Associations between vascular risk factors and brain MRI indices in UK Biobank

Abstract

Aims: Several factors are known to increase risk for cerebrovascular disease and dementia, but there is limited evidence on associations between multiple vascular risk factors (VRFs) and detailed aspects of brain macrostructure and microstructure in large community-dwelling populations across middle and older age.

Methods and results: Associations between VRFs (smoking, hypertension, pulse pressure, diabetes, hypercholesterolaemia, body mass index, and waist-hip ratio) and brain structural and diffusion MRI markers were examined in UK Biobank (N = 9722, age range 44-79 years). A larger number of VRFs was associated with greater brain atrophy, lower grey matter volume, and poorer white matter health. Effect sizes were small (brain structural R2 ≤1.8%). Higher aggregate vascular risk was related to multiple regional MRI hallmarks associated with dementia risk: lower frontal and temporal cortical volumes, lower subcortical volumes, higher white matter hyperintensity volumes, and poorer white matter microstructure in association and thalamic pathways. Smoking pack years, hypertension and diabetes showed the most consistent associations across all brain measures. Hypercholesterolaemia was not uniquely associated with any MRI marker.

Conclusion: Higher levels of VRFs were associated with poorer brain health across grey and white matter macrostructure and microstructure. Effects are mainly additive, converging upon frontal and temporal cortex, subcortical structures, and specific classes of white matter fibres. Though effect sizes were small, these results emphasize the vulnerability of brain health to vascular factors even in relatively healthy middle and older age, and the potential to partly ameliorate cognitive decline by addressing these malleable risk factors.

Keywords: Brain; Cortex; Diffusion; MRI; Vascular risk; White matter.

Figure 1

White matter tracts-of-interest (left panel) and subcortical structures (right panel) measured in the current study. AR, acoustic radiation; ATR, anterior thalamic radiation; Cing, cingulum (gyrus and parahippocampal); CST, corticospinal tract; Fmaj and Fmin (forceps major and minor); IFOF, inferior fronto-occipital fasciculus; ILF, inferior longitudinal fasciculus; MCP, middle cerebellar peduncle; ML, medial lemniscus; PTR, posterior thalamic radiation; SLF, superior longitudinal fasciculus; STR, superior thalamic radiation; Unc, uncinate fasciculus.

Figure 2

Associations between aggregate vascular risk and cortical volume (left panel), white matter tract-specific microstructure (centre panel showing right lateral and superior views), and subcortical volume (right panel showing, from top to bottom: anterior, lateral, and inferior views). Higher aggregate vascular risk is associated with significantly lower cortical volume, lower fractional anisotropy and higher mean diffusivity in the majority of white matter fibres, and lower subcortical volume, except for the amygdala (grey).

Figure 3

Significant associations (left: t-maps and right: FDR-corrected q-values) between cortical volume and vascular risk factors (modelled individually, alongside age, sex, ethnicity, head size, and scanner head position confounds). See Supplementary material online, Figure S3 for non-significant associations for pulse pressure and hypercholesterolaemia. T-maps are scaled with the same limits to aid comparison of relative effect size across risk factors.

Figure 4

Standardized betas of associations between individually modelled vascular risk factors and white matter tract fractional anisotropy (top panel), white matter tract mean diffusivity (centre panel), and subcortical volumes (bottom panel). Acc, accumbens; Amyg, amygdala; ATR, anterior thalamic radiation; BP, hypertension; Caud, caudate; Cing, cingulum (gyrus and parahippocampal); CST, corticospinal tract; Diab, diabetes; Fmaj and Fmin (forceps major and minor); HiChol, hypercholesterolaemia. AR, acoustic radiation; Hipp, hippocampus; IFOF, inferior fronto-occipital fasciculus; ILF, inferior longitudinal fasciculus; MCP, middle cerebellar peduncle; ML, medial lemniscus; Pall, pallidum; PP, pulse pressure; PTR, posterior thalamic radiation; Put, putamen; SLF, superior longitudinal fasciculus; STR, superior thalamic radiation; Thal, thalamus; Unc, uncinate fasciculus.