HAI Book 2025 - Flipbook - Page 505
Heston, Margo
Examining neighborhood disadvantage as a factor explaining amyloid onset
age
Margo B. Heston1,2, Rebecca E. Langhough1,3, Will R. Buckingham2, Jacob Morse1, Elena Ruiz
de Chavez1, Richard J. Chappell1,4, Sterling C. Johnson1,3, Sanjay Asthana1, Brian A. Gordon5,
Carey E. Gleason1, Lindsay R. Clark1, Barbara B. Bendlin1,2,3, W. Ryan Powell2, Megan L.
Zuelsdorff1,2,6, Tobey J. Betthauser1,2
Wisconsin Alzheimer’s Disease Research Center, University of Wisconsin School of Medicine and Public Health,
Madison, WI, US
2
Center for Health Disparities Research, University of Wisconsin School of Medicine and Public Health, Madison, WI,
US
3
Wisconsin Alzheimer’s Institute, University of Wisconsin School of Medicine and Public Health, Madison, WI, US
4
Department of Biostatistics and Medical Informatics, University of Wisconsin School of Medicine and Public Health,
Madison, WI, US
5
Knight Alzheimer’s Disease Research Center, Washington University School of Medicine in St. Louis, St. Louis, MO,
US
6
University of Wisconsin-Madison School of Nursing, Madison, WI, US
1
Background: Social factors including neighborhood disadvantage have shown associations with higher
postmortem amyloid burden, lower hippocampal volume, and accelerated trajectories of cortical thinning and
cognitive decline, but it is unknown if disadvantage affects when AD pathology begins. This work investigates
whether amyloid onset age differs between most and least disadvantaged neighborhood groups.
Methods: We included amyloid PET data from the Wisconsin ADRC (WADRC)/Wisconsin Registry for Alzheimer9s
Prevention (WRAP) and Washington University Knight ADRC (WashU) (Ns=787 and 738; Table 1). Cortical amyloid
was quantified averaging 11C-Pittsburgh compound B DVRLGA across AAL3-defined regions (WADRC) and 11C-PiB
SUVR30-60/18F-Florbetapir SUVR50-70 across Freesurfer-defined regions (WashU, cerebellar cortex reference);
WashU SUVRs were converted to Centiloids. We applied sampled iterative local approximation (SILA) across
cohorts separately to model amyloid accumulation and estimate individual A+ onset age (EAOA). A+ was defined as
global DVR>1.16 (CL>17.10) and CL>14.76 using GMM. Area deprivation index (ADI) measured neighborhood
disadvantage (2020 vintage), binarizing state decile rankings to compare EAOA between participants in the 30%
most and 70% least disadvantaged neighborhoods. Cox proportional hazards regression tested ADI effects on A+
risk, controlling for APOE, sex, and baseline age; Kaplan-Meier curves illustrated risk model effects.
Results: Cohort differences were observed for proportion of A+, baseline age, and composition across diagnosis,
sex and race, but not ADI (Table 1, Figure 1 bottom). Cox regression (Figure 2) indicated ADI was not associated
with A+ risk or EAOA (Figure 1 top, ps=.43,.09) while confirming APOE main effects. Female sex was associated
with earlier EAOA in WashU (p=.01; WADRC/WRAP p=.44), however ADI-by-sex effects were nonsignificant in
exploratory testing (ps=.57,.13).
Discussion: Results in two cohorts suggest neighborhood disadvantage does not influence A+ onset age, however
samples had low representation of highly disadvantaged neighborhoods. Future work in larger samples will
investigate whether ADI affects time from A+ onset to dementia.
HAI2025 - 505
