Two Windows That Never Reopen | The Starved Brain, Chapter 6

B vitamins · mild cognitive impairment

B Vitamins in Mild Cognitive Impairment: Measure Before It Is Late

Test B12 status in mild cognitive impairment before reaching for B vitamins: in VITACOG, high-dose B vitamins slowed brain atrophy most in people whose homocysteine was high at entry. In anyone over 60, serum B12 goes on the order with methylmalonic acid or holotranscobalamin.

Nearly the same B-vitamin doses slowed brain atrophy in mild cognitive impairment and did nothing in established Alzheimer’s disease. The variable was not the capsule. It was who was depleted, and when.

The chapter video, Two Windows That Never Reopen, frames B12 across two irreversible windows, the first thousand days and the last. For primary care, geriatrics and neurology-adjacent practice, the clinically useful half is the second. On B vitamins, mild cognitive impairment and established Alzheimer’s disease give opposite trial answers with nearly the same capsules, and the difference is who was enrolled and when. That is a screening argument before it is a treatment argument.

Do B vitamins help mild cognitive impairment or Alzheimer’s?

VITACOG randomized 271 people over 70 with mild cognitive impairment to two years of high-dose B vitamins or placebo; 168 completed paired MRI. Whole-brain atrophy ran 0.76% a year on treatment versus 1.08% on placebo (P = 0.001), and above a baseline homocysteine of 13 µmol/L it was 53% lower. A later analysis reported a cognitive benefit in participants whose homocysteine was high at entry.

The Alzheimer’s disease trial ran the other direction. It randomized 409 people with mild to moderate disease, MMSE 14–26, to folate 5 mg, B6 25 mg and B12 1 mg or placebo for 18 months. Homocysteine fell 2.42 versus 0.86 µmol/L. The ADAS-cog worsened 0.401 points a month on vitamins and 0.372 on placebo (P = .52), and depression was reported in 28% on vitamins against 18%.

Two readings fit, and the trial cannot separate them. Too late: the tissue was already lost. Wrong patients: everyone entered with normal folate, B12 and homocysteine, the subgroup VITACOG suggests does not respond. Both readings point to the same clinical move. A struggling neuron can be protected, a lost one cannot be rebuilt, and a patient who is not depleted cannot be repleted. Status has to be measured while the window is open, and treatment follows the measurement rather than the diagnosis.

Why do older adults become low in B12?

The common teaching that aging stomachs stop making acid does not hold up. Among 248 independently living older adults, basal gastric contents were acidic in 84%, and the authors noted that nearly 90% could acidify. The deficit concentrates in a subgroup: 31.5% of 359 adults aged 60 to 99 had atrophic gastritis by pepsinogen ratio, with low serum B12 rising stepwise with severity. Food-cobalamin malabsorption accounts for more than 60% of deficiency in older adults and pernicious anemia for 15–20%. After 60, roughly 6% are deficient by plasma B12 and closer to 20% are marginal.

Serum values understate the problem. In 548 Framingham survivors, methylmalonic acid was markedly raised in 11.3%, and the authors estimated metabolic deficiency at 12% or more, many with normal serum concentrations. The practice position is not to order serum B12 alone in anyone over 60; methylmalonic acid or holotranscobalamin goes on the same order.

Which medications lower B12 in older patients?

Acid suppression has a modest but real signal. In Kaiser Permanente data, two or more years of proton pump inhibitor supply carried OR 1.65, H2 blockers alone OR 1.25, and more than 1.5 pills a day OR 1.95, a dose gradient. Pooled data are less certain: one called its pooled OR too low to imply a clear association, and a 2025 meta-analysis of six studies found no difference in serum B12. The average effect is small, but an individual patient with atrophic gastritis is not an average.

The metformin question is settled by randomization. The Diabetes Prevention Program assigned 1,073 people to metformin and 1,082 to placebo. At five years, B12 at or below 203 pg/mL occurred in 4.3% versus 2.3%, low-plus-borderline in 19.1% versus 9.5%, and each year of use carried OR 1.13. Pooled meta-analytic odds came to 2.45 across 29 studies and 2.95 across 17 later studies in type 2 diabetes.

Intake, protein and the muscle compartment

Supply falls on the intake side too. Appetite declines, meat is often dropped first for dental reasons, the effort of cooking for one, or longstanding advice to limit red meat, and the anorexia of aging resists late correction once weight loss is established. Social isolation is a nutritional exposure in this population, not a footnote to it.

Demand moves the other way. PROT-AGE recommends at least 1.0–1.2 g/kg/day of protein for healthy adults over 65 and 1.2–1.5 g/kg/day in acute or chronic disease, excepting severe kidney disease with eGFR below 30, because older muscle shows declining anabolic response. Sarcopenia is defined as muscle failure that accrues across a lifetime. Chronic low-grade inflammation rises with age, and the same machinery driven by overnutrition is metaflammation, which is why the overfed, insulin-resistant adult runs this program early. Bioimpedance body composition quantifies the lean and fat compartments that weight and BMI blur together.

Who to test and what to pair it with

Measura [Cardiometabolic and Autonomic Health Analysis] is a measurement layer that reports to the ordering physician and leaves management with that physician. For this population the useful set is compact: laboratory panels specified to include methylmalonic acid or holotranscobalamin, homocysteine and folate; a cognitive assessment to document a baseline while the patient is still in the “normal for age” zone; and, because sarcopenia and cognitive change travel with falls, vestibular and balance testing as part of fall-risk work. The pairing logic is set out in cognitive assessment and fall prevention.

One gap is worth naming plainly. VITACOG had four participants with diabetes, so the insulin-resistant patient is effectively missing from the trial that worked. An absence from a trial is a blank, not a finding, and the blank is filled with a measurement rather than an assumption. The first window, preconception and pregnancy status, is assessed through obstetric and primary care; it is not part of Measura’s test library.

Documentation and workflow

Trigger criteria that follow from this evidence: age over 60, mild cognitive impairment or subjective cognitive complaint, “age-appropriate atrophy” on imaging, metformin at any duration, and two or more years of acid suppression. Encode them once so that screening does not depend on who is seeing the patient that day. The cognitive baseline, medication review and functional B12 result fit annual wellness visit integration, and the pattern behind a written baseline is discussed in what a cognitive baseline is for. The companion piece on functional B12 screening in metformin users covers the serum-marker evidence in depth, and every lifespan study, pregnancy data included, is collected in the Starved Brain evidence briefing for Chapter 6.

Frequently asked questions

Should every patient with mild cognitive impairment receive B vitamins?

The trial data argue for measuring first. In VITACOG the atrophy benefit was concentrated in patients with high baseline homocysteine and absent at or below 9.5 µmol/L, and the Alzheimer’s trial that enrolled replete patients found no effect on decline. Homocysteine and functional B12 status identify who is likely to respond. Selection guidance is on who to test.

Why did high-dose B vitamins fail in established Alzheimer’s disease?

Two explanations fit the data. The disease may have been too advanced for tissue protection to matter, or the enrolled patients may simply not have been depleted, since all entered with normal folate, B12 and homocysteine. The trial cannot distinguish them, and both favor testing status earlier. The reasoning for measuring before diagnosis is on clinical rationale.

Does long-term proton pump inhibitor use justify functional B12 testing?

The evidence is mixed on average: a large case-control study found a dose gradient, while pooled reviews found a small or absent effect. In an older patient who may also have atrophic gastritis or take metformin, the exposures stack, and measurement is a low-burden way to resolve the individual case. Encoding that trigger is covered in standing orders for screening.

How does this screening support quality documentation?

A documented cognitive baseline, a functional B12 result with its medication context, and a fall-risk assessment create a structured record that supports value-based care reporting and follow-up at the next visit. The measures themselves are clinical; documentation makes them auditable over time. How this maps to quality programs is described in HEDIS and value-based care.

Where does body composition testing fit for older adults?

Sarcopenia accrues silently across decades, and body weight can stay stable while muscle declines. Measuring lean and fat compartments gives a concrete baseline alongside protein intake, which PROT-AGE sets at 1.0–1.2 g/kg/day or more for healthy adults over 65. The measurement is descriptive and goes to the treating physician. Practical workflow is on staffing and workflow.

Can B vitamins slow brain shrinkage?

In mild cognitive impairment, they did in the trial that tested it. In VITACOG, 271 people over 70 took high-dose B vitamins or placebo for two years. Whole-brain shrinkage ran 0.76% a year on the vitamins against 1.08% on placebo, and it was 53% lower in those who started with homocysteine above 13 µmol/L. A later analysis reported a cognitive benefit in that high-homocysteine group.

Is serum B12 enough to test for deficiency after 60?

No. In 548 Framingham survivors, methylmalonic acid was markedly raised in 11.3%, and the authors estimated metabolic deficiency at 12% or more, many of them with normal serum concentrations. The practice position is that serum B12 is not ordered alone in anyone over 60: methylmalonic acid or holotranscobalamin goes on the same order, with homocysteine and folate read beside them.

Does metformin lower B12 in older adults?

Yes, and randomization settled it. The Diabetes Prevention Program assigned 1,073 people to metformin and 1,082 to placebo. At five years, low-plus-borderline B12 occurred in 19.1% against 9.5%, and each year of use carried an odds ratio of 1.13. That is why metformin at any duration sits on the trigger list for functional B12 testing, alongside age over 60 and a cognitive complaint.

Screen while the window is still open

Learn how the Measura protocol fits a practice, from trigger criteria for older patients to a documented cognitive baseline.

Send a message

4477 Woodson Rd, Suite 201, St. Louis, MO 63134. Monday to Friday, 9:00 a.m. to 5:00 p.m. Please do not send symptoms, diagnoses or images through a web form — a website form is not a secure medical channel. Send your name and number and we will call you back.

References

Smith, A. D., Smith, S. M., de Jager, C. A., Whitbread, P., Johnston, C., Agacinski, G., Oulhaj, A., Bradley, K. M., Jacoby, R., & Refsum, H. (2010). Homocysteine-lowering by B vitamins slows the rate of accelerated brain atrophy in mild cognitive impairment: A randomized controlled trial. PLoS ONE, 5(9), Article e12244. https://doi.org/10.1371/journal.pone.0012244
Aisen, P. S., Schneider, L. S., Sano, M., Diaz-Arrastia, R., van Dyck, C. H., Weiner, M. F., Bottiglieri, T., Jin, S., Stokes, K. T., Thomas, R. G., Thal, L. J., & Alzheimer Disease Cooperative Study. (2008). High-dose B vitamin supplementation and cognitive decline in Alzheimer disease: A randomized controlled trial. JAMA, 300(15), 1774–1783. https://doi.org/10.1001/jama.300.15.1774
Hurwitz, A., Brady, D. A., Schaal, S. E., Samloff, I. M., Dedon, J., & Ruhl, C. E. (1997). Gastric acidity in older adults. JAMA, 278(8), 659–662. https://doi.org/10.1001/jama.1997.03550080069041
Krasinski, S. D., Russell, R. M., Samloff, I. M., Jacob, R. A., Dallal, G. E., McGandy, R. B., & Hartz, S. C. (1986). Fundic atrophic gastritis in an elderly population: Effect on hemoglobin and several serum nutritional indicators. Journal of the American Geriatrics Society, 34(11), 800–806. https://doi.org/10.1111/j.1532-5415.1986.tb03985.x
Andrès, E., Loukili, N. H., Noel, E., Kaltenbach, G., Abdelgheni, M. B., Perrin, A. E., Noblet-Dick, M., Maloisel, F., Schlienger, J.-L., & Blicklé, J.-F. (2004). Vitamin B12 (cobalamin) deficiency in elderly patients. CMAJ, 171(3), 251–259. https://doi.org/10.1503/cmaj.1031155
Lindenbaum, J., Rosenberg, I. H., Wilson, P. W., Stabler, S. P., & Allen, R. H. (1994). Prevalence of cobalamin deficiency in the Framingham elderly population. The American Journal of Clinical Nutrition, 60(1), 2–11. https://doi.org/10.1093/ajcn/60.1.2
Lam, J. R., Schneider, J. L., Zhao, W., & Corley, D. A. (2013). Proton pump inhibitor and histamine 2 receptor antagonist use and vitamin B12 deficiency. JAMA, 310(22), 2435–2442. https://doi.org/10.1001/jama.2013.280490
Aroda, V. R., Edelstein, S. L., Goldberg, R. B., Knowler, W. C., Marcovina, S. M., Orchard, T. J., Bray, G. A., Schade, D. S., Temprosa, M. G., White, N. H., Crandall, J. P., & Diabetes Prevention Program Research Group. (2016). Long-term metformin use and vitamin B12 deficiency in the Diabetes Prevention Program Outcomes Study. The Journal of Clinical Endocrinology & Metabolism, 101(4), 1754–1761. https://doi.org/10.1210/jc.2015-3754
Bauer, J., Biolo, G., Cederholm, T., Cesari, M., Cruz-Jentoft, A. J., Morley, J. E., Phillips, S., Sieber, C., Stehle, P., Teta, D., Visvanathan, R., Volpi, E., & Boirie, Y. (2013). Evidence-based recommendations for optimal dietary protein intake in older people: A position paper from the PROT-AGE Study Group. Journal of the American Medical Directors Association, 14(8), 542–559. https://doi.org/10.1016/j.jamda.2013.05.021
Franceschi, C., Garagnani, P., Parini, P., Giuliani, C., & Santoro, A. (2018). Inflammaging: A new immune-metabolic viewpoint for age-related diseases. Nature Reviews Endocrinology, 14(10), 576–590. https://doi.org/10.1038/s41574-018-0059-4