AbstractBackgroundType 2 diabetes (DM) is a risk factor for Alzheimer's disease (AD) and other dementias. DM, which is more common in Mexican Americans (MA) than non‐Hispanic Whites (NHW), is associated with high blood glucose levels and often with hyperinsulinemia. Hippocampal volume is an early neuroimaging marker of AD. We examined the relationship of blood insulin with hippocampal subregion volumes in MA and NHW.MethodIn a sample of 1588 older adults without dementia (age: 50‐92 years; 826 MA, 762 NHW; 387 DM, 1201 non‐DM; 1344 cognitively normal, 244 mild cognitive impairment; 613M/975F) from the Health and Aging Brain Study ‐ Health Disparities (HABS‐HD) cohort, we assessed whether fasting plasma insulin and glucose levels were associated with the volume of left and right hippocampal body subregions (CA1, CA23DG, subiculum) using hippocampal high resolution 3T MRI scans (Siemens 3T). We further examined the significant insulin relationship with CA23DG volume by ethnicity and diabetes diagnostic group. We used a multiple linear regression model, covarying for age, sex, cognitive diagnosis, diabetic status, years of education, number of MRI slices segmented, and intracranial volume. All p values were FDR corrected for multiple comparisons across regions.ResultHigher insulin levels were associated with a smaller volume in right and left CA23DG (right: FDR p=0.015, left: FDR p=0.015) across all participants. Glucose levels were not significantly associated with hippocampal subregion volume (Table 1). When we further investigated the significant relationship between insulin and CA23DG volume by ethnic group and diabetic status, we found that higher insulin was associated with lower CA23DG volume in NHW (right: p=0.014, left: p=0.013), MA (right only: p=0.034, left: p=0.115), and in non‐diabetics (right: p=0.027, left trend: p=0.056) (Figure 1).ConclusionHigher insulin levels were associated with lower CA23DG volume in older adults without dementia, with the strength of the negative association between blood insulin levels and hippocampal subfield volume varying by diabetic status.