News · Brain & Mental Health
A midlife blood marker tracked dementia 25 years ahead
Across several cohorts, a protein called GDF15 measured in midlife lined up with dementia risk 15 to 25 years later. The link was stronger for vascular dementia than for Alzheimer's disease.
Based on multiple cohorts with 15 to 25 year follow-up, plus genetic and cell-based analyses
- GDF15 measured in midlife tracked dementia risk over 15 to 25 years of follow-up.
- The link was stronger for vascular dementia than for Alzheimer's disease.
- A genetic analysis supported a causal role rather than the protein being a bystander.
- It also tracked small vessel disease and markers of immune activation in the brain.
- No test exists for the public, and nothing here says lowering it would help.
Most dementia biomarkers are measured late, in people already worried about their memory. A study in Science Advances went the other way, reaching back into blood drawn in midlife and asking what it predicted a quarter of a century later.
The answer was a protein called GDF15, a stress signal released by tissues that are struggling.
What was found
Across several groups of participants, plasma GDF15 is associated with greater dementia risk over 15-to 25-year follow-up periods when measured in midlife.
The length of that follow-up is what makes it worth attention. A marker measured five years before diagnosis is often just detecting disease that has already started. Twenty-five years is long enough that reverse causation becomes a much harder explanation.
The kind of dementia mattered too, with stronger associations observed for vascular dementia compared to Alzheimer’s disease (AD), vascular dementia being the form caused by damage to the brain’s small blood vessels rather than by the protein plaques of Alzheimer’s.
The case that it is more than a marker
Most blood markers turn out to be bystanders: they rise because someone is unwell, not because they are doing anything. The study makes a case that this one is different, using three separate lines.
First, genetics. Two-sample Mendelian randomization supported plasma GDF15’s mechanistic role in AD and related dementias. Mendelian randomization means using the fact that gene variants are inherited at random, so a variant that raises a protein for life acts a little like a randomized trial run by nature.
Second, brain imaging and spinal fluid. Cohort studies linked it to cerebral small vessel disease, diffuse neurodegeneration, phosphorylated tau, and to a spinal fluid signature suggesting immune activation.
Third, cells. Exposing cultured immune cells to the protein changed biological pathways that the researchers then showed were themselves predictive of dementia risk.
None of those is decisive alone. Together they are a more serious argument than a correlation.
What this does not settle
Whether anything can be done about it. A protein can sit squarely on a causal pathway and still be a bad place to intervene, and GDF15 has jobs in appetite and metabolism that make blocking it an awkward proposition.
There is also an interpretive trap worth naming. GDF15 rises with almost every form of ill health: obesity, kidney disease, heart failure, inflammation, and notably with metformin, one of the most prescribed drugs in the world. A marker that responds to everything is hard to attribute to anything.
And the vascular finding, which is the most actionable part, is a comparison between two dementia subtypes that are themselves difficult to separate in life. Mixed pathology is the rule rather than the exception in older brains.
Why it is interesting anyway
The useful shape of this result is not a test anyone can order. It is a pointer at a pathway.
If a midlife stress signal from the body’s tissues is tracking dementia decades later, and tracking the vascular form most strongly, then the story it tells is one where the brain’s long-term fate is bound up with the health of the rest of the body far earlier than symptoms suggest.
That is not a new idea. It is one more piece of evidence for it, measured unusually early.
People also ask
What is GDF15?
A protein released into the blood by tissues under stress, part hormone and part inflammatory signal. Levels rise with age, with obesity, with kidney and heart disease, and notably with metformin use. It is already one of the strongest single-protein markers of general ill health, which is part of why interpreting it is hard: almost everything raises it.
Why does the vascular result matter more than the Alzheimer's one?
Because it points at a mechanism that is already treatable. Vascular dementia arises from damage to the small blood vessels of the brain, which is driven by blood pressure, diabetes, smoking and cholesterol. If GDF15 is tracking that pathway rather than amyloid, it is pointing at risk factors medicine can already act on.
What is Mendelian randomization, and why does it strengthen this?
It is a way of testing causation using genetics. People inherit gene variants at random, so if variants that raise a protein for life also raise disease risk, that is harder to explain by lifestyle or reverse causation than an ordinary correlation. It is not proof, and it carries its own assumptions, but it is a meaningful step above an association.
Can I get my GDF15 tested?
Not usefully. It is a research measure, not an approved clinical test, and there is no threshold that means anything for an individual. It also rises with many ordinary conditions, so a high reading would not tell you what was causing it. This is general information about a research finding, not a suggestion to seek a test.
If it is causal, could lowering it prevent dementia?
That does not follow, and nobody has tested it. A protein can sit on a causal pathway and still be the wrong place to intervene, and GDF15 has roles in appetite and metabolism that make blocking it a complicated proposition. The value of this work is in pointing at a pathway, not at a drug target ready to use.