Imaging-based indices of Neuropathology and gait speed decline in older adults: the atherosclerosis risk in communities study

Abstract

Imaging markers of cerebrovascular disease and Alzheimer's disease (AD) are implicated in mobility impairment in older adults, but few studies have examined these relationships longitudinally in a racially-diverse population-based sample. At Visit 5 (2011-13) of the ARIC Study, 1859 participants had usual pace gait speed (cm/s) assessed and brain MRI (mean age = 76.3, 28.5% Black) and PET (n = 343; mean age = 75.9, 42.6% Black) measures including total/regional brain volume (cm³), white matter hyperintensities (WMH; cm³), infarcts (present/absent), microbleeds (count) and global beta-amyloid (Aβ). Participants returned at Visit 6 (n = 1264, 2016-17) and Visit 7 (n = 1108, 2018-19) for follow-up gait speed assessments. We used linear regression to estimate effects of baseline infarct presence, higher microbleed count, and a one interquartile range (IQR) poorer measures of continuous predictors (-1 IQR total brain volume, temporal-parietal lobe meta region of interest(ROI); +1 IQR WMH volume, global Aβ SUVR) on cross-sectional gait speed and change in gait speed adjusting for age, sex, education, study site, APOE e4, estimated intracranial volume, BMI, and cardiovascular risk factors. Cross-sectionally, slower gait speed outcome was associated with higher WMH volume, -3.38 cm/s (95%CI:-4.71, -2.04), infarct presence, -5.60 cm/s (-7.69, -3.51), microbleed count, -2.20 cm/s (-3.20, -0.91), smaller total brain volume, -9.26 cm/s (-12.1, -6.43), and smaller temporal-parietal lobe ROI -6.28 cm/s (-8.28, -4.28). Longitudinally, faster gait speed outcome decline was associated with higher WMH volume, -0.27 cm/s/year, (-0.51, -0.03) and higher global Aβ SUVR, -0.62 cm/s/year (-1.20, -0.03). Both cerebrovascular and AD pathology may contribute to mobility decline commonly seen with aging.

Keywords: Amyloid; Cerebrovascular disease; Neuroimaging; Physical function.

Fig. 1

Gait Speed (cm/s) and Gait Speed Change (cm/s/year) by White Matter Hyperintensity (WMH) Volume (cm³) and Global beta-amyloid (Aβ; SUVR) and by Race. Histograms display proportions of sample by race at given value of WMH Volume and Global Aβ

Fig. 2

Adjusted cross-sectional associations of Neuropathology with Gait Speed (cm/s) by Race. Estimates β (95% confidence interval) correspond to differences in gait speed (cm/s) associated with a 1 interquartile range (IQR) change in continuous predictors (+1 IQR in WMH Volume, Aβ SUVR; −1 IQR in Brain Volume, AD Region Volume) and presence (Infarcts, Aβ(+)) or count (Microbleeds, up to 3+) of categorical predictors. ROI = region of interest, WMH = White Matter Hyperintensity, Aβ = beta-amyloid, SUVR = Standardized Uptake Value Ratio. Estimates adjusted for age, sex, education, study site, APOE e4, BMI, hypertension, diabetes, coronary heart disease, alcohol consumption, statin use, heart failure, and estimated total intracranial volume

Fig. 3

Adjusted longitudinal associations of Neuropathology with Gait Speed Change by Race. Estimates β (95% confidence interval) correspond to gait speed change (cm/s/year) associated with a 1 interquartile range (IQR) change in continuous predictors (+1 IQR in WMH Volume, Aβ SUVR; −1 IQR in Brain Volume, AD Region Volume) and presence (Infarcts, Aβ(+)) or count (Microbleeds, up to 3+) of categorical predictors. ROI = region of interest, WMH = White Matter Hyperintensity, Aβ = beta-amyloid, SUVR = Standardized Uptake Value Ratio. Estimates adjusted for age, sex, education, study site, APOE e4, BMI, hypertension, diabetes, coronary heart disease, alcohol consumption, statin use, heart failure, and estimated total intracranial volume