Highlights
– In a cross‑sectional cohort of 200 ambulatory patients with type 2 diabetes, metformin use was associated with significantly lower serum vitamin B12 (227.1 ± 96.9 vs 325.6 ± 176.8 pmol/L; p < 0.001) and a threefold increased adjusted odds of B12 deficiency.
– No differences were observed between metformin users and non‑users in serum vitamins A, B1 (thiamine), B6 (pyridoxine), B9 (folate), C or E, consistent with a selective effect on B12 absorption or metabolism.
– Findings support targeted surveillance for vitamin B12 deficiency in patients on long‑term metformin and reinforce mechanistic hypotheses linking metformin to impaired intestinal B12 uptake.
Background
Metformin is first‑line pharmacotherapy for type 2 diabetes because of its glycemic efficacy, safety profile, weight‑neutral or modest weight‑loss effects, cardiovascular benefits, and low cost. Longstanding clinical and epidemiologic observations have associated chronic metformin exposure with low serum vitamin B12 concentrations and symptomatic B12 deficiency in a subset of patients. Vitamin B12 is required for hematopoiesis and neurological function; deficiency can cause macrocytic anemia and irreversible neuropathy if unrecognized.
Most prior work focused on B12 alone or small cohorts. The study by Vigolo and colleagues (2025) adds to the literature by simultaneously measuring multiple vitamin concentrations (B12, A, B1, B6, B9, C and E) in a contemporary ambulatory type 2 diabetes population to test whether the effect of metformin is selective for B12 or represents a broader effect on intestinal vitamin absorption.
Study Design
Vigolo et al. conducted a cross‑sectional analysis in 200 ambulatory patients with type 2 diabetes. Serum concentrations of vitamins B12, A, B1, B6, B9 (folate), C and E were obtained. Patients were dichotomized by metformin use (current users versus non‑users). The main comparisons were mean vitamin concentrations between groups and the prevalence of biochemical vitamin deficiencies. Multivariable logistic regression was used to adjust for potential confounders and estimate the association between metformin exposure and B12 deficiency.
Key Findings
Metformin users had markedly lower mean serum B12 levels than non‑users (227.1 ± 96.9 vs 325.6 ± 176.8 pmol/L; p < 0.001). The prevalence of biochemical B12 deficiency in the cohort was 21.1%, and metformin use was associated with approximately a threefold higher adjusted odds of deficiency in multivariable analysis. By contrast, measured concentrations of vitamins A, B1, B6, B9 (folate), C and E were not significantly different between metformin users and non‑users.
These results suggest a selective association between metformin and lower circulating B12 rather than a generalized malabsorption or nutritional deficiency affecting multiple vitamins.
Effect sizes and clinical meaning
The absolute between‑group difference in mean B12 (~98 pmol/L) is clinically relevant: depending on assay and local reference ranges, values in the low 200s pmol/L may fall into the borderline or deficient range, and deficiency prevalence of ~20% is high for a treated outpatient population. The reported adjusted threefold increase in odds of deficiency indicates a strong association, although the cross‑sectional design limits causal inference.
Secondary observations and safety
No excess of other vitamin deficiencies were found, which lowers the likelihood that general poor intake, broad intestinal malabsorption, or selection bias fully account for the B12 effect. The study did not report clinical outcomes such as neuropathy, anemia, or functional markers of B12 deficiency (methylmalonic acid [MMA] or homocysteine), which are important for assessing clinical impact.
Mechanistic considerations and biological plausibility
Several mechanisms have been proposed to explain metformin‑related B12 deficiency. The leading hypothesis implicates impaired calcium‑dependent binding of the intrinsic factor–B12 complex to ileal cubilin receptors, thereby reducing active B12 absorption. Other proposed mechanisms include alterations in gut microbiota with increased bacterial B12 consumption and changes in bile salt metabolism or small bowel motility. The selectivity observed by Vigolo et al. (affecting B12 but not other fat‑ or water‑soluble vitamins) is consistent with a mechanism targeting the intrinsic factor–mediated ileal uptake pathway, which is specific to cobalamin.
It should be noted that serum total B12 may not always reflect tissue stores or functional deficiency. Elevated MMA or homocysteine is a more sensitive indicator of tissue B12 insufficiency; these markers were not reported in the study, which limits assessment of the functional significance of the biochemical reductions.
Clinical implications
For clinicians caring for patients with type 2 diabetes on metformin, particularly those on long‑term or high‑dose therapy, these findings reinforce existing concerns and argue for pragmatic strategies:
- Consider baseline and periodic measurement of vitamin B12 in patients on chronic metformin, especially those with symptoms consistent with neuropathy, macrocytic anemia, or with other risk factors for deficiency (e.g., older age, prior gastrectomy, strict vegan diet).
- Use functional testing (methylmalonic acid, homocysteine) when serum B12 is borderline (~150–300 pmol/L) or when clinical suspicion is high.
- Treat documented deficiency with standard cobalamin replacement (intramuscular or high‑dose oral cobalamin), and reassess clinically and biochemically. Oral high‑dose replacement is effective in many cases and may be preferred for long‑term management.
- Educate patients about the potential risk and symptoms of B12 deficiency and the rationale for testing rather than changing metformin therapy, because metformin’s benefits often outweigh the treatable risk of B12 deficiency.
Expert commentary and guideline context
Professional organizations have increasingly recognized the association between metformin and B12 deficiency. The American Diabetes Association suggests consideration of periodic B12 testing in patients on long‑term metformin or in those with peripheral neuropathy. The pattern reported by Vigolo et al., showing selectivity for B12 and not for other vitamins, aligns with prior prospective and observational studies (e.g., De Jager et al., Aroda et al.) and mechanistic work led by investigators such as Bauman that implicated altered ileal absorption.
However, gaps remain. The cross‑sectional design cannot prove causality or temporal sequence, the absence of functional markers limits inference about clinically meaningful deficiency, and potential confounders (dietary intake, duration and dose of metformin, concomitant medications such as proton pump inhibitors) require careful adjustment or stratified analyses. Randomized or longitudinal observational data that include MMA/homocysteine and clinical endpoints (neuropathy progression, anemia, cognitive outcomes) would offer stronger evidence for testing and treatment thresholds.
Limitations of the study
Key limitations include cross‑sectional design, lack of data on metformin dose and duration in the report summary, absence of functional biomarkers (MMA/homocysteine), limited information on dietary intake and other medications (e.g., PPIs, H2 blockers) that can affect B12, and lack of correlation with clinical outcomes. The cohort size (n=200) is moderate but not large enough to precisely estimate subgroup effects. Finally, assay variability and different laboratory reference ranges for B12 can influence prevalence estimates.
Research and practice gaps
Important unanswered questions include the dose‑response relationship between metformin and B12 decline, time to onset of deficiency, the proportion of patients with biochemical deficiency who develop clinical deficits, optimal monitoring intervals, and whether routine prophylactic supplementation is cost‑effective. Prospective studies with serial measurements, functional assays, and clinical outcomes are needed to define evidence‑based screening and management algorithms.
Conclusion
The study by Vigolo et al. strengthens evidence that metformin is associated with lower circulating vitamin B12 levels and a higher risk of biochemical deficiency while not affecting a range of other measured vitamins. The selectivity supports mechanistic models of impaired ileal B12 uptake. Clinicians should be alert to this association and consider appropriate testing, particularly for patients on long‑term or high‑dose metformin and those with symptoms compatible with B12 deficiency. Treatment of confirmed deficiency is straightforward and prevents potentially irreversible neurologic damage, allowing continued metformin therapy when indicated.
Funding and trial registration
The original article did not indicate funding or clinical trial registration in the provided excerpt. Readers should consult the full publication for declarations of funding, conflicts of interest, and ethical approvals.
References
1. Vigolo N, Toffalini A, Rolli N, Paviati E, Gelati M, Trombetta M, Danese E, Zoppini G. Metformin is associated with low levels of vitamin B12 with no effect on other vitamin levels. A selective action of metformin. Nutr Metab Cardiovasc Dis. 2025 Nov;35(11):104193. doi: 10.1016/j.numecd.2025.104193. Epub 2025 Jun 11. PMID: 40610296.
2. De Jager J, Kooy A, Lehert P, et al. Long term treatment with metformin in type 2 diabetes and vitamin B12 deficiency. Diabetes Care. 2010;33(2):266–271. (See Diabetes Care 2010 for details.)
3. Aroda VR, et al. Long‑term metformin use and vitamin B12 deficiency in patients with prediabetes and diabetes: findings from the Diabetes Prevention Program Outcomes Study. Diabetes Care. 2016;39(2):??–??. (See Diabetes Care 2016 for full report and outcomes.)
4. American Diabetes Association. Standards of Medical Care in Diabetes—2024. ADA; 2024. (Guidance includes consideration of B12 testing in patients with neuropathy or on long‑term metformin.)
5. Bauman WA, et al. Proposed mechanisms and evidence for metformin‑associated vitamin B12 malabsorption. (Seminal mechanistic reports exploring calcium‑dependent ileal absorption and intrinsic factor–receptor interactions.)
Note: For precise citation details (page numbers, issue), consult the primary publications and professional society guidance documents.
