Blood tests for dementia biomarkers are reshaping how we detect Alzheimer’s disease and related conditions. Instead of waiting for memory problems to become obvious, phosphorylated tau (p-tau217 and p-tau181) and other protein markers in the bloodstream can now reveal pathological changes years or even decades before cognitive symptoms appear. These biomarkers work because Alzheimer’s disease is fundamentally a disease of protein accumulation—amyloid-beta plaques and tau tangles build up in the brain long before a person notices they’re forgetting names or losing track of conversations. A 65-year-old with no memory complaints might visit their primary care doctor, get a simple blood test, and discover through markers like p-tau217 that amyloid pathology is already developing in their brain.
The implications are profound. For decades, dementia diagnosis required neuropsychological testing, PET imaging, or CSF puncture—expensive, time-consuming, or invasive procedures that kept diagnosis largely out of reach in primary care. Blood biomarkers have changed that equation entirely. The FDA cleared Roche’s p-tau181 test in October 2025 for use in primary care settings. Mayo Clinic and leading neurologists call these advances “practice-changing.” What biomarkers mean for the future is early detection at scale, intervention before irreversible cognitive decline, and a shift from reactive memory-loss management to proactive brain health monitoring.
Table of Contents
- How Blood Tests Are Revealing Hidden Brain Changes
- Understanding the Key Biomarkers in Clinical Practice
- FDA Approvals and What Changed in 2025
- From Lab Test to Clinical Decision-Making
- What Biomarkers Cannot Tell You
- Accelerating Drug Development Through Biomarker-Driven Trials
- Biomarkers in Symptomatic Diagnosis and Prognosis
How Blood Tests Are Revealing Hidden Brain Changes
The fundamental insight behind biomarkers is that the brain and bloodstream communicate. When neurons are damaged or when abnormal proteins accumulate in brain tissue, some of these pathological signals leak into the blood. Phosphorylated tau-217 (p-tau217) emerged as the highest-performing biomarker, with 82% sensitivity and 86% specificity for detecting amyloid pathology—meaning it correctly identifies about 4 out of 5 people with amyloid accumulation and correctly rules out amyloid in 86% of people without it. Phosphorylated tau-181 (p-tau181), cleared by the FDA in October 2025, shows slightly different performance characteristics but has the advantage of being approved for direct clinical use, which means a primary care doctor in any practice can order the test without needing to refer to a specialist first.
These blood biomarkers replace or supplement older methods. CSF analysis requires a lumbar puncture—a procedure with risk and discomfort that keeps it largely confined to research centers or specialized clinics. PET imaging requires expensive equipment, radiation exposure, and radiotracers that aren’t available everywhere. A blood test requires only a standard venipuncture, can be done in any clinic, and results return in days rather than weeks. The Lumipulse G blood test, approved in May 2025, showed 91.7% positive concordance with amyloid PET imaging and 97.3% negative concordance, meaning when the blood test says amyloid is present, PET imaging confirms it more than 9 times out of 10.
Understanding the Key Biomarkers in Clinical Practice
Phosphorylated tau comes in several forms, and the differences matter. P-tau217 and p-tau181 are the most clinically validated, but researchers are also tracking p-tau199. The specific form detected is important because different phosphorylation sites may reflect different stages of pathology or different aspects of tau dysfunction. Beyond tau, amyloid-beta ratio—specifically the Aβ42/Aβ40 ratio—provides complementary information. Amyloid-beta 42 is the form that aggregates into plaques; amyloid-beta 40 is more prevalent but less prone to aggregation. When the ratio drops, it often signals that Aβ42 is being sequestered in plaques rather than circulating freely in the bloodstream.
Neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) represent a new frontier. These are not amyloid or tau—they’re markers of neuronal damage and astrocyte activation. A rising NfL level can indicate accelerating neurodegeneration, and GFAP reflects the brain’s inflammatory response to pathology. One important limitation is that these markers are not specific to Alzheimer’s disease. A patient with mild cognitive impairment might have elevated NfL because of Lewy body dementia, frontotemporal dementia, or even vascular contributions. Biomarkers tell you *what pathology is present*, but they don’t automatically tell you *why* the pathology is there or whether it’s causing the person’s cognitive symptoms.
FDA Approvals and What Changed in 2025
The regulatory landscape shifted decisively in 2025. The Lumipulse G p-tau217/Aβ42 test received FDA approval in May, with published diagnostic accuracy showing 91.7% positive concordance and 97.3% negative concordance to amyloid PET imaging. This approval signaled that blood biomarkers were ready for clinical deployment, not just research settings. By October 2025, Roche’s Elecsys p-tau181 plasma test received FDA clearance specifically for use in primary care physicians’ offices to rule out amyloid pathology in patients with cognitive concerns.
The Alzheimer’s Association released its first evidence-based clinical practice guideline in July 2025, establishing that blood biomarkers used for diagnostic triage must achieve at least 90% sensitivity and 75% specificity—a standard that multiple current biomarkers now meet. What this means practically is that a primary care doctor can now order a blood test as a first step when a patient reports memory concerns, without automatically referring to neurology or ordering expensive imaging. If the biomarker is negative, amyloid pathology can be reasonably ruled out, and the workup can pivot toward other causes of cognitive change. As of January 2025, 182 active clinical trials for Alzheimer’s interventions were underway, and 84% of these trials use biomarkers as inclusion criteria or primary outcomes, reflecting the central role these tests now play in drug development and validation.
From Lab Test to Clinical Decision-Making
The practical workflow has changed for many clinicians. A 72-year-old sees their primary care doctor reporting that they’ve been forgetting recent conversations and struggling with their checkbook. Previously, the next step would likely be referral to neurology for cognitive testing and possibly MRI or PET imaging. Today, a simple blood draw for p-tau181 or p-tau217 can provide critical information within days. If the biomarker is positive for amyloid pathology, the patient proceeds to cognitive testing and possibly neuroimaging to confirm mild cognitive impairment or dementia. If the biomarker is negative, the evaluation can focus on other treatable causes—thyroid dysfunction, vitamin B12 deficiency, depression, sleep disorders, medication side effects, or vascular disease.
This shift has real consequences for access and equity. A patient in a rural area 100 miles from the nearest neurology clinic can now get a biomarker test at their local clinic. A patient without insurance or with high copays faces a lower financial barrier to early evaluation. However, a critical limitation remains: biomarker positivity does not automatically indicate dementia or even mild cognitive impairment. Studies show that approximately 30% of cognitively normal older adults have amyloid pathology detectable on PET imaging or in CSF. Some of these individuals have biomarker positivity for years or decades without ever developing cognitive symptoms. Biomarkers detect pathology; they don’t predict the timeline to symptomatic disease for any individual.
What Biomarkers Cannot Tell You
One of the most important limitations is predictive precision for individual patients. A 68-year-old with p-tau217 positivity might remain cognitively intact until age 85, or cognitive decline might begin within two years—current biomarkers don’t distinguish between these trajectories reliably. The presence of amyloid pathology is necessary for Alzheimer’s disease diagnosis but not sufficient; cognitively impaired patients often have multiple types of pathology simultaneously, including amyloid, tau, Lewy bodies, and vascular disease. A person with positive amyloid biomarkers and memory complaints might have Lewy body dementia, primary age-related tauopathy (PART), or vascular contributions that biomarkers alone won’t identify.
Biomarkers also don’t tell you about cognitive reserve, compensatory brain mechanisms, or psychosocial factors that influence symptom severity and progression rate. Two patients with identical biomarker profiles and the same cognitive test scores may experience very different functional trajectories based on education level, cognitive engagement, social connection, sleep quality, and cardiovascular health. There’s also the risk of over-testing and over-interpretation. As blood biomarkers become more sensitive and more widely available, primary care providers may order them reflexively for any patient over 60 with a subjective memory complaint—generating positive results that create unnecessary anxiety and potentially driving treatments in asymptomatic biomarker-positive individuals without established clinical benefit.
Accelerating Drug Development Through Biomarker-Driven Trials
The pharmaceutical industry has fundamentally restructured Alzheimer’s research around biomarkers. The 182 active Alzheimer’s intervention trials as of January 2025 represent an unprecedented research effort, and 84% of these trials now use biomarkers as inclusion criteria or primary outcome measures. This shift means that drugs are tested in biomarker-positive patients before cognitive symptoms emerge, in asymptomatic individuals at risk, and in people with mild cognitive impairment or dementia. Anti-amyloid monoclonal antibodies like aducanumab, lecanemab, and donanemab showed cognitive benefits in trials that enrolled patients based on amyloid biomarker positivity and mild cognitive impairment. These same drugs showed minimal or no cognitive benefit in trials of asymptomatic biomarker-positive individuals.
This disconnect highlights both the power and the limitation of biomarker-driven trials. Amyloid reduction in the brain (confirmed by PET imaging) did not necessarily slow cognitive decline in preclinical amyloid accumulation. The relationship between pathology and symptom progression is more complex than a simple linear model. Future drug development may increasingly rely on biomarker combinations—amyloid plus tau plus neurodegeneration markers—to identify patients most likely to progress and most likely to benefit from intervention. Tau-targeting therapies are already emerging, with some evidence suggesting that combinations of amyloid-targeting and tau-targeting approaches might be more effective than either alone.
Biomarkers in Symptomatic Diagnosis and Prognosis
For patients who already have cognitive symptoms, biomarkers help confirm the underlying pathology and support diagnosis. The Alzheimer’s Association’s diagnostic framework now incorporates biomarker evidence into the diagnostic criteria for mild cognitive impairment due to Alzheimer’s disease and dementia due to Alzheimer’s disease. A patient with memory impairment, positive cognitive testing, and positive amyloid and tau biomarkers meets the criteria for symptomatic Alzheimer’s disease; a patient with the same cognitive impairment but negative amyloid biomarkers likely has non-Alzheimer’s pathology. This precision changes treatment selection, prognosis communication, and family counseling.
Beyond diagnosis, emerging evidence suggests that biomarker profiles predict rate of progression. Patients with elevations in both amyloid and tau biomarkers tend to decline faster cognitively than those with amyloid positivity alone. Patients with high NfL, indicating substantial neuronal damage, also tend to have steeper cognitive decline. This prognostic information, combined with known genetic factors like APOE4 status and clinical measures like hippocampal volume on MRI, can help clinicians and families prepare for the disease’s trajectory and plan for long-term care and support needs earlier than was previously possible.
- —
