Abstract Number: Esoc2026a599 Normative Model-Defined White Matter Hyperintensities Severity Reveals Age-Independent Disease Burden and Mortality Risk

Abstract Background and aims Covert white matter hyperintensities (WMHs) are common, yet age referenced evaluation thresholds are lacking. We therefore developed a normative grading model and validated its clinical relevance using cross-sectional measures and longitudinal mortality outcomes. Methods We constructed an age and sex-adjusted normative WMH model using hierarchical Bayesian regression trained on 2,544 healthy adults aged 50-92 years. The model was applied to an independent cohort of 1,237 community-dwelling, stroke- and dementia- free adults. Model validity was evaluated using multivariate linear regression for cross-sectional associations and Cox proportional-hazards analyses for mortality outcomes. Results A total of 1,237 ILAS participants (62.7 ± 8.8 years; 52.9% women) were stratified into no/minimal, mild, moderate, and severe WMHs based on quartiles of normative model–derived Z-scores. The severe WMH group was not older but slightly younger than the no/minimal group. Higher WMH severity was associated with more adverse vascular profiles, including greater prevalence of hypertension and diabetes. Each one-level increase in WMH severity was linked to poorer cognition: lower MMSE (β= –0.179; FDR-p= 0.034), verbal memory (β= –0.121; FDR-p= 0.034), and visuospatial construction (β= –0.330; FDR-p= 0.049), and to lower global and hippocampal and thalamic GMVs. Over 9.4 years of follow-up, severe WMH predicted nearly twofold higher risk of all-cause mortality (HR= 1.90; 95% CI= 1.16–3.09) versus the no/minimal WMH. Conclusions This first validation study demonstrates that normative WMH grading captures age-independent pathological WMH burden and identifies individuals at increased morta lity risk, with particular sensitivity to younger individuals who would be underestimated by conventional approaches. Conflict of interest The authors have no disclosures relevant to the current work.