Dementia Is the UK's
Biggest Killer.
Nearly Half of Cases
May Be Preventable.

One in three. Why the number is not inevitable.

The Alzheimer's Society statistic is real and it is alarming. But it is not a verdict. The lifetime risk of dementia reflects the trajectory of an ageing population with a specific pattern of metabolic and lifestyle exposures — not an inescapable genetic fate.

The APOE4 genetic variant — the most well-known Alzheimer's risk gene — raises lifetime risk, but about a quarter of people carry it, and many people who develop dementia don't. Genes load the dice; they don't settle the outcome, and the modifiable factors below apply to carriers too.

The Lancet Commission on dementia prevention, updated in 2024, identified 14 modifiable risk factors that together account for about 45% of dementia cases worldwide (Livingston 2024). Not a marginal contribution: nearly half of cases are potentially preventable or delayable, through factors that can be addressed with the right information at the right time.

Alzheimer's as Type 3 Diabetes — the insulin resistance connection

The Type 3 Diabetes hypothesis is not universally accepted and remains a framework rather than a settled clinical diagnosis. But the epidemiological evidence is consistent: people with Type 2 diabetes have approximately double the risk of developing Alzheimer's disease. People with insulin resistance — who have not yet crossed the diagnostic threshold for diabetes — have elevated risk proportional to the degree of resistance. Whether improving insulin sensitivity in midlife lowers dementia risk later is plausible, but trials haven't shown it yet.

This is why the sliding timeline argument from the dam-bursting post applies directly here. The HOMA-IR was 1.8 at 40. The fasting insulin was elevated at 45. The HbA1c was 38 at 50. The brain was accumulating the consequences of impaired insulin signalling for twenty years before the first memory complaint at 70. The metabolic dysfunction and the neurological decline are not separate stories — they are the same story told in two different organs.

The 14 modifiable risk factors — and what you can measure

Homocysteine — a risk marker almost nobody measures

Of the functional markers linked to dementia, homocysteine has one of the more consistent observational links: in the Framingham study, Alzheimer's risk nearly doubled above 14 µmol/L (Seshadri 2002). Whether lowering it prevents dementia is a separate question, and the answer so far is not established: a Cochrane review of B-vitamin trials in people with mild cognitive impairment found little or no effect on memory or thinking over 6 to 24 months (McCleery 2018, Cochrane).

Homocysteine is a metabolic intermediate produced from methionine during the methylation cycle. When the cycle runs efficiently, homocysteine is remethylated back to methionine (requiring folate and B12) or diverted to cysteine via transsulphuration (requiring B6). When these pathways are insufficient — due to MTHFR variants, B-vitamin deficiency, or excessive methionine intake — homocysteine accumulates.

Proposed mechanisms, mostly from laboratory studies, include damage to the blood-brain barrier, oxidative stress and tau phosphorylation (the process behind Alzheimer's tangles). In Framingham, each 1-standard-deviation rise in homocysteine carried a relative risk of 1.8 for Alzheimer's (Seshadri 2002). That is an association, graded observational.

The VITACOG trial, a randomised controlled trial published in PLoS ONE in 2010, found that B vitamins (folic acid, B12, B6) slowed brain shrinkage in people with mild cognitive impairment; in those with homocysteine above 13 µmol/L, the rate of shrinkage was 53% lower (Smith 2010). Brain shrinkage is a scan measure, not dementia itself. A 2016 re-analysis found B vitamins slowed cognitive decline only in people whose omega-3 levels were already in the upper-normal range (Oulhaj 2016). That is a subgroup finding in one trial: it suggests, rather than establishes, that the two work together, and the authors called for a trial of the combination.

Homocysteine is measured on the Functional Platinum blood panel; it isn’t part of routine NHS blood tests. A result of 12 µmol/L sits inside many laboratory ranges, which often run up to 15, but above my working flag of about 11. It is a reason to look further, not a diagnosis.

The omega-3 index — DHA and the brain that built itself from fat

The brain is approximately 60% fat by dry weight. The most abundant structural fat in neural tissue is DHA — docosahexaenoic acid, the long-chain omega-3 fatty acid found in oily fish and, in smaller amounts, algae. DHA is not a supplement that might be nice to take. It is a structural component of every neuronal membrane in the brain, and its availability during neural development, maintenance, and repair determines neuronal function at the most fundamental level.

A low omega-3 index (the percentage of EPA and DHA in red blood cell membranes) is linked in observational studies to poorer brain health. In the Framingham Heart Study, people in the top quarter for blood DHA had 47% lower risk of dementia than the other three quarters (Schaefer 2006). That is observational: it doesn’t show that taking omega-3 prevents dementia.

The VITACOG re-analysis (B vitamins helped mainly in people with good omega-3 status; Oulhaj 2016) is a reason to look at homocysteine and omega-3 together. It isn’t proof that taking both prevents dementia.

Sleep — the glymphatic system and amyloid clearance

One of the most significant advances in dementia neuroscience in the past decade has been the discovery of the glymphatic system — a brain-specific waste clearance mechanism that operates primarily during slow-wave sleep. In mice, the space between brain cells widened by about 60% during sleep, letting cerebrospinal fluid flush through and clear waste, including amyloid beta, the protein that accumulates in Alzheimer’s plaques (Xie 2013). That was shown in animals; human work is still catching up.

The implication is direct: chronic sleep deprivation — defined in most research as consistently less than 7 hours — impairs glymphatic clearance of amyloid beta and accelerates its accumulation. In 20 healthy people, one night without sleep raised amyloid in two brain regions on PET scans (Shokri-Kojori 2018). Whether that adds up over years into dementia is plausible, but not proven.

Sleep is not a lifestyle preference. In the context of brain health, it is the nightly cleaning cycle that keeps the brain clear of the protein debris that Alzheimer's disease is built from. Functional assessment of sleep quality — via DUTCH Plus (which measures melatonin and cortisol awakening response as circadian markers) and clinical sleep history — is a core component of any brain health assessment.

The markers worth knowing — and how strong the evidence is

The MTHFR connection — why the folic acid fortification story matters for brain health

The folic acid fortification debate is directly relevant to dementia risk in a way that has not been widely discussed. The MTHFR C677T variant, when inherited from both parents, lowers the activity of the enzyme that makes methylfolate, the form used to recycle homocysteine. People with it tend to have higher homocysteine, especially when folate intake is low. Whether fortified flour leaves them accumulating unmetabolised folic acid while homocysteine stays high is debated rather than settled.

The practical point holds either way: measure homocysteine rather than assume it from genotype, diet or supplements. If it is raised, the next step is finding out why (B12, folate, kidney function, thyroid, medicines), with your GP where needed.

What a brain health assessment looks like at 45

A functional brain health assessment at 45 is not about finding dementia. It is about finding the precursors of dementia — the metabolic and nutritional patterns that, if left unaddressed, will accumulate into neurological risk over the following decades. At 45, those patterns are largely reversible. At 75, they may not be.

The assessment combines the Functional Platinum blood panel — for homocysteine, fasting insulin, HOMA-IR, hs-CRP and vitamin D — with the Metabolomix+ organic acids test for 8-OHdG, MMA (functional B12), FIGLU (functional folate), and the omega-3 fatty acids add-on for DHA status. The DUTCH Plus adds the sleep-relevant melatonin and cortisol awakening response picture. Read together, they give a picture of the modifiable metabolic factors. None of them diagnoses or predicts dementia on its own.

The intervention picture — what the evidence supports

Roughly in order of evidence strength, strongest first:

Blood pressure: midlife high blood pressure is one of the Lancet Commission’s risk factors (Livingston 2024). Raised blood pressure damages small blood vessels in the brain and contributes to vascular dementia. Have it checked, and treated if it’s high.

Hearing, vision and cholesterol: in the 2024 Commission, untreated hearing loss and high LDL cholesterol each account for about 7% of cases, the largest shares of any factor; untreated vision loss was added as a new factor. Hearing or vision loss in midlife should be assessed and treated, not put up with.

Activity, weight and diabetes: physical inactivity, obesity and diabetes are all Commission factors. Resistance training, regular aerobic exercise and diet quality improve insulin sensitivity. If the Type 3 Diabetes idea is right, that also helps the brain; trials haven’t yet shown it prevents dementia.

Smoking, alcohol, social contact, air quality: all Commission factors, and all modifiable to some degree.

Sleep: 7–9 hours consistently. The Commission doesn’t list sleep as a factor, but the mechanism (amyloid clearance during sleep) is plausible and supported by animal and short-term human studies.

Homocysteine and omega-3 (weaker evidence): raised homocysteine and low omega-3 are linked to dementia in observational studies, and B vitamins slowed brain shrinkage in one trial, but they haven’t been shown to protect memory. If homocysteine is raised, find out why first. I don’t give doses here: they depend on the cause, and omega-3 above 1 g a day raises the risk of atrial fibrillation (Gencer 2021).

Related reading