Memory and dementia: what is established, and what is sold

Dementia prevention is a field where the supplements have been tried extensively and have mostly failed, while the unglamorous interventions have accumulated evidence.

Ginkgo biloba. Large randomised trials — including a trial of over 3,000 people over a median of six years — found that ginkgo did not prevent dementia or cognitive decline. For cognitive function in healthy older adults, the evidence is similarly negative or equivocal. Ginkgo does have a signal for tinnitus and peripheral arterial disease in small trials, and it increases bleeding risk. The dementia claim is the one that should be given up.

Curcumin has biological plausibility and small trials; the larger and better trials in Alzheimer's disease have not shown cognitive benefit. Omega-3 — the trials in dementia prevention have been negative, though the observational associations were positive; this is the classic pattern of a claim that survives on observational data and dies in trials.

What has the best evidence for dementia risk. Midlife hypertension, diabetes, obesity, smoking, hearing loss, depression, physical inactivity, and low education are the modifiable risk factors with the strongest evidence, and modelling suggests that a substantial share of cases is attributable to them. Treating midlife blood pressure is the single strongest pharmacological intervention — a randomised trial found that intensive blood-pressure control reduced the incidence of mild cognitive impairment. Hearing aids for hearing loss may reduce dementia risk, which is a surprising and important finding. Exercise has randomised evidence for cognitive function.

For memory complaints in a healthy person. The strongest evidence is for exercise, sleep, treating depression, treating sleep apnoea, correcting B12 and thyroid, hearing and vision correction, and staying socially and cognitively engaged. Bacopa and lion's mane have small trials with modest and short-lived effects on memory measures.

The framing that matters. A new, progressive memory problem in a person is a medical presentation, not a supplement decision. Some causes are treatable — B12 deficiency, thyroid disease, depression, medication effects, sleep apnoea, normal-pressure hydrocephalus, subdural haematoma — and they are found by assessment. Giving someone with treatable cognitive impairment a ginkgo trial costs them the thing that was fixable.