Introduction and Objective: Proteomics can provide insight into mechanisms linking diabetes to dementia risk. We identified plasma central nervous system (CNS) proteins associated with incident dementia in older adults with diabetes. Methods: We conducted a prospective cohort study of 487 ARIC Study participants with diabetes (age 58-84y, 47% women, 34% Black) in 2005-06. We quantified the relative abundance of 131 CNS-relevant plasma proteins (Alamar Bioscience). We used Cox regression to test protein associations (per 1 SD) with incident dementia. Statistical significance was based on FDR0.05. We generated C statistics for models including 12 covariates and dementia-associated proteins. Results: Over ~15 years of follow-up, there were 135 new dementia cases. We identified 16 dementia-associated proteins, including biomarkers of amyloid-β and tau pathology (e.g. pTau217), neurodegeneration (e.g. NEFL), and inflammation (e.g. IL33) (Figure). Brain-derived pTau217 and pTau181 were associated with a 2.3 (95% CI: 2.0, 2.8) and 2.7-fold (95% CI: 2.2, 3.4) higher dementia risk per 1 SD, respectively. Models including all covariates and 16 proteins had a C statistic of 0.83, with a ∆C statistic of 0.08 (P=2e-5) beyond the covariates alone. Conclusion: We identified proteins on a novel platform that were strongly associated with incident dementia in persons with diabetes, signaling mechanistic insight and opportunity for risk stratification. Disclosure M.R. Rooney: None. J. Chen: None. B. Thyagarajan: None. K. Walker: Other - Founder, Shareholder; Current; Centia Bio. R.F. Gottesman: None. S. Sedaghat: None. P.L. Lutsey: None. R.C. Hoogeveen: Consultant; Current; Denka Seiken. C.M. Ballantyne: Consultant; Current; AstraZeneca, Eli Lilly and Company, Merck Ended; Siemens.
ROONEY et al. (Mon,) studied this question.