Scientists have identified the first DNA variant associated with a higher susceptibility or increased liability of developing B12 deficiency when a person is treated with metformin.

The analysis identified a genome-wide non-synonymous SNP in the cubilin gene (CUBN; rs1801222/p.S253F) that was significantly associated with metformin-induced vitamin B12 deficiency. This finding was replicated in three Scottish cohorts, in the Diabetes Prevention Program Outcomes Study (DPPOS) cohort, and in a small clinical cohort from Liverpool. Vitamin B12 deficiency occurred in 0.84–1.20% of individuals who were not exposed to metformin, regardless of their rs1801222 genotype. However, a large interaction with metformin use was observed, with vitamin B12 deficiency developing at 6.02% in GG, 7.96% in GA and 12.84% in AA genotype groups.

Co-author Ewan Pearson: “The 14% of the white population with the AA genotype are 2.5 times more likely to require B12 replacement GG. Equates to 10% requiring B12 replacement by 11 years after starting metformin with AA genotype vs 21 years for GG. A large effect!” But the SNP occurs in nearly 30% of Europeans, so many people with the SNP won’t develop a deficiency.

It should be noted that the primary analysis was conducted among participants in the UK Biobank. Individuals were identified with a diagnosis of vitamin B12 deficiency and/or a record of B12 injection prescriptions, and metformin use was extensively documented. However, clinical diagnosis of B12 deficiency is often only done on the basis of abnormal B12 measurements, and -for instance- measurements of methylmalonic acid are rarely done. Furthermore, people who are using (multi)vitamin supplementation may not so easily be recognized as truly B12-deficient. So, the incidence of B12 deficiency may be higher than reported in the paper. Nevertheless, this is great work in the beautiful dataset of UK Biobank, and confirmed in other cohorts, such as Generation Scotland, GoDARTS, and SHARE, and DPPOS. Researchers are eagerly awaiting the time when UK Biobank data will also be enriched with serum B12, MMA, and homocysteine measurements for all participants.

Full source: https://link.springer.com/article/10.1007/s00125-025-06655-5