WHITE MATTER HYPERINTENSITIES ARE NOT RELATED TO COGNITION IN OLDER-OLD PATIENTS

Abstract

Patients who had a cerebrovascular event are at higher risk of developing vascular dementia, especially in older age. However, some patients appear resilient to cognitive impairment even at advanced age. One possible explanation is that they have resistance to small vessel disease, particularly white matter hyperintensities (WMH) detectable on MRI. We studied the relevance of WMH to cognition in relation to age in patients who had a TIA/non-disabling stroke, particularly the older-old (>80). 570 consecutive patients with recent TIA/non-disabling stroke from the Oxford Vascular Study underwent multimodal MRI and cognitive assessment with the Montreal Cognitive Assessment scale (MoCA). They were divided into two age groups (<80, N=461; ≥80, N=109) and three cognitive groups according to their MoCA score: no vascular cognitive impairment (NoCI, MoCA>24, N=384), mild vascular cognitive impairment (MildVCI, MoCA=20–24, N=143) or severe vascular cognitive impairment (SevereVCI, MoCA<20, N=43). WMH were automatically segmented on FLAIR images using BIANCA (Griffanti et al., Neuroimage 2016) and the relative WMH volumes calculated. A two-way between groups factorial ANOVA was performed on the WMH volumes to test the interaction between age and cognition. Voxel-wise correlational analyses were then performed on the binarised, normalized, smoothed maps of WMH using non-parametric permutation test in FSL to explore if lower MoCA score would be associated with higher probability of WMH in different localization in the younger relative to the older-old patients. A 2x3 between-group factorial ANOVA on WMH volumes showed significant main effects of age (F=29.04, p<0.001) and cognitive group (F=5.71, p=0.004), and a significant age/cognition interaction (F=4.67,p=0.010), in that differences across cognitive groups were significant in the younger but not in the older-old group. Accordingly, the voxel-wise analysis showed a significant negative correlation (voxel-corrected p<0.05) between WMH probability and MoCA score in the younger group only, mainly in periventricular frontal areas (figure). WMH in patients aged ≥80 years with TIA/minor stroke is unrelated to cognition suggesting that the presence of WMH (i) does not undermine resilience to dementia in the older-old and (ii) should not prompt consideration of a diagnosis of vascular cognitive impairment in the older old. Voxel-wise analysis of WMH. Negative correlation between WMH and MoCA score for subjects < 80 years (red-yellow) and for older-old (> 80 years, blue-light blue)