Alzheimer's disease neuropathology plasma biomarkers and cognition in midlife: a community-based cohort study.

BACKGROUND: Alzheimer's disease neuropathology, characterised by amyloid β (Aβ) and phosphorylated-tau (p-tau) protein accumulation, has primarily been assessed with biomarkers in clinical samples of older adults. Less is known about plasma biomarkers of Alzheimer's disease neuropathology and their associations with cognitive outcomes in midlife in diverse community-based samples. Our goal was to address these gaps. METHODS: In this cohort study, we analysed participants who were retained in the US Coronary Artery Risk Development in Young Adults (CARDIA) Study with available plasma biomarkers at year 35 (2020-22). We excluded participants without cognitive measures and individuals with probable dementia. Cognition in five domains was measured with standardised tests at years 30 and 35; accelerated cognitive decline in each domain was defined as a 5-year decline at least 1·5 SD greater than the cohort mean change. Plasma Aβ42, Aβ40, and p-tau217 concentrations were assayed with the use of the Fujirebio Lumipulse G1200 analyser and used to calculate the p-tau217-to-Aβ42 ratio (p-tau217/Aβ42) and Aβ42-to-Aβ40 ratio (Aβ42/40). Alzheimer's disease neuropathology status (ie, negative, intermediate, or positive) was defined based on amyloid PET-validated cutpoints for each biomarker (p-tau217/Aβ42, p-tau217, and Aβ42/40). Associations of Alzheimer's disease neuropathology with cognition (Z scores) and accelerated decline were evaluated with the use of multivariable linear and logistic regression. FINDINGS: From the 2248 CARDIA participants who completed the year 35 visit, we randomly selected 1500 participants for plasma biomarker measurement. We excluded three participants with poor biomarker assay quality, 143 without cognitive measures, and four with probable dementia resulting in a final cohort of 1350. The mean participant age was 61 years (SD 3·6, range 53·0-69·0); 779 (58%) participants were women, 571 (42%) were men, 613 (45%) were Black, and 737 (55%) were White. Alzheimer's disease neuropathology positivity was present in 86 (6%) participants based on p-tau217/Aβ42, 196 (15%) based on Aβ42/40, and 48 (4%) based on p-tau217, and was associated with worse performance on processing speed (standardised cognitive difference comparing Alzheimer's disease neuropathology positive to negative for Aβ42/40, p-tau217, and p-tau217/Aβ42 -0·54 to -0·25; p values 0·0001 to 0·0048) and executive function (-0·42 to -0·19; p values 0·0070 to 0·049). Alzheimer's disease neuropathology positivity was also associated with increased odds of accelerated decline on verbal memory (Aβ42/40: odds ratio 4·31, 95% CI 1·71-10·9, p-tau217/Aβ42: 2·44, 1·16-5·13) and processing speed (p-tau217: 3·98, 1·71-9·3; p-tau217/Aβ42: 3·35, 1·77-6·35) compared with Alzheimer's disease neuropathology negativity. There was no association for global cognition or fluency. Although not consistent, some effect modification was observed, with stronger associations among women and Black participants and individuals with APOE ∈4. INTERPRETATION: Alzheimer's disease neuropathology is relatively uncommon in midlife but associated with worse cognitive performance and accelerated decline and might have stronger association among some groups. Early Alzheimer's disease neuropathology detection with the use of plasma biomarkers might enable timely prevention and intervention in midlife adults including risk reduction and pharmacological therapies. FUNDING: National Heart, Lung, and Blood Institute (75N92023D00002, 75N92023D00003, 75N92023D00004, 75N92023D00005, and 75N92023D00006), National Institute on Aging (R01AG063887, R01AG091431, R35AG071916, and K99AG083211), and the Alzheimer's Association (AARFD-23-1150636).