We read with interest the excellent article by Ting and colleagues1 about risk factors of vitamin B12 deficiency in patients receiving metformin. We would like to offer some findings from our own experience. To date, more than 30 patients with metformin-associated vitamin B12 deficiency have been registered in a cohort study of cobalamin deficiency in the University Hospital of Strasbourg, Strasbourg, France (partial data were previously published2). An analysis of these patients (median age, 66 years) proves some of the comments by Ting et al.1 First, the clinical severity of metformin-associated vitamin B12 deficiency is moderate, with 90% of minor hematological abnormalities (median hemoglobin levels and mean erythrocyte cell volume, 11.5 g/dL and 95.4 fL, respectively) and 30% of mild peripheral sensitive neuropathy. However, there were 2 cases (5.7%) of symptomatic anemia and pancytopenia (requiring transfusions). The mean ± SD serum vitamin B12 and total homocysteine levels were 156 ± 31 pg/mL (range, 87-200 pg/mL) (115.1 ± 22.9 pmol/L [range, 64-148 pmol/L) and 2.0 ± 0.8 mg/L (range, 2-3 mg/L) (14.8 ± 5.7 μmol/L [range, 13-22 μmol/L]). Second, an analysis of our cases suggests that the principal mechanism of vitamin B12 deficiency in patients receiving metformin is undoubtedly food-cobalamin malabsorption. In fact, of 34 patients' responses to the diagnostic criteria we had previously published,3 1 patient had a malabsorption with chronic diarrhea.2 In our opinion, this mechanism of food-cobalamin malabsorption may explain the constant efficacy of oral crystalline cyanocobalamin therapy that we have observed as a benefit of calcium supplementation. To our knowledge, discontinuation of metformin therapy remains controversial, but our results suggest that it may be unnecessary.