QRISK2 AND FRAMINGHAM CARDIOVASCULAR RISK SCORES SIGNIFICANTLY CORRELATE WITH IMAGING BIOMARKERS OF PRECLINICAL AD: FINDINGS FROM THE HARVARD AGING BRAIN STUDY

Abstract

Neuropathological studies suggest many cases of late onset dementia may be due to mixed pathologies, most commonly a combination of Alzheimer's disease (AD) pathology and cerebrovascular disease. The amount of white matter disease (WM hyperintensities, WMH) is often used as a proxy measure for the impact of vascular disease on the brain, but lacks the extensive validation of commonly used vascular risk (CV Risk) scores such as the QRISK2, Framingham Heart Study (FHS) cardiovascular disease (FHS-CVD) and stroke (FHS-Stroke) metrics. Accordingly, we examined the cross-sectional relationship of these CV risk metrics to commonly used imaging biomarkers, including 18F- fludeoxyglucose (FDG) PET, Pittsburgh Compound B (PiB) PET, hippocampal volume, ventricle size, and WMH. MRI, PiB PET, and FDG PET data from 195 clinically normal older participants in the Harvard Aging Brain Study (HABS) were compared to CV risk scores calculated from available demographic, blood pressure, and cholesterol data. Linear regression models (corrected for age and sex) were used to assess relationships between CV Risk and imaging biomarkers and with the Preclinical Alzheimer's Cognitive Composite (PACC). QRISK2 and FHS-CVD risk scores were significantly correlated with AD-typical hypometabolism on FDG PET (Figure 1). Higher QRISK2 was associated with lower hippocampal volume. FHS-CVD and FHS-Stroke risk scores (but not QRISK2) were significantly correlated with WMH. FHS-CVD and FHS-Stroke risk scores also correlated with lateral ventricle size, a measure of central atrophy. Higher FHS-CVD predicted significantly lower scores on the PACC. We observed no significant relationships between any CV risk scores and PET measures of amyloid burden. We observed significant (but modest) cross-sectional relationships between well-validated measures of CV risk and imaging biomarkers commonly used to stratify risk of decline in preclinical AD, but no relationship between elevated CV risk and amyloid burden. Though significant relationships between CV risk scores and WMH were present, these were also relatively weak, suggesting measures of WMH may have limited overlap with vascular risk as assessed in other medical contexts. These results support considering CV risk in studies of preclinical AD, and suggest WMH measures may not completely account for CV risk. Associations Between Cardiovascular Risk Scores and Cross-Sectional AD Biomarkers in the Harvard Aging Brain Study. The heatmap shown depicts partial correlations between cardiovascular risk scores and baseline imaging and cognitive measures, controlling forage and sex. (CVD = Cardiovascular Disease; CHD = Coronary Heart Disease; CVD Death and CHD Death refer to risk of death from CVD or CHD, respectively; *, **, and *** correspond to p < 0.05, < 0.005, and < 0.0005, respectively).