Two modifiable habits at midlife — staying physically active and keeping blood sugar under control — may help protect the brain decades later.
Scientists at UT Health San Antonio followed 402 participants and found that higher physical activity and better-controlled blood sugar in midlife were associated with slower cognitive decline years afterward. The findings were published in Alzheimer’s & Dementia.
Why midlife is the interesting window
The choice of midlife rather than old age reflects a shift in how dementia risk is understood.
The pathological processes underlying dementia begin decades before symptoms. By the time cognitive decline is apparent, substantial and largely irreversible damage has accumulated, which is part of why treatments given at that stage achieve so little.
That reframes prevention as something belonging to the forties and fifties rather than the seventies. It also means studies must follow people for many years to link an exposure at midlife to an outcome much later — which is why such studies are rare and why existing cohorts are valuable.
Why these two factors specifically
Both connect to brain health through vascular and metabolic routes that are increasingly well characterised.
Blood sugar control matters because chronically elevated glucose damages small blood vessels, including those supplying the brain. Diabetes is an established dementia risk factor, and the relationship appears to extend below the diabetic threshold — suggesting a gradient rather than a cutoff.
Physical activity operates through several plausible mechanisms: improving vascular health and cerebral blood flow, improving insulin sensitivity, reducing inflammation, and possibly stimulating factors supporting neuronal maintenance. It is also correlated with almost everything else that is good for people, which is precisely the interpretive problem.
The confounding, stated plainly
This is an observational study, so it can show links but cannot prove that exercise or blood-sugar control directly cause better brain ageing.
People who exercise and manage their metabolic health differ in other ways that influence cognition. They tend to have more education, higher income, better access to healthcare, less smoking, better sleep, more social engagement and fewer untreated conditions — and education in particular is one of the strongest known predictors of cognitive resilience.
Statistical adjustment handles what is measured. It cannot handle the general tendency toward health-promoting behaviour that no variable captures directly.
The reverse-causation possibility
There is a subtler problem specific to this design, and it is worth naming.
Early neurodegenerative change can affect behaviour before it affects test scores. Apathy, reduced initiative and subtle executive difficulty appear early, and someone in the earliest preclinical phase may exercise less and manage their diabetes less consistently because of that.
On that reading, low midlife activity is an early symptom rather than a cause — and studies measuring exposure at midlife and outcomes later cannot easily distinguish the two.
Why the finding is still worth having
The results add to a large and consistent body of evidence tying cardiovascular and metabolic health to brain health — and consistency across many studies with different designs and populations does carry weight, even when each individually has the same limitations.
More importantly, both factors are modifiable. Most established dementia risk factors are not: age, genetics, family history. A risk factor that can be changed is worth pursuing even at lower confidence, because the cost of acting is low and the alternative is doing nothing.
Physical activity and glucose control are also independently justified. Neither requires a dementia benefit to be worth doing — the cardiovascular case alone is overwhelming — so acting on this evidence carries essentially no downside risk.
What would resolve it
Randomised trials of multidomain lifestyle intervention have been conducted and produced mixed results, generally showing improvement on cognitive testing over a few years rather than demonstrated dementia prevention.
The design problem is severe: preventing dementia requires intervening in midlife and following people for decades, at a scale and duration no trial has managed. Observational evidence may remain the best available for a long time, which argues for reading it carefully rather than dismissing it.
How the researchers frame it
They present the findings as further support for treating heart-and-metabolism health as brain health — without promising that any single habit prevents dementia.
The modifiable risk factor framework
This study contributes to a body of work that has been assembled into a formal framework, and the framework is worth knowing.
Major reviews have identified a set of potentially modifiable dementia risk factors spanning the life course — including education in early life, hearing loss, hypertension, obesity, physical inactivity, diabetes, smoking, excessive alcohol, depression, social isolation, air pollution and traumatic brain injury — and estimated that addressing all of them could prevent or delay a substantial proportion of dementia cases.
Two things about that estimate deserve emphasis. It is derived from observational associations, so it inherits every causal limitation of the underlying evidence and should be read as an upper bound on what is achievable rather than a target.
And several factors cluster in the same people, so they cannot simply be added together — someone who is inactive is more likely to have diabetes, hypertension and obesity, and treating one may capture much of the benefit attributed to the others.
That framing is appropriately calibrated, and it is the useful takeaway: the things already recommended for cardiovascular reasons appear to matter for the brain too, and midlife is when they matter most. This summarises observational research and is not medical advice.