Metformin B12 deficiency · Screening

Metformin B12 Deficiency: Screening Beyond the Serum Level

Metformin roughly doubles the risk of B12 deficiency, and serum B12 alone routinely misses it. Screening metformin users with a functional marker such as methylmalonic acid or homocysteine, or with holotranscobalamin, closes that gap.

The patient most likely to be losing B12 is the metformin user, and that patient was nearly absent from the trial showing B vitamins slow brain atrophy.

The chapter video, Your Lab Says Normal. Your Brain Is Shrinking., is aimed at patients, but its sharpest implication is for the clinician holding the requisition. Metformin B12 deficiency is common, dose- and duration-dependent, and routinely missed when serum B12 is the only marker ordered. The insulin-resistant patient on metformin is also nearly absent from the one randomized trial showing that B vitamins slow brain atrophy. That combination is a screening gap, and it closes with better ordering, not new technology.

Is a normal serum B12 enough to rule out deficiency?

In Oxford’s OPTIMA cohort, 107 cognitively normal adults aged 61–87 were scanned annually for five years. None had plasma B12 below 150 pmol/L. The bottom tertile, below 308 pmol/L, carried an adjusted OR of 6.17 (95% CI 1.25–30.47) for membership in the fastest-atrophying third; low holotranscobalamin, below 54 pmol/L, gave OR 5.99. Methylmalonic acid and homocysteine trended higher in fast losers but did not survive adjustment in a sample that small. In the Stockholm cohort of 501 adults aged 60 and over, each SD of higher B12 and holotranscobalamin predicted slower whole-brain loss and each SD of homocysteine predicted faster loss.

The nulls deserve equal space. An Austrian cohort found no association between any B12 marker and atrophy, and the Dutch B-PROOF substudy found none with volume, though it scanned once and so measured volume rather than rate. The practice position is that the low-normal band matters clinically. That rests on the rate cohorts, not on any settled consensus, and it is stated here as a position.

The size of that band in the general population is not trivial. Among 12,335 U.S. adults, 2.2% had serum B12 below 148 pmol/L and a further 13% sat between 148 and 222 pmol/L. For cognitive complaints specifically, total B12 discriminated deficiency with an AUC of 0.54, while holotranscobalamin reached 0.69. The test most often accepted as the answer performed close to chance for the indication that brings older patients in.

How much does metformin lower B12, and do acid blockers add to it?

Across 31 studies, metformin roughly doubled the risk of biochemical B12 deficiency (RR 2.09, 95% CI 1.49–2.93), with a mean fall of 63.70 pmol/L that scaled with dose and duration. In a randomized trial of 390 insulin-treated patients followed 4.3 years, metformin lowered B12 19% against placebo, one new deficiency per 13.8 patients treated. Five-year DPP follow-up put low-or-borderline status at 19.1% with metformin against 9.5% with placebo, each year of use carried OR 1.13, and neuropathy was more common in users with low levels.

Acid suppression layers on top. In Kaiser Permanente data, two or more years of a proton pump inhibitor carried OR 1.65 for a deficiency diagnosis; a 2023 meta-analysis pooled a smaller 1.42 and judged the association unclear. Treat it as a contributor that stacks with metformin in the same older patient. None of this argues against metformin. It argues that a functional B12 result belongs on every metformin chart.

The upstream driver is the food environment. A diet built on refined energy is low in both B12 and preformed EPA and DHA while it drives the insulin resistance that earns the metformin prescription. The system produces the deficiency, then prescribes the drug that deepens it, and the visit length leaves no room to connect the two.

What VITACOG does and does not tell you

In VITACOG, 271 patients with mild cognitive impairment, all aged 70 and over, were randomized to folic acid 0.8 mg, B12 0.5 mg and B6 20 mg daily or placebo for two years. Whole-brain atrophy ran 0.76% a year on treatment against 1.08% on placebo, 29.6% slower. Above a baseline homocysteine of 13.0 µmol/L the slowing was 53.3%; at or below 9.5 µmol/L there was none. Voxel-based analysis localized the protection to Alzheimer-vulnerable gray matter.

The limits matter for your panel. Diabetes was rare: 10 placebo and 4 active participants in the imaging group, with faster atrophy that looked unresponsive on numbers too small to read. The trial was powered for atrophy, not cognition. The closest replication, in 279 Hong Kong outpatients, found no difference in its cognitive primary outcome and had no MRI endpoint. The investigators hold declared patents on this indication. And a retrospective subgroup analysis found the vitamins slowed atrophy 40.0% only in the top omega-3 tertile, with no significant effect in the bottom tertile.

Dr. Padda treats B vitamins and omega-3 as a pair on the strength of those subgroups, a mirror-image result in the OmegAD trial and a tracer study, against one contradicting trial and no trial that randomized both. That is practicing ahead of the evidence, and the practice says so in writing rather than presenting the pairing as proven.

What does an abnormal functional B12 result change?

A raised methylmalonic acid or low holotranscobalamin with a normal serum B12 reclassifies the patient from reassured to insufficient, which changes the next decision. Homocysteine sorts who is likely to respond: the VITACOG investigators propose at or below 10 µmol/L as probably safe and 11 or above as possibly justifying intervention, a proposal from the same patent-holding group. Folate status has to be read beside B12. Once B12 fell below 148 pmol/L in national survey data, rising serum folate went with higher homocysteine and methylmalonic acid, and low B12 with high folate carried OR 2.6 for cognitive impairment in 1,459 older Americans. Across 25 trials, folic acid lowered homocysteine 23% and added B12 a further 7%, while three trials using 15 mg a day saw more cardiovascular events. More vitamin is not more benefit.

Pairing the panel with cognition and nerve function

Measura [Cardiometabolic and Autonomic Health Analysis] supplies the measurement layer and returns findings to the ordering physician; interpretation and management stay in your hands. Three pieces fit this patient. Laboratory panels carry whatever markers you specify, so the order can name methylmalonic acid, holotranscobalamin, homocysteine and folate beside serum B12. A cognitive assessment sets a documented baseline before any decline is clinically obvious, which is the only way to show later that a trajectory changed. Sudomotor testing quantifies small-fiber sudomotor output at the hands and feet, a distal autonomic readout that matters here because neuropathy tracked with low B12 among metformin users. It does not measure B12; it shows the nerve terrain in the same patient. The pairing with gait and falls risk is covered under cognitive assessment and fall prevention.

Making it reproducible in the workflow

Screening that depends on memory does not happen. A written protocol does. A reasonable trigger list from this evidence: any patient on metformin, two or more years of acid suppression, age over 60 with a cognitive complaint, and anyone whose serum B12 sits in the low-normal band. Build those into standing orders for screening so staff can queue functional markers without a separate decision each time; the rationale for that approach is laid out in making screening reproducible. The cognitive baseline and medication review both belong in annual wellness visit documentation. The companion post on B vitamins and the timing of mild cognitive impairment carries the argument into later life, and every cited study, with its limits, is collected in the Starved Brain evidence briefing for Chapter 5.

Frequently asked questions

Which patients should get functional B12 markers rather than serum B12 alone?

The evidence supports prioritizing metformin users, patients with two or more years of acid suppression, older adults with cognitive complaints and anyone in the low-normal serum band. In those groups total B12 discriminates poorly, and methylmalonic acid, homocysteine and holotranscobalamin show whether the vitamin is working at the tissue level. Detailed selection criteria are on who to test.

Does a low B12 on metformin mean the drug should change?

Not on that basis. The metformin data describe a biochemical effect that is dose- and duration-dependent, and the sources frame it as a reason to monitor B12, not a case against the drug. Glycemic management and repletion decisions remain with the treating physician, weighed against the patient’s full metabolic picture. The broader reasoning behind measuring first is on clinical rationale.

Is the omega-3 interaction strong enough to change practice?

It is post hoc and hypothesis-generating: a retrospective subgroup of one trial, unadjusted for multiplicity, using plasma rather than red-cell omega-3. A mirror-image result in OmegAD supports it, and the FACIT folic acid analysis runs the other way. It is a reasonable argument for measuring both nutrient systems together rather than a proven rule. How paired markers appear in results is covered in interpreting the report.

How should functional B12 results be documented?

Record the marker, the value, where it sits relative to the reference interval, the medication context such as metformin or acid suppression, and the plan. Pairing the entry with the cognitive baseline lets the chart show a trajectory at the next review instead of isolated numbers. Structured result delivery into the record is described on getting results into the record.

Does metformin cause B12 deficiency?

Yes, and it is common. Across 31 studies, metformin roughly doubled the risk of biochemical B12 deficiency. In five-year Diabetes Prevention Program follow-up, low or borderline B12 ran at 19.1% with metformin against 9.5% with placebo, and neuropathy was more common in users with low levels. Serum B12 alone routinely misses it, which is why methylmalonic acid, homocysteine or holotranscobalamin belongs on the order.

How quickly can metformin cause B12 deficiency?

It builds with dose and time rather than overnight. Across 31 studies, the fall in B12 scaled with both. In a randomized trial that followed insulin-treated patients for 4.3 years, metformin lowered B12 19% against placebo. In Diabetes Prevention Program follow-up, each year of metformin use raised the odds of low or borderline B12 further. The longer the prescription, the stronger the case for a functional B12 result on the chart.

Should I take B12 if I take metformin?

Measure first, then decide. A functional B12 result belongs on every metformin chart: methylmalonic acid, homocysteine or holotranscobalamin, read beside serum B12 and folate. A raised methylmalonic acid or a low holotranscobalamin with a normal serum B12 changes the next decision. More vitamin is not more benefit, repletion stays with the treating physician, and none of this is a reason to stop metformin.

References

• Vogiatzoglou, A., Refsum, H., Johnston, C., Smith, S. M., Bradley, K. M., de Jager, C., Budge, M. M., & Smith, A. D. (2008). Vitamin B12 status and rate of brain volume loss in community-dwelling elderly. Neurology, 71(11), 826–832. https://doi.org/10.1212/01.wnl.0000325581.26991.f2
• Mineva, E. M., Sternberg, M. R., Bailey, R. L., Storandt, R. J., & Pfeiffer, C. M. (2021). Fewer US adults had low or transitional vitamin B12 status based on the novel combined indicator of vitamin B12 status compared with individual, conventional markers, NHANES 1999-2004. The American Journal of Clinical Nutrition, 114(3), 1070-1079. https://doi.org/10.1093/ajcn/nqab122
• Yang, W., Cai, X., Wu, H., & Ji, L. (2019b). Associations between metformin use and vitamin B12 levels, anemia, and neuropathy in patients with diabetes: A meta-analysis. Journal of Diabetes, 11(9), 729–743. https://doi.org/10.1111/1753-0407.12900
• de Jager, J., Kooy, A., Lehert, P., Wulffelé, M. G., van der Kolk, J., Bets, D., Verburg, J., Donker, A. J. M., & Stehouwer, C. D. A. (2010). Long term treatment with metformin in patients with type 2 diabetes and risk of vitamin B-12 deficiency: Randomised placebo controlled trial. BMJ, 340, c2181. https://doi.org/10.1136/bmj.c2181
• 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., for the 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
• 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), e12244. https://doi.org/10.1371/journal.pone.0012244
• Jernerén, F., Elshorbagy, A. K., Oulhaj, A., Smith, S. M., Refsum, H., & Smith, A. D. (2015). Brain atrophy in cognitively impaired elderly: The importance of long-chain ω-3 fatty acids and B vitamin status in a randomized controlled trial. The American Journal of Clinical Nutrition, 102(1), 215–221. https://doi.org/10.3945/ajcn.114.103283
• Murphy, M. J., Brandie, F., Ebare, M., Harrison, M., Dow, E., Bartlett, W. A., & Craig, D. (2021). Personalising laboratory medicine in the ‘real world’: Assessing clinical utility, by clinical indication, of serum total B12 and Active-B12 (holotranscobalamin) in the diagnosis of vitamin B12 deficiency. Annals of Clinical Biochemistry, 58(5), 445–451. https://doi.org/10.1177/00045632211003605
• Morris, M. S., Jacques, P. F., Rosenberg, I. H., & Selhub, J. (2007). Folate and vitamin B-12 status in relation to anemia, macrocytosis, and cognitive impairment in older Americans in the age of folic acid fortification. The American Journal of Clinical Nutrition, 85(1), 193–200. https://doi.org/10.1093/ajcn/85.1.193
• 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

Related reading

• Statins and Dementia: Measuring Cognition in Patients on Statins
• B Vitamins in Mild Cognitive Impairment: Measure Before It Is Late
• Laboratory Panels

Medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP, medical director of Measura. Last reviewed September 12, 2026.