Abstract
IMPORTANCE: Hippocampal sclerosis (HS), defined as severe neuronal loss and astrogliosis in coronal sections of the midhippocampus cornu ammonis 1 or subiculum, is an important pathology associated with limbic-predominant age-related transactive response DNA-binding protein 43 encephalopathy neuropathological change (LATE-NC), Alzheimer disease, and dementia. The association of diet with HS or HS with LATE-NC in humans remains underexplored. OBJECTIVE: To investigate the association of the Mediterranean-Dietary Approaches to Stop Hypertension Intervention for Neurodegenerative Delay (MIND) diet with HS and HS with LATE-NC. DESIGN, SETTING, AND PARTICIPANTS: This cohort study included autopsied participants from the ongoing Rush Memory and Aging Project cohort study. Participants with dietary and neuropathological data from 2004 to 2024 were included. Data were analyzed from April 3, 2024, to May, 13, 2025. EXPOSURES: Mean MIND diet scores (range, 0-15; higher score indicates better diet), computed from all validated food frequency questionnaires administered annually for up to 18 years before death. MAIN OUTCOMES AND MEASURES: The outcomes of interest were presence of HS, assessed using hematoxylin and eosin staining, and presence of LATE-NC, detected by TDP-43 immunohistochemistry in 8 brain regions. For a subset of 300 participants, hippocampal neuronal loss severity was scored using a semiquantitative scale from 0, indicating none, to 5, severe, and categorized as none to mild, moderate, and severe neuronal loss. RESULTS: Among 809 participants (mean [SD] age at death, 91.3 [6.1] years; 538 [72%] female; mean [SD] follow-up, 7.2 [4.4] years), HS was present in 82 (10.1%) participants; 71 participants (9%) had both HS and LATE-NC, and in the scored subset, 43 participants (14%) had moderate and 35 participants (12%) had severe hippocampal neuronal loss. Higher MIND diet scores were associated with lower odds of HS (odds ratio [OR], 0.78; 95% CI, 0.65 to 0.95) and HS with LATE-NC (OR, 0.79; 95% CI, 0.64 to 0.97) after controlling for age at death, sex, education, total calories, APOE-ε4 status, AD, and vascular pathologies. In mediation analyses, the MIND diet was associated with less dementia at the time of death (β = -0.26; 95% CI, -0.36 to -0.15; P < .001), with an indirect association of 21% through HS (β = -0.05; 95% CI, -0.10 to -0.01; P = .02). A higher MIND diet score was associated with less hippocampal neuronal loss (P for trend = .01). CONCLUSIONS AND RELEVANCE: This cohort study of autopsied participants found that MIND diet adherence during follow-up was associated with a lower likelihood of HS, HS with LATE-NC, and hippocampal neuronal loss. The association of diet with dementia was partially mediated by its association with HS. These findings suggest that the MIND diet may reduce adverse brain health outcomes.