The World Health Organization's 2026 guidelines state that up to 45% of dementia risk is attributable to modifiable risk factors — things like hearing loss, high blood pressure, and physical inactivity rather than genes or age. That figure is a population-level theoretical ceiling, not a personal guarantee: it describes what might happen if an entire population fully eliminated 14 risk factors, which no population has ever done. For adults over 65, the practical content of the guidelines matters more than the headline number. WHO makes specific recommendations for this age group — cognitive training, social activity, hearing aids — and specific recommendations against things many older adults are already paying for, including supplements taken solely to prevent dementia.
Medical information disclaimer: This article is for general educational purposes only and does not provide medical advice, diagnosis, or treatment. Always consult a physician or other qualified health professional about symptoms, medications, tests, or treatment decisions.
Table of Contents
- Where the 45% number comes from
- What "up to 45%" does not promise
- WHO's specific advice for adults over 65
- How strong is the evidence behind each recommendation?
- Which factors carry the most weight late in life
- The scale behind the guidelines
- Frequently Asked Questions
Where the 45% number comes from
WHO published the second edition of *Risk reduction of cognitive decline and dementia* on 15 July 2026, under the release title "New WHO guidelines: up to 45% of dementia risk could be prevented or delayed". The guidelines cover 19 potentially modifiable risk factors plus multidomain interventions, and update earlier advice on diet, cognitive activity, social activity, hearing loss, obesity, hypertension and diabetes. The 45% itself did not originate with WHO. It comes from the 2024 Lancet standing Commission on dementia prevention (Livingston et al., *The Lancet*, 10 August 2024), which calculated a combined population attributable fraction of 45.3% across 14 modifiable factors.
That was an increase from the 12 factors named in 2020, with high midlife LDL cholesterol and untreated vision loss added. The phrase to hold onto is "population attributable fraction," or PAF. A PAF estimates the share of cases in a population that would not occur if a risk factor were absent from that population entirely. It is a statement about populations, not about any individual's odds.
What "up to 45%" does not promise
The Commission's method assumes each risk factor causes dementia and that each could be completely eliminated. The 2020 report says plainly that the authors "assumed a causal relationship" when running the calculation. Nobody eliminates hearing loss, hypertension, depression, air pollution and low education from a whole population, so 45% is a ceiling that describes the outer edge of what prevention could theoretically reach. The number also does not transfer between populations.
Research in *The Lancet Regional Health – Americas* in 2025 applied the 2024 relative risks to Brazil and produced a combined PAF of 59.5%, while an analysis of older American Indians found that education, diabetes, hypertension, depression and isolation alone accounted for 60% — with several Commission factors showing no association at all in that group. A PAF rises when a risk factor is more common locally, which is why the same arithmetic yields very different headline percentages. Two practical implications follow. A press release quoting 45% in one country and 60% in another is not describing a disagreement about biology; it is describing different populations. And a reader over 65 cannot subtract 45% from their own risk, because the figure was never calculated about them.
WHO's specific advice for adults over 65
WHO's recommendations in the over-65 range are narrower than the general prevention literature, and some are explicitly negative. According to the WHO release, the guidelines recommend: The supplement recommendation is worth sitting with. It applies to supplements taken for the specific purpose of preventing dementia, in the absence of a diagnosed deficiency.
It does not override a clinician's instruction to treat a confirmed B12 deficiency or any other diagnosed condition. The 2026 edition also adds a first-ever recommendation to reduce exposure to ambient and household air pollution, graded conditional. "Conditional" is WHO's weaker grade: it signals that the recommendation should apply to most people but that the balance of benefits, burdens and evidence quality leaves room for individual circumstances.
- Cognitive training or stimulation, for people with normal cognition *or* with mild cognitive impairment
- Social activity, on the same basis
- Hearing aids, which may be offered as part of risk reduction
- Against vitamin B, vitamin E, omega-3 or multivitamin supplements taken solely to prevent dementia, where no deficiency has been diagnosed
How strong is the evidence behind each recommendation?
Grades matter more than headlines here, because several 2026 recommendations rest on thin evidence. WHO's new recommendation against menopausal hormone therapy for reducing dementia risk in postmenopausal women aged 65 and older is rated "conditional" in strength on "very low" certainty of evidence, as recorded in the guideline's evidence tables. A conditional recommendation on very low certainty means WHO is advising a direction, not reporting settled proof. The strongest randomized evidence in this age group is US POINTER, published in *JAMA* on 28 July 2025.
It enrolled 2,111 sedentary adults at elevated risk across five US sites and compared a structured multidomain program — exercise, the MIND diet, cognitive and social engagement, cardiovascular monitoring — against a self-guided version of the same ideas. Over two years, the structured arm produced a statistically significant but modest advantage in global cognition. Read that result carefully. Both arms were doing something; the comparison was structure versus self-direction, not intervention versus nothing. "Statistically significant but modest" describes a real average difference in test scores, not a demonstrated reduction in dementia diagnoses.
Which factors carry the most weight late in life
Hearing loss carries the single largest late-life attributable share in the 2024 Commission, at roughly 7%. An independent 2024 meta-analysis in *Lancet Healthy Longevity* reached a similar figure, putting hearing loss at 7.2% and ranking low education highest at 9.3%. For someone already over 65, the two numbers point in different directions.
Low education is a midlife-and-earlier factor that has already been set. Hearing loss is the largest factor still open to action, and it is the one WHO addresses directly by saying hearing aids may be offered as part of risk reduction. An untreated hearing problem is the most actionable item on this list for most older adults. That makes the ordering of steps fairly clear for a reader in this age bracket:.
- Get hearing tested, and treat loss that is found
- Have blood pressure, diabetes and cholesterol managed with a clinician, since all three appear among the modifiable factors
- Keep cognitive and social activity up, which WHO recommends even after mild cognitive impairment is diagnosed
- Reduce indoor air pollution exposure where that is within reach
- Stop paying for supplements bought purely as dementia prevention, absent a diagnosed deficiency
The scale behind the guidelines
The reason WHO is issuing guidance at all is the size of the problem. The WHO dementia fact sheet records 57 million people living with dementia in 2021, with over 60% in low- and middle-income countries, and nearly 10 million new cases a year.
The cost figure explains who bears the weight. Dementia cost US$1.3 trillion globally in 2019, and roughly half of that was borne by unpaid family carers. That is also why the guidelines' population framing cuts both ways: a 45% ceiling that never materializes for one person can still represent an enormous shift if it moves even a few percentage points across 57 million cases and the families supporting them.
Frequently Asked Questions
Does the 45% figure mean I can cut my own dementia risk by 45%?
No. It is a population attributable fraction, which estimates the share of cases in a whole population that would not occur if 14 risk factors were eliminated entirely. It was not calculated as an individual risk reduction, and it assumes complete elimination of factors like hearing loss and hypertension.
Should I stop taking omega-3 or a multivitamin?
WHO advises against taking vitamin B, vitamin E, omega-3 or multivitamin supplements *solely* to prevent dementia when no deficiency has been diagnosed. If a clinician prescribed one to treat a confirmed deficiency or another condition, that is a separate question to raise with them.
Is it too late to act if I already have mild cognitive impairment?
WHO recommends cognitive training or stimulation and social activity for people with mild cognitive impairment as well as for people with normal cognition, so the advice extends past a diagnosis of MCI.
Why do different countries report different percentages?
Because a population attributable fraction depends on how common each risk factor is locally. Applying the 2024 relative risks to Brazil produced 59.5%, and an analysis of older American Indians attributed 60% to just five factors, with several Commission factors showing no association in that group.
You Might Also Like
- WHO’s Second Dementia Risk-Reduction Guideline in 2026: What the Evidence Supports and Does Not Promise for Adults Over 65
- Cross-Cultural Dementia Risk Programs After LatAm-FINGERS 2026: How Adults Over 65 Can Build a Realistic Brain Health Plan
- Subjective Cognitive Decline Data for Public Health Action in 2026: Access and Equity Questions for Adults in Midlife
