A. I. Boeriu, T. Gu, B. Fulton-Howard, E. M. Lucero, P. Garg, A. Sharp, J. A. Shortt, C. A. Gignoux, F. J. Rajabli, A. J. Griswold, K. Yaffe, S. J. Andrews
INTRODUCTION: APOE*{varepsilon}4 is the strongest genetic risk factor for Alzheimer's disease (AD), yet its effect varies across ancestral populations. As plasma biomarkers increasingly inform AD diagnosis, failure to account for both APOE genotype and genetic ancestry could lead to diagnostic misclassification. METHODS: In 2,733 participants from the Health and Aging Brain Study Health Disparities (HABS-HD), we evaluated whether genetically-inferred ancestry modifies APOE associations with AD endophenotypes. Regression models were applied across three frameworks: ancestry-stratified analyses, compositional modeling of global ancestry, and local ancestry analyses at the APOE locus. RESULTS: APOE {varepsilon}4+ was associated with higher pTau217 and pTau181, lower A{beta}42/A{beta}40, and worse global cognition (CDR-SB), significant within ancestry groups after FDR correction. The {varepsilon}4+ effect on pTau217 was ~2.5-fold larger in EUR-like than AFR-like participants. DISCUSSION: African ancestry attenuates APOE*{varepsilon}4+ associations with pTau217, highlighting the need to consider both APOE genotype and ancestry in biomarker-based AD evaluations.