Abstract
Red cell distribution width (RDW), reflecting erythrocyte size variability, and serum vitamin D are emerging biomarkers for dementia. We examined the independent and joint associations of RDW and 25-hydroxyvitamin D [25(OH)D] with incident dementia and the potential mediating role of plasma metabolomic profiles in a large prospective cohort study. We analyzed data from 162,606 UK Biobank participants aged ≥ 50 years and dementia-free with RDW and 25(OH)D measurements at baseline. Incident dementia was identified via electronic health records for ≤ 15 years. Cox proportional hazards models adjusted for sociodemographic, cardiovascular, and genetic factors. Mediation was assessed using generalized structural equation models based on 15 metabolomic principal components (zMETAB1-zMETAB15) derived from 249 NMR-based plasma metabolites. Higher RDW was associated with increased dementia risk (HR = 1.05; 95% CI: 1.02-1.08) and a atherogenic lipid signature (zMETAB1: Cholesteryl Esters in VLDL), while higher 25(OH)D was associated with lower risk (HR = 0.89; 95% CI: 0.86-0.92) and more favorable metabolomic profiles (zMETAB7: higher Omega-3 Fatty Acids; zMETAB12: lower Saturated Fatty Acids). zMETAB1, zMETAB2 (Free Cholesterol in Small HDL), and zMETAB4 (Phospholipids in Small HDL) significantly mediated RDW's effect on dementia (up to 8.8%), whereas ~ 10.8% of the association with vitamin D was mediated by zMETAB1, with additional contributions from zMETAB7 and zMETAB12. Omega-3 fatty acids partially mediated both associations, while RDW and 25(OH)D did not interact significantly in relation to dementia risk. RDW and vitamin D exert opposing, non-interactive influences on dementia risk. Lipid metabolism, particularly omega-3 fatty acids, partially mediates both associations, highlighting a common metabolic pathway that could be targeted for prevention. These findings suggest that interventions aimed at increasing omega-3 intake may simultaneously modulate the effects of both RDW and vitamin D on dementia risk, offering a promising translational strategy for early risk reduction.</p>