ABSTRACT
OBJECTIVE: The bidirectional correlation between low bone mineral density (BMD) and frailty, despite its extensive documentation, still lacks a conclusive understanding. The objective of this Mendelian randomization (MR) study is to investigate the bidirectional causal relationship between BMD and frailty. METHODS: We utilized summary statistics data for BMD at different skeletal sites-including heel BMD (e-BMD, N = 40,613), forearm BMD (FA-BMD, N = 8,143), femoral neck BMD (FN-BMD, N = 32,735), and lumbar spine BMD (LS-BMD, N = 28,489), alongside frailty index (FI, N = 175,226) data in participants of European ancestry. MR analysis in our study was conducted using well-established analytical methods, including inverse variance weighted (IVW), weighted median (WM), and MR-Egger approaches. RESULTS: We observed negative causal estimates between genetically predicted e-BMD (IVW ß = - 0.020, 95% confidence interval (CI) = - 0.038, - 0.002, P = 0.029) and FA-BMD (IVW ß = -0.035, 95% CI = -0.066, -0.004, P = 0.028) with FI. However, the results did not reach statistical significance after applying the Bonferroni correction, with a significance threshold set at P < 0.0125 (0.05/4). There was no causal effect of FN-BMD (IVW ß = - 0.024, 95% CI = -0.052, 0.004, P = 0.088) and LS-BMD (IVW ß = - 0.005, 95% CI = -0.034, 0.024, P = 0.749) on FI. In the reverse Mendelian randomization (MR) analysis, we observed no causal effect of FI on BMD at various skeletal sites. CONCLUSION: Our study provides support for the hypothesis that low BMD may be a potential causal risk factor for frailty, but further research is needed to confirm this relationship. However, our findings did not confirm reverse causality.