, the significant effects of frontal thickness on the word Stroop task remained significant in TWMHs (p ¼ 0.030) and DWMHs (p ¼ 0.030) (Table 3).

3. Discussion The present study was the first to investigate the relationships among WMHs, cortical atrophy, and cognitive dysfunction in nondemented patients with PD. The major findings of the current study were that (1) PD patients with high-grade WMHs showed poorer performance on frontal lobe-based cognitive tasks, (2) the topography of cortical thinning related to WMHs was extensive and primarily occurred in the frontal region, (3) the influence of WMHs on cognitive performance or cortical thinning was prominent in patients with DWMHs relative to those with PWMHs, and (4) according to the DWMH score, DWMH-associated frontal thickness had an independent effect on executive dysfunction. These data suggest that in patients with PD, DWMHs are closely coupled with decreased cortical thickness in the frontal areas and may lead to declines in executive function. Ample evidence has demonstrated that, in addition to decreased WM volume and dilated ventricles, WMHs are accompanied by cortical atrophy. Using voxel-based morphometric analysis, Wen and colleagues found that WMHs are associated with reductions of GM in elderly asymptomatic individuals [23]. A number of underlying mechanisms have been suggested as possible mediators of this relationship [24,25]. First, the volume loss of GM may be secondary to that of WMHs, because primary WM pathology can lead to neuronal cell loss. Second, WMHs may be a reflection of WM changes that are secondary to cortical neuronal loss. Finally, it is possible that both WMHs and cortical atrophy are the result of a shared pathological process, such as ischemic insults. In this study, the relationship between WMHs and cortical atrophy was evaluated in patients with PD, showing that patients with severe pathologies exhibited significant cortical thinning compared with those with mild WMHs. These findings indicate that the

376

J.H. Ham et al. / Parkinsonism and Related Disorders 21 (2015) 372e377 Table 3 Relationship among white matter hyperintensities (WMH), cortical thickness in frontal area, and frontal executive tasks. Word stroop task b (standard error)

p

Semantic fluency

p

b (standard error)

Total WMH-associated cortical thickness in frontal area 21.698 (7.555) 0.005 Model 1a 19.303 (8.738) 0.030 Model 2b Deep WMH-associated cortical thickness in frontal area a Model 1 24.330 (7.286) 0.001 4.610 (2.197) Model 2b 19.149 (8.633) 0.030 3.001 (2.602)

0.039 0.252

Bold: statistically significant (p < 0.05). a Multiple regression analysis controlled age, gender, education, MMSE, and ICV. b Multiple regression analysis controlled age, gender, education, MMSE, ICV and each type of WMH.

pathological processes underlying the relationship between WMHs and cortical atrophy may be operating consistently across various disease entities, including PD pathologies. In this study, the areas of cortical thinning were more extensive in patients with DWMHs and involved the entire frontal areas, such as the dorsolateral frontal region, anterior cingulate cortex, and orbitofrontal region, as well as restricted temporoparietal areas. These findings suggest that the specific influence of WMHs on cortical atrophy differs depending on the anatomical distribution of the WMHs. Regarding the pathological aspects, PWMHs appear to be related to myelin pallor or rarefaction without any other convincing evidence of ischemia. Furthermore, WMHs within 3 mm of the ventricles were non-ischemic in origin, and irregular PWMHs were correlated with ischemic damage, which are more likely to be classified as chronic hemodynamic insufficiency [26]. In contrast, DWMHs reflect an increased severity of ischemic damage caused by small vessel disease [27]. Additionally, DWMHs are significantly associated with risk factors such as age and vascular variables, including hypertension, diabetes, and smoking [27], while PWMHs are associated with carotid atherosclerosis [28]. Moreover, plasma homocysteine levels are reported to be a determinant of DWMHs but not PWMHs [29]. In this regard, it is possible that PD-related cerebrovascular risks, such as immobility, supine hypertension, and levodopa-related hyperhomocysteinemia, may be significant conditions that lead to the development of DWMHs. Thus, cortical neuronal damage is more extensive in the presence of DWMHs relative to PWMHs, because ischemic insults promote the oligomerization and aggregation of alpha-synuclein, a key protein underlying the pathogenesis of PD. In turn, this may increase the pathological burden of PD [30]. However, cerebrovascular risk factors in PD are multifactorial, because a lower incidence of smoking and decreased glucose, cholesterol, and blood pressure levels following treatment with levodopa may exert a protective influence. Furthermore, according to previous studies, the relationship of the WMH subtype with cortical atrophy is inconsistent and appears to not be specific to the disease entity [14,23,31]. Therefore, further study is needed to resolve whether the relationship between DWMHs and cortical thinning is relatively specific to PD pathology. Surmounting evidence suggests that the burden of WMHs may negatively impact cognitive performance, and that they are a major risk factor for ongoing cognitive dysfunction in patients with PD

Fig. 1. Statistical maps of cortical thickness according to the locations of white matter hyperintensities (WMHs) in patients with Parkinson's disease. After adjusting for age, gender, years of education, disease duration, and ICV, patients with severe total WMHs (TWMH) exhibited cortical thinning in the dorsolateral prefrontal cortex, orbitofrontal cortex, anterior cingulate cortex, left anterior temporal area, and the right superior and medial temporal areas compared with patients with low-grade TWMH (A). The pattern

of cortical thinning in patients with severe deep WMHs (DWMH) relative to those with low-grade DWMH was similar to that of patients with high-grade TWMH (B). In patients with periventricular WMHs (PWMH), cortical thinning in the high-grade group relative to the low-grade group was localized into the dorsolateral prefrontal cortex (C). The colouring of the surface corresponds to the mean thickness in mm. The color scale bar shows the difference in mean cortical thickness between high-grade and lowgrade groups, with red indicating more severe cortical thinning. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

J.H. Ham et al. / Parkinsonism and Related Disorders 21 (2015) 372e377

[13]. The present study demonstrated in detail the specific influence of WMH subtypes on cognitive performance and the association of WMHs with decreasing cortical thickness. Interestingly, DWMH-associated mean cortical thickness exhibited a significant correlation with frontal lobe-based cognition, even though this relationship was dependent on the type of WMH. Furthermore, the present study showed that frontal lobe thickness related to DWMHs was significantly correlated with frontal executive dysfunction and was independent of DWMHs. These results suggest that DWMHs in patients with PD may primarily induce cognitive impairments in association with cortical thinning in frontal areas. Accordingly, in addition to the fact that DWMHs interrupt frontosubcortical circuits and/or cholinergic projections from the basal forebrain, the present data suggest that frontal atrophy is an independent contributor to executive dysfunction in patients with PD with high-grade WMHs. There are several limitations in the present study that should be addressed. First, this is a retrospective study from a single hospital with a relatively limited sample size, which may have influenced the determination of the study population. Second, although the scoring method used for WMHs has been widely used and was performed by blinded investigators, this visual scaling method is less objective than a volumetric analysis and less sensitive to detect microstructural abnormalities of white matter than a diffusion tensor imaging. Third, this was a cross-sectional study and, thus, the current findings regarding DWMH-associated cortical thickness as a predictor of ongoing cognitive decline should be interpreted cautiously in clinical situations. Finally, we did not include non-PD subjects as a control group and thus, we could not conclude that the topographic pattern of cortical thinning related to WMH is specific to PD. A further study including control subjects and other neurodegenerative diseases is needed to clarify this issue. In summary, the current data demonstrated that DWMHs in patients with PD primarily affected cognitive impairments in association with cortical thinning in frontal areas. These data suggest that the modulation of risk factors responsible for DWMHs is an important treatment strategy to prevent cognitive dysfunction in patients with PD. Disclosures of conflicts of interest All authors declare no conflicts of interest. Acknowledgement This study was supported by a grant of the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare and Family Affairs, Republic of Korea (A121942). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.parkreldis.2015.01.015. References [1] Aarsland D, Andersen K, Larsen JP, Lolk A, Nielsen H, Kragh-Sorensen P. Risk of dementia in Parkinson's disease: a community-based, prospective study. Neurology 2001;56(6):730e6. [2] Aarsland D, Bronnick K, Larsen JP, Tysnes OB, Alves G. Cognitive impairment in incident, untreated Parkinson disease: the Norwegian ParkWest study. Neurology 2009;72(13):1121e6. [3] Emre M. Dementia associated with Parkinson's disease. Lancet Neurol 2003;2(4):229e37. [4] Compta Y, Parkkinen L, O'Sullivan SS, Vandrovcova J, Holton JL, Collins C, et al. Lewy- and Alzheimer-type pathologies in Parkinson's disease dementia: which is more important? Brain 2011;134(Pt 5):1493e505.

377

[5] Prasad K, Wiryasaputra L, Ng A, Kandiah N. White matter disease independently predicts progression from mild cognitive impairment to Alzheimer's disease in a clinic cohort. Dement Geriatr Cogn Disord 2011;31(6):431e4. [6] Fazekas F, Chawluk JB, Alavi A, Hurtig HI, Zimmerman RA. MR signal abnormalities at 1.5 T in Alzheimer's dementia and normal aging. AJR Am J Roentgenol 1987;149(2):351e6. [7] Bombois S, Debette S, Delbeuck X, Bruandet A, Lepoittevin S, Delmaire C, et al. Prevalence of subcortical vascular lesions and association with executive function in mild cognitive impairment subtypes. Stroke 2007;38(9):2595e7. [8] van den Heuvel DM, ten Dam VH, de Craen AJ, Admiraal-Behloul F, Olofsen H, Bollen EL, et al. Increase in periventricular white matter hyperintensities parallels decline in mental processing speed in a non-demented elderly population. J Neurol Neurosurg Psychiatry 2006;77(2):149e53. [9] Delano-Wood L, Abeles N, Sacco JM, Wierenga CE, Horne NR, Bozoki A. Regional white matter pathology in mild cognitive impairment: differential influence of lesion type on neuropsychological functioning. Stroke 2008;39(3):794e9. [10] Soriano-Raya JJ, Miralbell J, Lopez-Cancio E, Bargallo N, Arenillas JF, Barrios M, et al. Deep versus periventricular white matter lesions and cognitive function in a community sample of middle-aged participants. J Int Neuropsychol Soc 2012;18(5):874e85. [11] Lee SJ, Kim JS, Yoo JY, Song IU, Kim BS, Jung SL, et al. Influence of white matter hyperintensities on the cognition of patients with Parkinson disease. Alzheimer Dis Assoc Disord 2010;24(3):227e33. [12] Beyer MK, Aarsland D, Greve OJ, Larsen JP. Visual rating of white matter hyperintensities in Parkinson's disease. Mov Disord 2006;21(2):223e9. [13] Slawek J, Wieczorek D, Derejko M, Dubaniewicz M, Brockhuis B, Sitek E, et al. The influence of vascular risk factors and white matter hyperintensities on the degree of cognitive impairment in Parkinson's disease. Neurol Neurochir Pol 2008;42(6):505e12. [14] Ha SY, Youn YC, Kim S, Hsiung GY, Ahn SW, Shin HW, et al. A voxel-based morphometric study of cortical gray matter volume changes in Alzheimer's disease with white matter hyperintensities. J Clin Neurosci 2012;19(11): 1506e10. [15] Seo SW, Ahn J, Yoon U, Im K, Lee J, Tae Kim S, et al. Cortical thinning in vascular mild cognitive impairment and vascular dementia of subcortical type. J Neuroimaging 2010;20(1):37e45. [16] Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992;55(3):181e4. [17] Kang Y, Na D. Seoul neuropsychological Screening battery. Incheon: Republic of Korea Human Brain Research & Consulting Corp; 2003. [18] Scheltens P, Barkhof F, Leys D, Pruvo JP, Nauta JJ, Vermersch P, et al. A semiquantative rating scale for the assessment of signal hyperintensities on magnetic resonance imaging. J Neurol Sci 1993;114(1):7e12. [19] Collins DL, Neelin P, Peters TM, Evans AC. Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space. J Comput Assist Tomogr 1994;18(2):192e205. [20] MacDonald D, Kabani N, Avis D, Evans AC. Automated 3-D extraction of inner and outer surfaces of cerebral cortex from MRI. Neuroimage 2000;12(3): 340e56. [21] Robbins S, Evans AC, Collins DL, Whitesides S. Tuning and comparing spatial normalization methods. Med Image Anal 2004;8(3):311e23. [22] Im K, Lee JM, Yoon U, Shin YW, Hong SB, Kim IY, et al. Fractal dimension in human cortical surface: multiple regression analysis with cortical thickness, sulcal depth, and folding area. Hum Brain Mapp 2006;27(12):994e1003. [23] Wen W, Sachdev PS, Chen X, Anstey K. Gray matter reduction is correlated with white matter hyperintensity volume: a voxel-based morphometric study in a large epidemiological sample. Neuroimage 2006;29(4):1031e9. [24] Jones TA, Schallert T. Subcortical deterioration after cortical damage: effects of diazepam and relation to recovery of function. Behav Brain Res 1992;51(1): 1e13. [25] Cowey A, Alexander I, Stoerig P. Transneuronal retrograde degeneration of retinal ganglion cells and optic tract in hemianopic monkeys and humans. Brain 2011;134(Pt 7):2149e57. [26] Kim KW, MacFall JR, Payne ME. Classification of white matter lesions on magnetic resonance imaging in elderly persons. Biol Psychiatry 2008;64(4): 273e80. [27] Fazekas F, Kleinert R, Offenbacher H, Schmidt R, Kleinert G, Payer F, et al. Pathologic correlates of incidental MRI white matter signal hyperintensities. Neurology 1993;43(9):1683e9. [28] de Leeuw FE, de Groot JC, Bots ML, Witteman JC, Oudkerk M, Hofman A, et al. Carotid atherosclerosis and cerebral white matter lesions in a population based magnetic resonance imaging study. J Neurol 2000;247(4):291e6. [29] Sachdev P, Parslow R, Salonikas C, Lux O, Wen W, Kumar R, et al. Homocysteine and the brain in midadult life: evidence for an increased risk of leukoaraiosis in men. Arch Neurol 2004;61(9):1369e76. [30] Unal-Cevik I, Gursoy-Ozdemir Y, Yemisci M, Lule S, Gurer G, Can A, et al. Alpha-synuclein aggregation induced by brief ischemia negatively impacts neuronal survival in vivo: a study in [A30P]alpha-synuclein transgenic mouse. J Cereb Blood Flow Metab 2011;31(3):913e23. [31] DeCarli C, Fletcher E, Ramey V, Harvey D, Jagust WJ. Anatomical mapping of white matter hyperintensities (WMH) exploring the relationships between periventricular WMH, deep WMH, and total WMH burden. Stroke 2005;36(1): 50e5.