Around 45% of Dementia Linked to Modifiable Risks: What the Lancet Figure Means

The Lancet Commission on dementia estimates that around 40% of dementia worldwide is associated with 12 potentially modifiable risk factors, a figure its 2024 update raised to about 45% with 14 factors. That is a modelled estimate of how much dementia might be prevented or delayed across whole populations. It is not a verdict on why any one person developed dementia, and the share tends to be higher in the countries and communities with the fewest resources.

Population attributable fractions are often reported as if they were a personal checklist. They are not. Many people who did everything the list suggests still develop dementia, and many risk factors, such as limited schooling in childhood, untreated hearing loss and air pollution, are shaped far more by where and when a person grew up than by their own choices. Read properly, the figures say less about individual blame and more about public action: education, hearing and sight care, blood pressure control, cleaner air and fairer access to health services.

Around two in five dementia cases are linked to factors that could, in principle, change

The 2020 report of the Lancet Commission on dementia prevention, intervention and care named 12 modifiable risk factors across the life course: less education, hearing loss, traumatic brain injury, hypertension, excessive alcohol use, obesity, smoking, depression, social isolation, physical inactivity, diabetes and air pollution. Together, it concluded, they "account for around 40% of worldwide dementias, which consequently could theoretically be prevented or delayed", with an overall weighted population attributable fraction (PAF) of 39.7%^[1].

The largest single shares in that 2020 global estimate were hearing loss (8.2%), less education (7.1%) and smoking (5.2%), followed by depression (3.9%), social isolation (3.5%) and traumatic brain injury (3.4%)^[1]. The Commission's 2024 update added two more factors, untreated vision loss and high LDL cholesterol, and the share attributed to the 14 factors rose to about 45% globally^[2].

Two caveats matter. First, the Commission was explicit that "we have assumed a causal relationship between risk factors and dementia"^[1]. The PAF is the share of cases that would be expected to disappear if a factor were removed entirely and if the association were fully causal. Second, the factors overlap, so the Commission weights each one for its shared variance with the others. That is why the individual shares add up to the overall figure rather than to more. A PAF is a planning estimate of potential, not a count of cases already caused.

There is real-world support for the idea that risk can move at population level: the 2020 report noted that "the age-specific incidence of dementia has fallen in many countries, probably because of improvements in education, nutrition, health care, and lifestyle changes"^[1]. Rising numbers of people with dementia (see our piece on prevalence and projections) mostly reflect populations living longer rather than a rising risk at each age, at least in the high-income countries where incidence has been tracked over time.

The potential for prevention is largest where resources are smallest

The global figure hides wide differences. A 2025 ecological analysis in BMJ Public Health, which combined global health databases for adults, estimated that 34.03% of dementia cases worldwide were potentially preventable, ranging from 42.16% in South Asia, the highest of any region, to 23.25% in Western Europe, the lowest. Across countries, the overall PAF was negatively correlated with gross national income (r = −0.39): the poorer the country, the larger the preventable share tended to be^[4]. Its method differs from the Commission's, so its 34% should not be set directly against the Commission's 40–45%.

Survey-based estimates point the same way. Using 10/66 Dementia Research Group surveys of people aged 65 and over (mostly 2004–06), and nine of the Commission's factors, Mukadam and colleagues estimated weighted PAFs of 39.5% in China, 41.2% in India and 55.8% in their Latin American sample^[5]. Less education contributed 10.8%, 13.6% and 10.9% respectively, against 7.5% in the worldwide estimate of the time, and hypertension 6.4%, 4.0% and 9.3%, against 2.0%^[5]. The 2020 Commission concluded that "preventative interventions might yield the largest dementia reductions in LMIC"^[1].

In Brazil, a 2025 analysis of 9,949 people aged 50 and over in the ELSI-Brazil survey (2019–21) put the overall PAF for all 14 factors at 59.5%, one of the highest national estimates published. The largest contributions were less education (9.5%), untreated vision loss (9.2%) and midlife depression (6.3%), and the estimate was similar across racial groups and regions^[2]. A six-country analysis of the WHO SAGE surveys (19,278 adults aged 50 and over in China, Ghana, India, Mexico, Russia and South Africa; ten factors) estimated 37.6% overall, from 22.9% in Ghana to 38.3% in China, with less education contributing 8.3%^[6].

WHO counts over 60% of people with dementia as living in low- and middle-income countries^[3]. These are also the countries where the trial evidence on prevention is thinnest, a gap WHO's own guideline developers have flagged^[7].

Which factors matter most changes from country to country

The same 14 factors carry very different weight in different places, because how common each one is varies. That is the most useful thing national estimates offer: they show where public action would have the most potential.

Figures come from different surveys, years, age groups and factor lists, so the overall PAFs are not directly comparable and should not be ranked as a league table.

Japan's 2026 estimate shows how a country's profile shapes its priorities. Less education contributed only 1.5%, because few older adults in the data (7.0%) had less education, while hearing loss contributed 6.7% (prevalence 57.8%) and physical inactivity 6.0% (prevalence 75.9%)^[10]. The authors noted that "only 15% of those aware of their hearing impairment use hearing aids", and estimated that, on the model's assumptions, reducing all 14 factors by 10% could prevent around 208,000 cases, and by 20% around 407,000^[10]. They also cautioned that "a direct comparison" of their 38.9% with other countries' published estimates (Australia 40.6%, the US 41.0%, New Zealand 47.7%, Canada 49.2%, Italy 39.6%) "is problematic", because methods and factor lists differ^[10].

In Turkey, using the 2022 Turkey Health Survey, the overall weighted PAF was 49.9%^[12]. In Norway's HUNT4 70+ study of 9,745 people aged 70 and over, 1,525 of whom had dementia, the 14 factors accounted for an estimated 50.9%^[8]. In Brazil, vision loss sat almost level with education, a reminder that the 2024 additions are not marginal in every country^[2]. We look at hearing and sight more closely in our piece on hearing loss, vision loss and dementia.

Modifiable risk follows disadvantage: income, ethnicity and sex

Within countries, the preventable share is not spread evenly. Across the six SAGE countries, the overall PAF was 38.1% in the poorest income fifth against 30.9% in the richest^[6]. The pattern was not simple, though: the cardiometabolic share was 61% higher in the wealthiest group, so poorer and richer households face different mixes of risk, not just more or less of it^[6].

In UK Biobank, a cohort of 344,793 adults followed for a median of 14.1 years (5,515 of whom developed dementia), the 14 factors accounted for 40.70% overall. Less education was the largest preventable fraction in the low and medium socioeconomic groups (24.6% and 18.3%), while in the high socioeconomic group the largest single fraction was smoking (5.7%). Hypertension was linked to dementia only in the low socioeconomic group^[13]. UK Biobank is a volunteer cohort, not a representative sample, so these figures describe that cohort rather than the nation.

Differences by ethnicity in these studies are differences in how common the risk factors are, which reflects unequal circumstances across the life course, not anything inherent to a group. In New Zealand the weighted PAF was 47.6% for Europeans, 51.4% for Māori, 50.8% for Pacific peoples and 40.8% for Asian New Zealanders, with obesity contributing most for Māori (7.3%) and Pacific peoples (8.9%)^[9]. In the US, the share of Alzheimer's disease and related dementias associated with eight factors was 40% among Black adults, 39% among American Indian and Alaska Native adults, 34% among Hispanic adults, 29% among White adults and 16% among Asian adults^[11]. Our pieces on ethnicity and dementia and deprivation and dementia cover these gaps in more depth.

Sex differences run in different directions in different places. The estimated share was higher for women in Brazil (61.1% vs 58.2%) and Turkey (54.2% vs 44.3%), but higher for men in the US (35.9% vs 30.1%)^[2][12][11]. This is consistent with risk factors such as education reflecting the opportunities a generation of women or men were given, not something fixed by sex.

Trials show modest, mixed effects, so far

PAFs come from observational data. Whether changing a risk factor actually prevents dementia has to be tested in trials, and those results are more cautious. WHO's guideline developers described "the mismatch between observational and intervention-based evidence" as a central lesson of the first WHO guidelines, and concluded that "multidomain approaches targeting simultaneously multiple risk factors and tailored at both individual and population level, are likely to be most effective"^[7].

Two large trials illustrate this. Finland's FINGER trial randomised 1,260 people aged 60–77 at increased risk to a two-year programme of diet, exercise, cognitive training and vascular risk monitoring, or to a control group. The between-group difference in the change of the overall cognitive test score was 0.022 per year (95% CI 0.002–0.042, p = 0.030), a small but statistically significant benefit on cognition, not a measure of dementia cases prevented^[14].

The US ACHIEVE trial randomised 977 adults aged 70–84 with hearing loss to a hearing intervention or a health education control. "The hearing intervention did not reduce 3-year cognitive decline in the primary analysis of the total cohort" (difference 0.002, 95% CI −0.077 to 0.081). The effect differed between the two groups recruited, and the authors concluded it "might reduce cognitive change over 3 years in populations of older adults at increased risk for cognitive decline"^[15].

None of this undermines the case for treating hearing loss, controlling blood pressure or widening education: each is worthwhile for health and quality of life in its own right. It does mean that the 45% figure should be presented as a ceiling on potential under strong assumptions, not a promise, and that trials in low- and middle-income countries, where the potential appears largest, are badly needed^[7].

What the evidence asks of us

The risk-factor evidence is, above all, an argument for public policy. Most of the factors with the largest shares, such as education, hearing and vision care, blood pressure, air quality and physical activity, are set by schools, health systems, cities and workplaces long before anyone is old. What the figures ask of governments, health services and funders:

• Treat dementia risk reduction as public health, not personal responsibility. WHO's second edition of its risk-reduction guidelines (2026), "reflecting advances in research since the first edition published in 2019", covers physical, cognitive and social activity, tobacco, alcohol, diet, obesity, diabetes, hypertension, dyslipidaemia, hearing loss, depression, stroke, traumatic brain injury, vision impairment, sleep, HIV and air pollution, and stresses "the importance of integrated, equitable and multisectoral action"^[16][17].
• Close the hearing and sight gap. Hearing loss is one of the largest single shares in global and Japanese estimates, and untreated vision loss is one of the largest in Brazil's, yet in Japan only 15% of people aware of a hearing impairment use hearing aids^[1][10][2]. Affordable assessment, hearing aids and cataract and eye care belong in any dementia strategy.
• Target the populations with the most to gain. Preventable shares are higher in poorer countries, poorer households, and Māori, Pacific, Black and American Indian and Alaska Native communities in the studies above^{[4][6][9][11]}. Programmes that reach the already advantaged first will widen the gap.
• Fund trials and data where they are missing. WHO's guideline developers flagged "the limited evidence in low-middle income countries"^[7], where over 60% of people with dementia live^[3]. Countries working under WHO's Global action plan on the public health response to dementia, now extended to 2031, need national estimates of their own^[3].
• Never use prevention to stigmatise. A person living with dementia did not fail a checklist. Messages about risk should sit alongside, never in place of, investment in diagnosis, support and care for the millions already living with the condition.

Explore the rest of our dementia and health research library for diagnosis, care and support.

Language: this piece uses person-first language ("people living with dementia", "people with dementia"), as most dementia organisations and people with lived experience prefer, and uses clinical terms such as "Alzheimer's disease" or "mild cognitive impairment" only where a study, dataset or diagnosis refers to that specific condition.

Why we're publishing this

The "up to 45% preventable" headline is widely quoted and often misread, sometimes in ways that leave people living with dementia and their families feeling blamed. Dementia organisations need a clear, sourced account of what the figure means, how it differs between countries and communities, and what it asks of policymakers. The case it makes is for fairer public health, and it belongs to the people and organisations working for it.

How to read this data

• PAFs are modelled. Each combines how common a risk factor is in a population with a relative risk taken from published studies, and assumes the association is causal. They estimate potential prevention or delay at population level and say nothing about the cause of any individual's dementia.
• Weighted vs unweighted. The Commission's method weights each factor for its overlap with the others, so the shares add up to the overall figure. The figures quoted here are weighted unless stated.
• Estimates are not directly comparable. They use different surveys and cohorts, years (10/66 data from 2004–06; Brazil, Norway, New Zealand, the US and Turkey from 2017–22), age bands (50+, 65+, 70+ or all adults), and factor lists (8, 9, 10, 12 or 14 factors). The BMJ Public Health global estimate uses an ecological method on global databases. Differences between countries partly reflect method, not only risk.
• Survey vs cohort. Most national estimates use cross-sectional surveys, often with self-reported risk factors; the Norwegian and UK Biobank estimates come from cohorts with recorded dementia outcomes.
• The 2024 Commission. The 2024 report's full text gives a weighted PAF of 45.3% for all 14 factors (rounded to 45%), with hearing loss (7.0%) and high LDL cholesterol (6.9%) the largest single shares, followed by social isolation (4.6%) and less education (4.5%)^[18]. The body cites the figure as summarised in the peer-reviewed Brazilian analysis that applied the 2024 method^[2]; the 2020 figures come from the 2020 report's full text^[1].
• Trials measure cognition, not dementia. FINGER and ACHIEVE report changes in cognitive test scores over two to three years, not dementia diagnoses.

Use this data

Free to cite with attribution to Understanding Aging. Charities, carers' groups, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research and statistics for general information. It is not medical, care or financial advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's memory or health, please speak to a GP or another qualified health professional.

Sources

1. Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. The Lancet, 2020. europepmc.org/article/PMC/PMC7392084
2. Suemoto CK, Borelli WV, Calandri IL, et al. The potential for dementia prevention in Brazil: a population attributable fraction calculation for 14 modifiable risk factors. Lancet Regional Health – Americas, 2025. europepmc.org/article/PMC/PMC12355584
3. World Health Organization. Dementia: fact sheet. WHO, 2026. who.int/news-room/fact-sheets/detail/dementia
4. Giannoni-Luza S, Urso D, Logroscino G. Global, regional and national population attributable fraction of potentially modifiable risk factors for dementia in the adult population: an analysis using global health databases. BMJ Public Health, 2025. europepmc.org/article/MED/41394021
5. Mukadam N, Sommerlad A, Huntley J, Livingston G. Population attributable fractions for risk factors for dementia in low-income and middle-income countries: an analysis using cross-sectional survey data. Lancet Global Health, 2019. europepmc.org/article/MED/31000129
6. Feter N, Feter JS, Borelli WV, et al. Potentially modifiable risk factors for dementia in six low-income and middle-income countries: a multinational, population-based survey. Maturitas, 2024. europepmc.org/article/MED/38484600
7. Stephen R, Barbera M, Peters R, et al. Development of the first WHO guidelines for risk reduction of cognitive decline and dementia: lessons learned and future directions. Frontiers in Neurology, 2021. europepmc.org/article/MED/34764935
8. Ellingjord-Dale M, Strand BH, Skirbekk V, et al. Potentially modifiable risk factors for dementia in Norway (HUNT4 70+): a retrospective cohort study. Lancet Healthy Longevity, 2025. europepmc.org/article/MED/41475703
9. Ma'u E, Cullum S, Cheung G, Livingston G, Mukadam N. Differences in the potential for dementia prevention between major ethnic groups within one country: a cross sectional analysis of population attributable fraction of potentially modifiable risk factors in New Zealand. Lancet Regional Health – Western Pacific, 2021. europepmc.org/article/MED/34527984
10. Wasano K, Jørgensen K. The potential for dementia prevention in Japan: a population attributable fraction calculation for 14 modifiable risk factors and estimates of the impact of risk factor reductions. Lancet Regional Health – Western Pacific, 2026. europepmc.org/article/PMC/PMC12869287
11. Nianogo RA, Rosenwohl-Mack A, Yaffe K, et al. Risk factors associated with Alzheimer disease and related dementias by sex and race and ethnicity in the US. JAMA Neurology, 2022. europepmc.org/article/MED/35532912
12. Gulle BT, Tozduman B. Population attributable fraction of modifiable risk factors for dementia in Turkey. Dementia (London), 2026. europepmc.org/article/MED/40583126
13. Zhang X, Si K, Chen N, et al. Socioeconomic disparities in modifiable risk burden and dementia prevention: a prospective cohort study. Journal of Nutrition, Health and Aging, 2026. europepmc.org/article/MED/42349092
14. Ngandu T, Lehtisalo J, Solomon A, et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial. The Lancet, 2015. europepmc.org/article/MED/25771249
15. Lin FR, Pike JR, Albert MS, et al. Hearing intervention versus health education control to reduce cognitive decline in older adults with hearing loss in the USA (ACHIEVE): a multicentre, randomised controlled trial. The Lancet, 2023. europepmc.org/article/MED/37478886
16. World Health Organization. Risk reduction of cognitive decline and dementia: WHO guidelines, second edition. WHO, 2026. europepmc.org/article/MED/42555903
17. World Health Organization. Risk reduction of cognitive decline and dementia: WHO guidelines. WHO, 2019. who.int/publications/i/item/9789241550543
18. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. The Lancet, 2024. doi.org/10.1016/S0140-6736(24)01296-0