Abstract INTRODUCTION Both short and long sleep duration have been associated with Alzheimer's disease (AD) risk, yet the nature of the non‐linear associations between sleep quantity and blood‐based biomarkers of AD and neurodegeneration remains understudied. METHODS Among 2410 Framingham Heart Study participants (mean age: 70.0 ± 8.45 years; 55.2% female), we examined associations between self‐reported sleep duration and plasma phosphorylated tau (p‐tau)181, total tau, neurofilament light chain, and glial fibrillary acidic protein using restricted cubic splines (RCS) and categorical comparisons (≤ 6, > 6–< 9 reference; ≥ 9 hours). RESULTS RCS revealed non‐linear associations between sleep duration and p‐tau181 (overall p = 0.005; non‐linearity p = 0.002), with higher levels at ≥ 8.5 hours, after adjusting for age, sex, apolipoprotein E ε4 genotype, sleep apnea, depression, and kidney function. Categorical comparisons showed no association in adjusted models. DISCUSSION Sleep duration exhibits robust non‐linear associations with p‐tau181. Findings underscore the importance of non‐linear modeling in relating sleep to blood‐based biomarkers. Longitudinal studies are needed to clarify temporal relationships and mechanistic pathways.
Young et al. (Fri,) studied this question.