Blood-Based Biomarkers as Predictors of All-Cause Dementia : A 13-Year Follow-Up in the Whitehall II Cohort Study
We are pleased to share the publication of a new article examining the prediction accuracy of blood-based biomarkers for the long-term risk of dementia. Whilst these biomarkers are strongly correlated with amyloid positivity in studies on memory clinic patients seen in research settings, their long-term performance in the general population remains unclear. We therefore assessed the prediction accuracy of these blood biomarkers for incident all-cause dementia on their own and when used alongside known risk factors.
To do this, we used the Whitehall II longitudinal cohort (N=5,225), established in 1985 among civil servants and includes a clinical examination every 4-5 years. Measurements of amyloid-beta (Aβ) 42 and 40, glial fibrillary acidic protein (GFAP), neurofilament light (NfL), and phosphorylated tau 217 (p-tau217) were derived from blood samples collected at the 2012–2013 wave and assayed in 2024. Cases of all-cause dementia were identified through linkage with medical records up to 2025 (median follow-up of 12.2 years).
We observed that participants with values in the worst quartile of biomarker levels had a significantly higher risk of developing dementia. P-tau217 showed the highest predictive ability and the lowest was for amyloid-beta 42/40 ratio. Although these blood biomarkers have excellent specificity and very good negative predictive value, their sensitivity and positive predictive value were modest. When all four are incorporated into a model, they offer predictive performance similar to that of a model containing sociodemographic, behavioural and health-related risk factors. Adding the biomarkers to the risk-factor model yielded the best predictive ability.
These results show that the long-term predictive ability of blood biomarkers for incident dementia in older adults in the general population was modest due to poor sensitivity, and performing similarly to a model based on known risk factors.
- The article is available here: https://www.neurology.org/doi/10.1212/WNL.0000000000218681?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed
By Archana Singh-Manoux
