Impact of white matter hyperintensities on domain-specific cognition in Southeast Asians

Abstract

Background: Dementia affects 55 million people worldwide, with 60% of the burden in Asia. White Matter Hyperintensities (WMH) is a key marker of small vessel disease, with high prevalence in Asian populations with prodromal and clinical dementia. WMH is described as: Deep White Matter Hyperintensities (DWMH), Periventricular Hyperintensities (PVH) or Fazekas-Total (DWMH and PVH).

However, association between WMH topography and performance in specific cognitive domains remains unexplored. Thus, this study aims to characterise the impact of Fazekas-Total, DWMH and PVH on different cognitive domains.

Method: 304 participants (mean age 60.6, mean education years 14.2, mean Montreal Cognitive Assessment 25.99, 43.8% males) from Biomarkers and Cognition Study, Singapore (Dementia Research Centre (Singapore)) met the inclusion criteria.

Eight domains of cognition were tested: global cognition, learning and memory, language, executive function, complex-attention, perceptual-motor, social-cognition, and processing speed.

Normality tests, correlation analysis and stepwise regression (with Benjamini and Hochberg False Discovery Rate correction) were performed to understand association between Fazekas-Total, DWMH, PVH on domains of cognition tested, accounting for age, education, and gender.

Two approaches were used for data analysis: (1) by grouping participants as Mild Cognitive Impairment (MCI), Subjective Cognitive Decline (SCD) and Cognitively Normal (CN) as per the National Institute on Aging-Alzheimer’s Association (NIA-AA) criteria and (2) by grouping participants’ WMH as confluent (DWMH ≥2 and PVH ≥3) and non-confluent.

Result: Higher Fazekas-Total was associated with slower processing speed (p = 0.0255, R = -0.039) in prodromal participants (MCI and SCD). Higher PVH was associated with slower processing speed in MCI (p = 0.017, R = -0.32) and CN (p = 0.017, R = -0.663) participants. Higher PVH was significantly associated with poorer learning and memory (p = 0.045, R = -0.043) in SCD participants. Higher DWMH was significantly associated with poorer learning and memory (p = 0.017, R = -0.765) in CN participants.

Conclusion These results demonstrate differential association between PVH, DWMH and cognitive domains, thus the location of WMH determines the affected cognitive domains. Therefore, it is worthwhile to assess PVH and DWMH in clinical cognitive outcomes separately and to understand the upstream pathobiology of DWMH and PVH.

Furthermore, higher Fazekas-Total was strongly associated with slower processing speed in prodromal participants and could be a marker of cognitive decline.