Combined dementia-risk biomarkers in Parkinson's disease: A prospective longitudinal study

Abstract

Neuropsychological (mostly posterior-cortical) deficits, quantitative magnetic resonance imaging (MRI) atrophy patterns, and low cerebrospinal fluid (CSF) levels of amyloid-β have been separately related to worsening cognition in Parkinson's disease (PD). However, these biomarkers have not been longitudinally assessed in combination as PD-dementia predictors. In this prospective longitudinal study, 27 non-demented PD patients underwent CSF, neuropsychological and 3-T brain-MRI studies at baseline and were re-assessed 18 months later in terms of progression to dementia (primary outcome) and longitudinal neuropsychological and cortical thickness changes (secondary outcomes). At follow-up 11 patients (41%) had progressed to dementia. Lower CSF amyloid-β, worse verbal learning, semantic fluency and visuoperceptual scores, and thinner superior-frontal/anterior cingulate and precentral regions were significant baseline dementia predictors in binary logistic regressions as quantitative and/or dichotomised traits. All participants without baseline biomarker abnormalities remained non-demented whereas all with abnormalities in each biomarker type progressed to dementia, with intermediate risk for those showing abnormalities in a single to two biomarker types (p = 0.006). Both the dementia-outcome and low baseline CSF amyloid-β were prospectively associated with limbic and posterior-cortical neuropsychological decline and frontal, limbic and posterior-cortical thinning from baseline to follow-up. These findings suggest that the combination of CSF amyloid-β, neuropsychological and cortical thickness biomarkers might provide a basis for dementia-risk stratification and progression monitoring in PD.

Introduction

Despite claims of inevitability of dementia in Parkinson's disease (PD), the rate to and the severity of PD-dementia are largely variable and some patients remain non-demented after long follow-up [1]. There are hints that specific mechanisms may account for such variability, with amyloid-β pathology [2] and posterior-cortical dysfunction [3] having been related to progression to dementia. The identification of risk biomarkers in PD would be relevant for better prognosis knowledge and to know whom to and when to test strategies aimed at preventing dementia. On the other hand, progression-biomarkers might become surrogate endpoints of response to any such strategies [4].

Neuropsychological cognitive assessment has been extensively explored in PD, with growing attention being paid to posterior-cortical deficits [3]. PD-dementia biomarkers may also come from MRI-quantification of brain atrophy, but longitudinal studies are scanty. In one such study using voxel-based morphometry (VBM), longitudinal grey matter volume reductions in limbic, paralimbic and posterior-cortical areas were seen in PD with and without dementia [5]. Another study has identified a validated Alzheimer-type atrophy MRI pattern as an independent predictor of cognitive impairment in PD [6]. Molecular biology constitutes another source of PD-dementia biomarkers, with low cerebrospinal fluid (CSF) levels of amyloid-β (Aβ) having been cross-sectionally associated with semantic fluency and memory deficits [7], [8] and longitudinally with decline in neuropsychological performance [9]. However, neither of the MRI or CSF longitudinal studies had progression to dementia as outcome.

With the current study we aimed to explore the relative and combined longitudinal CSF, neuropsychological and quantitative MRI associations with PD progression to dementia. We hypothesised that low CSF Aβ and limbic and posterior-cortical neuropsychological and MRI changes herald PD-dementia, and that progressive neuropsychological and structural alterations occur longitudinally.