AbstractBackgroundStrong psychosocial health has been linked to preserved cognition and functional independence in older adults. Associations between late‐life psychosocial health and dementia risk have been established in older adults, but it remains less clear how one’s social interactions in mid‐life may relate to dementia risk in late life. We hypothesized that social engagement in mid‐life will be negatively associated with the risk of dementia in late‐life.MethodWe included data from 13,216 dementia‐free participants at midlife (mean age: 57.0 years, SD: 5.7 years, 54.9% female, 23.8% Black) from the Atherosclerosis Risk in Communities (ARIC) cohort study. Social support and isolation were assessed via interviewer‐administered questionnaires (visit 2: 1990‐1992). Based upon categorization of both factors, participants were classified as having high, intermediate, or low mid‐life social engagement. Incident dementia cases were identified from visit 2 through December 31, 2019 with ongoing surveillance (e.g., in‐person neurocognitive testing, informant interviews, and hospitalization codes). Primary analyses used Cox proportional hazards regression models and were adjusted for demographics, APOE ε4, occupational/marital status, depressive symptoms, and vascular risk factors measured at visit 2. Secondary analyses examined whether sex and/or race modified the association between mid‐life social engagement and dementia risk with formal tests for interaction.ResultAmong 13,261 participants, 2,555 developed dementia (median follow‐up: 23.7 years). Participants with high (hazard ratio (HR): 0.77, 95% CI: 0.68 ‐ 0.86) or intermediate social engagement (HR: 0.84, 95% CI: 0.74 – 0.95) in mid‐life had a lower risk for developing dementia, compared to participants with low social engagement in mid‐life (Figure 1, Table 1). Sex modified this effect (p‐value for interaction = 0.042) whereas race did not. In males, only those with high social engagement in mid‐life had a lower risk of dementia; the association between males with intermediate social engagement in mid‐life and dementia risk was not significant (Table 2).ConclusionGreater social engagement in mid‐life is associated with a lower risk of dementia. This finding provides further evidence that mid‐life may be a critical time window for dementia prevention. Longitudinal studies evaluating the mechanisms by which psychosocial factors modify the likelihood of incident dementia are needed.