The Reversible Cause That Usually Isn't

Every dementia workup checks B12 and vitamin D, and it should. Correcting them rarely reverses memory loss, but a missed B12 deficiency can cause permanent nerve damage.

Can a B12 deficiency cause memory problems?

Yes. B12 deficiency can produce confusion, low mood, and nerve damage, and it is one of the few things on a dementia workup that is straightforwardly treatable.

Vitamin B12 is needed to build myelin, the insulation around nerve fibers. Without enough of it, nerves in the spinal cord and periphery degenerate — a condition called subacute combined degeneration. The symptoms are numbness and tingling in the feet, an unsteady walk, weakness, and in some people irritability, depression, or confusion.

Older adults are more vulnerable for a mechanical reason. Extracting B12 from food requires stomach acid and an intact stomach lining, and both decline with age. Someone can eat a perfectly adequate diet and still absorb very little.

So the test belongs on the workup. Every reputable diagnostic protocol includes it, and any clinician who skipped it would be criticized for it. The question families are usually asking is a different one.

How often does treating a deficiency actually reverse dementia?

Rarely. Around 9% of people evaluated for dementia have a potentially reversible cause found; about 0.6% actually get better when it is treated.

A meta-analysis pooled 39 studies covering roughly 7,000 people who had been worked up for dementia. Potentially reversible causes — B12 deficiency, thyroid disease, depression, and others — were identified in about 9% of them. When those causes were treated and the people were followed, only 0.6% improved, partially or fully.

The gap between 9% and 0.6% is the whole problem. A low B12 found in someone with dementia is usually a second condition sitting alongside the first, not the explanation for it. Both are real. Only one of them is going away.

A smaller series makes the same point at closer range. Among 181 patients meeting diagnostic criteria for dementia, 25% were found to be B12 deficient. Nineteen of those were followed after starting B12 treatment on its own, and three returned to a normal score on the Mini-Mental State Examination.

Then why is it worth testing at all?

Because the deficiency causes damage of its own that is preventable, and because treating it costs almost nothing.

Three reasons hold up.

The neurological damage is separate and it is permanent. Untreated B12 deficiency damages the spinal cord. Someone with dementia who also loses position sense in her feet is at far greater risk of falling, and a fall is one of the most consequential events in this disease. Correcting the B12 does not restore memory, but it can protect what is left of her balance.

Deficiency worsens function on top of dementia. Fatigue, low mood and anemia all make cognition look worse than the underlying disease alone would. Removing that layer will not change the diagnosis and can still change the day.

The intervention is cheap and low-risk. A B12 injection or a high-dose oral tablet has few downsides. Almost nothing else in this illness has that ratio.

What the test is not is a lottery ticket. Treat it because it is worth treating, and set the expectation accordingly.

Her B12 came back "normal." Does that settle it?

Not always. Results in the low-normal range can still reflect a deficiency at the cellular level, and two extra tests can tell the difference.

Serum B12 measures what is circulating, not what is reaching cells. In the low-normal range, people can be functionally deficient with a result their lab flags as fine.

Two additional markers help. Methylmalonic acid builds up in cells when B12 is short and spills into the blood, which makes it a functional indicator of true deficiency. Homocysteine rises for the same reason, though it also rises with folate deficiency and kidney disease, so it is less specific. Research groups increasingly recommend measuring more than one marker, because a single number misclassifies people in both directions.

If symptoms point at B12 — numbness, an unsteady walk, a burning tongue, anemia — and the level came back at the bottom of the normal range, asking whether methylmalonic acid should be checked is a reasonable question.

What is draining her B12 in the first place?

Often a medication she has taken for years. Metformin and acid-reducing drugs are the two most common.

Metformin, the most prescribed drug for type 2 diabetes, interferes with B12 absorption in the small intestine. The effect builds with years of use.

Proton pump inhibitors — omeprazole (Prilosec), pantoprazole (Protonix), esomeprazole (Nexium) — and H2 blockers like famotidine reduce the stomach acid needed to release B12 from food. Long-term users show measurable changes in B12 status, including elevated methylmalonic acid.

Taken together, the risk is greater than either alone, and studies find the combination hits people aged 65 and over harder than younger patients. Both drugs are extremely common in exactly this population, and neither is typically flagged as a nutritional issue when it is started.

Other causes: pernicious anemia, which is an autoimmune loss of the protein needed to absorb B12; atrophic gastritis; previous gastric or bowel surgery; and a diet with very little animal protein, which becomes more likely in dementia as appetite and food preferences change.

Does a B12 supplement protect the brain if she is not deficient?

No. Supplementation reliably lowers homocysteine and reliably fails to improve cognition.

This is where the supplement aisle and the trial data separate.

A meta-analysis of 23 randomized controlled trials in adults over 50 compared B vitamins and folate against placebo. Homocysteine dropped significantly in the supplemented groups. Cognitive scores showed no significant difference.

A two-year randomized trial in 2,919 older adults selected for elevated homocysteine — the group most likely to benefit, in theory — did not find improved cognitive function, though it reported a slightly slower rate of decline on some measures. In adults with subclinical deficiency, supplementation did not meaningfully improve cognitive or neurological outcomes.

The lesson is narrow and worth being precise about: correct a deficiency because it is a deficiency. Do not take B12 as a memory drug, and be skeptical of any product sold on that basis.

What about vitamin D?

The observational evidence looks alarming and the randomized evidence is null. Both need to be held at once.

In a cohort study, adults who were severely vitamin D deficient had an adjusted hazard ratio of 2.22 for developing Alzheimer's disease compared with those whose levels were sufficient, with risk rising notably below a threshold of 50 nmol/L. That finding is widely quoted, and it is a real association.

Then there is the trial. Two ancillary studies of the VITAL randomized trial gave community-dwelling adults aged 60 and over either 2,000 IU of vitamin D3 daily or placebo and followed cognition for two to three years — 3,424 people assessed by telephone and 794 in person. The pooled difference in annual rate of cognitive decline was 0.01, which is nothing.

Two honest caveats keep this from being a clean verdict. VITAL enrolled generally healthy people and did not select for deficiency, so it did not test whether correcting a genuine deficiency helps. And two to three years is short against a disease that develops over decades.

Vitamin D deficiency is worth correcting in an older adult — it matters for bone density, muscle strength and falls, and falls matter enormously here. Just not as a treatment for memory.

What should I ask for?

Ask for the tests, ask what the plan is if they come back abnormal, and ask what the medication list is doing to her.

• "Has her B12 been checked, and what was the number?" Not "was it normal" — the actual value.
• "If it was low-normal and she has neurological symptoms, should we check methylmalonic acid?"
• "Is she on metformin, omeprazole, pantoprazole, esomeprazole or famotidine?" If yes, ask whether the acid reducer still has a reason to exist.
• "If we correct this, what improvement would you expect, and by when?" A specific answer is useful. "We'll see" means the deficiency is being treated on its own merits, which is fine — but say so.
• "Has thyroid been checked as well?" It sits in the same category: worth testing, rarely the answer.
• "Is depression being considered?" It is the potentially reversible cause with the best odds of actually improving something, and it is the one most often missed in an older adult.