Abstract
AIMS: To evaluate whether diabetes and prediabetes are associated with impaired cognitive performance among older adults and examine depressive symptoms as a mediator. METHODS: We used cross-sectional data from the Einstein Aging Study, a systematically recruited, community-based cohort study of diverse older adults (N = 794; Age Mean (SD) = 78.9 (5.3); 64.4% Non-Hispanic White, 28.7% Non-Hispanic Black, 5.7% Hispanic). Diabetes status was established via self-reported diagnosis, prescribed medications, and fasting blood glucose. Depressive symptoms were assessed using the Geriatric Depression Scale. Cognitive tests included Digit Symbol, Trails-B, Free Recall, Category Fluency, Boston Naming, and Block Design. Linear regression and mediation analyses were applied. RESULTS: Compared to those without diabetes, diabetes was associated with worse performance on all cognitive tests (ps < 0.05), except Trails-B (p = 0.53), and increased depressive symptoms (p < 0.01). For diabetes, mediation via increased depressive symptoms was observed for Free Recall (p = 0.044), Category Fluency (p = 0.033), and Boston Naming (p = 0.048). CONCLUSIONS: Diabetes was consistently associated with worse cognitive performance and increased depressive symptoms among this older cohort, while prediabetes was not. Mediation findings suggest depressive symptoms may be a biobehavioral pathway linking diabetes and cognition, though the temporal sequence is unclear. If causal, addressing both diabetes and depressive symptoms among older adults may protect cognitive function.