Role of Social, Clinical, and Contextual Factors in Explaining MRI Markers of White Matter Injury and Change in Objectively‐Assessed Cognition by Race/Ethnicity: The Multi‐Ethnic Study of Atherosclerosis

Abstract

AbstractBackgroundRacial and ethnic disparities in Alzheimer’s disease and related dementias (ADRD) are well documented. The mechanisms through which race and ethnicity as markers of systemic racism are associated with brain health are lacking.MethodWe used data from the Multi‐Ethnic Study of Atherosclerosis (MESA) and a causal mediation framework to determine whether the association between race/ethnicity (measured in Exam 1, 2000‐2002) and brain health is mediated through pathways represented by social (education level, health insurance (yes/no)), clinical (hypertension, diabetes, and depression symptoms), and contextual factors (neighborhood socioeconomic status) measured in Exam 1. Brain health was measured by 3T magnetic resonance imaging (MRI, N = 966) and change in cognitive scores (N = 2002). Participants completed the Cognitive Abilities Screening Instrument (CASI), Digit Symbol Coding (DSC), and Digit Span (DS) cognitive tests in Exams 5 (2010‐2012) and 6 (2016‐2018). Participants underwent MRI in 2018‐2019 to assess white matter (WM) injury assessed by WM hyperintensity volume (WMHV) and fractional anisotropy (FA). All models were adjusted for age, gender, education, marital status, income, and for WMHV models, total intracranial volume.ResultMean (SD) age at baseline was 58 (9) years, 29% participants self‐reported Black, 17% Hispanic, 11% Chinese and 43% White. Black and Hispanic participants experienced a larger decrease in CASI (‐3.3 points (95%CI: ‐4.1, ‐2.5); ‐4.7 (‐5.7, ‐3.7) respectively), DSC (‐2.2 (‐3.6,‐0.8); ‐5.3 (‐7.0,‐3.6)), and DS scores (‐0.4 (‐0.7,‐0.2); ‐1.1 (‐1.4,‐0.8)) from Exam5 to 6 than White participants, after accounting for covariates and baseline cognitive score. Black participants had greater WMHV (21% (1%, 50%)), and lower FA (‐.006 (‐0.01,‐0.002)) than White participants. We observed that the association between race/ethnicity as markers of racism and CASI change was mediated by education (proportion mediated 13% (1%, 21%)), and that the association with DSC change was mediated by health insurance (proportion mediated 9% (4%, 24%)) but not by any other social, clinical or contextual factor. The association between race/ethnicity and WM injury markers and change in DS was not mediated by social, clinical nor contextual factors.ConclusionEducation and health insurance appear as factors that may partially explain the association between race/ethnicity and cognitive decline.