Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission

This is not a single experiment but a comprehensive evidence synthesis — a systematic review and meta-analysis of hundreds of studies published since the Commission's 2020 report. The Commission examined:

**Risk factors:** 14 potentially modifiable factors across the lifespan (less education, hearing loss, high LDL cholesterol, depression, traumatic brain injury, physical inactivity, diabetes, smoking, hypertension, obesity, excessive alcohol, air pollution, social isolation, and uncorrected vision loss)

**Interventions:** Pharmacological treatments (cholinesterase inhibitors, anti-amyloid antibodies), non-pharmacological interventions (hearing aids, cognitive stimulation, multicomponent psychosocial programs), and population-level policies (air quality regulations, smoking bans, noise reduction)

**Outcomes:** Dementia incidence, cognitive decline, quality of life, hospitalisation rates, and carer burden

The Commission performed new meta-analyses specifically for hearing impairment and depression as risk factors for future dementia, and recalculated population attributable fractions (PAFs) — the proportion of dementia cases that would be eliminated if a risk factor were removed entirely.

The report synthesises evidence from multiple study types including:

**Prospective cohort studies:** Ranging from thousands to hundreds of thousands of participants, followed for 5–30+ years, across 40+ countries including high-income (UK, USA, Japan, Finland, France) and low/middle-income countries (India, Brazil, Uganda, China)

**Randomised controlled trials (RCTs):** Including the FINGER trial (n=1,260, Finland, aged 60–77), SPRINT MIND trial (n=9,361, USA, aged 50+), and multiple anti-amyloid antibody trials (n=1,000–2,000 per trial)

**Meta-analyses:** Pooling data from 10–50+ individual studies per risk factor

**Population-based surveys:** For prevalence estimates of risk factors globally (e.g., WHO Global Health Observatory, Global Burden of Disease studies)

Specific populations studied include community-dwelling older adults, people with mild cognitive impairment, people with diagnosed dementia, family carers, and individuals from minoritised ethnic groups and lower socioeconomic backgrounds.

The Commission used multiple measurement approaches depending on the risk factor and outcome:

**Dementia diagnosis:** Clinical assessment using DSM-5 or ICD-10 criteria, plus biomarker confirmation (amyloid PET, CSF amyloid/tau, blood biomarkers) in some studies

**Cognitive function:** Mini-Mental State Examination (MMSE, 0–30, lower = worse), Montreal Cognitive Assessment (MoCA, 0–30), and domain-specific tests (memory, executive function, processing speed)

**Risk factor measurement:** Self-report questionnaires (smoking, alcohol, physical activity), clinical measurements (blood pressure, BMI, LDL cholesterol), audiometry (hearing loss), eye charts (vision), and air pollution monitoring (PM2.5 levels)

**Biomarkers:** Amyloid PET scans, CSF amyloid-β42 and phosphorylated tau, blood-based biomarkers (p-tau217, p-tau181, neurofilament light)

**Quality of life:** DEMQOL (Dementia Quality of Life), EQ-5D, and carer-reported measures

**Study design:** This is an expert consensus report and evidence synthesis, not a single empirical study. The Commission conducted systematic reviews and meta-analyses of published literature, then used a Delphi consensus process among 24 international experts to agree on recommendations.

**Meta-analytic approach:** For each risk factor, the Commission calculated a new population attributable fraction (PAF) using the formula: PAF = (P × (RR − 1)) / (1 + (P × (RR − 1))), where P = prevalence of the risk factor and RR = relative risk from meta-analysed studies. They used random-effects models to pool effect sizes across studies.

**Key methodological features:**

**Lifecourse perspective:** Risk factors were assigned to specific life stages (early-life: education; mid-life: hearing loss, hypertension, obesity, high cholesterol; late-life: smoking, depression, social isolation, physical inactivity, diabetes, air pollution, alcohol)

**Adjustment for confounding:** Most included studies adjusted for age, sex, education, and socioeconomic status, though residual confounding remains possible

**Sensitivity analyses:** The Commission tested whether results held when excluding studies with high risk of bias

**Genetic risk analysis:** New evidence examined whether risk modification works equally in APOE ε4 carriers (highest genetic risk) versus non-carriers

**What this design can and cannot prove:**

**Can prove:** Associations between risk factors and dementia at the population level; the proportion of dementia cases theoretically attributable to each factor; consistency of findings across different populations and study designs

**Cannot prove:** Causality for individual risk factors (observational studies cannot rule out reverse causation or residual confounding); that intervening on a risk factor will definitely reduce dementia risk in any given individual; the exact magnitude of benefit from specific interventions (requires RCTs)

**Major strengths:** Comprehensive scope, rigorous meta-analytic methods, explicit handling of uncertainty, inclusion of LMIC data

**Major weaknesses:** PAFs assume risk factors are independent and causally linked to dementia (likely overestimates prevention potential); prevalence data for some risk factors are sparse in LMICs; the report cannot account for interactions between risk factors; some risk factors (e.g., depression) may be early symptoms rather than causes of dementia

**New modifiable risk factors (added since 2020):**

**Uncorrected vision loss:** PAF = 2.0% (95% CI 1.0–3.0%) — meaning 2% of dementia cases are attributable to vision impairment

**High LDL cholesterol:** PAF = 7.0% (95% CI 4.0–10.0%) — the largest single modifiable risk factor after education and hearing loss

**Updated PAFs for all 14 risk factors (ranked by contribution):**

Less education (early life): PAF = 5.0%

Hearing loss (mid-life): PAF = 7.0%

High LDL cholesterol (mid-life): PAF = 7.0%

Social isolation (late life): PAF = 4.0%

Traumatic brain injury (any age): PAF = 3.0%

Air pollution (late life): PAF = 3.0%

Hypertension (mid-life): PAF = 2.0%

Physical inactivity (late life): PAF = 2.0%

Diabetes (late life): PAF = 2.0%

Excessive alcohol (mid-life): PAF = 2.0%

Obesity (mid-life): PAF = 1.0%

Smoking (late life): PAF = 1.0%

Depression (late life): PAF = 4.0%

Uncorrected vision loss (late life): PAF = 2.0%

**Combined PAF:** 45% (95% CI 38–52%) of dementia cases are attributable to these 14 factors combined.

**Hearing aids and dementia risk:**

In a meta-analysis of 3 studies (n=137,484), hearing aid use was associated with 19% lower risk of dementia (HR 0.81, 95% CI 0.76–0.87) compared with untreated hearing loss

The effect was stronger in those with additional risk factors (HR 0.73, 95% CI 0.65–0.82)

**Anti-amyloid antibody treatments:**

Three phase 3 trials (donanemab, lecanemab, aducanumab) showed 27–35% slower cognitive decline over 18 months on the Clinical Dementia Rating-Sum of Boxes (CDR-SB) scale

Absolute benefit: approximately 0.5–1.0 points on CDR-SB (0–18 scale, higher = worse) — a small effect

Side effects: amyloid-related imaging abnormalities (ARIA) occurred in 12–40% of treated participants, with symptomatic ARIA in 2–8%

**Cholinesterase inhibitors:**

New evidence supports long-term benefit (2–3 years) in Alzheimer's disease and Lewy body dementia, not just short-term effects

Effect size: approximately 2–3 points on MMSE (0–30 scale) over 6–12 months

**Multicomponent interventions:**

The FINGER trial (2-year multidomain intervention: diet, exercise, cognitive training, vascular risk management) showed 25% improvement in cognitive function (composite z-score) compared with control (p=0.03)

Effect was similar in APOE ε4 carriers and non-carriers

**Population-level impact:** If all 14 risk factors were eliminated entirely, approximately 45% of dementia cases would not occur. In real-world terms, this means that for every 100 people who would develop dementia, 45 could potentially avoid it through risk factor modification.

**Individual-level impact:** The effect of any single risk factor is modest. For example:

Treating hearing loss reduces dementia risk by about 19% — meaning if 100 people with untreated hearing loss would develop dementia, about 81 would still develop it even with hearing aids

Lowering LDL cholesterol from high to normal might reduce risk by about 7% at the population level

The anti-amyloid antibody effect is roughly equivalent to slowing cognitive decline by 4–6 months over an 18-month treatment period

**Comparison to other interventions:** The effect of addressing all 14 risk factors (45% reduction) is substantially larger than any single drug treatment (27–35% slowing of decline over 18 months). However, the drug effects are measured in people who already have dementia, while risk factor modification aims to prevent dementia entirely.

**Acknowledged by the authors:**

PAFs assume risk factors are causal and independent — if two risk factors share a common cause, the combined PAF overestimates prevention potential

Prevalence data for some risk factors (especially in LMICs) are based on limited surveys

The report cannot account for all interactions between risk factors

Much evidence comes from high-income countries; generalisability to LMICs is uncertain

Observational studies cannot prove causality — some risk factors (e.g., depression, social isolation) may be early symptoms of dementia rather than causes

**Critical reader observations:**

The 45% PAF is a theoretical maximum — achieving even half of this would require massive population-level policy changes over decades

The report does not provide a clear hierarchy of which risk factors to address first for individuals

Anti-amyloid antibody evidence comes from highly selected trial populations (healthy, few comorbidities) — real-world effectiveness may be lower

The report is authored by experts with potential conflicts of interest (some have consulted for pharmaceutical companies developing anti-amyloid drugs)

No cost-effectiveness analysis is provided for the full prevention package

The report does not address whether risk factor modification works equally well across different genetic backgrounds beyond APOE ε4

For someone running their own n=1 experiment:

**What to test (specific intervention and dose):**

**Hearing:** Get a hearing test if you're over 40 or notice any difficulty. If hearing loss is detected, use hearing aids consistently (8+ hours/day)

**LDL cholesterol:** Target LDL < 2.6 mmol/L (100 mg/dL) through diet (reduce saturated fat, increase fibre) or medication if needed

**Blood pressure:** Maintain systolic BP ≤ 130 mmHg from age 40. Measure weekly at home

**Cognitive stimulation:** Engage in cognitively demanding activities (learning a language, musical instrument, complex games) for 30+ minutes daily

**Physical activity:** 150 minutes/week of moderate-intensity aerobic exercise (brisk walking, cycling) plus 2 sessions/week of resistance training

**Social connection:** At least 1 in-person social interaction per week (not just digital)

**Vision:** Regular eye exams every 2 years after age 40; correct refractive errors promptly

**Alcohol:** Limit to ≤ 14 units/week (UK guidelines) or ≤ 1 drink/day (US guidelines)

**Smoking:** Complete cessation — no safe level

**Minimum meaningful duration:**

**Vascular risk factors (BP, cholesterol, weight):** 6–12 months to see measurable changes; 2–5 years to see cognitive benefit

**Hearing aids:** 3–6 months for full adaptation; cognitive benefits may take 1–2 years

**Physical activity:** 6 months for measurable cognitive improvements

**Cognitive training:** 3–6 months for domain-specific improvements

**Multicomponent intervention:** 2 years (based on FINGER trial duration)

**What to measure (specific metrics):**

**Primary outcome:** Cognitive function — use a validated online tool (e.g., Cogniciti, BrainHQ, or the Montreal Cognitive Assessment self-administered version) every 3 months

**Secondary outcomes:**

- Blood pressure (systolic/diastolic) — weekly

- LDL cholesterol — every 6 months

- Weight/BMI — weekly

- Hearing — annual audiogram

- Physical activity — minutes/week logged

- Social contact — number of in-person interactions per week

- Sleep quality — Pittsburgh Sleep Quality Index (PSQI) monthly

- Mood — PHQ-9 (depression) monthly

**Tracking:** Use a spreadsheet or app to log all metrics weekly; note any life changes (illness, stress, medication changes)

**Key confounds to control for:**

**Age:** Cognitive decline accelerates with age; compare yourself to age-matched norms, not your younger self

**Education:** Higher education provides cognitive reserve; adjust expectations accordingly

**Genetics:** APOE ε4 status (can be tested via 23andMe or direct-to-consumer genetic tests) — if you're a carrier, you need to be more aggressive with risk factor modification

**Medical conditions:** Diabetes, thyroid disorders, vitamin B12 deficiency, sleep apnoea — all affect cognition and should be ruled out or treated

**Medications:** Anticholinergics, benzodiazepines, antihistamines can impair cognition

**Stress and sleep:** Both strongly affect cognitive performance; track these as covariates

**Practice effects:** Cognitive tests improve with repetition; use alternate versions or allow 3+ months between tests

**What a positive result would look like:**

**Stable or improving cognitive scores** over 12–24 months (most people decline 0.1–0.2 SD per year after age 60)

**Improved vascular markers:** BP dropping from 140/90 to 125/80; LDL from 4.0 to 2.5 mmol/L; weight loss of 5–10% if overweight

**Subjective improvement:** Feeling sharper, fewer "brain fog" episodes, better word-finding

**Realistic expectation:** You are unlikely to see dramatic improvements — the goal is to slow or prevent decline, not reverse aging. A positive result is maintaining cognitive function while peers decline

**Caveat for n=1 experiments:** You cannot prove causation from a single-person experiment. If you improve multiple risk factors simultaneously and see cognitive improvement, you won't know which factor caused the change. Consider testing one intervention at a time (e.g., hearing aids first, then exercise, then diet) with 6-month intervals between changes.