Blood Tests for Alzheimer: What p-tau217 Can and Cannot Do

The short answer

A blood test for p-tau217 can now help a specialist judge whether Alzheimer's disease pathology is likely present in a person who already has memory or thinking problems, and its greatest strength is telling clinicians when Alzheimer's is unlikely to be the cause. In May 2025 the FDA cleared the first such blood test for diagnostic use, and in July 2025 the Alzheimer's Association published its first clinical practice guideline for these tests. Both make the same boundaries clear: this is a tool for symptomatic patients evaluated in specialized care, not a screening test for healthy adults, and not a diagnosis on its own.

This article is educational and is not medical advice.

What p-tau217 is measuring

Phosphorylated tau 217, or p-tau217, is a form of the tau protein that changes early in the Alzheimer's disease process. It tracks closely with the buildup of amyloid plaques in the brain, which is why a plasma measurement can stand in, imperfectly, for what a PET scan or a spinal-fluid test would show. Most high-performing assays report p-tau217 either alone, as a ratio to non-phosphorylated tau, or combined with the amyloid-beta 42/40 ratio. The point is not the protein itself but what it signals: the presence or absence of Alzheimer's-type pathology in someone whose symptoms already raise the question.

What the FDA actually cleared

On May 16, 2025, the FDA cleared the Lumipulse G pTau217/Beta-Amyloid 1-42 Plasma Ratio, made by Fujirebio, as the first blood test authorized to aid in diagnosing Alzheimer's disease. The intended population is narrow and specific: patients aged 55 and older who are already showing cognitive symptoms and being evaluated for Alzheimer's. In the clearance data, 91.7% of people with a positive result had amyloid confirmed by PET or spinal-fluid testing, and 97.3% of people with a negative result were confirmed negative. The FDA was explicit that the test is meant to be used alongside other clinical information and is not a standalone diagnosis, and that false positive and false negative results do occur.

Ruling out is easier than ruling in

The most useful way to read those two numbers is that they do different jobs. A high negative predictive value, around 97% in the clearance data, means a negative result is very good at ruling Alzheimer's pathology out. That has real clinical value, because it can spare a patient invasive or costly confirmatory testing and redirect attention to other treatable causes of cognitive symptoms. A positive result is informative but less definitive. As a 2025 analysis in the Journal of Clinical Investigation put it, the test carries high utility in ruling out Alzheimer's as the source of symptoms and useful, though weaker, utility in suggesting that pathology is present.

That same analysis makes a point worth holding onto: a test like this will be clearly informative for roughly four out of five patients, while about one in five fall into an intermediate or indeterminate range. Many newer testing strategies use two cutpoints rather than one, sorting results into low, high, and an in-between zone that still needs a PET scan or spinal-fluid test to resolve. That intermediate group remains the least studied part of the picture.

What the 2025 guideline says, and does not say

The Alzheimer's Association guideline, led by Sebastian Palmqvist and colleagues and published in Alzheimer's & Dementia, is deliberately conservative. It applies only to specialized care, meaning clinicians who spend a substantial share of their time on cognitive disorders, and only to people who already have objective cognitive impairment such as mild cognitive impairment or dementia. It does not address screening, primary care use, or people with no cognitive concerns.

The guideline sets two performance bars. For a triaging test, one used to flag who needs further workup, it recommends at least 90% sensitivity and at least 75% specificity, with positive results confirmed by PET or spinal-fluid testing. For a confirmatory test that could substitute for those methods, it recommends at least 90% for both sensitivity and specificity. These were issued as conditional recommendations backed by low-certainty evidence, which is the panel's own signal that the field is early.

Two cautions in the guideline deserve emphasis. First, the panel found wide variation across the many assays it reviewed, and warned that many commercially available blood tests do not meet these thresholds, especially when a single cutoff is used. Second, several conditions can distort results, including severe kidney disease, recent head injury, and other factors, so a number cannot be read in isolation.

Where this leaves patients and clinicians

The honest framing is that a p-tau217 blood test changes the workflow more than it changes the diagnosis. Used in the right person, one already being evaluated for cognitive symptoms by a specialist, it can reduce the need for lumbar puncture or PET imaging and can rule Alzheimer's out with considerable confidence. Used in the wrong context, an anxious but asymptomatic adult, a direct-to-consumer purchase, or a primary care visit with no cognitive workup, it invites misinterpretation, false reassurance, and false alarm. A biomarker describes pathology; it does not, by itself, describe a person's future or dictate care.