Can I test my risk of Alzheimer's with a blood test (p-tau217)?
The p-tau217 blood test is an accurate tool if you have memory or cognitive complaints and a doctor is supervising you. As a self-test or screening tool without complaints, it is not suitable.
The p-tau217 blood test detects with high accuracy whether there are abnormal protein accumulations in your brain, a hallmark sign of Alzheimer's. Combined with a second marker (Aβ42), the test achieves an accuracy score of 0.963 to 0.966. That is comparable to a lumbar puncture or brain scan, but via a simple blood draw.1
In people who do have memory or cognitive complaints, diagnostic accuracy is 88 to 92 percent. For comparison: general practitioners without the test correctly identified only 61 percent of cases, while specialists managed no more than 73 percent. With the test, the accuracy of general practitioners rose to 91 percent. Useful, then, but it does not replace a full assessment by a specialist.2
Not every commercial blood test meets the quality threshold required. The 2025 Alzheimer's Association guidelines demand a minimum of 90 percent on both sensitivity and specificity at a single fixed cut-off value. Many tests on the market do not clear that bar. So ask your doctor explicitly whether the test being used meets those requirements.3
Do you want to test yourself without any complaints? There are serious caveats. The guidelines are intended for people who do have objective complaints in a medical setting, not for population-wide screening. In people without a diagnosis, the single test produced an inconclusive result in 31 percent of cases; the combination test still came in at 16.5 percent inconclusive. Moreover, such an accumulation in the brain does not automatically mean you will develop Alzheimer's: it can be present for years without consequences.1,3
The test also helps in distinguishing other forms of dementia, but it is not infallible. In short: have it requested and interpreted by a doctor, preferably a specialist, and never use it outside of a clinical context.
Based on two diagnostic validation studies (PMID 40156286, 39068545) and one consensus guideline from the Alzheimer's Association 2025 (PMID 40729527). The claim about distinguishing other forms of dementia had no confirmed PMID and has therefore been presented cautiously. No randomised studies are available on clinical outcomes such as disease course or survival.