Homocysteine, B vitamins and memory: who actually benefits?

Short answer: High homocysteine is a genuine marker of higher dementia risk — that part is well established. But lowering it with B vitamins has been tested in more than 22,000 people, and across those trials it did not improve memory or thinking. One small trial did slow brain shrinkage, and only in a specific group: people with mild cognitive impairment whose homocysteine was already high. That narrow finding is now a live disagreement among researchers, and it has never been tested in a trial designed to settle it. So for most people without a diagnosed deficiency, B vitamins are not a memory intervention. For a smaller group, the question is genuinely open — and it's a question for a clinician, not a supplement aisle.

This is educational, not medical advice. Homocysteine testing, B12 deficiency and folic acid dosing are clinical matters. If you have symptoms, a known deficiency, kidney disease, or you take methotrexate, metformin or acid-reducing medication, discuss testing and any supplement with your clinician rather than self-prescribing.

What homocysteine actually is

Homocysteine is an amino acid your body makes as a by-product of normal metabolism. It's normally cleared quickly, and the clearing machinery runs on three B vitamins: B12, folate (B9) and B6. When one runs short, homocysteine accumulates in the blood.

That's why the two topics are welded together in every article you'll read — homocysteine is largely a readout of B-vitamin status. It's also why "lower your homocysteine" sounds like it should be a simple fix. Take the vitamins, the number falls. It does fall. The hard part is what happens next.

The association is real — and it's strong

The landmark result came from the Framingham Study. Researchers followed 1,092 dementia-free adults (average age 76) for eight years, during which 111 developed dementia and 83 were diagnosed with Alzheimer's disease.

For each one-standard-deviation rise in homocysteine, the risk of dementia rose by a factor of 1.4 (95% CI 1.1–1.9), and of Alzheimer's disease by 1.8 (95% CI 1.3–2.5). Above 14 µmol/L, the risk of Alzheimer's disease nearly doubled. Crucially, the association held even when researchers measured homocysteine eight years before the follow-up window began (relative risk 1.6, 1.2–2.1) — so it wasn't simply that early dementia was raising homocysteine (Seshadri et al., New England Journal of Medicine, 2002).

That's a solid, well-replicated association — but an observational one, and a marker that travels with a disease isn't necessarily a cause of it. Homocysteine rises with poor kidney function, low B-vitamin intake, smoking, certain medications and age, all of which have their own links to brain health. The only way to find out whether the homocysteine itself matters is to lower it on purpose and see what happens.

Which is exactly what researchers did.

The trial that made headlines: VITACOG

In 2010, an Oxford-led team ran VITACOG: 271 adults aged 70 and over with mild cognitive impairment, randomised for two years to placebo or a high-dose combination of folic acid (0.8 mg), vitamin B6 (20 mg) and vitamin B12 (0.5 mg). The outcome was the rate of brain shrinkage on MRI, measured in 168 participants who completed the imaging protocol.

The treated group's brains shrank at 0.76% per year versus 1.08% on placebo — a 29.6% lower rate of atrophy (P=0.001). And the effect was not spread evenly. Among participants in the highest homocysteine quartile (above 13 µmol/L), the reduction reached 53%. Among those in the lowest quartile (9.5 µmol/L or below), there was no effect at all (Smith et al., PLOS ONE, 2010).

A later re-analysis of the same 168 people added a second condition. The benefit appeared only in participants who already had good omega-3 fatty acid status; in those with low omega-3 levels, the vitamins did nothing measurable (Jernerén et al., American Journal of Clinical Nutrition, 2015).

That's a striking result with three caveats stacked on it. It was a single, relatively small trial. Its endpoint was brain atrophy on a scan, not dementia — slower shrinkage is encouraging, but it's a stand-in for the outcome that matters. And the findings that make it exciting are subgroups: a quartile here, a nutrient-status split there. Subgroups generate hypotheses; they don't confirm them.

The result that outweighs it: 22,000 people

Four years later, the B-Vitamin Treatment Trialists' Collaboration pooled the cognitive data from 11 randomised trials covering roughly 22,000 participants.

B vitamins did what they are supposed to do: homocysteine fell by 28%. And the effect on cognition was not significant — not on global cognitive function, not on any individual cognitive domain, and not on the rate of cognitive ageing (Clarke et al., American Journal of Clinical Nutrition, 2014).

That is the number that governs the field, and it is why every major guideline lands where it does. When a small trial and a 22,000-person pooled analysis disagree, the default reading is that the small trial found something specific to its participants, or found noise. Not always — but usually.

The most recent synthesis doesn't overturn it. In December 2025, a meta-analysis of 17 randomised trials in 5,275 older adults reported "high-certainty evidence that vitamin B6, B9, or B12 supplementation has a very small benefit on global cognitive function" — effect size g = 0.110 (95% CI 0.034–0.186) — which did not differ between people with intact cognition, mild cognitive impairment, or dementia (Nutrition Reviews, 2025).

An effect size of 0.11 is a nudge: detectable across thousands of people, never noticeable in your own life. It's real. It's also not what anyone is hoping for when they buy a B-complex for their memory.

So why do experts still argue about it?

Because a group of them think the trials asked the wrong question.

In June 2025, six researchers — Joshua Miller, Andrew McCaddon, Jin-Tai Yu, Babak Hooshmand, Helga Refsum and A. David Smith — published a piece in the Journal of Alzheimer's Disease arguing that dismissing homocysteine-lowering is a major missed opportunity. Their case is specific rather than sweeping: benefit should be expected only where baseline homocysteine is elevated (they propose above about 11.3 µmol/L), where omega-3 status is adequate, and where the person is at an early stage — mild cognitive impairment rather than established dementia (Miller et al., J Alzheimers Dis, 2025).

Their argument is that the large pooled trials mostly enrolled people meeting none of those conditions — adequately nourished participants with normal homocysteine — and that giving B vitamins to someone who doesn't need them and finding no effect isn't evidence they don't work in someone who does.

Two things are worth knowing about that argument. It's a serious one from serious people, several of whom ran the original trial. It's also a position paper, not new evidence, written by the researchers most invested in the hypothesis — a conflict worth naming plainly. The disagreement is real and unresolved, and the trial that would settle it (enrolling only people with high homocysteine, adequate omega-3 and early cognitive symptoms) has not been run.

That's the honest state of the science: not "B vitamins work", not "B vitamins are debunked", but a narrow, plausible, unproven hypothesis awaiting a trial nobody has funded.

What the guidelines say

Both major reference points land in the same place. The World Health Organization's 2026 risk-reduction guidelines recommend against vitamin B supplements for reducing dementia risk in people without a diagnosed deficiency — alongside vitamin E, omega-3 and multivitamins — on the grounds that there isn't evidence the potential benefits outweigh possible harms (more in what the WHO now says about lowering dementia risk). Alzheimer's Research UK puts it in one sentence: folic acid and B12 supplements lower homocysteine but don't affect memory and thinking in people without a B12 deficiency.

Neither is saying homocysteine doesn't matter. Both are saying the supplement doesn't fix it. It's the same pattern we've found every time we've examined a specific product — see do brain supplements actually work? and turmeric and dementia.

Where B vitamins genuinely do matter

There's one clear exception, and it's medical rather than a wellness purchase: an actual deficiency.

Low B12 is one of the genuinely reversible causes of memory problems, brain fog, confusion and nerve symptoms — which is why a proper cognitive workup checks for it before anything else. Higher-risk groups include adults over 60, vegans and long-term vegetarians, people who've had gastric or intestinal surgery, and people taking metformin or long-term acid-reducing medication. If that describes you, the useful action is a blood test ordered by your clinician, not a supplement bought on spec — see low vitamin B12 and memory.

The distinction matters because the two situations look identical from the supplement aisle and are completely different underneath. Treating a diagnosed deficiency corrects a problem. Supplementing without one treats a number — and the trials say treating the number doesn't move the outcome.

Should you get your homocysteine tested?

For most people, testing changes nothing you'd actually do. If it came back high, the evidence doesn't support B vitamins as a memory intervention, and the realistic response — improve diet, check for a B12 or folate deficiency, address kidney function, stop smoking — is largely what you'd be advised to do anyway.

It can be reasonable inside a clinical picture: alongside investigation of cognitive symptoms, suspected deficiency, unexplained vascular disease, or a relevant family history. That's a conversation with your doctor, who can read the number against everything else about you. A homocysteine result in isolation, ordered from a direct-to-consumer lab, mostly generates anxiety and a supplement purchase.

What to do instead

If your reason for reading this was "I want to lower my dementia risk," the evidence points somewhere less exotic and far better supported.

A study published this month makes the point plainly. Researchers followed 12,409 American adults (average age 56) for about 26 years and found those who reached midlife without high blood pressure, type 2 diabetes or smoking averaged nearly 13 more years free of dementia than those carrying all three — roughly 30 years versus 17 (Hu et al., Neurology Open Access, 5 Aug 2026). The authors are explicit that this is an association rather than proof of cause, and that risk factors were measured only once. But three ordinary, clinically manageable things — not a supplement — carried that much distance. All three are worth discussing with your clinician: we've covered blood pressure and blood sugar separately.

That fits the broader picture: around 45% of dementia cases worldwide are attributed to modifiable risk factors — a population-level estimate under ideal conditions, not a personal guarantee — spanning blood pressure, hearing, blood sugar, activity, alcohol, smoking, sleep, social connection, vision and cholesterol. Which of those is loose in your life isn't something a B-complex can tell you.

How we research and review our content →

This article is educational and is not medical advice, diagnosis, or treatment. The studies cited — the 2002 Framingham NEJM analysis, the 2010 VITACOG trial and its 2015 omega-3 re-analysis, the 2014 pooled analysis of 11 trials in 22,000 participants, the December 2025 Nutrition Reviews meta-analysis, and the 2025 Journal of Alzheimer's Disease position paper — describe associations and trial results in general populations, not Solenna specifically. The 2002 Framingham finding is observational and does not establish that homocysteine causes dementia; the VITACOG subgroup findings are hypothesis-generating and have not been confirmed in a trial designed to test them. Solenna is not a medical device and does not diagnose, prevent, treat, or cure Alzheimer's disease or any other condition. Risk reduction means lowering probability, not eliminating it; individual results vary and no outcome is guaranteed. Do not start, stop, or change any supplement or medication on the basis of this article, and do not use it to interpret a blood test result — speak with a qualified healthcare professional.