Studio Toriumi
The most important thing first. No method reliably prevents dementia. Plenty of products claim otherwise; none of them have evidence that can promise certainty. That is still not a reason for pessimism. The Lancet Commission (2024) reports that around 45% of dementia cases worldwide are associated with 14 modifiable risk factors. This page deals only with the part you can move, in order of how much it matters.
Modifiable — the share the 14 factors relate to Age is the dominant factor, and it cannot be changed
Skip this and you will misread everything below. Three things.
Incidence climbs steeply with age. However good your habits are, living longer raises the risk. Nobody gets around that.
Genetics is similar. Carrying APOE ε4 raises the likelihood of Alzheimer's disease — but many carriers never develop it, and many non-carriers do. The probability shifts. Nothing is decided.
It is a population attributable fraction (PAF): the share of cases that would theoretically disappear if a factor were entirely removed from the whole population.
So it does not mean "treating my hearing loss cuts my risk by 7%". For an individual, the honest statement is only that the probability goes down somewhat. Be sceptical of any article or advert that blurs this distinction while quoting the numbers.
Most of the 14 come from observational studies. Is it that inactive people are more likely to develop dementia, or that early-stage disease makes people less active? Observation alone can't fully separate those — this is called reverse causation.
That said, some have now been tested in trials. The section on evidence strength further down separates what has been tested from what has only been observed. Not flattening those together is the point of this page.
Figures are population attributable fractions (Lancet Commission, 2024). Larger means more impact at the population level — not the size of the effect for you personally. They sum to roughly 45%.
Early life up to about 45
Less education
Cognitive reserve built early is thought to help compensate for later changes in the brain
Midlife about 45–65 — the window where action pays
Hearing loss
The largest single factor. Not hearing well means less conversation, and the input to the brain thins out
High LDL cholesterol
Newly added in 2024. Thought to act through vascular damage
Depression
Sometimes a prodromal symptom, sometimes a risk factor. Either way it warrants treatment
Traumatic brain injury
Repeated impacts matter most. Helmets and fall prevention are the practical response
Physical inactivity
The number looks modest, but exercise lowers blood pressure, glucose, weight and depression at the same time, so the real effect is larger
Diabetes
Sustained high blood glucose damages both vessels and nerves
Smoking
Risk begins falling from the point you stop. It is never too late
Hypertension (midlife)
Midlife control is what matters. Starting in old age makes the effect harder to detect
Obesity
Probably acting together with blood pressure, glucose and lipids rather than alone
Excessive alcohol
More than about 21 units a week. The claim that small amounts are protective has weakened considerably
Later life 65 onward
Social isolation
Talking to people is itself both a cognitive load and a source of input
Air pollution
Not fully avoidable individually, but where you live and how you ventilate move it somewhat
Untreated vision loss
Added in 2024. The point is not to leave treatable causes — cataracts above all — untreated
Plenty of studies link poor sleep and sleep apnoea to dementia, but sleep is not among the Lancet Commission's 14, because the direction of causation isn't settled — early changes may disrupt sleep rather than the other way round.
Sleep apnoea is still worth treating, and better sleep improves depression, weight and blood pressure, all of which are on the list. Worth doing — just not something we can claim is one of the 14. Keeping those separate is the policy here.
Set your age, then tick what applies. This produces no risk score. That is a clinician's job, and it isn't calculable for an individual anyway. What you get is the list of factors you can still move, ordered by impact. Nothing leaves this page.
Up to 45build now
45–65highest yield
65+maintain
Highlighted chips are the factors for your current life stage. Nothing here is harmed by starting early.
Everyone already knows about diet and exercise. These are the high-ranking ones that rarely come up.
Hearing aids almost never come up in conversations about dementia. Yet this is the biggest of the 14. Poor hearing means less conversation, then fewer outings, and it chains directly into social isolation (5%).
In 2023 the ACHIEVE trial reported that hearing aids did not significantly slow cognitive decline across the whole sample, but reduced decline by about 48% in the subgroup already at high risk. It does not work for everyone — but it is one of the few interventions with trial evidence at all.
LDL was newly added to the 14 in 2024. It shares something important with blood pressure: midlife is when control pays off. Lowering these only in old age makes the effect much harder to demonstrate.
For blood pressure, the SPRINT MIND trial found that intensive lowering reduced the incidence of mild cognitive impairment (the effect on dementia itself did not reach statistical significance). Again — one of the few areas with an actual trial.
Almost everything else here you can handle by yourself. This one you can't. And it collapses the moment you retire — the more your relationships run through work, the harder this hits later.
"Make friends" is too large a goal. It is enough to have a standing reason to see people once a week: a hobby group, something local, a place you're a regular. The recurring slot is the point, not the depth.
"Good for prevention" covers wildly different levels of proof. Most articles flatten them together. This one doesn't.
ACHIEVE trial. Slowed cognitive decline in a high-risk subgroup; no significant effect across the full sample. Both halves of that belong in the summary.
SPRINT MIND reduced mild cognitive impairment. The effect on dementia itself is not established.
Finland's FINGER trial: exercise, diet, cognitive training and vascular risk management together for two years produced a small but significant cognitive benefit. Grounds for thinking "several at once" beats "one thing".
Extremely consistent association, though reverse causation can't be fully excluded. No reason not to do it — it moves several other factors simultaneously.
Broadly supports the cognitive-reserve hypothesis. Whether it can be built through deliberate intervention is still being tested.
Well established for cardiovascular prevention. The dementia benefit is considered a reasonable extension of that.
Good observational results. But a randomised trial of the MIND diet published in 2023 found no difference between groups over three years. Temper expectations accordingly.
Many reported associations. Excluded from the 14 because the direction of causation is unresolved. Worth treating on general health grounds regardless.
The current view is that you get better at the trained task, but transfer to everyday cognition or to dementia incidence is doubtful. Play them if you enjoy them; don't build your strategy on them.
No supplement has been shown to prevent dementia. Ginkgo biloba failed in a large trial (GEM); vitamin E and omega-3 have not demonstrated preventive effect either. This is exactly why this site links no supplements at all.
Coconut oil, special waters, unusual electrical stimulation devices. Dementia combines high anxiety with hard-to-verify claims, which is precisely the environment these products grow in. Certainty in the wording is itself a warning sign.
Catching it early matters as much as prevention. At the stage of mild cognitive impairment (MCI), some cases return to normal, and the drugs that exist work differently early than late.
Forgetting things happens with age. The distinguishing pattern looks like this.
You don't need a specialist centre first.
Lecanemab and similar agents targeting the underlying pathology of Alzheimer's disease are now available. But they are drugs that slow progression — not cures, and not preventives. Eligibility is limited to early disease, side effects include brain swelling and microhaemorrhage, and regular imaging is required. Go through both the promise and the limits with the treating physician.
This field attracts products and claims that feed on anxiety. The words "guaranteed", "certain", "all you need" are on their own sufficient reason to doubt the evidence behind them.
What actually has support is unglamorous. Check your hearing. Measure your blood pressure. Walk. See people. Not exciting — but currently the surest hand available.
This is not medical guidance. Consult a physician for diagnosis and treatment. The figures come from published research and do not describe any individual's probability of developing dementia.
Research moves. In a few years both the count and the weights will have changed. When that happens, this page gets rewritten.
Principal sources — Livingston et al., Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission, The Lancet, 2024 / Lin et al., ACHIEVE trial, The Lancet, 2023 / SPRINT MIND Investigators, JAMA, 2019 / Ngandu et al., FINGER trial, The Lancet, 2015 / DeKosky et al., GEM study, JAMA, 2008