Electronic database searches were undertaken (1st January 1957–1st July 2013) using the Cochrane library, Scopus, CINAHL, Grey literature databases, Pub Med Central, NICE Clinical Guidelines UK, and ongoing clinical trials. Included studies were of any study design, with data from patients with type 2 diabetes of any age or gender, taking any dose or duration of metformin. Planned primary outcomes were serum vitamin B12 levels, % prevalence or incidence of vitamin B12 deficiency and risk of vitamin B12 deficiency.
Twenty-six papers were included in the review. Ten out of 17 observational studies showed statistically significantly lower levels of vitamin B12 in patients on metformin than not on metformin. Meta-analysis performed on four trials demonstrated a statistically significant overall mean B12 reducing effect of metformin of 57 pmol/L [WMD (fixed) = –0.57 (95% CI: –35 to –79 pmol/L)] after 6 weeks to 3 months of use.
The evidence from this review demonstrates an association between metformin usage and lower levels of vitamin B12 by 57 pmol/L, which leads to frank deficiency or borderline status in some patients with type 2 diabetes. This suggests that it is prudent to monitor B12 levels in these patients who are at increased risk of deficiency.