The Modifiable Share

Somewhere between the fatalism of "dementia is genetic" and the optimism of "dementia is optional" sits a number: roughly 40–45%. That is the share of dementia the Lancet Commission's arithmetic estimates is potentially preventable through modifiable risk factors. This page explains how that number is built, what the fourteen factors are worth, and — just as important — what the arithmetic cannot tell you about your own odds.

The Question Behind the Number

The brain-aging primer ends where this page begins: with the question of how much of the decline is fate and how much is choice. The most rigorous answer on record comes from the Lancet Commission on dementia prevention, intervention, and care, which has twice assembled the world's cohort evidence into a single arithmetic — first with 12 risk factors (Livingston et al., The Lancet, 2020), then with 14 (Livingston et al., The Lancet, 2024). The Alzheimer's prevention topic owns the deep dive on each factor and what to do about it; this page owns the arithmetic itself — how the number is built, what it is worth, and where it stops being useful.

How the Commission Counts

The method is the population attributable fraction (PAF): for each factor, researchers combine how common it is with how strongly it associates with dementia, then adjust for the fact that factors overlap — the same person often carries several. A factor that is rare but potent can contribute as much as one that is common but weak. The output is a percentage: how much of all dementia in a population would, in principle, disappear if that factor were removed. Three properties of the method are worth stating plainly before the numbers:

• 🧮 It is population math. A PAF says nothing about any single person's odds. It describes the population-wide headroom for prevention.
• 🔗 It assumes the associations are causal — the strongest assumption in the whole enterprise, and true to varying degrees for different factors.
• ⏱️ It is a maximum, not a promise. "Potentially preventable" assumes full, perfect removal of every factor — a ceiling the real world approaches slowly.

The 2020 Ledger: Twelve Factors, ~40%

The 2020 Commission's headline: roughly 40% of dementia worldwide is potentially attributable to 12 modifiable factors, spread across the lifespan — from education in childhood to air pollution in late life. The weighted contributions, as published:

Two features of the 2020 ledger deserve a second look before the update below. First, the tail matters as a group: air pollution, diabetes, obesity, and excess alcohol weigh in at 1–2% each, which sounds trivial until you remember the method — a small PAF on a disease that afflicts tens of millions of people is still a large number of lives. Second, the factors distribute across the whole lifespan, which is the commission's quiet argument: prevention is not a retirement project. It begins with education in childhood and runs through midlife hearing, blood pressure, and lipids, all the way to late-life air quality and social contact.

The 2024 Update: Two New Factors

The 2024 revision added two factors and raised the estimate to roughly 45%. Untreated vision loss entered at about 2% — the mechanism being less visual input, less activity, and more isolation. High LDL cholesterol entered as one of the largest single factors at about 7% — a striking placement for a blood-lipid number, and a direct extension of the vascular half of aging this series documented earlier. The update also re-weighted the existing factors downward in places, because adding correlated factors forces the overlap adjustment to do more work — a small reminder that these numbers are estimates of a moving target, not measurements of a fixed one.

The Ledger, Line by Line

What the Arithmetic Cannot Say

The 40–45% number travels badly from the population to the person, for reasons worth stating as clearly as the number itself:

• 🔗 It is not your 40%. PAFs are group-level ceilings. An individual's addressable share could be near zero — or far higher — depending on which factors they actually carry.
• 🌀 Some factors run both directions. Depression and social withdrawal can be early symptoms of dementia as much as causes — the honest reading is "treat the depression and see," not "depression caused this."
• ⏳ Timing is part of the price. Midlife hypertension removed at 70 does not refund its 25-year toll in full. The arithmetic assumes factors are addressed in their window — the reason the quarterly audit starts decades early.
• 🧪 The trial evidence is thinner than the cohort evidence. The FINGER trial's multidomain package produced modest but real cognitive benefits in at-risk adults (Ngandu et al., The Lancet, 2015) — encouraging, but a long way from confirming the full 45% in practice.

From Percentages to Practice

If the arithmetic were a triage list, it would read like this: the hearing-and-vascular pair (hearing loss, LDL, hypertension, diabetes, smoking) carries the most weight and has the most actionable levers; the engagement pair (education, cognitive and social activity) builds the reserve that makes whatever remains cost less — the three accounts working in concert; and the isolation trio (depression, social isolation, sensory loss) is where the purpose-and-brain topic earns its place in this pillar. None of it requires an exotic supplement or a retreat. It requires the unglamorous compound interest of ordinary habits, started in the window when the arithmetic still works — which, for the midlife factors, is now.

Related Topics