Associations of aortic and carotid artery health with cerebrovascular markers and cognition in older adults from the Whitehall II imaging study

Abstract

Background: Cardiovascular disease has been associated with an increased dementia risk, but the underlying mechanisms for this heart-brain link are unclear. This study sought to examine associations between aortic and carotid artery structure with cerebrovascular reactivity (CVR), white matter hyperintensities (WMHs), and cognition in later-life.

Methods: One hundred sixty three participants (25.8% female) from the Whitehall II Imaging cohort completed two examinations (M ± SD age 68.2 ± 4.4 at Wave-1 and 76.9 ± 4.5 at Wave-2) of neuropsychological assessments and 3T brain magnetic resonance imaging (MRI) FLAIR scans to quantify WMHs. Wave-2 additionally included vascular sonography of the aorta and carotid artery, and 3T functional MRI scans to measure CVR (mean % change BOLD signal change during a CO₂ challenge). Wave-2 factor scores of aortic and carotid arterial diameters, stiffness, and compliance were the exposure variables. Midlife Framingham Cardiovascular Risk Score (FRS) measured before Wave-1 was a potential effect modifier. WMH volume, grey matter CVR, cognitive factor scores (episodic memory, working memory, executive function, visuospatial memory, fluency, lexical retrieval) at Wave-2, and changes in WMH and cognition between Wave-1 and Wave-2 were used as outcome variables.

Results: Larger aortic diameter (ß = 0.38, SE = 0.11) and greater aortic stiffness (ß = 0.27, SE = 0.10) were associated with larger carotid diameter, independently of body size. Higher midlife FRS was associated with larger aortic and carotid diameters and increased carotid stiffness in old age. We observed notable artery-brain associations, such that larger aortic (ß = 0.17, SE = 0.06) and carotid diameters (ß = 0.11, SE = 0.05) were associated with larger WMH lesion volumes at Wave-2. Larger aortic diameter (ß = 0.08 SE = 0.03) and lower carotid compliance (ß = - 0.06, SE = 0.02) at Wave-2 were also associated with greater longitudinal increases in WMH volumes over the preceding 9 years. Higher stiffness and lower compliance of the aorta and carotid were associated with worse cognitive outcomes across a range of domains, and these associations were moderated by midlife FRS. Larger carotid diameter was associated with higher cerebrovascular reactivity (ß = 0.02, SE = 0.01), suggesting a potential compensatory pathway.

Conclusions: Adverse structural and functional changes in the aorta and carotid artery were inter-related and associated with vascular brain lesions, cerebrovascular reactivity, and poorer cognitive outcomes in older age.

Keywords: Aorta; Cardiovascular risk factors; Carotid; Cognition; Dementia; Longitudinal cohort; Magnetic resonance imaging; Ultrasound.

Fig. 1

Association of aorta and carotid phenotypes with midlife Framingham risk (blue: low risk group, orange: moderate-high risk). *indicates p_corr < 0.05, corrected for multiple comparisons

Fig. 2

Statistically significant associations between aorta structure with A white matter hyperintensities (WMHs) at MRI-Wave 2, B change in WMHs, and C lexical retrieval at Wave-2. Low and moderate-high midlife FRS are represented in blue and orange, respectively and there were no moderating effects of FRS on these associations. Solid lines represent regression lines with 95% confidence intervals. Black dashed lines represent mean regression lines across cardiovascular risk groups

Fig. 3

Statistically significant moderating effects of FRS on the association between A aortic compliance and change in episodic memory and B aortic stiffness and change in working memory. Low and moderate-high midlife FRS are represented in blue and orange, respectively. Solid lines represent regression lines with 95% confidence intervals

Fig. 4

Statistically significant associations between carotid measures and MRI (A–E) and neuropsychological measures (F–H). Low and moderate-high midlife FRS are represented in blue and orange, respectively. Solid lines represent regression lines with 95% confidence intervals. Black dashed lines represent mean regression lines across cardiovascular risk groups. For Delta Fluency (H): more negative values correspond to greater longitudinal decline

Fig. 5

Statistically significant moderating effects of FRS on the association between A carotid compliance and lexical retrieval at follow-up, B carotid stiffness and fluency at follow-up, C carotid diameter and longitudinal change in visuospatial memory, and D carotid stiffness and longitudinal change in fluency scores. Low and moderate-high midlife FRS are represented in blue and orange, respectively. For Δ Visuospatial (C): more positive values correspond to greater longitudinal decline, whereas for Δ Fluency (D): more negative values correspond to greater longitudinal decline. Solid lines represent regression lines with 95% confidence intervals