Long-term metformin treatment and risk of peripheral neuropathy in older Veterans

Aim: Our objective was to assess whether increased duration of metformin therapy is associated with incident peripheral neuropathy (PN) in older Veterans with diabetes. Methods: Using national Veterans Affairs registry data from 2002 to 2015, we examined Veterans (50 + years) with diabetes. Long-term metformin therapy was defined as prescription ≥ 500 mg/day, filled for ≥ 6 consecutive months. Metformin therapy duration was examined both as continuous and categorical measures. Incident PN was defined by medical chart review. We estimated unadjusted and adjusted (variables selected a priori) odds ratios (OR) and 95% confidence intervals (CI) using logistic regression. Results: The study included n = 210,004 individuals (mean ± SD: age: 66.2 ± 8.4 yrs, 96% male) prescribed metformin for 47.0 ± 34.0 months. Nineteen percent developed PN during follow-up. After adjusting for age, body mass index, duration of time receiving health care within the VA, smoking status, alcohol abuse, and vitamin B12 testing and treatment, the number of months of metformin treatment was associated with elevated odds for incident PN (aOR (metformin treatment - continuous) = 1.009 (95% CI = 1.009, 1.010); aOR (metformin treatment – categorical (ref: 6-<18 months): 18-<44.1 months = 1.57 (1.51–1.63), 44.1-<61 months = 2.05 (1.97–2.14), 61 + months = 2.69 (2.58–2.79), all p-values < 0.0001). Conclusion: Our study suggests that Veterans treated for at least 18 months with metformin are approximately 2–3 times more likely to develop PN than those treated at least six, but<18 months. Future studies are needed to determine whether the association we found may be due to a decline in vitamin B12 status following metformin initiation.