Dementia prevention refers to lifestyle and medical factors associated with reduced dementia risk, and two markers tracked in Longevity Coach IQ connect to genuinely cited dementia research, homocysteine and midlife blood pressure, worth understanding with an honest distinction between them, one is a useful risk signal, the other a more directly actionable lever.
- Every 5 µmol/L homocysteine increase is associated with roughly 15% higher Alzheimer-type dementia risk (RR 1.15, 95% CI 1.04–1.26).
- B-vitamin supplementation to lower homocysteine hasn't improved cognitive outcomes in trials.
- Cumulative midlife blood pressure has a well-established link to dementia risk (Lancet Commission).
- Blood pressure management is the more directly actionable of the two levers.
The homocysteine link, and its caveat
An updated dose-response meta-analysis of prospective cohort studies found every 5 µmol/L increase in homocysteine, an amino acid measurable on standard bloodwork, associated with a relative risk of 1.15 (95% CI 1.04–1.26) for Alzheimer-type dementia. Roughly a 15% increase, across five studies, with a clear linear dose-response relationship.
Two details matter and usually get lost. The association held for Alzheimer-type dementia specifically, not for all-cause or vascular dementia, where that analysis found no appreciable association and attributed it to publication bias and low statistical power. And you will see much larger figures quoted elsewhere, up to around 50% per 5 µmol/L, from earlier analyses. The more recent dose-response work finds a smaller effect confined to a narrower outcome, and that disagreement is worth knowing rather than resolving by picking the bigger number.
The caveat is what makes this the cautionary example. Trials have tested whether lowering homocysteine with B vitamins improves cognition, and the answer in general populations is no.
Clarke and colleagues pooled 11 trials with cognitive data on 22,000 individuals, across roughly five years of supplementation. Allocation to B vitamins lowered homocysteine by 26% and produced no significant effect on global cognitive function, with a z-score difference of −0.01 (95% CI −0.03 to 0.02), and no significant effect on any individual cognitive domain.
That is the argument of this page in one result. The trials successfully did the thing they set out to do. The cognition did not follow.
One counterpoint keeps the picture honest, and it does not overturn the conclusion. In VITACOG, 271 people over 70 with mild cognitive impairment were randomised to high-dose folic acid, B12 and B6 or placebo for 24 months. Whole-brain atrophy ran at 0.76% per year in the treatment group against 1.08% in placebo (P = 0.001), and in participants with baseline homocysteine above 13 µmol/L the atrophy rate was 53% lower (P = 0.001).
So: B vitamins reliably lower homocysteine. In unselected populations that has not translated into cognitive benefit. In a high-risk subgroup with elevated baseline homocysteine there is trial evidence of an effect on brain atrophy, which is an imaging endpoint rather than a clinical one.
Elevated homocysteine remains a useful risk signal worth knowing about, rather than a directly modifiable cause you can supplement your way out of. Stating the nuance makes that conclusion more credible, not less.
Midlife blood pressure, the stronger lever
Cumulative blood pressure specifically during midlife has a well-established link to later dementia risk, and the Lancet Commission carries hypertension among its modifiable midlife risk factors. It is a relationship considered more directly actionable than the homocysteine association, since blood pressure management, through lifestyle intervention or medication when appropriate, has robust trial evidence supporting its effectiveness, unlike the homocysteine-lowering trials.
The 14 modifiable risk factors, by life stage
The Lancet Commission on dementia prevention, updated in 2024, estimated that around 45 percent of dementia cases worldwide could in principle be prevented or delayed by addressing 14 modifiable risk factors. The 2024 update added two to the previous list of twelve: high LDL cholesterol in midlife and untreated vision loss in later life.
Some individual fractions are worth naming, because they are larger than people expect. Hearing loss is 7%. The two additions in 2024 were high LDL cholesterol in midlife at 7% and untreated vision loss in later life at 2%.
One methodological detail belongs on a page built around evidence honesty. The attributable fractions were calculated using data from the Norwegian Trøndelag Health Study. A figure presented as applying worldwide is derived from one Norwegian cohort's risk-factor prevalences. The Commission also acknowledges explicitly that some of the associations might be only partly causal.
Attempts to replicate it elsewhere show how much that matters. In a study of older American Indians, education, diabetes, hypertension, depression and isolation alone accounted for 60% of the attributable fraction, while other Lancet factors showed no association. An Ethiopian analysis using the same relative risks produced an overall figure of 51.4%. In a US longitudinal cohort replicating the full model, only hearing loss, diabetes and hypertension remained statistically significant, and midlife hypertension in that cohort was associated with decreased dementia odds.
None of that discredits the Commission. It means the headline number is a modelled estimate whose components vary substantially by population, which is more useful to understand than the number itself.
Two caveats belong with that headline figure before the list. It is a population-attributable estimate assuming every factor is eliminated everywhere, not a personal probability. And attributable fractions aren’t additive in a simple way, because the factors overlap substantially in the same people.
Early life
- Less education. Fewer years of formal education, acting largely through reduced cognitive reserve rather than through knowledge itself.
Midlife
- Hearing loss. Among the largest single contributors in the model, and unusually actionable given that hearing aids exist.
- High LDL cholesterol. Added in 2024 on the basis of large cohort studies and Mendelian randomisation, and one of the two largest contributors alongside hearing loss.
- Hypertension. Midlife blood pressure is the lever with the clearest mechanistic path, through cerebrovascular damage.
- Obesity. Particularly midlife obesity, which interacts with the metabolic factors below.
- Excessive alcohol. Defined in the Commission's model as more than 21 units weekly.
- Traumatic brain injury. Cumulative exposure, relevant to contact sport and occupational risk.
- Physical inactivity. Consistently associated in observational data, with more mixed results in prevention trials.
Later life
- Smoking. Acting through both vascular and direct neurotoxic pathways.
- Depression. The direction of causation remains debated, since early dementia can present as depression.
- Social isolation. One of the larger contributors, and among the least medicalised.
- Diabetes. Connected to the same metabolic processes covered in insulin resistance.
- Air pollution. Particulate exposure, largely outside individual control.
- Untreated vision loss. Added in 2024, and like hearing loss, unusually treatable for a dementia risk factor.
The pattern worth noticing is how many are cardiovascular or metabolic. Blood pressure, LDL, diabetes, obesity, smoking and inactivity all appear on this list and all appear in cardiovascular disease risk. The overlap isn’t coincidence, and it means the same work addresses both.
Risk signal vs. directly modifiable cause
| Marker | Association Strength | Directly Modifiable? |
|---|---|---|
| Homocysteine | ~50% higher risk per 5 µmol/L increase | Lowering it hasn't shown benefit in trials |
| Midlife blood pressure | Well-established link | Yes, robust trial evidence for management |
What if you carry APOE4?
Anyone reading that 45% of dementia risk is potentially modifiable will immediately want to know what happens if they carry the risk variant.
APOE4 is the single most consequential result a longevity-motivated genetic test is likely to return, and there is no treatment that changes the risk the variant itself confers. The genes entry covers what the numbers actually are, including the point that around half of people carrying two copies never develop Alzheimer's dementia.
The connecting argument is the one both pages make independently. Inherited risk you cannot change raises the value of the factors you can, because a fixed disadvantage means the remaining variables carry more weight. Carrying APOE4 is a reason to take the modifiable factors on this page more seriously and earlier, not a reason to conclude they are beside the point.
Modifiable versus merely linked
Brain health content lists supplements and interventions without distinguishing which have trial evidence of benefit from which are correlated risk markers.
That distinction decides where your effort belongs. Midlife blood pressure is the strongest modifiable lever with real trial support behind it.
Homocysteine is the cautionary example. A strong observational link, and trials that lowered it with B vitamins without improving cognition.
It tracks both markers with this honest evidence distinction preserved, not oversimplified.
Sources
Key references for the claims on this page. Where a figure is attributed to a specific study or body, it is named here.
- Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. The Lancet, 2024. Source of the 45% figure, the 14 modifiable risk factors and the individual attributable fractions quoted. View paper
- Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. The Lancet, 2020;396(10248):413–446. The previous 12-factor, 40% estimate. PMID 32738937
- Clarke R, Bennett D, Parish S, et al. Effects of homocysteine lowering with B vitamins on cognitive aging: meta-analysis of 11 trials with cognitive data on 22,000 individuals. American Journal of Clinical Nutrition. Source of the 26% homocysteine reduction and the null cognitive result. PMC4095663
- Smith AD, Smith SM, de Jager CA, et al. Homocysteine-lowering by B vitamins slows the rate of accelerated brain atrophy in mild cognitive impairment: a randomized controlled trial (VITACOG). PLoS ONE, 2010. Source of the atrophy figures in the high-risk subgroup. PMC2935890
Frequently asked
How strong is the homocysteine-dementia link?
Weaker and narrower than commonly quoted. An updated dose-response meta-analysis found a relative risk of 1.15 (95% CI 1.04–1.26) per 5 µmol/L increase, and only for Alzheimer-type dementia. It found no appreciable association with all-cause or vascular dementia. Larger figures circulate from earlier analyses.
Does lowering homocysteine with B-vitamins help?
Trials haven't shown improved cognitive outcomes despite the observational association.
How does blood pressure relate to dementia risk?
Cumulative midlife blood pressure has a well-established link to later dementia risk, and the Lancet Commission carries hypertension among its modifiable midlife risk factors.
How much dementia risk is actually modifiable?
The 2024 Lancet Commission estimated that around 45% of dementia cases worldwide could in principle be prevented or delayed by addressing 14 modifiable risk factors across life. That is a population-level figure about theoretical maximums, not a personal guarantee, and it assumes every factor is eliminated.
Which risk factors carry the most weight?
In the 2024 Commission's model, hearing loss and high LDL cholesterol in midlife each account for the largest single shares, with less education, hypertension, obesity, smoking, depression, physical inactivity, diabetes, alcohol, head injury, air pollution, social isolation and untreated vision loss making up the rest.
Does exercise reduce dementia risk?
Observational data consistently associate physical activity with lower risk, and physical inactivity appears on the Commission's list. Randomised trials aimed specifically at preventing dementia have produced more mixed results, partly because they need to run for a very long time to show anything.
Does a family history make it inevitable?
No. Carrying higher-risk genetic variants such as APOE4 raises probability without determining outcome, and many carriers never develop dementia. Genetic risk also doesn’t remove the modifiable factors; if anything it raises the return on addressing the ones you can control.
Is there a test that tells me if I'm heading toward it?
Nothing available to consumers gives a reliable individual answer. Blood-based biomarkers for Alzheimer's pathology are advancing quickly in research settings, but interpreting them in someone without symptoms isn’t yet established practice. Treating any consumer test as a verdict is premature.