Equity and Transportability of Plasma ATN Phenotypes in a Population-Representative U.S. Aging Cohort.

Plasma biomarkers are transforming Alzheimer's disease research, yet their performance in diverse, population-representative settings remains largely unknown. Using data from 4,427 adults in the nationally representative Health and Retirement Study, this work evaluates whether amyloid, tau, and neurodegeneration (ATN) biomarkers transport equitably across demographic groups. Population-weighted analyses reveal that tau is the only biomarker showing a stable association with cognition at the population level, whereas amyloid and neurodegeneration markers lose significance after accounting for sampling design. Fairness analyses uncover striking disparities: sensitivity for detecting cognitive impairment is more than twofold higher in White than Black participants, with Black women exhibiting the lowest detection rates. Education further modifies biomarker-cognition relationships, producing paradoxical positive amyloid associations and amplified neurodegeneration effects in structurally disadvantaged groups. These patterns persist after adjustment for vascular comorbidities, indicating that equity gaps arise from more than differential cerebrovascular burden. Together, these findings show that plasma ATN biomarkers do not generalize uniformly across the U.S. population and may systematically underperform in minoritized and socioeconomically disadvantaged groups. The results highlight the need for population-based validation and fairness-aware calibration before deploying blood-based biomarkers in clinical or public health settings.