New blood tests can now show whether a person is likely to have amyloid plaques in the brain, a hallmark of Alzheimer's disease. Most measure a protein called p-tau217. Some combine it with amyloid-beta proteins. In the US, the Food and Drug Administration (FDA) cleared the first such test in May 2025. By August 2026 it had cleared four. Two p-tau217 tests received European CE marks in May 2026, and a third in July 2026.
These tests are for adults who already have memory or thinking symptoms, as part of a full assessment. They are not screening tests for healthy people. A positive result means amyloid is likely. It does not, on its own, mean your symptoms are caused by Alzheimer's, or that you will develop dementia. Some results land in an in-between zone and need a scan or spinal fluid test to settle.
This guide covers what has been approved, how accurate the tests are, and what they cannot tell you.
Alzheimer's disease involves two proteins: amyloid-beta, which clumps into plaques, and tau, which forms tangles. Our article on what happens in the brain in Alzheimer's explains both.
Until recently, checking for amyloid in a living person meant an amyloid PET scan or a lumbar puncture (spinal tap). The blood tests measure related proteins in plasma instead:
The FDA says the first cleared test's result "is correlated to the presence or absence of amyloid plaques". That is what these tests estimate: amyloid in the brain at the time of the test.
Sources: FDA, Roche on pTau181, Roche on the CE mark, Roche on FDA clearance, Alzheimer Europe on Fujirebio, C2N and Beckman Coulter. The Alzheimer's Association confirmed Elecsys pTau217 as the fourth FDA-cleared blood test.
A CE mark means a test has met EU safety and performance rules and can be sold across the EU. Alzheimer Europe explains this. A CE mark does not mean a health service has started offering the test.
The FDA and the makers stress that none of these tests should be used on its own. The FDA and C2N also say plainly that they are not screening tests.
Accuracy is reported in different ways, and much of it comes from manufacturers. Here is what we found, with the population and comparison test for each.
Lumipulse ratio, FDA review. In 499 adults with cognitive impairment, results were compared with amyloid PET or spinal fluid tests. Of people with a positive blood result, 91.7% had amyloid. Of people with a negative result, 97.3% did not. Under 20% got an indeterminate result.
PrecivityAD2, company data. In 1,142 people with signs or symptoms of cognitive decline, C2N reports a 97.6% positive predictive value for a "Positive" result. It reports a 93.1% negative predictive value for a "Negative" result. The comparison was amyloid PET or spinal fluid. A "Likely Positive" result occurred in 17.3% of people, and only 77.3% of them had amyloid.
Elecsys pTau181, company data. In 312 people, in a lower-risk group meant to reflect primary care, Roche reports a 97.9% negative predictive value. This test is cleared to help rule amyloid out, not in.
Elecsys pTau217. Roche's releases describe "high agreement with amyloid PET imaging" but publish no figures. We have not reported any.
Independent research, Sweden. A 2024 JAMA study tested 1,213 patients with cognitive symptoms in primary and specialist care. Half had Alzheimer's pathology on spinal fluid testing. The blood test used mass spectrometry and combined p-tau217 with amyloid ratios. Analysed as samples arrived, it was right 88% of the time when positive and 90% when negative in primary care. In specialist care, both figures were 91%. GPs' own accuracy, after examination, cognitive testing and a CT scan, was 61%.
Independent research, Ireland. At the Tallaght Institute of Memory and Cognition, researchers tested 148 people with mild cognitive impairment or mild dementia. The study compared plasma p-tau217 with spinal fluid. The area under the curve was 0.92. Using two thresholds might have avoided more than half of lumbar punctures, the authors estimated.
For pooled accuracy across studies, our Alzheimer's article covers a 2026 meta-analysis against PET.
How much a result tells you depends on how likely amyloid was before the test. The Alzheimer's Association guideline gives an example. Take a test with 90% sensitivity and 90% specificity:
In a low-risk group, many positives are false. In a high-risk group, more negatives are false. This is one reason guidelines tie testing to people with symptoms, assessed by a clinician first.
Several of these tests report three or more result bands, not just two:
In July 2025 the Alzheimer's Association published the first clinical practice guideline on these tests. It covers people with objective cognitive impairment, seen in specialist memory care. Its main points:
The guideline lists conditions that can affect results. These include severe chronic kidney disease, acute brain injury, motor neurone disease (ALS) and some medicines. It says primary care will get its own guideline later.
The Alzheimer's Association guideline "does not extend to cognitively unimpaired individuals". The FDA says the first test "is not intended as a screening or stand-alone diagnostic test".
Amyloid without symptoms is common. A meta-analysis of 2,914 cognitively normal people found amyloid in 10% at age 50 and 44% at age 90. For someone with no symptoms, a positive result would raise a question these tests cannot answer: what happens next, and when.
Roche states that safety and effectiveness "have not been established for predicting development of dementia". It says the same about monitoring treatment.
The FDA also spells out the harms of wrong results. A false positive could lead to the wrong diagnosis, distress and unnecessary treatment. A false negative could delay the right diagnosis.
Lecanemab and donanemab are antibody drugs that target amyloid. The US lecanemab label says doctors must "confirm the presence of amyloid beta pathology prior to initiating treatment". The FDA says, of the first blood test, that other evaluations or tests "should be used for determining treatment options".
Most people with symptoms would not qualify for these drugs anyway. A Swedish primary care study of 607 patients found 13% to 14% met the criteria. What these drugs achieve is covered in our Alzheimer's article, linked above. Whether they are available to you depends on where you live. Ask your specialist.
In February 2025, the Irish researchers behind the Tallaght study said the test "is not available currently in Ireland". Three p-tau217 tests have since received EU CE marks. We could not find a public announcement that the HSE now offers one routinely. If you are under a memory service, ask whether it uses blood biomarkers. The usual route to a memory service is a GP referral.
In the UK, p-tau217 is being studied in NHS memory services. The ADAPT trial covers up to 20 community memory clinics. It plans to include 1,100 people newly referred to memory services. It will look at whether giving people their result changes diagnosis, care and costs. We did not find evidence of routine NHS use outside research.
FDA-cleared tests are ordered by clinicians. Each has a set age and care setting, as in the list above.
A result can be hard news, even when expected. The Alzheimer Society of Ireland's helpline is 1800 341 341. If you feel overwhelmed or have thoughts of harming yourself, call Samaritans free on 116 123. It is open day and night in Ireland and the UK. In the US, call or text 988.
You can ask. But current guidance says the test should follow a full clinical assessment of symptoms, not replace it. In Ireland and the UK, it is not yet a routine GP test as far as we could find.
No. It means amyloid is likely present. Amyloid is common in older people without symptoms. Roche, for one, says its test has not been shown to predict who will develop dementia.
Not with these tests, if you have no symptoms. The Alzheimer's Association guideline does not cover cognitively unimpaired people, and the FDA says the first cleared test is not for screening. If you are worried, talk to your GP. Our guide to memory worries may help.
Some come close in people with symptoms, but not all. The Alzheimer's Association says many commercial tests miss its thresholds. Intermediate results still need PET or spinal fluid testing.
This article is general information, not medical advice. It cannot tell you whether a test is right for you. Talk to your GP or memory specialist.
Reviewed by: Proco Team
Last updated: November 2026