Dissecting Alzheimer's disease heritability across populations

Abstract

Introduction: Late-onset Alzheimer's disease (LOAD) is highly heritable; however, its estimated incidence across populations remains unclear.

Methods: We computed family-based heritability leveraging Alzheimer's Disease Sequencing Project pedigrees from non-Hispanic White (404 pedigrees), non-Hispanic Black (13 pedigrees), Dominican (100 pedigrees), and Dutch isolate (10 pedigrees), with four models incorporating age, sex, apolipoproten E epsilon4 (APOE ε4), and contributing study using two methods.

Results: Heritability estimates varied by method, model, and study populations. Statistical Analysis for Genetic Epidemiology (S.A.G.E.) estimates were highest for Dutch isolate (78.3%), followed by non-Hispanic Blacks (39.1%), Dominicans (31.7%), and non-Hispanic Whites (29.1%), adjusted for age and sex. APOE adjustment reduced estimates (4.9% on average), while study adjustment primarily affected groups that included multiple studies. Sequential Oligogenic Linkage Analysis Routines (SOLAR-Eclipse) estimates were higher (45.2% to 80.2%) than S.A.G.E. (20.4% to 80.9%) but behaved in parallel, except for the Dutch isolate.

Discussion: LOAD heritability estimates are dependent on study population and may reflect or indicate differences in LOAD risk by population.

Keywords: APOE ε4; cohort; family‐based; heritability; multiple populations.