Can an Alzheimer’s blood test predict cognitive decline?

Yes, in part. In a pooled analysis of 8,582 older adults, higher blood levels of p-tau217 were linked to a higher risk of mild cognitive impairment or dementia, and the link was stronger in people who carry the APOE-ε4 gene. Among people who started with no thinking problems, differences between higher and lower levels showed up 3 to 4 years after the blood draw, before symptoms began.

The Alzheimer’s blood test in this study measured p-tau217, a form of the protein tau. It is a marker of the brain changes seen in Alzheimer’s disease, and it can rise before obvious symptoms appear. The question behind the study is simple. Two people can have the same p-tau217 level but different genes. Does knowing whether someone carries APOE-ε4 add anything to the blood result when estimating who will develop problems with memory and thinking, and when?

What the data show

When people first joined, those with higher p-tau217 levels had about 77% higher odds of already having cognitive impairment (odds ratio 1.77). The true figure is very likely somewhere between 42% and 120% higher. The link was stronger in APOE-ε4 carriers, who had about 125% higher odds (odds ratio 2.25), than in non-carriers, who had about 52% higher odds (odds ratio 1.52).

The more useful result comes from the 4,569 people who had no cognitive impairment at the start. For each standard step up in p-tau217 (one standard deviation, a statistical yardstick for how spread out the levels were), the risk of later developing cognitive impairment rose about 41% (hazard ratio 1.41). That increase is very likely between 22% and 64%. In APOE-ε4 carriers, each step meant about 76% higher risk (hazard ratio 1.76), compared with about 26% higher risk in non-carriers (hazard ratio 1.26). All of these are relative comparisons between people with higher and lower levels.

Why the gene changes the timeline

The study also looked at when problems showed up, not only whether they did. Each one-step rise in p-tau217 meant the time until cognitive impairment was about 24% shorter in APOE-ε4 carriers, compared with 13% shorter in non-carriers. In other words, the same rise in the blood marker pointed to a faster timeline when the gene was present.

People with different p-tau217 levels began to separate 3 to 4 years after the blood test, and that separation came before symptoms started. So the blood marker carried information about future risk years before anyone had symptoms, and the gene result helped sharpen that forecast. That timing is what sets this apart from a test that only confirms a problem after it is obvious.

Dr. Kumar’s Take

A single lab number rarely tells the whole story, and this study shows why. A p-tau217 result means something different depending on your genes. For an APOE-ε4 carrier, the same rise in the marker went with a much larger jump in risk and a shorter runway to symptoms than it did for a non-carrier.

I see two sides to this. A blood draw plus a gene result that can flag risk years before symptoms is a real step forward from waiting until memory loss is obvious. But knowing your risk only helps if you can act on it. The authors frame the value as guiding monitoring and treatment in carefully selected at-risk people, and I think that framing is right. This is a tool for people who already have a reason to worry, such as a family history of dementia, not a screening test for everyone over 60.

How the study was done

Researchers pooled individual records from seven prospective cohorts, meaning groups of people followed forward over time. The cohorts were in Canada, the Dominican Republic, and the United States, and data were collected from 1992 to 2025. Participants came through academic research centers and community settings. Some had no thinking problems but were at higher risk of dementia, and others already had mild cognitive impairment or dementia. Everyone included had a p-tau217 blood measurement, a clinical assessment, and an APOE gene result. The average age was 70, and about two in three were women (65.9%). The team used standard survival statistics and a machine learning method called random survival forests to judge timing and how well the marker predicted outcomes.

Limits and caveats

This was an observational study. It shows that higher p-tau217 goes along with higher risk, not that the protein causes the decline or that lowering it would help. The cohorts were recruited in different ways across more than three decades, which can add noise when records are pooled. The main outcome combined mild cognitive impairment and dementia, so a “case” here could range from mild changes to full dementia. The authors say the test “might help,” which is appropriately modest language. Several authors also report consulting fees or grants from drug and diagnostics companies, and one author reports a provisional patent application.

Practical takeaways

If you have a family history of dementia or know you carry APOE-ε4, the researchers see a p-tau217 blood test read alongside your gene result as a way to estimate risk and timing, so it is a fair topic to raise with your doctor.
A higher p-tau217 level signals higher risk in the years ahead, not a diagnosis, because this study measured risk across groups rather than certainty for any one person.
Read the blood result and the gene result together, because in this study the same p-tau217 rise pointed to a faster timeline in APOE-ε4 carriers than in non-carriers.

FAQs

What is p-tau217, and why can it be measured before symptoms?

P-tau217 is a form of the protein tau with a phosphate group attached at one specific spot, called threonine 217. It is measured in plasma, the liquid part of the blood. The researchers describe it as a marker of the brain changes of Alzheimer’s disease, and they note it can rise before obvious symptoms occur. That early rise is what makes it useful: it offers a look at risk while a person still feels fine. In this study, it was treated as a continuous number, so the analysis tracked how risk changed as levels went up rather than sorting people into “positive” and “negative.”

Did the study include people from different racial and ethnic groups?

Yes. Of the 8,582 people in the main analysis, 1,380 (16.1%) were Black, 4,056 (47.3%) were non-Hispanic White, and 3,146 (36.7%) were of Hispanic or other ethnic origin. Recruitment took place in the Dominican Republic as well as Canada and the United States. The risk figures in this article describe the combined group. That mix makes the pooled result more relevant to a wider range of people than a study drawn from a single population.

Why were only about half the participants used to predict future decline?

Of 8,873 people screened with blood and clinical data, 8,582 were included in the first analysis, which compared p-tau217 levels with thinking status at a single point in time. Only 4,569 (53.2%) were used to study future decline. To be in that group, a person had to have no cognitive impairment at the start and at least one follow-up visit. That design makes sense, because you cannot measure who develops a problem if they already have it. It also means the prediction figures come from people who were still unimpaired when their blood was drawn, which is the group that matters most to anyone considering a test before symptoms.

Bottom line

In this pooled study of 8,582 older adults, higher levels of the blood marker p-tau217 were linked to a higher risk of mild cognitive impairment or dementia. Among people without symptoms, each step up meant about 41% higher risk of developing it. The APOE-ε4 gene made that link stronger and the timeline shorter, and differences appeared 3 to 4 years after testing, before symptoms. An Alzheimer’s blood test read alongside a gene result may help at-risk people and their doctors plan monitoring, but it estimates risk rather than making a diagnosis.

Read the full study

Can an Alzheimer's blood test predict cognitive decline?