Fourteen. That's how many dementia risk factors a major 2024 analysis identified as modifiable — things a person or a health system can actually change, as opposed to age or genetics. Addressing all 14 across a lifetime could, in theory, prevent or delay about 45% of dementia cases worldwide.
The commission groups the 14 by the life stage where each matters most:
- Less education (early life)
- Hearing loss (midlife)
- High LDL cholesterol (midlife)
- Depression (midlife)
- Traumatic brain injury (midlife)
- Physical inactivity (midlife)
- Diabetes (midlife)
- Smoking (midlife)
- Hypertension (midlife)
- Obesity (midlife)
- Excessive alcohol use (midlife)
- Social isolation (later life)
- Air pollution (later life)
- Untreated vision loss (later life)
Where did the 45% figure come from?
A standing commission of dementia researchers, convened by The Lancet and led by UCL psychiatrist Gill Livingston, has been updating this list since 2017. The 2024 report combined decades of published evidence — relative risks from meta-analyses, paired with how common each risk factor is in the population — into a single estimate for each factor's contribution, then combined all 14 into one number. That combined number is a population attributable fraction: roughly how much dementia, across a whole population, traces back to these factors being present rather than absent.
It is not a prediction for any one person. It describes what happens in aggregate if a risk factor this common in a population were eliminated from it — a different question from "how much does quitting smoking lower my personal risk."
What's new on the list since the last update?
The commission's earlier reports, in 2017 and 2020, used 9 and then 12 factors. The 2024 update added two: untreated vision loss and high LDL cholesterol, after newer evidence tied both to dementia risk strongly enough to meet the commission's bar for inclusion. It also re-ran the numbers for hearing loss and depression using newer meta-analyses, which changed how large a share of cases each was estimated to explain.
Which single factor matters most?
Hearing loss is usually singled out because it is extremely common in older adults and detectable with a basic test. In the 2024 report's own estimate, uncorrected hearing loss carries a population attributable fraction of about 7% — among the largest of the 14 taken individually. That does not mean treating hearing loss prevents 7% of dementia cases outright; it's an estimate of how much dementia, across the population studied, is associated with hearing loss being present and untreated rather than absent.
Is this the same as saying dementia is preventable?
No — and the commission is explicit about the gap. Even if every one of the 14 factors were eliminated, the remaining 55% of risk traces to things this framework doesn't cover: age itself, genetics (including the APOE gene), and causes research hasn't identified yet. The 45% is also calculated assuming each factor's relationship to dementia is causal, which is better established for some factors (like hearing loss and hypertension) than for others — an assumption the commission itself flags as a limitation of any attributable-fraction calculation like this one.
What the research says
Livingston and colleagues' 2024 standing commission report, published in The Lancet, estimated that 14 modifiable risk factors together account for about 45% of dementia risk at the population level — up from the 40% estimated for 12 factors in the group's 2020 report.¹ Two factors, untreated vision loss and high LDL cholesterol, were added to the list for the first time based on evidence that had matured since the prior review. The estimate is built from an observational evidence base: the underlying studies linking each factor to dementia are cohort and case-control data, not randomized trials where researchers assigned people to develop or avoid hearing loss or high cholesterol, so causation for any single factor is inferred, not proven. The 45% also compounds uncertainty across 14 separate estimates, each with its own margin of error, which the published figure does not fully convey as a single headline number.
What you can do this week
- Get hearing checked if you suspect any loss. A baseline hearing test is quick, and treating hearing loss (including with hearing aids) is one of the more actionable items on this list — unlike education level or early-life factors, it's modifiable right now.
- Ask your clinician about your blood pressure and LDL cholesterol numbers specifically, not just whether they're "normal" — both are on the list, and both have established, treatable ranges a clinician can act on.
- If weight meets clinical criteria for treatment, say so to a clinician. Obesity and type 2 diabetes are both on the 14-factor list, and Pallas prescribes compounded semaglutide and tirzepatide when a US-licensed clinician decides it's appropriate. Whether treating weight with medication changes dementia risk has not been established; the list describes risk factors, not treatments that have been shown to prevent dementia.
No single action here is the one that matters — the commission's own framing is additive, the same way keeping more of five general health habits adds up to more years.
