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

The short answer#

If you already have memory or thinking problems, a p-tau217 blood test can help a specialist judge whether Alzheimer's disease is the likely cause, and it does its best work when the answer is no. In May 2025 the FDA cleared the first such test for diagnostic use, and in July 2025 the Alzheimer's Association issued its first clinical practice guideline for these tests. Read together, they draw the same fence around the technology: it is for symptomatic patients being worked up in specialty care, not a wellness screen for people who feel fine, and never a diagnosis standing on its own.

Key points#

• A p-tau217 result is strongest as a rule-out. A negative test makes Alzheimer's pathology unlikely with roughly 97% confidence in the clearance data.
• A positive result is suggestive but weaker, so guidelines ask for a confirming PET scan or spinal-fluid test before it changes management.
• Roughly one patient in five lands in an indeterminate middle that a single blood draw cannot settle.
• The cleared use is narrow: adults 55 and older who already have cognitive symptoms, evaluated by clinicians who focus on memory disorders.
• Kidney disease, recent head injury, and other conditions can shift the number, so it is never read in isolation.

Why a protein in the blood can mirror the brain#

Tau is a structural protein inside neurons. In Alzheimer's disease it becomes abnormally phosphorylated, and one variant, phosphorylated tau 217 or p-tau217, starts to rise early, tracking the accumulation of amyloid plaques years before symptoms peak. Because a small fraction of it spills into the bloodstream, a plasma level can act as a proxy for what a PET scan or a lumbar puncture would otherwise reveal. Most of the better assays do not report the protein alone. They combine it into a ratio, often p-tau217 against the amyloid-beta 42/40 measurement, which sharpens the signal. The number is not the point in itself. It is a stand-in for a yes-or-no question: is Alzheimer's-type pathology present in a person whose symptoms already prompted the question.

What the FDA actually cleared, and for whom#

On May 16, 2025, the FDA cleared the Lumipulse G pTau217/Beta-Amyloid 1-42 Plasma Ratio as the first blood test authorized to aid in diagnosing Alzheimer's disease. The label is deliberately tight. It covers adults aged 55 and older who are already showing cognitive symptoms and being evaluated for the condition, not the worried well and not a routine check-up add-on. In the supporting data, about 92% of people with a positive result had amyloid confirmed on PET or spinal-fluid testing, and about 97% of people with a negative result were confirmed to have none. The agency was blunt on two points: the test is meant to sit alongside the rest of the clinical picture rather than replace it, and both false positives and false negatives happen.

The asymmetry that matters most#

Those two percentages look similar, but they do different jobs, and confusing them is the most common way you will see this test misread. The high figure for negatives, near 97%, means a clean result is very good at ruling Alzheimer's pathology out. That is genuinely useful. It can spare you a lumbar puncture or an expensive scan and push the workup toward other, sometimes reversible, drivers of cognitive change, such as thyroid disease, medication effects, depression, or sleep disorders. A positive result carries less weight in the other direction. As one 2025 analysis in the Journal of Clinical Investigation framed it, the assay is strong for ruling the disease out and useful but softer for ruling it in.

A helpful way to picture it: a negative test is like a smoke detector that stays silent, which is fairly reassuring, while a positive test is like the alarm going off, which warrants a look but is not proof of a fire. That is why the guidance treats a positive plasma result as a prompt for confirmatory imaging or spinal fluid rather than an endpoint.

The indeterminate middle#

The same JCI analysis flags a detail that rarely survives the headlines: a test like this is clearly informative for about four patients in five, while roughly one in five falls into an intermediate band. Many modern strategies use two thresholds instead of one, sorting results into low, high, and an in-between zone that still needs a PET scan or spinal-fluid test to resolve. That middle group is also the least studied, so if you land there you have not been given a wrong answer so much as an unfinished one.

The guardrails the guideline sets#

The Alzheimer's Association guideline, published in Alzheimer's and Dementia, is cautious by design. It applies only in specialized care, meaning clinicians who devote a substantial share of their practice to cognitive disorders, and only to people who already have measurable impairment, such as mild cognitive impairment or dementia. If that does not describe you, the guideline does not cover the test you are asking about. It does not endorse screening, primary-care use, or testing in people with no cognitive concerns.

It also sets performance bars. For a triaging test, one used to decide who needs a fuller workup, it asks for at least 90% sensitivity and at least 75% specificity, with positives confirmed by imaging or spinal fluid. For a confirmatory test that could stand in for those methods, it asks for at least 90% on both. These were issued as conditional recommendations resting on low-certainty evidence, the panel's own way of saying the field is still young. The guideline adds two warnings worth keeping in view: assays vary widely and many commercial tests do not clear these thresholds, especially with a single cutoff, and several conditions including severe kidney disease and recent head injury can distort the reading.

Where this leaves patients and clinicians#

The most accurate summary is that a p-tau217 blood test reshapes the workup more than it delivers the diagnosis. In the right person, someone already being evaluated for memory symptoms by a specialist, it can trim the need for a spinal tap or a PET scan and can rule Alzheimer's out with real confidence. In the wrong setting, an anxious but symptom-free adult, a mail-order kit, or a primary-care visit with no cognitive assessment behind it, the same number invites false reassurance and false alarm. A biomarker describes pathology. It does not, by itself, describe your future or decide your care.