Blood proteins hint at who gains most from a lifestyle program for the aging brain
Researchers measured about 120 brain-related proteins in the blood of 1,158 people from the Finnish FINGER lifestyle trial. Some proteins were linked to faster decline in thinking, the program shifted a few of them, and some seemed to predict who benefited. An exploratory analysis, not yet peer-reviewed, whose findings need to be repeated.
The Finnish FINGER trial was the first large trial to show that a combined lifestyle program (diet, exercise, brain training and checks of blood pressure, cholesterol and blood sugar) can help thinking skills in older people at raised risk of dementia. But nobody knows exactly what the program changes in the body, or why some people benefit more than others.
A new analysis looks for clues in the blood. It is a preprint, not yet peer-reviewed, and the authors call it exploratory: many of its findings were weak enough that they could be chance, and they need to be confirmed in other groups of people.
What they did
FINGER enrolled 1,260 Finns aged 60 to 77 whom a dementia risk score placed at raised risk, but whose thinking was still around normal for their age. Half were randomly assigned to two years of the full program; the other half received regular health advice. Psychologists, who did not know who was in which group, tested memory, attention and speed of thinking at the start, after one year and after two.
Blood samples had been frozen. The team measured about 120 proteins in samples from 1,158 participants, using a very sensitive test aimed at proteins linked to the brain, the immune system, blood vessels and metabolism. An earlier analysis had looked at the best-known Alzheimer's markers in the same blood and found that the program did not change them (/post/2026-10-10-finger-lifestyle-trial-alzheimers-blood-markers/). This one looked at the rest of the panel.
What they found
- Some proteins went with faster decline, in both groups. People with higher starting levels of 35 proteins tended to do worse on the thinking tests over two years. Ten of these links held up after a correction for testing so many proteins at once, among them TREM2, a protein of the brain's immune cells, and TNF and IL18, two signals of inflammation.
- The program shifted eight proteins, though only at the less strict level of statistical significance, without the correction for multiple testing. The clearest were CRH, a stress hormone made in the brain, which rose more in the program group, and MME, an enzyme, which fell. The authors suggest the MME drop may reflect better metabolic health. The CRH rise could be a healthy adaptation of metabolism and the immune system, or a sign of more stress; they say this needs further study.
- Starting levels of eleven proteins seemed to change how much people gained. Higher levels of proteins such as KDR (a receptor involved in growing blood vessels), CRH, BDNF (which supports nerve cells) and CCL2 went with more benefit from the program; higher TREM2 went with less. A machine-learning analysis picked out CRH, PTN and CCL2 again.
- Among people who stuck to the program, 283 of 363 (78%) improved on the tests over two years, and 80 did not. Fifteen starting proteins differed between the two groups.
- Some further links differed between men and women, and between carriers and non-carriers of APOE4, the best-known gene variant that raises the risk of Alzheimer's disease.
What to keep in mind
The authors list the limits themselves:
- Exploratory. The protein analyses were not planned in advance around specific proteins, and many of the results on the program's effects did not survive the correction for multiple testing.
- Too few people for some questions. Even with more than 1,000 participants, the study had limited power to find differences between subgroups, such as men and women or APOE4 carriers.
- No outside check. The machine-learning models showed signs of overfitting (fitting chance patterns in the data), and they were not tested in a separate group of people.
- Finnish, at-risk volunteers only. The results may not apply to more diverse populations.
- The responder analysis is tricky. It only included people who stuck closely to the program, and sticking with it may itself be related to things that affect thinking.
Why it matters
Lifestyle programs like FINGER help thinking on average, but not everyone gains equally. If a blood test could one day show who is likely to benefit, and which biological pathways the program works through, prevention could be better targeted. This study does not deliver that test. It offers a short list of candidate proteins, involved in blood vessels, nerve support, the immune system and stress hormones, that other trials can now check.