Symptom-Guided Blood Biomarker Testing for Alzheimer's Disease

Introduction Blood-based biomarkers can accurately detect Alzheimer's disease (AD) pathology, but symptom-guided testing approaches are understudied. This study evaluated the utility of early symptom and biomarker changes in predicting cognitive decline due to AD. Methods We analyzed data from the Alzheimer's Disease Neuroimaging Initiative-2 (development, n=603) and -3 (validation, n=378). The primary outcome was 4-year progression to mild cognitive impairment or dementia associated with amyloid-beta neuropathology. Machine learning estimated baseline symptom burden from subjective ratings. Logistic regression calculated area under the curve (AUC), relative risk, and time-to-event. Results Symptom burden predicted clinical progression (AUC=0.79), which further improved with blood biomarkers (AUC=0.90). Participants with elevated symptoms and biomarkers had a progression incidence of 72.1% and a 19-fold higher progression risk. Findings replicated in validation. Conclusion Findings underscore the potential of pairing symptoms and biomarkers to more efficiently and accurately detect AD cases at risk of clinical progression. Trial Parent study registrations: NCT01231971; NCT02854033.