Two-sample Mendelian randomization studies revealed a causal relationship between insulin use and osteoporosis: An observational study

Abstract

Objective: To investigate causal associations between diabetes, insulin treatment and osteoporosis using LDSC analysis with a 2-way Mendelian randomization study.

Methods: LDSC analysis was used to estimate the likelihood-scale heritability of the genome-wide association study used with genetic correlation between the 2 genome-wide association study used. Then a 2-sample Mendelian randomization study was performed using 3 methods including inverse variance weighted, MR Egger, and weighted median.

Results: The genetic correlation between diabetes, insulin treatment (h2_Z = 3.70, P = 2.16e-4), osteoporosis (h2_Z = 4.93, h2_p = 8.13e-7) and genes was significant. There was a significant genetic correlation (rg = 0.122, P = 0.0211). There was a causal association between diabetes, insulin treatment and osteoporosis [P = 0.003754, OR (95%CI) = 0.998876 (0.998116-0.999636)], while no causal association existed between osteoporosis and insulin use (P = 0.998116-0.999636) causal association existed (P = 0.333244).

Conclusion: There was a strong genetic correlation between diabetes, insulin treatment and osteoporosis, a causal association between diabetes, insulin treatment and osteoporosis, and no causal association between osteoporosis and diabetes, insulin treatment.

Figure 1.

(A) Leave-one-out sensitivity test. (B) Forest plot. (C) Scatter plot. (D) Funnel plot.