The WHO's new dementia guidelines (2026): what actually changed

Short answer: On 15 July 2026 the World Health Organization published the second edition of its guidelines on reducing the risk of cognitive decline and dementia — the first update since 2019. Two recommendations are genuinely new: reducing exposure to air pollution, and offering tailored multidomain programmes. Five further areas were newly reviewed — sleep, stroke, traumatic brain injury, vision impairment and HIV — but WHO judged the evidence insufficient to recommend specific interventions in those areas for the specific purpose of reducing dementia risk. The guidance is now organised around four things you can act on — healthy behaviours, managing health conditions, reducing environmental exposure, and multidomain programmes that tackle several factors at once. Just as importantly, WHO added an explicit don't bother list: vitamins B and E, omega-3 and multivitamins are not recommended for reducing dementia risk in people without a diagnosed deficiency. The headline figure — a population-level estimate that up to 45% of dementia worldwide is attributable to modifiable factors — stands.

This is educational, not medical advice. Several of the recommendations below concern medical conditions and medications — blood pressure, cholesterol, diabetes, hormone therapy, aspirin. Nothing here is a reason to start or stop a treatment. Those are conversations with your clinician.

Why a guideline update beats another study

Individual studies arrive weekly and contradict each other constantly — as anyone who has followed the coffee, alcohol or sugary- and diet-drinks literature can attest. A WHO guideline is a different object: what happens when a body with no product to sell reads the whole evidence base, grades it, and states what it is confident enough to recommend. The 2026 edition uses GRADE, the standard framework for rating certainty of evidence and strength of recommendation, which is why many of its recommendations carry an explicit certainty rating (EMJ summary).

That matters for reading it honestly: many recommendations are conditional, not strong. The evidence supports offering them, not guaranteeing them. Anyone quoting these guidelines as proof that a given habit will keep you from getting dementia is misreading them. The flip side is that a document graded this carefully will always lag the literature: the zoster-vaccine research, where large studies consistently link vaccination to lower dementia risk but no randomised trial has reported yet, is exactly the kind of live question that doesn't yet appear among the recommendations.

What's new since 2019

The 2019 edition covered familiar ground: physical activity, tobacco, diet, alcohol, cognitive training, weight, blood pressure, diabetes, cholesterol, hearing, depression and social activity. The second edition strengthens or updates the interventions carried over from 2019 — including cognitive activity, social activity, physical activity, healthy diet, tobacco cessation, reducing harmful alcohol use, and management of obesity, diabetes, hypertension, dyslipidaemia and hearing loss — and adds two genuinely new recommendations (WHO, 15 July 2026; FBHI summary):

• Air pollution — reducing exposure to household and ambient air pollution may reduce risk. This is the first environmental factor WHO has formally brought into dementia risk reduction.
• Tailored multidomain interventions — programmes that address several risk factors at once, offered as a conditional recommendation.

Five further areas were reviewed for the first time and did not produce recommendations: sleep, stroke, traumatic brain injury, vision impairment and HIV. In each case WHO judged the evidence insufficient to recommend specific interventions for the specific purpose of reducing dementia risk. (It reached the same conclusion for management of depression.) That is a statement about the strength of the trial evidence, not a finding that these things are irrelevant — and it changes nothing about the ordinary clinical reasons to look after them. We've covered two of them in their own right: sleep and dementia risk and vision loss and dementia.

Air pollution breaks the pattern, and is worth pausing on. Every other recommendation in the guidelines is something you do or manage; this one happens to you, and the meaningful levers are largely policy-level. Individually the options are modest — indoor air quality, avoiding exercise beside heavy traffic at peak times — but its inclusion signals that dementia risk is not purely a matter of personal discipline.

The four categories, in plain language

WHO groups the recommendations into four sets (WHO guidelines, second edition):

1 · Healthy behaviours. Physical activity, a healthy diet, cognitive activity, social activity, stopping tobacco use, reducing harmful alcohol use. The 2026 edition puts noticeably more emphasis on cognitive activity — learning new skills, mentally demanding pursuits — as distinct from commercial brain-training. (Our read on that distinction hasn't changed: do brain games actually work?) On diet, WHO stays at the level of the overall pattern and names no individual foods; the food-by-food evidence — which foods carry the signal, and how strong each one is — sits underneath that recommendation.

2 · Managing health conditions. Obesity, diabetes, hypertension, dyslipidaemia (high cholesterol) and hearing loss. Hearing aids may be offered as part of risk reduction. This is the category with the most robust evidence behind it, and — not coincidentally — the one that requires a clinician rather than a resolution. See blood pressure and dementia, blood sugar and dementia, cholesterol and dementia and hearing loss and dementia.

3 · Reducing environmental exposure. Air pollution.

4 · Multidomain interventions. Programmes that target three or more risk factors at once. This is the category that didn't really exist as a recommendation in 2019, and it may be the most important change in the document.

The most useful part: what WHO now recommends against

Guidelines are usually read for what they endorse. This one is more valuable for what it rules out.

WHO does not recommend, for the purpose of reducing dementia risk in people without a diagnosed deficiency (EMJ):

• Vitamin B supplements
• Vitamin E supplements
• Omega-3 polyunsaturated fatty acids
• Multivitamins and mineral supplements

The stated rationale is that there isn't evidence that potential benefits outweigh possible harms. WHO also advises against hormone replacement therapy, cholinesterase inhibitors, metformin and aspirin taken solely for dementia prevention.

That last group deserves care. Plenty of people take HRT, metformin or aspirin for good reasons, and the guidance is not telling them to stop — it is saying that dementia prevention is not a valid reason to start one. If you take any of these, that decision belongs with your prescriber, and it should be made on the indication it was prescribed for.

For the supplement aisle, though, the message is unusually blunt for a WHO document, and it aligns with what we've found every time we've examined a specific product — see do brain supplements actually work? and vitamin B12 and memory. Turmeric isn't on WHO's list either way, and its own evidence lands in the same place. The one genuine exception is treating an actual diagnosed deficiency, which is a medical matter, not a wellness purchase.

About that 45% figure

The guidelines retain the population-level estimate that up to 45% of dementia cases worldwide could be prevented or delayed by addressing modifiable risk factors — the figure popularised by the 2024 Lancet Commission and now carried into WHO guidance.

It is a real, carefully derived number, and it is routinely misread. Three things it does not mean:

• It is not a personal guarantee. It's a population-attributable fraction — an estimate of how much dementia across an entire population might be avoided if every one of these risk factors were eliminated everywhere. No individual gets a 45% discount for good behaviour.
• It doesn't mean 45% is achievable in practice. It's the ceiling under ideal conditions, not a forecast.
• It doesn't mean the other 55% is your fate. Age and genetics carry real weight, but "not modifiable" is not the same as "already decided."

The honest version is the useful one: a substantial share of dementia risk is influenced by things that can change, and which of those things matter most is different for each person. That last clause is where the guidelines stop and a personal assessment starts.

The quiet headline: structure helps you work the list

The fourth category — multidomain interventions — is the change most likely to matter for what you actually do on a Tuesday.

The strongest recent trial of this approach is US POINTER, a two-year randomised trial across five US sites with 2,111 older adults at risk of cognitive decline. Everyone was encouraged toward the same things — physical activity, better diet, cognitive and social engagement, cardiovascular monitoring. The difference was how: one group followed a structured programme with schedules, sessions, support and accountability; the other was self-guided.

Both improved. The structured group improved more on global cognition, consistently across age, sex, ethnicity, cardiovascular status and APOE4 genotype (Baker et al., JAMA, 2025; Alzheimer's Association summary). The effect sizes were modest — this is not a cure, and the self-guided arm did fine too.

But the direction is instructive, and it's what public-health advice keeps failing to act on. Being told the list is not the intervention. Working the list, with structure, is the intervention. Almost everyone reading this already knows exercise, sleep and blood pressure matter. Very few have a system for acting on the two or three that are actually loose in their own life.

What to do with this

• Get the clinical numbers measured. Blood pressure, blood sugar (HbA1c), cholesterol. Three of the four categories run through them, and you cannot manage what you haven't measured. Ask your clinician for these.
• Get hearing checked — hearing loss management is a WHO recommendation, and hearing aids may be offered as part of risk reduction. Get vision checked too: WHO judged the dementia-specific evidence not yet sufficient to recommend it for that purpose, but untreated vision loss is one of the 2024 Lancet Commission's 14 factors, and it is correctable.
• Don't buy supplements for this. WHO has now said plainly that vitamin B, vitamin E, omega-3 and multivitamins are not recommended for reducing dementia risk absent a diagnosed deficiency. If you already take one — for a deficiency, or because a clinician advised it — that stays a conversation with them, not something to stop on the basis of an article.
• Don't ignore bad sleep — but know what WHO actually said. Sleep was reviewed for the first time in this edition, and WHO judged the evidence not yet sufficient to recommend specific sleep interventions for reducing dementia risk. That is a statement about the trial evidence, not a reason to tolerate poor sleep: persistent snoring, gasping or unrefreshing sleep is worth raising with a clinician, and sleep apnoea is treatable.
• Pick a structure, not a resolution. Given US POINTER, structure added a real but modest increment on top of the behaviours themselves — the self-guided group still improved substantially. Structure is a multiplier on doing the work, not a substitute for it.

Where Solenna fits

These guidelines are a list of levers that apply to populations. They can't tell you which ones are loose in your life — and that's the gap between reading good advice and doing anything with it.

How we research and review our content →

This article is educational and is not medical advice, diagnosis, or treatment. The sources cited — including the WHO guidelines on risk reduction of cognitive decline and dementia (second edition, 15 July 2026), the 45% population-attributable estimate popularised by the 2024 Lancet Commission, and the US POINTER randomised trial reported in JAMA in 2025 — describe risk factors, population estimates and group-level results in general, not Solenna specifically; the 45% figure is a population estimate, not an individual guarantee. Solenna is not a medical device and does not diagnose, prevent, treat, or cure Alzheimer's disease or any form of dementia; risk reduction means lowering probability, not eliminating it, and no outcome is guaranteed. Do not start or stop any medication or supplement — including hormone replacement therapy, metformin, aspirin, cholinesterase inhibitors, vitamin B, vitamin E, omega-3 or multivitamins — on the basis of this article; speak with a qualified healthcare professional.