Education, white matter hyperintensities, and cognition in diverse cohorts of older adults

Abstract

AbstractBackgroundHigher education predicts better late‐life cognition among older adults, but mechanisms are poorly understood. We examined whether education modified the pathway between vascular injury and late‐life cognition.MethodAnalyses pooled imaging data from the multiethnic Kaiser Healthy Aging and Diverse Life Experience Study (KHANDLE; n=219) and the Study of Healthy Aging in African Americans (STAR; n=132), two parallel studies of older adult long‐term Kaiser Permanente members. Self‐reported education was considered as continuous years of completed education. White matter hyperintensities (WMH) were obtained through 3T MRI. Cerebral WMH volume to total cerebral volume ratio was log transformed and regressed on age; residuals were z‐standardized for measure of age‐adjusted WMH volume. Cognitive domains of verbal episodic memory (VRMEM), semantic memory (SEM), and executive function (EF) were measured using the Spanish and English Neuropsychological Assessment Scales (SENAS) and z‐standardized. Domain‐stratified regression models examined (1) independent effect of WMH adjusting for education, and (2) effect modification by education on the relationship between WMH and cognition. All models were adjusted for age, race/ethnicity, gender, and study.ResultParticipants were on average 73.2 years of age with 14.7 years of education, with 57.8% of the sample identifying as female and 49% of the sample identifying as Black. Education was significantly associated with all late‐life cognitive domains but for WMH volume (B: ‐0.02, 95% CI: ‐0.06, 0.02, p‐value: 0.38). In models adjusting for education and all covariates, effect of WMH on cognition was largest for EF (B: ‐0.09, 95% CI: ‐0.17, ‐0.01, p‐value: 0.03) followed by VRMEM (B: ‐0.06, 95% CI: ‐0.15, 0.03, p‐value: 0.18) and SEM (B: 0.00, 95% CI: ‐0.08, 0.08, p‐value: 0.97). Effect modification was suggested for VRMEM (WMH x education B: ‐0.03, 95% CI: ‐0.06, 0.00, p‐value: 0.09) but not EF (WMH x education B: ‐0.01, 95% CI: ‐0.04, 0.03, p‐value: 0.74) or SEM (WMH x education B: 0.00, 95% CI: ‐0.04, 0.03, p‐value: 0.79).ConclusionThe relationship between WMH and late‐life cognition differed by cognitive domain with the strongest effect of WMH independent of education for executive functioning, and effect modification by education for the relationship between WMH verbal episodic memory.