Brain health and metabolism

Vitamin B12: the most common reversible cause of cognitive problems

Vitamin B12 (cobalamin) is needed to make myelin, the insulation around nerve fibres, and to recycle homocysteine. Deficiency causes memory problems, numbness, unsteadiness, low mood and, if untreated, permanent nerve damage. It is common after 60 because stomach acid falls with age, and it is caused or worsened by metformin, acid-suppressing drugs, vegan diets, gastric surgery and pernicious anaemia.

The standard "normal" range may be set too low. A June 2025 study in Annals of Neurology of healthy older adults found that lower B12 levels, even within the normal range, were associated with slower nerve conduction in the visual pathway, slower processing speed and more white-matter damage on MRI, while the functionally active fraction of B12 tracked the effect most closely. The authors argued that the definition of deficiency should be revisited for older people.

How B12 metabolism works

1. Dietary B12 (meat, fish, eggs, dairy, fortified foods) is freed by stomach acid and bound by intrinsic factor, a protein made in the stomach.
2. The complex is absorbed in the last part of the small intestine; without intrinsic factor (pernicious anaemia) or with low acid, little gets in.
3. In cells, B12 and folate together convert homocysteine back to methionine. The MTHFR enzyme supplies the folate in the right form.
4. When B12 or folate is short, homocysteine rises. High homocysteine is linked to stroke, heart disease, brain atrophy and dementia.

Testing and treatment

• Ask for serum B12 together with methylmalonic acid or homocysteine, which detect deficiency at the cell level when the serum level looks borderline.
• Treatment is tablets of 1,000 micrograms daily for most people, or injections where absorption is impaired. Response is usually quick for mood and energy, slower for nerves.
• In the VITACOG trial, high-dose B12, folic acid and B6 slowed brain atrophy in people with mild cognitive impairment who had raised homocysteine, and did nothing for those with normal homocysteine. Treat the deficiency; do not supplement blindly.

Blood sugar and insulin

Type 2 diabetes roughly doubles the risk of dementia, and insulin resistance in midlife is associated with worse cognition decades later. The brain is sometimes said to develop its own insulin resistance, which is why drugs for diabetes have been tested as Alzheimer's treatments.

What the GLP-1 trials found

• Liraglutide (ELAD, phase 2b, published in Nature Medicine, January 2026). In 204 people with mild to moderate Alzheimer's, a year of liraglutide did not change the primary outcome, brain glucose metabolism on PET. One secondary cognitive measure favoured the drug, while daily function and the clinician's global rating did not differ. The drug was safe in people without diabetes.
• Semaglutide (evoke and evoke+, phase 3). Novo Nordisk's two large trials in early Alzheimer's did not slow cognitive decline, a result reported in Nature Medicine in March 2026 and widely regarded as closing the question for this drug class in symptomatic disease, even though the drug lowered blood markers of inflammation.
• Observational data. A July 2025 study in JAMA Network Open found lower rates of neurodegeneration and stroke in people with diabetes and obesity taking semaglutide or tirzepatide, which is consistent with vascular benefit rather than a direct effect on Alzheimer's.

The practical conclusion is unchanged: preventing and controlling diabetes protects the brain, and GLP-1 drugs are not Alzheimer's treatments.

Thyroid, iron and other reversible causes

Every memory assessment should include this screen. A surprising share of "early dementia" referrals turn out to be thyroid, B12, medication or sleep problems that improve with treatment.

Diet patterns with evidence

The MIND diet (Mediterranean plus specific brain-protective foods such as berries and leafy greens) is the pattern used in the US POINTER trial and is associated with slower decline in cohort studies. No supplement has matched it. Omega-3, vitamin E, ginkgo and most "brain" supplements have failed in randomised trials; B vitamins help only when homocysteine is high.