Adding an Alzheimer’s blood biomarker panel to a full specialist workup changed the suspected underlying cause of a patient’s memory problem in 23 of 450 people, about 5%, according to a prospective diagnostic study published August 3, 2026 in JAMA Network Open. The study ran from September 2023 to October 2024 at three academic memory clinics in the Netherlands, where results were revealed during weekly team meetings after the standard workup was complete.
Key takeaways
- The blood panel changed the syndrome diagnosis in 6 patients (1%) and the suspected underlying cause in 23 patients (5%).
- Doctors’ median confidence in their own diagnosis rose from 80% to 90%, though it fell for 68 patients (15%).
What the study found
According to the paper, “A total of 450 patients (mean [SD] age, 66 [10] years; 183 [41%] female; mean [SD] MMSE score, 25 [5]) were enrolled.” The blood panel measured three proteins: phosphorylated tau 181, glial fibrillary acidic protein, and neurofilament light chain.
After the blood results were disclosed, the syndrome diagnosis was revised in 6 patients (1%) and the primary suspected cause in 23 patients (5%).
Confidence moved more than diagnosis did. Median diagnostic confidence went from 80% to 90%, with less than a 1 in 1,000 chance of being coincidence (P less than .001). It increased for 207 patients (46%), stayed the same for 175 (39%), and decreased for 68 (15%).
About half the group, 234 patients (52%), also had both spinal fluid testing and an amyloid PET scan. Measured against those, the blood panel gave a high probability result for 76% of the patients who did have brain amyloid, and a low probability result for 85% of those who did not. The remainder landed in the middle: an intermediate result in 24% of the amyloid positive patients and 39% of the amyloid negative patients.
The authors concluded that the panel “was associated with altered etiologic diagnoses in a few patients and was associated with increased diagnostic confidence overall.”
Dr. Kumar’s take
Put this next to the news out of the Alzheimer’s Association International Conference this year, where a blood test in more than 1,300 patients in Sweden lifted primary care physicians’ diagnostic accuracy from 65% to 93%, and specialists’ accuracy from 74% to 89%. Both results are real, and together they say more than either alone. Even in the Dutch patients, who had already been through clinical assessment, neuropsychological testing, and a brain MRI, the panel still moved something: median confidence rose from 80% to 90%, it went up for 46% of patients, and the suspected cause changed for 5%.
The two tests are not the same product. The Swedish test measures amyloid beta and phosphorylated tau. The Dutch panel measured phosphorylated tau 181 plus two markers of brain injury and inflammation. Accuracy figures do not transfer between them, so do not assume the test your doctor can order is the one in the headline you read.
The intermediate zone is the practical limit here. Nearly 4 in 10 patients who truly had no brain amyloid still got a result that pointed neither way, and about 1 in 4 who did have amyloid got the same non-answer. That is not a broken test, it is one that sorts a large minority of people into “still needs more testing.”
The 15% whose doctors became less confident matter too. A test that shifts confidence far more than it shifts diagnoses can settle real uncertainty, or it can add certainty that is not earned. This study cannot tell those apart, because it measured what clinicians believed, not whether they were right.
For background on how these panels perform head to head, see my earlier write-up on blood-based ATN biomarkers in Alzheimer’s and frontotemporal dementia.
What it means for you
If a primary care doctor offers you an Alzheimer’s blood test for memory symptoms, that is reasonable, and it is where the evidence for a large accuracy gain is strongest. Treat a high probability result as a reason for a specialist referral, not as a diagnosis.
If you have already been through a full memory clinic evaluation, expect the blood test to confirm rather than change what you were told. On these numbers, that is what happened for 95% of patients.
Expect a gray-zone answer to be common. An intermediate result is not a failure of the test, and it does not mean you have Alzheimer’s disease.
This was an observational study at three academic centers in one country, and the patients were young for a dementia population, with a mean age of 66. Results in a community clinic with older patients may look different. Memory changes that worry you are still worth an evaluation, and reversible causes, including sleep loss, medications, and thyroid and vitamin problems, are worth ruling out first.