Abstract
The aim of this study was to investigate the effect of common antidiabetic drugs on BMD by two-sample Mendelian randomization (MR). The single nucleotide polymorphisms that were strongly associated with insulin, metformin, rosiglitazone and gliclazide were extracted as instrumental variables (IVs) for MR analysis. The inverse variance weighted (IVW) method was used as the primary MR method to assess the causal effect of antidiabetic drugs on BMD, and other MR methods, including Weighted median, MR Egger and Weighted mode, were used for complementary analysis. Reliability and stability were assessed by the leave-one-out test. In the present work, IVW estimation of the causal effect of insulin on heel BMD demonstrated that there was a null effect of insulin on heel BMD (β = 0.765; se = 0.971; P = 0.430), while metformin treatment had a positive effect on heel BMD (β = 1.414; se = 0.460; P = 2.118*10−3). The causal relationship between rosiglitazone and heel BMD analysed by IVW suggested that there was a null effect of rosiglitazone on heel BMD (β = -0.526; se = 1.744; P = 0.763), but the causal effect of gliclazide on heel BMD evaluated by IVW demonstrated that there was a positive effect of gliclazide on heel BMD (β = 2.671; se = 1.340; P = 0.046). In summary, the present work showed that metformin and gliclazide have a role in reducing BMD loss in patients with diabetes and are recommended for BMD loss prevention in diabetes.