Brain Atrophy And White Matter Injury In Relation To Cardiovascular Risk Factors And Race/Ethnicity: The Multi-Ethnic Study Of Atherosclerosis

Introduction: Cardiovascular risk factors are associated with cognitive decline and dementia. Brain magnetic resonance imaging (MRI) provides sensitive measurement of cerebral atrophy and small vessel disease, reflecting multiple pathologies leading to dementia. However, large brain MRI studies include primarily white participants. We investigated associations in the diverse Multi-Ethnic Study of Atherosclerosis (MESA). Hypothesis: Cardiovascular and sociodemographic risk factors are associated with brain morphology and white matter injury in a racially and ethnically diverse population. Methods: In MESA, brain MRI was performed in 2018-2019 with automated measurement of total brain volume, gray and white matter (GM, WM) volume, and measures of WM injury including WM lesion volume, WM fractional anisotropy, and WM apparent diffusion coefficient. In cross-sectional analyses, we assessed the association of race/ethnicity with MRI measures, with and without adjustment for cardiovascular risk factors, education, and socioeconomic status. In a multivariable model, we assessed the association of cardiovascular risk factors with brain MRI measures. All analyses of volumes, including WM lesion volume, were adjusted for total intracranial volume. Results: MRI data were complete in 1,051 participants; 40% were white, 15% Chinese-American, 25% African-American, and 20% Hispanic. Mean (standard deviation, SD) age was 73 (8) years and 53% of participants were women. Adjusted for age and sex, African-American participants had slightly greater total brain and WM volume than white participants. Adjusted for age and sex, African-American participants had on average more WM injury than whites as measured by higher WM lesion volume (46.7% higher, 95% CI: 19.9, 79.4%) and lower fractional anisotropy (-0.20 SD, 95% CI: -0.34, -0.05); these associations were attenuated after additional adjustment for cardiovascular risk factors and socioeconomic status (24.3% higher WM lesion volume, 95% CI: 0.0, 54.3; -0.06 SD fractional anisotropy, 95% CI: -0.22, 0.09). Conversely, all non-white race/ethnic groups had slightly less WM injury than white participants as estimated by apparent diffusion coefficient. Overall, greater age, diabetes, current smoking, high systolic blood pressure, and treated hypertension were strongly associated with more WM injury; in addition, age and diabetes were strongly associated with lower brain volumes. Conclusions: We found little evidence of differences in measures of brain atrophy and WM injury by race/ethnicity after adjustment for cardiovascular risk factors and socioeconomic status. Findings of differences by race/ethnicity in apparent diffusion coefficient are intriguing and need further investigation. Consistent with previous studies, age, diabetes, current smoking and hypertension were strongly and consistently associated with WM injury.