Prevention · August 2026
Five ways to protect your brain that are not “eat well and exercise”
Everyone has heard the obvious two. These are the ones that surprise people — and one of them is a vaccine.
The Lancet Commission's 2024 report identified fourteen modifiable risk factors that, if eliminated, might prevent nearly half of dementia cases worldwide. Some of the fourteen are the ones you would guess. Several are not, and those are the ones worth talking about.
1. Get the shingles vaccine. This is the most striking recent finding in the field, and it comes from an unusually clean study design. In Wales, eligibility for the shingles vaccine was set by exact date of birth: people born on or after 2 September 1933 could have it, people born days earlier could not, ever. That is a natural experiment — two groups who differ in essentially nothing except whether they were allowed the shot.
Vaccination reduced the probability of a new dementia diagnosis by 3.5 percentage points over seven years — a 20% relative reduction. The effect was stronger in women.
Eyting et al., Nature, 2025
Most claims of this kind are correlations, and correlations about vaccines and health behaviour are notoriously unreliable — people who get vaccinated tend to do other healthy things. The date-of-birth cutoff removes that problem. It is not proof of mechanism, but it is far stronger than the usual observational evidence.
2. Treat hearing loss. And get your eyes checked. Both hearing loss and vision loss are on the Lancet fourteen. They are among the easiest to act on, and among the least likely to come up in an appointment that is ostensibly about memory. If someone is withdrawing from conversation, the first question worth asking is whether they can hear it.
3. Move — and lift, not only walk. The U.S. POINTER trial is the first large randomised US trial to show a structured lifestyle programme protecting cognitive function. Its actual exercise prescription was 30–35 minutes of moderate-to-intense aerobic activity four times a week, plus strength and flexibility work twice a week. The strength half is the part people skip. Participants in the structured group finished with cognitive scores equivalent to people up to two years younger.
4. Mind the vascular set in midlife. High blood pressure, diabetes, high cholesterol, smoking. In the ARIC cohort — 12,409 adults measured between 45 and 65 and followed an average of 26 years — people with none of high blood pressure, diabetes or smoking went about 30 years before dementia appeared. People with all three went about 17. A gap of nearly thirteen dementia-free years.
5. Reduce air-pollution exposure. Also one of the fourteen, and the one nobody names when asked to list a dementia risk factor. Treat it as a vascular exposure, not only a respiratory one.
And one we will not oversell. “Good stress versus bad stress,” novelty-seeking, particular brain-training regimens: the mechanisms are plausible and the trial evidence is thin. Cognitive and social engagement is on the Lancet list. A specific puzzle is not. We would rather say that than let one weak item devalue the other five.
Sources. Eyting M, et al. A natural experiment on the effect of herpes zoster vaccination on dementia.
Nature 2025.
nature.com ·
Lancet Commission 2024 modifiable risk factors and U.S. POINTER, both via Alzheimer's Association,
2026 Alzheimer's Disease Facts and Figures, pages 19 and 98–100 ·
ARIC midlife vascular risk,
Neurology Open Access 2026 ·
Popular summary: World Economic Forum,
“Dementia is rising fast — 5 surprising ways to protect your brain.”
Early detection · August 2026
The twenty-year head start nobody is using
Alzheimer's begins two decades before symptoms. The tests that can see it are only allowed once symptoms show. That gap is not a technicality — it is the whole problem.
The Alzheimer's Association states it plainly on page 4 of its 2026 report: “Alzheimer's begins 20 years or more before memory loss and other symptoms develop.” The report goes further on page 8, timing each brain change separately in people whose age of onset is genetically predictable:
- Markers of damaged neurons rise a mean of 22 years before symptoms
- Beta-amyloid rises up to 22 years before — mean 18.9
- Abnormally folded tau increases as early as two decades before tangles appear
- The brain's glucose metabolism falls up to 18 years before
- Brain atrophy begins up to 13 years before
So the warning is long. And the blood tests now cleared to detect that biology are indicated for people already showing signs, symptoms or complaints of cognitive decline. The biology is findable years before anyone is permitted to look for it, and the thing that unlocks looking is documented symptoms.
Which is where it falls apart. Among Americans 45 and older who noticed their own memory or thinking getting worse, 54% never told a health care professional. Of older Americans living with mild cognitive impairment — the stage the Association says is where “treatment and planning may be most effective” — only 8% receive a diagnosis.
It is not that clinicians are not trying. The Association's own workforce chapter is blunt about why:
“Cognitive assessments take time and training to administer, interpret, document and follow up on, which can make them hard to implement in primary care practice settings.”
2026 Alzheimer's Disease Facts and Figures, page 63
Thirty-nine percent of primary care physicians say they are never or only sometimes comfortable making a dementia diagnosis. Half say the profession is not prepared for the numbers coming.
None of that is fixed by a better test. It is fixed by someone writing things down, on a schedule, before the appointment — which is a thing a family can do and a clinic largely cannot.
Sources. Alzheimer's Association,
2026 Alzheimer's Disease Facts and Figures, pages 4, 8, 28, 29, 63 and 64.
doi.org/10.1002/alz.71345
The IQCODE is the instrument built for exactly this — sixteen questions answered by someone who knows the person well, about change over ten years.
See how it works and how it is scored
Living with it · August 2026
What actually slows it down after a diagnosis
The honest version: no lifestyle change stops Alzheimer's. Several change how the years go, and one of them is boringly practical.
People ask what to do after a diagnosis, and the answers online tend to swing between miracle and nothing. The useful middle is that a set of ordinary things have real evidence behind them for function, mood and independence — and that they overlap almost entirely with what protects the heart.
- Physical activity. At least 150 minutes a week of moderate aerobic activity — brisk walking, swimming — plus strength work twice weekly, which is the part POINTER included and most advice omits.
- Diet. Vegetables, fruit, whole grains, olive oil, fish; less sugar and less processed food. POINTER used the MIND diet, a Mediterranean–DASH hybrid built for this purpose.
- Mental and social engagement. Not a specific puzzle — conversation, learning, and staying in company. Infrequent social contact is one of the Lancet fourteen.
- Sleep, and get apnoea treated. Seven to eight hours. Untreated sleep apnoea is one of the conditions that most convincingly mimics cognitive decline, and treating it sometimes resolves what looked like dementia.
- Health monitoring. Regular check-ins on blood pressure, weight and lab results. This was one of POINTER's four domains, not an afterthought to the other three.
The part worth saying out loud: a number of things that look like dementia are not. The Association notes that subjective cognitive decline “is often an indicator of other, treatable conditions, such as sleep apnea, depression or anxiety.” Depression in an older adult mimics memory trouble closely. So does an anticholinergic sleep aid bought over the counter. So does an untreated thyroid, or a B12 level on the floor.
That is the case for measuring mood alongside thinking rather than instead of it — and for measuring both more than once. A single result cannot tell you which one moved. Two results a quarter apart can.
Safety · August 2026
Before anyone starts an antipsychotic, ask which dementia is suspected
One question, asked in the right order, and it has to come before the prescription rather than after it.
Agitation is the symptom families most often describe as the hardest. Until 2026 the only drug approved for agitation in Alzheimer's dementia was an antipsychotic — and every antipsychotic carries an FDA boxed warning for increased mortality in older adults with dementia-related psychosis.
On 30 April 2026, Auvelity (dextromethorphan and bupropion, from Axsome Therapeutics) became the first approved treatment for Alzheimer's agitation that is not an antipsychotic. In trials it substantially reduced the risk of agitation relapse, with twelve-month safety data showing no observed increase in falls, sedation or cognitive decline.
There is a second reason the question matters. In dementia with Lewy bodies, conventional antipsychotics can be dangerous — marked worsening of rigidity, sedation and confusion, and increased mortality risk. And no screening instrument can tell which form of dementia is present. The Association's own autopsy data makes the point starkly: of 447 people believed to have Alzheimer's dementia when they died, only 3% had those brain changes alone. 82% had Alzheimer's plus at least one other cause.
Mixed pathology is the norm. “Which one is it?” is often the wrong question — but “which one is suspected, before this medicine starts?” is the right one.
None of this is a reason to refuse a medicine. It is a reason to ask a question in the right order, and to bring a dated behavioural record to the appointment — because agitation has no laboratory value, and a record of when it started and how often it happens is the only evidence anyone has.
Sources. Alzheimer's Association coverage of the 30 April 2026 FDA approval of Auvelity for agitation associated with dementia due to Alzheimer's disease · boxed warnings and Lewy body sensitivity, and the 3% / 82% autopsy findings, 2026 Alzheimer's Disease Facts and Figures, page 8.