Differential Lifestyle and Metabolic Risk Factor Associations in Early- Versus Late-Onset Alzheimer's Disease in US Women

Vasuk Gautam^1,4, Smita S. Ghare^1,2,3,4, Stephen Furmanek⁴, Paula M. Chilton^1,2,4, Fengyuan Li^1,2,4, Chelsea Song⁴, Brandon C.Dennis⁴, Lisa B.E. Shields⁴, Andrew Thaliath⁴, Rachel N. Hart⁴, Kenneth G. Pugh⁴, Lynnie Meyer⁴, Christopher B Shields⁴, and Shirish Barve^1,2,3,4

¹Department of Medicine, ²UofL Alcohol Research Center, ³UofL Hepatobiology COBRE, University of Louisville, School of Medicine, Louisville, KY 40202; ⁴Norton Norton Neuroscience Institute Memory Center, Norton Healthcare, Louisville, KY 40202.

 

Background: Women exhibit a higher prevalence of Alzheimer’s disease (AD) and more rapid cognitive decline than men. Existing research on women-specific risk factors has largely focused on late-onset AD (LOAD), leaving early-onset AD (EOAD) understudied. The present work addresses this gap by determining lifestyle and behavioral risk factors relevant for EOAD in women.

Methods: We conducted a retrospective observational study using de-identified U.S. electronic health records from the Epic-Cosmos database. AD cases in women were identified by ICD-10 code G30 between January 1, 2020, and December 31, 2024. Age-matched controls without AD were randomly selected at a 1:1 ratio. AD subtypes were classified as early-onset (<65 years) or late-onset (≥65 years). AD-related lifestyle and metabolic risk factors associated with LOAD and EOAD were evaluated employing multinomial logistic regression analysis.   

Results: Among 2,358,897 U.S. women, 449,431 (19.0%) had AD, including 14,991 (3.3%) with EOAD and 434, 440 (96.7%) with LOAD. A comparison between EOAD and LOAD showed distinct patterns of comorbidity associations. Specifically, women with EOAD demonstrated significantly higher adjusted odds ratios (aORs) for lifestyle-risk factors including alcohol use disorder (aOR: 1.75, 95% CI: 1.72-1.78), tobacco use (aOR: 2.59, 95% CI: 2.56-2.62), obesity (aOR: 2.51, 95% CI: 2.48-2.54), depression (aOR: 1.70, 95% CI: 1.68-1.71), and sleep apnea (aOR: 1.77, 95% CI: 1.75-1.79). In contrast, EOAD demonstrated weaker associations for high blood pressure (aOR 0.67, 95% CI 0.66–0.68), hyperlipidemia (aOR 0.50, 95% CI 0.49–0.51), atrial fibrillation (aOR 0.32, 95% CI 0.32–0.33), hearing loss (aOR 0.52, 95% CI 0.51–0.63), and vision impairment (aOR 0.49, 95% CI 0.48–0.50), indicating significantly greater risk of vascular-metabolic and sensory impairments in LOAD. Moreover, sensitivity analysis with age-matched controls further demonstrated an increase in magnitude of effect for the lifestyle and behavioral risk factors linked with EOAD.

 

Conclusion: Evaluation of Large-scale Epic-Cosmos database highlights distinct differences between women with EOAD and LOAD. Particularly, lifestyle and behavioral risk factors were strongly linked with EOAD, whereas vascular-metabolic comorbidities and sensory impairment were associated with LOAD. Integrating these insights into clinical practice could enhance personalized risk assessment and reduce EOAD Alzheimer’s disease burden in women.