The study performed by Coen Stehouwer, MD, of Maastricht University in the Netherlands and colleagues online in BMJ noted that with 4.3 years of treatment in a randomized, placebo-controlled trial, the hazard ratio for developing vitamin B₁₂ deficiency while on metformin was 5.5 relative to placebo (95% CI 1.6 to 19.1). One of the main takeaways the researchers concluded was that the deficiency persists and grows over time.  Shorter studies showing slight relation to vitamin B12 deficiency and metformin had been previously recorded, but this is the first long-term study that truly reveals the association.

Throughout this study, they analyzed data on 390 type 2 diabetes patients randomized to metformin or placebo in a study named HOME. This study was designed to examine the effects of metformin on metabolic parameters and micro- and macrovascular diabetic complications.

To being, the patients in the study were initially treated with insulin and no other medications during a 12-week start-up phase. Metformin or placebo was then added to insulin treatment for 52 months.

The average patients chosen for this study had been on insulin treatment to manage their diabetes for around 7 years and had been diabetic for 13 years with a hemoglobin level at 7.9%. The typical age for participants of this study was middle age around 61.

Calculating the data from the 52-month study concluded that although the vitamin B12 levels declined throughout the entire study, the most significant drop occurred during the first few months.

From a mean of about 355 pmol/L at the end of the run-in phase, B₁₂ levels dropped to about 305 pmol/L after four months, and from there to 280 pmol/L over the next four years.

The study also revealed that falling vitamin B₁₂ levels were associated with increased serum homocysteine. When last measured, patients with a deficiency (defined as less than 150 pmol/L) had mean homocysteine levels of 23.7 ?mol/L versus 18.1 ?mol/L among those with vitamin B₁₂ levels of 150 pmol/L, classified as low, and 14.9 ?mol/L for those with normal vitamin B₁₂.

These differences were proven statistically significant, with P=0.003 for deficient versus low, and P=0.005 for low versus normal.

It should be noted that the researchers indicated there was no significant relationship between homocysteine levels and treatment assignments in the trial, but that the “relatively low incidence of vitamin B₁₂ deficiency” probably accounted for the finding.

“As treatment with metformin continues, however, we expect that vitamin B₁₂ levels will continue to decrease, making increases in homocysteine concentrations inevitable in time,” the researchers added.

In conclusion, these associations remained significant after adjusting for age, gender, insulin dose, smoking status, duration of diabetes, and previous metformin treatment.

Citing earlier research, Stehouwer and colleagues wrote that metformin appears to reduce B₁₂ levels by inhibiting dietary absorption in the intestine. Calcium supplements can reverse the malabsorption.

Stehouwer and colleagues argued that the study results “provide a strong case for routine assessment of vitamin B-12 levels during long-term treatment with metformin.”