Blood test flagged Alzheimer's-linked memory decline up to 10 years ahead

A large international study suggests that one blood test can help estimate which older adults will develop memory and thinking problems tied to Alzheimer’s disease, years before any symptoms show up.¹ Among people with very high levels of a blood marker called p-tau217, the estimated chance of developing cognitive impairment was 38% within five years and 78% within ten.¹ The researchers stressed that the 10-year figure rests on much less data than the 5-year one.¹

The work was led by scientists at the Mass General Brigham Neuroscience Institute, with teams on three continents. It was presented at the Alzheimer’s Association International Conference and published at the same time in the medical journal JAMA, according to a release from the hospital system.¹

What the researchers did

The study looked at p-tau217, short for phosphorylated tau 217. It is a form of the tau protein that can be measured in a blood sample and is linked to Alzheimer’s disease.¹ A test measuring this marker recently received clearance from US regulators, which has raised hopes that a simple blood draw could make Alzheimer’s detection easier to get.¹ The open question was whether such a test can say anything useful about people whose memory is still normal.¹

To find out, the team combined data from six earlier observational studies and clinical trials run in North America, Japan and Australia.¹ Together they included 2,684 older adults who showed no signs of cognitive impairment when they joined.¹ At the start, each person gave a blood sample that was tested for p-tau217 and had a PET brain scan, a type of imaging that can show protein buildup in the brain.¹ Memory and thinking were then checked every year.¹

The time span was long. Some people joined one of the studies as early as 2004, and the latest check-ups took place in 2025.¹ Over that period, about 478 participants went on to develop cognitive impairment.¹ When the team matched those outcomes against the original blood results, higher p-tau217 levels went with a higher chance of later problems.¹

The evidence so far

A key finding was that the blood marker added information beyond other known risk factors. It stayed linked to later impairment even after the researchers took into account amyloid plaques, the abnormal protein deposits that PET scans can detect.¹ It also added predictive value beyond known genetic risk, such as the APOE4 gene variant.¹

Lead author Rachel F. Buckley, a cognitive neuroscientist at the Mass General Brigham Neuroscience Institute, described what sets the work apart:

“What really sets this work apart is that it estimates an individual’s level of risk for cognitive impairment. We harmonized data across six cohorts, creating a large and varied data set, and still found consistent results showing how p-tau217 informs risk over time.”

In other words, the study tries to move from “this person has Alzheimer’s-type changes in the blood” to “this is roughly how likely they are to develop symptoms within a set number of years.”¹

Smaller studies point the same way. A two-year study of 76 people who felt their memory was slipping, published in the Journal of Neurology, found that those who tested positive for p-tau217 declined on a delayed story-recall task and a visual search task, and that p-tau217 tracked change more consistently than two other blood markers.² Earlier work presented by the Alzheimer’s Association noted that p-tau can build up before patients show signs of impairment, and that rising p-tau217 over time goes along with worsening thinking and brain shrinkage.³

What it does not show yet

This is observational research. It shows that a blood marker goes along with later decline in groups of people. It did not test any treatment, and it cannot show that knowing your result would change what happens to you.

The headline 10-year number is the least certain part. The researchers said substantially fewer observations were available for 10 years and beyond, which makes those longer-term predictions less reliable.¹ The estimates also apply to people with very high marker levels, not to everyone who has a test.¹

The team also cautioned that blood tests are not currently recommended for routine Alzheimer’s screening in people without symptoms.¹ The main reason they gave is that no treatment has yet been shown to change the course of the disease in otherwise healthy people found to be at higher risk.¹ In practice, they said, medical advice for such a person would not change based on the result today.¹

Who is behind it

The study was led from the Mass General Brigham Neuroscience Institute and drew on data from six studies and trials in North America, Japan and Australia.¹ The release Biohack reviewed does not name the funders or describe any company ties of the authors, and the full JAMA paper was not among the sources for this article.

The Alzheimer’s Association, which runs the conference where the results were presented, has funded some of the blood-test research it showcases. It said research reported at its 2024 conference was funded in part by the Association.³ Readers should keep in mind that the main source here is the research institution’s own summary of its work.

How it fits the field

Blood tests for Alzheimer’s have mostly been studied in people who already have symptoms. In a 2024 study of 1,213 patients with memory complaints, a commercial blood test that combines a p-tau217 measure with an amyloid measure was around 90% accurate at identifying Alzheimer’s disease.³ Without the test, primary care doctors in that study were 63% accurate and specialists 73%.³ At the time, the Alzheimer’s Association said doctors should use blood tests together with memory testing and other biomarkers, and should follow its recommendations on appropriate use.³

Claims about predicting Alzheimer’s years ahead are not new. In 2021, a company test based on a different protein, a variant of p53, was reported in a not-yet-peer-reviewed analysis to predict Alzheimer’s onset up to six years before diagnosis in a 482-person Australian cohort.⁴ That claim was promoted by the chair of the company’s own medical advisory board, who said further validation studies were underway.⁴ The new study differs in that it pools several independent groups and gives risk estimates over time, but it too will need others to check it.

Researchers are also testing p-tau217 as a way to judge drugs. A study in the journal Neurology, using 716 people from four groups, concluded that changes in blood p-tau217 could serve as a solid measure of whether an Alzheimer’s drug is working in trials.⁵ It estimated a trial would need at least 610 people per group to detect a 25% drug effect in cognitively normal people with amyloid buildup, and fewer if it picked people with intermediate amyloid levels.⁵ Biohack recently covered a drug that lowered brain plaque but did not slow decline, a reminder that changing a biomarker and helping patients are not the same thing.

What to watch next

• Whether other research groups reproduce the 5-year and 10-year risk estimates, especially with longer follow-up that firms up the 10-year number.¹
• Whether prevention trials in people with high p-tau217 show any treatment can delay symptoms, which the researchers named as the missing piece before screening healthy people makes sense.¹
• How professional guidelines handle blood tests for people without symptoms, given that current advice from the study team is not to use them that way.¹
• Whether p-tau217 becomes a standard way to measure drug effects in Alzheimer’s trials.⁵