The cross-sectional and longitudinal relationships between white matter hyperintensities and dementia in patients with Parkinson's disease: A retrospective analysis of 132 patients in a single center

Objective: To clarify the cross-sectional and longitudinal relationships between white matter hyper-intensities (WMH) and dementia in Parkinson's disease (PD) patients. Methods: One hundred thirty-two PD patients were included. Using medical records, the patient data including Hoehn and Yahr stage, postural instability, neuropsychological tests and magnetic resonance imaging were analyzed. The degree of WMH was rated according to a modified Fazekas scale. The relationship between the variables and dementia was analyzed using the independent t-test, the chisquare test, logistic regression analysis and the Cox proportional hazard model. Results: The mean age of the study patients (35 males and 97 females) was 71.6 years (range, 45-93 years). The baseline WMH was associated not only cross-sectionally with the contemporary prevalence of dementia but also longitudinally with subsequent occurrence of dementia in the univariate analysis. These relationships became attenuated and statistically insignificant in the multivariate analysis after adjusting for confounders. It was postural instability that consistently predicted dementia in both the cross-sectional and the longitudinal data. Conclusions: Our study showed that baseline WMH was not independently associated with dementia, and instead postural instability revealed at first examination can be a more reliable predictor of dementia in PD patients. (C) 2015 Elsevier Ireland Ltd. All rights reserved.