Blood Tests for Alzheimer's Detection Get Their Most Rigorous Review Yet

Confirming Alzheimer's disease pathology in living patients has historically required either PET neuroimaging or cerebrospinal fluid (CSF) analysis—procedures that are costly, invasive, and difficult to access broadly. Blood-based biomarkers promise a simpler, cheaper alternative, but until now no comprehensive, evidence-graded systematic review had rigorously evaluated their real-world diagnostic performance across the full range of available assays in the population most likely to receive them: cognitively impaired individuals presenting to specialist memory or neurology clinics.

This systematic review and meta-analysis searched MEDLINE, Embase, and Cochrane Library from January 2019 to November 2024, identifying 49 observational studies that evaluated 31 distinct BBM tests against reference standards of CSF AD biomarkers, amyloid PET, or neuropathology. Eligible participants had mild cognitive impairment (MCI) or dementia and were seen in specialized care settings. The analytes assessed included plasma phosphorylated tau isoforms (p-tau181, p-tau217, %p-tau217, p-tau231) and the amyloid ratio Aβ42/Aβ40. Pooled sensitivity, specificity, positive and negative likelihood ratios, and post-test probabilities across a range of pre-test prevalences were calculated following Cochrane Handbook guidance. Risk of bias was assessed with the QUADAS-2 tool and certainty of evidence graded using GRADE.

Performance varied dramatically by analyte and assay platform. Using a single pre-specified cut-point per test, pooled sensitivity ranged from 49.3% (95% CI: 41.2–57.4) to 91.4% (95% CI: 86.6–94.6), and pooled specificity from 61.5% (95% CI: 45.6–75.3) to 96.7% (95% CI: 87.8–99.2). The magnitude of change in post-test probability following a positive or negative result also varied greatly depending on both the specific test and the pre-test prevalence of amyloid positivity in the clinical population. The highest-performing tests (notably certain p-tau217 assays) showed clinically meaningful shifts in post-test probability, while lower-performing tests provided limited diagnostic value beyond prior clinical assessment.

Critically, most included studies were judged at high risk of bias, particularly in patient selection (e.g., non-consecutive enrollment, unclear inclusion criteria), index test conduct (e.g., unblinded interpretation), and reference standard quality. Certainty of evidence ranged from moderate down to very low across tests, meaning current estimates could change substantially as more rigorous data emerge. The review explicitly cautions that BBM results must be interpreted in the context of the specific assay used, integrated with comprehensive clinical assessment, and not applied interchangeably across platforms.

These findings served as the evidentiary foundation for the Alzheimer's Association's new clinical practice guidelines on BBM use in specialty care. The review underscores that no single blood test performs uniformly and that assay-specific performance data must guide clinical decision-making. As this field is evolving rapidly—with new assays, regulatory approvals, and clinical data emerging continuously—the authors stress that ongoing evidence surveillance and guideline updating will be essential to maintain clinical relevance.