Turmeric and dementia risk: does curcumin actually help?

Short answer: Probably not in a capsule. Turmeric's active compound, curcumin, looks impressive in lab dishes and in mice — but no clinical trial has ever tested curcumin for dementia risk reduction, and the trials that tested it for memory have been small or negative. The population studies that get quoted are about people who eat curry, not people who take supplements, and they can't prove cause and effect. In July 2026 the World Health Organization's updated guidelines did not recommend vitamin B, vitamin E, omega-3 or multivitamin/mineral supplements for reducing dementia risk in people without a diagnosed deficiency, citing a lack of evidence that any benefit would outweigh possible harms. Turmeric was not among the supplements named. Cooking with turmeric is a perfectly good thing to do. Buying it as brain insurance is not supported by the evidence — and high-dose supplements carry real, documented risks.

This is educational, not medical advice. Turmeric supplements interact with common medications and have been linked to liver injury in case reports. If you take a blood thinner, diabetes medication, or have any liver condition, discuss supplements with your clinician or pharmacist before starting.

Why turmeric got its reputation

The excitement is genuinely well-founded — in the laboratory.

Curcumin, the yellow pigment in turmeric, is an antioxidant and anti-inflammatory compound. In cell studies and animal models it does striking things: it reduces oxidative damage, dampens inflammation, lowers beta-amyloid and plaque burden in mice bred to develop Alzheimer's-like pathology, and improves memory and learning in rodents. A 2025 systematic review of 29 preclinical rodent studies found improved cognitive performance in roughly 80% of them (Int J Mol Sci, 2025).

Then there's the epidemiology that launched a thousand supplement ads. In 2006, researchers studying 1,010 non-demented older adults in Singapore found that those who ate curry "occasionally" or "often" scored better on a standard cognitive test than those who "never or rarely" did (Ng et al., American Journal of Epidemiology).

That combination — a plausible mechanism plus a real-world correlation — is how a spice becomes a "superfood."

What actually happened when they tested it in people

Here the story changes.

In people who already have Alzheimer's, curcumin has failed. Two randomised controlled trials — Baum and colleagues (6 months) and Ringman and colleagues (24 weeks) — found no benefit. The nonprofit evidence-rating group ADDF Cognitive Vitality, which is about as sympathetic an assessor as supplements get, summarises the whole field bluntly: there are zero meta-analyses, only four randomised trials, and no trial has tested curcumin for dementia prevention at all. Trials on cognition "have not reported substantial benefits even with curcumin optimized for absorption" (ADDF Cognitive Vitality — Curcumin).

In healthy older adults, results are small and inconsistent. The most-cited positive study is an 18-month UCLA trial: 40 non-demented adults took either a high-absorption curcumin (Theracurmin, 90 mg twice daily) or placebo. The curcumin group improved on verbal memory and attention, and a PET scan measure of protein deposits fell in one brain region (Small et al., American Journal of Geriatric Psychiatry, 2018).

Encouraging — but read the size again. Forty people, split into two groups of about twenty, at one site, using a PET tracer that isn't a validated measure of Alzheimer's pathology. A study that small can produce a real-looking result by chance, and it has never been replicated at scale. It is a reason to run a bigger trial, not a reason to buy a supplement.

The fact the supplement pages leave out

In 2017, a team of medicinal chemists published a review of curcumin in the Journal of Medicinal Chemistry that has become the reference point for anyone assessing this compound honestly. Their conclusion: curcumin is an unstable, reactive, poorly absorbed molecule that shows up as a false positive in laboratory screens so reliably that the authors classify it as a PAINS candidate — a pan-assay interference compound, meaning it appears active against almost anything you test it on precisely because it interferes with the test. Their summary of the human evidence, after more than 120 clinical trials across many diseases: "No double-blinded, placebo controlled clinical trial of curcumin has been successful" (Nelson et al., J Med Chem 2017).

That single sentence explains the whole pattern — spectacular in a dish, unimpressive in a person.

The newest study — and how to read it honestly

There is fresh evidence, and it deserves a fair hearing. In July 2025, researchers followed 2,920 older adults in the Singapore Longitudinal Ageing Study for about 4.5 years. Compared with people who never or rarely ate curry, frequent eaters had less cognitive decline (odds ratio 0.68 for "often", 95% CI 0.48–0.95) — though the daily-consumption estimate for that same outcome was not statistically significant (0.66, 95% CI 0.41–1.06), and the trend across consumption levels was null once the analysis was adjusted (p = 0.15). In a separate analysis of the 2,446 participants who were cognitively normal at baseline, daily eaters had a lower rate of new mild cognitive impairment or dementia — an odds ratio of 0.21 (95% CI 0.09–0.49), resting on 8 new cases among 222 daily curry eaters (Lu et al., Nutrients, 2025).

An 0.21 odds ratio sounds spectacular. Four reasons to hold it loosely:

• It's observational. People who eat home-cooked curry several times a week differ from people who don't in dozens of ways — diet overall, cooking at home, income, social life, culture. No adjustment catches all of that.
• The follow-up is short and the case count is small. The authors say so themselves; dementia takes decades to develop, and with 8 cases in the daily-curry group, a handful either way swings the odds ratio dramatically.
• It measured frequency, not dose. Nobody knows how much curcumin anyone actually consumed.
• It's about food, not capsules. Curry means turmeric plus onions, garlic, ginger, chilli, black pepper and olive or coconut oil, eaten as part of a whole dietary pattern — closer to what our MIND diet guide describes than to a pill.

The reasonable conclusion is the modest one: turmeric-rich cooking may be part of a brain-healthy dietary pattern. That is not the same claim as "curcumin capsules lower your dementia risk," and the research does not support the second one.

What the WHO said this month

On 15 July 2026, the World Health Organization published the second edition of its guidelines on reducing the risk of cognitive decline and dementia — its first update since 2019. Alongside recommendations on physical activity, diet, blood pressure, diabetes, cholesterol, hearing loss, smoking, alcohol, cognitive and social activity and air pollution, it does not recommend vitamin B, vitamin E, omega-3 or multivitamin/mineral supplementation for dementia risk reduction in people without a diagnosed deficiency, citing a lack of evidence that potential benefits outweigh possible harms (WHO, July 2026).

Turmeric isn't among the supplements named. But the direction of travel is clear: after reviewing the evidence, WHO did not endorse a single supplement for dementia risk reduction, and put its recommendations into diet, activity, blood pressure, cholesterol, diabetes, hearing, smoking, alcohol, cognitive and social activity and air pollution instead. That's the same conclusion we reached in do brain supplements actually work?

If you're going to take it anyway: the safety part nobody covers

Most pages on this topic treat turmeric as harmless because it's a food. High-dose supplements are not food, and there are three documented issues worth knowing:

• Liver injury. Turmeric-associated liver injury has been documented, and reports have become more frequent. The US Drug-Induced Liver Injury Network has published a series of 10 cases, 6 of them enrolled since 2017 (Halegoua-DeMarzio et al., American Journal of Medicine, 2022). Ten cases is a small number, and worth keeping in proportion. The authors suggest that formulations adding piperine (black pepper extract) to boost absorption may also increase the amount reaching the liver.
• Bleeding risk and drug interactions. Curcumin has mild anticoagulant activity. There are reports of raised INR — a serious bleeding risk — in people taking warfarin who added a turmeric product (Welsh Medicines Advice Service). It can also interact with diabetes medications and some painkillers.
• Lead contamination. Ground turmeric has been documented as a source of lead exposure in the US, with multiple brand recalls, because lead chromate has been used to brighten the colour of the raw spice (Cowell et al., Public Health Reports, 2017). Independent testing has found lead more often in turmeric powder products than in purified curcumin extracts.

Practical version: cooking with turmeric from a reputable brand is fine. No dose or formulation of curcumin has been shown to reduce dementia risk, so there is no supplement version of this we can recommend for that purpose. If you're taking one anyway, that decision belongs with your clinician or pharmacist — especially if you take a blood thinner or diabetes medication, or have any liver problem. Third-party testing addresses the contamination issue and nothing else; it doesn't make the benefit appear.

Where this fits in the bigger picture

The uncomfortable truth about brain-health supplements is that they are popular precisely because they're easy. A capsule asks nothing of you. The levers that actually have evidence behind them — blood pressure, hearing, blood sugar, activity, sleep, alcohol, smoking, social connection, vision — ask more, and together they account for a large share of dementia risk that is genuinely modifiable.

The honest question isn't "should I take turmeric?" It's "which of the proven levers is currently loose in my life?" — and that answer is different for everyone.

That's what Solenna's free 3-minute risk profile is for: instead of guessing, you get a clear, evidence-based picture of your modifiable risk factors and the specific steps that move them. No supplements to buy. Check your dementia risk profile — free.

How we research and review our content →

This article is educational and is not medical advice, diagnosis, or treatment. Sources cited — including the 2006 and 2025 Singapore curry cohorts, the 2018 Theracurmin trial, the ADDF Cognitive Vitality curcumin rating, Nelson et al.'s 2017 medicinal-chemistry review, the WHO's 2026 dementia risk-reduction guidelines and the turmeric liver-injury, warfarin-interaction and lead-contamination reports — describe research findings in general, not Solenna specifically; the curry studies are observational and cannot establish cause and effect. Solenna does not diagnose, prevent, treat, or cure Alzheimer's disease or any form of dementia; risk reduction means lowering probability, not eliminating it, and no outcome is guaranteed. Supplements can interact with prescription medication — speak with a qualified clinician or pharmacist before starting, stopping, or changing any supplement or medicine.