Louis ED, Brickman AM, DeCarli C, Small SA, Marder K, Schupf N, Brown TR. Quantitative Brain Measurements in Community-Dwelling Elderly Persons With Mild Parkinsonian Signs. Arch Neurol. 2008;65(12):1649-1654. doi:10.1001/archneurol.2008.504
Copyright 2008 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2008
Mild parkinsonian signs (MPS) are a marker of incident dementia. They have been linked with cerebrovascular disease, which can be evaluated using magnetic resonance imaging (MRI). Also, if MPS are a marker for developing Alzheimer-type changes, hippocampal volume on MRI might be diminished in individuals with MPS.
To examine white matter hyperintensity (WMH) volume and total hippocampal volume in elderly individuals with and without MPS.
Community-dwelling elderly persons in northern Manhattan (New York), New York, underwent neurologic examination and brain MRI. The WMH volume (derived from fluid-attenuated inversion recovery–weighted MRIs using a semiautomated thresholding approach) and total hippocampal volume (derived manually) were expressed relative to total cranial volume.
Mild parkinsonian signs were present in 111 of 666 participants (16.7%). Mean (SD) relative WMH volume was larger in participants with MPS vs those without MPS (1.70 [1.28] vs 1.17 [1.18]; P < .001). In a multivariate logistic regression analysis adjusting for age, sex, race/ethnicity, years of educational achievement, and depression, relative WMH volume was associated with MPS (odds ratio, 1.26; 95% confidence interval, 1.08-1.47; P = .004). In both unadjusted and adjusted analyses, total relative hippocampal volume was similar in participants with MPS vs those without MPS regardless of cognitive status.
In this MRI study of community-dwelling elderly persons, WMH volume was associated with MPS and total relative hippocampal volume was not. These data raise the possibility that vascular disease could have a role in the development of MPS.
Signs of tremor, rigidity, and bradykinesia are commonly found during the clinical examination of older persons without known neurologic disease.1- 4 It is unclear whether the emergence of these mild parkinsonian signs (MPS) reflects the development of vascular disease in the brain, the early development of degenerative disorders in the brain (eg, Alzheimer disease [AD] or the presence of Lewy bodies), or an age-associated decline in nigrostriatal dopaminergic activity.5
There is a growing body of literature linking MPS to vascular risk factors and vascular disease.5 In several studies, severity of white matter hyperintensity (WMH) has been linked to impaired gait and balance in elderly persons,6,7 although MPS per se have not been studied in detail.6 If MPS are a marker for vascular disease, one could hypothesize that WMH volume would be larger in individuals with MPS.
Hippocampal volume is diminished in patients with early AD.8,9 Pathologic changes in the hippocampus are among the earliest manifestations of AD.10 Mild parkinsonian signs are associated with mild cognitive impairment (MCI)11 and are a marker for the subsequent development of incident dementia.2,12 If MPS are indeed a marker for developing AD-type changes in the brain, one could hypothesize that hippocampal volume would be smaller in individuals with MPS.
The present study was conducted in community-dwelling older persons living in Washington Heights–Inwood, northern Manhattan (New York), New York. In 2003, we began systematically acquiring structural magnetic resonance (MR) imaging (MRI) in active participants without dementia in the study cohort. The purpose of the present analyses was to test the hypotheses that (1) WMH volume is larger in elderly persons with MPS than in their counterparts without MPS and (2) total hippocampal volume is reduced in elderly persons with MPS compared with their counterparts without MPS.
A total of 2776 individuals participated in a prospective study of aging and dementia in Medicare-eligible residents of northern Manhattan aged 65 years or older (Washington/Hamilton Heights-Inwood Columbia Aging Project II), as described elsewhere.13,14 Recruitment, informed consent, and study procedures were approved by the Columbia University Institutional Review Board.
The imaging protocol was concurrent with the second follow-up visit of the Washington/Hamilton Heights-Inwood Columbia Aging Project II cohort (n = 2113).14 Magnetic resonance imaging was performed in 769 participants; eligibility and participation have been described elsewhere.14 We further excluded 50 participants with incomplete MRI data, 19 with incomplete neurologic examination data, and 34 with incomplete neuropsychologic test data. Of the remaining 666 participants in the final sample, none had Parkinson disease or Parkinson-plus syndrome, as have been defined previously.11
A trained research assistant collected demographic information and administered a structured interview of health history (eg, hypertension, stroke, diabetes mellitus, each by self-report). Each participant also underwent a standardized neurologic examination that included an abbreviated (10-item) version of the Motor Examination portion of the Unified Parkinson Disease Rating Scale (UPDRS),11 which comprises evaluations of speech, facial expression, tremor at rest (in any body region), rigidity (rated separately in the neck, right arm, left arm, right leg, and left leg), changes in posture, and body (axial) bradykinesia.11 Each of the 10 items was rated on a scale of 0 to 4. A parkinsonian sign score (range: 0, no parkinsonian signs, to 40, maximum score) was calculated for each participant. The general medical physicians who administered this modified Motor Examination portion of the UPDRS were trained using a structured protocol.8 Interrater reliability of ratings was adequate (weighted κ statistics for ratings of speech, facial expression, tremor at rest, posture, and axial bradykinesia, 0.65-0.90), and agreement (percentage of concordance) with a movement disorder neurologist's (E.D.L.) ratings was 79%.11
As in previous analyses,11 MPS were defined as present when any of the following conditions was met: (1) 2 or more UPDRS ratings of 1, (2) 1 UPDRS rating of 2 or higher, or (3) UPDRS rest tremor rating of 1. The MPS were stratified into 3 subtypes (axial function, ie, changes in speech, facial expression, or posture, and axial bradykinesia; rigidity; and tremor) on the basis of factor analysis.15 An abnormality in axial function was considered present when there was either a UPDRS rating of 1 in 2 axial function items or 1 UPDRS rating of 2 or higher. Rigidity was considered present if there was either a UPDRS rating of 1 in 2 or more rigidity items or 1 UPDRS rating of 2 or higher. Tremor was considered present when there was a UPDRS rest tremor rating of 1.11
As previously described,16 all participants underwent a standardized neuropsychologic test battery17 and were considered to have dementia if they met established criteria.18 Participants with MCI were stratified into 2 groups: those with isolated impairment in memory or impairment in memory and at least 1 other cognitive domain (MCI + M) and those with no impairment in memory but impairment in at least 1 other cognitive domain (MCI − M). Depressive symptoms were assessed and rated with a 9-item version of the Center for Epidemiological Studies Depression Scale.19 On the basis of a previously established cut point,20 a score of 4 or higher was coded as presence of depression.
The MRIs were acquired using an Intera 1.5-T scanner (Philips Electronics North America Corp, Andover, Massachusetts) at Columbia University Medical Center and transferred electronically to the University of California at Davis for morphometric analysis in the Imaging of Dementia and Aging Laboratory.14 T1-weighted images (repetition time, 20 milliseconds; echo time, 2.1 milliseconds; field of view, 240 cm; 256 × 160 matrix with 1.3-mm section thickness) acquired in the axial plane and resectioned coronally were used to quantify hippocampal volume (repetition time, 20 milliseconds; echo time, 2.1 milliseconds; field of view, 240 cm; 256 × 160 matrix with 1.3-mm section thickness). For measures of total cranial volume and WMH volume, fluid-attenuated inverse recovery–weighted images (repetition time, 11 000 milliseconds; echo time, 144 milliseconds; inversion time, 2800 milliseconds; field of view, 25 cm; number of excitations, 2; 256 × 192 matrix with 3-mm section thickness) were acquired in the axial orientation.
User-operated image analysis was performed using an Ultra 5 Workstation (Sun Microsystems, Santa Clara, California) and Quantum 6.2 software. Hippocampal volume was measured using manual tracing as described previously.14 Intrarater reliability determined for both the right and the left hippocampus using this method was good, with intraclass correlation coefficients of 0.98 for the right hippocampus and 0.96 for the left hippocampus. For the present analyses, we reported total relative hippocampal volume (ie, total hippocampal volume expressed as a percentage of total cranial volume). In this data set, delayed recall21 was associated with total relative hippocampal volume even after adjusting for age (β = 5.20; P = .003).
Total cranial and WMH volumes were derived from fluid-attenuated inverse recovery–weighted images using a 2-step process as previously described.14,22,23 An operator manually traced the dura mater within the cranial vault, including the middle cranial fossa but not the posterior fossa and cerebellum. Total intracranial volume was defined as the number of voxels contained within the manual tracings multiplied by the voxel dimensions and section thickness. Nonuniformities in image intensity were removed,24 and 2 gaussian probability functions, representing brain matter and cerebrospinal fluid, were fitted to the skull-stripped image. After brain matter was isolated, a single gaussian distribution was fitted to image data and a segmentation threshold for WMH was set a priori at 3.5 SDs in pixel intensity above the mean of the fitted distribution of brain matter. Erosion of 2 exterior image pixels was applied to the brain matter image before modeling to remove partial volume effects and ventricular ependyma at WMH determination. The WMH volume was calculated as the sum of the remaining voxels multiplied by the voxel dimensions and section thickness. For the present analyses, we reported total relative WMH volume (ie, total WMH volume expressed as a percentage of total cranial volume).
As previously described,25 the presence or absence of cerebral infarction on MRI was determined for all subjects from the size, location, and imaging characteristics of the lesion. Only lesions 3 mm or larger qualified for classification as cerebral infarcts.
Analyses were performed using commercially available software (SPSS version 15.0; SPSS, Inc, Chicago, Illinois). In participants with MPS, quartiles were established on the basis of the parkinsonian sign score: no MPS; MPS lowest quartile, parkinsonian sign score 1; MPS second quartile, parkinsonian sign score 2; MPS third quartile, parkinsonian sign score between 3 and 6; or MPS highest quartile, parkinsonian sign score of 7 or higher. At regression analyses, we began with unadjusted models and then considered a number of potential confounders if they were associated with the main variables in univariate analyses. Because relative WMH volume was not normally distributed, univariate tests were nonparametric (eg, Mann-Whitney test), and this variable was log transformed in linear regression analyses.
There were 666 participants. The parkinsonian sign score was 1 or higher in 154 participants (23.1%), and 111 participants (16.7%) had MPS. The MPS were associated with older age, race/ethnicity, fewer years of educational achievement, and higher prevalence of self-reported stroke (Table 1). Dementia was more prevalent in participants with MPS vs those without MPS (P < .001; Table 1). Nineteen of 45 participants with dementia (42.2%) vs 73 of 466 participants with normal cognition (ie, without dementia or MCI) (15.7%) had MPS (χ2 = 19.61; P < .001). The 666 participants included 466 (70.0%) with normal cognition, 68 (10.2%) with MCI − M, 87 (13.1%) with MCI + M, and 45 (6.7%) with dementia. Mean (SD) time between the clinical examination and MRI was short (1.8 [4.9] months; median, 0 months) and was not correlated with relative WMH volume or total relative hippocampal volume.
Correlates of relative WMH volume are given in Table 2. In a multivariate linear regression analysis, log relative WMH volume was associated with age (β = .04; P < .001) and race/ethnicity (β = .16; P = .002) but not with sex, years of educational achievement, or depression. Mean (SD) relative WMH volume was larger in participants with dementia compared with those with normal cognition (1.75 [1.41] vs 1.15 [1.09]; Mann-Whitney z = 3.44; P = .001).
Relative WMH volume was larger in participants with MPS vs those without MPS (1.70 [1.28]; n = 111 vs 1.17 [1.18]; n = 555; Mann-Whitney z = 4.60; P < .001). When we excluded 69 participants who reported having had a stroke, relative WMH volume remained larger in participants with MPS vs those without MPS (1.63 [1.25] vs 1.16 [1.17]; Mann-Whitney z = 3.90; P < .001). When we further excluded 171 participants in whom MRI demonstrated evidence of stroke, this difference persisted (1.32 [1.15] vs 0.99 [0.90]; Mann-Whitney z = 2.04; P = .04). In a multivariate logistic regression analysis that included all participants and adjusted for age, sex, race/ethnicity, years of educational achievement, and depression, relative WMH volume was associated with MPS (dependent variable; odds ratio [OR], 1.26; 95% confidence interval [CI], 1.08-1.47; P = .004).
In each of the 4 cognitive status strata, relative WMH volume was larger in participants with MPS vs those without MPS, although, in both unadjusted and adjusted analyses, the difference reached significance only in participants with normal cognition (OR, 1.27; 95% CI, 1.03-1.55; P = .02; Table 3). We hypothesized that basal ganglia infarction might be associated with MPS. However, the percentage of participants with MPS who had basal ganglia infarcts at MRI was not significantly higher than the percentage of participants without MPS who had such infarcts (9 [8.1%] vs 29 [5.2%]; χ2 = 1.43; P = .23).
In the 466 participants with normal cognition, MPS were further stratified into quartiles. Relative WMH volume was associated with MPS quartile: 1.08 (1.07), no MPS; 1.30 (0.98), lowest MPS quartile; 1.40 (1.26), second MPS quartile; 1.41 (1.12), third MPS quartile; and 2.28 (1.14), highest MPS quartile (Kruskal-Wallis χ2 = 19.19; P = .001). At linear regression analysis, MPS quartile (independent variable) was associated with log relative WMH volume (β = .17; P < .001). In the 466 participants with normal cognition, MPS were stratified into 3 subtypes: axial function, rigidity, and tremor. Relative WMH volume was larger in participants with vs those without axial dysfunction (1.50 [1.15] vs 1.12 [1.08]; Mann-Whitney z = 2.40; P = .02), rigidity (1.53 [1.10] vs 1.11 [1.08]; Mann-Whitney z = 2.98; P = .003), and tremor (1.64 [1.17] vs 1.13 [1.08]; Mann-Whitney z = 2.45; P = .01). In a multivariate logistic regression analysis adjusting for age, sex, race/ethnicity, years of educational achievement, and depression, the association between relative WMH volume and MPS subtype (dependent variable) was similar for each MPS subtype (axial function: OR, 1.19; 95% CI, 0.91-1.57; P = .21; rigidity: 1.24; 0.98-1.56; P = .07; tremor: 1.30; 0.96-1.76; P = .09).
The correlates of total relative hippocampal volume are given in Table 2. In a multivariate linear regression analysis, total relative hippocampal volume was not associated with age, sex, race/ethnicity, years of educational achievement, or depression. Mean (SD) total relative hippocampal volume was smaller in the 45 participants with dementia vs the 466 participants with normal cognition (0.25 [0.06] vs 0.29 [0.06]; t = 4.14; P < .001).
In each of the 4 cognitive status strata, in both unadjusted analyses and in adjusted logistic regression analyses, total relative hippocampal volume was similar in participants with MPS vs those without MPS (Table 3). In each of the 4 cognitive status strata, there was no correlation between total relative hippocampal volume and parkinsonian sign score: normal cognition: r = 0.06, P = .22; MCI − M: r = 0.05, P = .67; MCI + M: r = −0.10, P = .38; and dementia: r = 0.13, P = .39.
In the 466 participants with normal cognition, MPS were further stratified into quartiles. There was no association between MPS quartile and total relative hippocampal volume (mean [SD]): no MPS, 0.29 (0.06); lowest MPS quartile, 0.28 (0.07); second MPS quartile, 0.30 (0.06); third MPS quartile, 0.29 (0.06); and highest MPS quartile, 0.31 (0.06) (analysis of variance: F = 0.49; P = .75). The MPS were stratified into 3 subtypes: axial function, rigidity, and tremor. Total relative hippocampal volume was similar in participants with vs those without axial dysfunction (0.29 [0.06] vs 0.29 [0.06]; t = 0.22; P = .83), rigidity (0.29 [0.06] vs 0.30 [0.06]; t = 0.63; P = .18), and tremor (0.29 [0.06] vs 0.30 [0.05]; t = 0.28; P = .78). In multivariate logistic regression analyses that adjusted for age, sex, race/ethnicity, years of educational achievement, depression, and stroke, total relative hippocampal volume was not associated with each MPS subtype (dependent variables in different models).
We found a robust association between larger relative WMH volume and presence of MPS. Furthermore, there was a dosing effect whereby stepwise increases in relative WMH volume were observed in higher MPS quartiles. Several epidemiologic studies in elderly persons with MPS have demonstrated associations between vascular risk factors or vascular disease and MPS.26- 28 In a sizable number of MRI studies, subcortical WMH has been linked to impaired gait and balance in elderly persons,6,7,29 although MPS per se and their relationship to these hyperintensities have not been studied in detail.6 Hence, the present study extends previous research by linking WMH to different kinds of motor dysfunction. Further along the spectrum of parkinsonism than MPS (ie, a more severe syndrome) is vascular parkinsonism, which also has vascular lesions as its underpinnings.30 Our MRI data further raise the possibility that vascular disease could have a role in the development of MPS. However, this conclusion should be approached with caution. While WMHs are associated with cerebrovascular mechanisms, one cannot necessarily equate the two.
The links between MPS in elderly persons without dementia and the development of AD are based on epidemiologic evidence. Longitudinal studies2,12,31 have demonstrated that individuals with MPS are more likely than their counterparts without MPS to develop dementia. Most of these individuals with MPS develop AD rather than other forms of dementia.2,12 Limited postmortem evidence also suggests a possible link between AD-type pathologic features and one manifestation of MPS, that is, gait disturbance. In a study of deceased persons without dementia in the Religious Orders Study,32 the number of neurofibrillary tangles in the substantia nigra was related to gait disturbance. To further examine the links between MPS and AD, we examined total relative hippocampal volume in community-dwelling elderly persons with MPS vs those without MPS. We did not find an association between total relative hippocampal volume and MPS or any of the MPS subtypes. This finding seems to argue against the notion that MPS themselves are a manifestation of early AD. However, these findings should be interpreted with caution because hippocampal volume is affected by factors other than AD, AD affects many brain regions other than the hippocampus, and the hippocampus is not generally thought to have a role in MPS.
We did not find an association between hippocampal volume and age. There is sizable literature on the relationship between hippocampal volume and age, with contradictory results; in some studies, there is an aging effect,33- 35 whereas, in others, there is not.36- 38
The strengths of the present study include the population-based design and the large sample size. This is also the first attempt, to our knowledge, to examine structural changes in the hippocampus in elderly persons without dementia with MPS. The study's limitations include the cross-sectional design, which prevented examination of associations between clinical course and longitudinal imaging changes. In addition, we did not perform postmortem examinations and were not able to correlate our findings with findings from microscopic pathologic studies. As reported previously,14 individuals who refused participation in the MRI study were a year older and more likely to be women and white, which may have introduced some biases. In our analyses, we adjusted for these demographic factors.
In conclusion, in this study of community-dwelling elderly persons, we noted that relative WMH volume is associated with MPS and total relative hippocampal volume is not. These data raise the possibility that vascular disease could have a role in the development of MPS.
Correspondence: Elan D. Louis, MD, Unit 198, Department of Neurology, Columbia University, 710 W 168th St, New York, NY 10032 (EDL2@columbia.edu).
Accepted for Publication: April 2, 2008.
Author Contributions:Study concept and design: Louis, Brickman, Small, Schupf, and Brown. Acquisition of data: DeCarli, Small, and Brown. Analysis and interpretation of data: Louis, Brickman, DeCarli, Small, Marder, Schupf, and Brown. Drafting of the manuscript: Louis. Critical revision of the manuscript for important intellectual content: Louis, Brickman, DeCarli, Small, Marder, Schupf, and Brown. Statistical analysis: Louis, Brickman, and Schupf. Obtained funding: Louis, Brickman, Small, Schupf, and Brown. Administrative, technical, and material support: Louis, DeCarli, Small, Schupf, and Brown. Study supervision: Louis.
Financial Disclosure: None reported.
Funding/Support: This study was supported by federal grants P01 AG07232 and R01 NS42859 (Dr Louis) from the National Institutes of Health.