Randomised controlled trialcited by the Tayebat archivefull text checked★ key study A-ODD-04 in-depth review
Metformin and vitamin B12
Long term treatment with metformin in patients with type 2 diabetes and risk of vitamin B-12 deficiency: randomised placebo controlled trial.
Overview
de Jager J et al.. BMJ 2010 · PMID 20488910 · DOI 10.1136/bmj.c2181
Over 4 years metformin lowered B12 by 19%, causing deficiency in about 1 of every 14 users.
Does the study support the specific claim? Yes · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Metformin and vitamin B12 deficiency
- How was the study done?
Multicentre double-blind placebo-controlled RCT (HOME trial follow-up)
- How many people, animals or samples were studied?
390; Type 2 diabetes treated with insulin
- What was actually found?
B12 -19%; NNH 13.8.
- How large was the observed effect?
-19% B12; +7.2 pp deficiency
- What do the results mean in everyday life?
Check B12 regularly on metformin.
- What are the study's limitations?
Secondary analysis; groups differed in age and smoking at baseline; one third without final blood sample; only blood levels, no clinical symptoms measured; insulin-treated patients.
- What does it say about Elawady's specific claim?
Supports C-ODD-07.
- Do the results apply to Germany and Europe?
Dutch data.
- What can explicitly not be concluded from this study?
That metformin is net harmful.
Scientific in-depth review
1. Full reference
- Authors
- de Jager J, Kooy A, Lehert P, Wulffelé MG, van der Kolk J, Bets D, Verburg J, Donker AJ, Stehouwer CD
- Journal
- BMJ
- Year
- 2010
- Identifier
- PMID 20488910 DOI 10.1136/bmj.c2181 · PMC2874129
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
HOME trial supported by Altana, Lifescan, Merck Sante, Merck Sharp & Dohme and Novo Nordisk (sponsors stated to have no role); authors declare financial support for the submitted work from Merck Sharp & Dohme and no other relevant relationships.
2. Research question
Metformin and vitamin B12 deficiency
3. Study design and population
- Study design
- Randomised controlled trial – Secondary analysis of the HOME multicentre double-blind placebo-controlled RCT (NCT00375388) of metformin added to insulin; vitamin B12/folate/homocysteine measured in stored serum at baseline and 4, 17, 30, 43 and 52 months.
- Population
Patients aged 30-80 with insulin-treated type 2 diabetes at three non-academic hospital outpatient clinics; mean diabetes duration 12-14 years, mean insulin ~61-64 IU/day.
- Country
Netherlands
- Period
4.3 years (16-week short-term plus 48-month long-term phase).
- Sample size
390 randomised (196 metformin, 194 placebo); 385 analysed for B12 (5 excluded for assay discrepancy; 8 supplemented patients excluded after 16 weeks); 277 (72%) on study drug at end; final samples for 256 (127 vs 129).
- Inclusion and exclusion criteria
Insulin-treated type 2 diabetes aged 30-80; concomitant glucose-lowering medication stopped in 12-week run-in (details in earlier HOME publications).
4. Methods and endpoints
- Exposure / intervention
Metformin 850 mg three times daily (actual mean 2,050 mg/day) added to insulin.
- Comparator
Placebo
- Endpoints
Percentage change in vitamin B12, folate, homocysteine; incidence of B12 deficiency (<150 pmol/L) and low B12 (150-220 pmol/L).
- Statistical methods
ITT with last observation carried forward; t tests on log-transformed values; hazard ratios; linear mixed models with adjustment (age, sex, diabetes duration, smoking, BMI, insulin dose etc.); per-protocol and mixed-model imputation sensitivity analyses.
5. Actual results
- Results
Versus placebo: B12 -19% (95% CI -24% to -14%; P<0.001); folate -5% (-10% to -0.4%; P=0.033; not significant after adjustment for BMI and smoking); homocysteine +5% (-1% to 11%; P=0.091). B12 deficiency at study end 19 (9.9%) vs 5 (2.7%): risk difference 7.2 percentage points (2.3-12.1), NNH 13.8 per 4.3 years (43.5 to 8.3); HR 5.5 (1.6-19.1). Low B12: 35 (18.2%) vs 13 (7.0%), difference 11.2 points (4.6-17.9), NNH 8.9; HR 3.0 (1.3-6.6). B12-lowering effect increased with time (P=0.023). Homocysteine 23.7 umol/L with deficiency vs 14.9 with normal B12.
- Absolute risks
B12 deficiency 9.9% vs 2.7% at 4.3 years (absolute difference 7.2 percentage points; NNH 13.8).
- Authors' conclusion
Long-term metformin increases the risk of vitamin B12 deficiency, which results in raised homocysteine; deficiency is preventable; regular B12 measurement during long-term metformin should be strongly considered.
6. Strengths
- Randomised, double-blind, placebo-controlled with 4.3 years follow-up and repeated measurements
- Objective laboratory outcomes with reported assay precision
- Absolute risks and NNH reported; robust to per-protocol and mixed-model analyses
7. Weaknesses
- Secondary analysis; trial powered for other endpoints
- Baseline imbalance despite randomisation (metformin group older, 64 vs 59 years; fewer smokers 19% vs 30%)
- 28% discontinued study drug; final samples for only 66%; LOCF imputation
- Total B12 only (no methylmalonic acid/holotranscobalamin); no clinical outcomes (neuropathy, anaemia)
- Insulin-treated population with long diabetes duration; dietary counselling may limit generalisability
- Partly industry funded (incl. metformin manufacturer MSD)
8. Risk of bias
- Orienting tool
- RoB 2
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Single-reviewer RoB 2-oriented judgement from the full text; formal assessment needs two independent reviewers and the protocol. Changed from 'low' (abstract) to 'some_concerns' due to baseline imbalance, missing final samples with LOCF and secondary-analysis status; the B12-lowering effect itself is consistent with other trials (e.g. DPPOS) and is considered robust.
| Domain | Judgement | Rationale |
|---|---|---|
| D1 randomisation process | some concerns | Computer randomisation with identical-looking boxes, but notable baseline imbalances in age and smoking. |
| D2 deviations from intended interventions | low | Double-blind placebo; ITT analysis. |
| D3 missing outcome data | some concerns | Final samples for 256/390; LOCF; more discontinuations for adverse effects with metformin (30 vs 16); sensitivity analyses consistent. |
| D4 measurement of the outcome | low | Laboratory assays on stored samples; unlikely to be influenced by allocation. |
| D5 selection of the reported result | some concerns | Secondary outcomes of a registered trial; several time points and cut-offs; results consistent across analyses. |
Full-text review (25 October 2026): Sample flow (277 on drug, 256 final samples), baseline imbalances, deficiency counts (9.9% vs 2.7%), HRs (5.5; 3.0), time interaction, folate adjustment and homocysteine by B12 category added; funding completed (Altana, Lifescan, Merck Sante, MSD, Novo Nordisk); RoB 2 overall low -> some_concerns; bias_risk updated.
9. Assessment of the statistical interpretation
Clear, progressive B12 lowering with a modest absolute increase in biochemical deficiency (NNH 13.8 over 4.3 years; wide CI 8.3-43.5). Effects on homocysteine overall not significant; clinical consequences not measured.
10. Relevance to the specific claim
Supports C-ODD-07.
11. Independent contrary findings
Benefits of metformin: I-ODD-01.
12. Transferability to Germany and Europe
Dutch data.
13. Overall methodological judgement
Correctly quoted by the archive; a real, monitorable and treatable side effect of metformin. The authors recommend B12 monitoring, not avoiding metformin.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | correctly reported |
14. Open questions
None beyond those stated.
15. Linked fact checks
- Every major class of glucose-lowering drug (metformin, sulfonylureas, thiazolidinediones, DPP-4 inhibitors, SGLT2 inhibitors, GLP-1 agonists) carries a documented signal of serious harm in the peer-reviewed literature. – Studies supporting the claim
- Long-term metformin lowers vitamin B12 and increases B12 deficiency (de Jager RCT: -19%, +7.2 percentage points deficiency over 4.3 years; DPPOS confirms). – Studies supporting the claim
Use by the Tayebat archive
Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.
- What the archive says about the study
de Jager RCT in 390 insulin-treated type 2 diabetics: metformin cut serum B12 by 19% and increased absolute deficiency by 7.2 percentage points over 4.3 years (NNH 13.8).
- A. What was actually studied?
Metformin and vitamin B12 deficiency
- B. What did the authors find?
B12 -19%; NNH 13.8.
- C. What conclusions do the authors themselves draw?
Long-term metformin increases B12 deficiency, raising homocysteine; regular B12 measurement should be considered.
- D. What does the archive derive from it?
Used to show that metformin, named by Elawady, is harmful.
- E. Is this derivation scientifically justified?
Numbers correct; the authors' conclusion is monitoring, not avoiding metformin.
- Adequacy of the interpretation (by the archive)
- correctly reported