The effect of plasma lipids and lipid lowering interventions on bone mineral density: a Mendelian randomization study

Abstract

Statin treatment increases bone mineral density (BMD) and reduces fracture risk, but the underlying mechanism is unclear. We used Mendelian randomization (MR) to assess whether this relation is explained by a specific effect in response to statin use, or by a general effect of lipid-lowering. We utilized 400 single nucleotide polymorphisms (SNPs) robustly associated with plasma lipid levels and results from a heel BMD GWAS (derived from quantitative ultrasound) in 426,824 individuals from the UK Biobank. We performed univariate and multivariable MR analyses of low-density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C) and triglyceride levels on BMD. To test whether the effect of statins on BMD was mediated by lowering lipid levels, MR was repeated with and without SNPs in the HMGCR region, the gene targeted by statins. Univariate MR analyses provided evidence for a causal effect of LDL-C on BMD (β = -0.060; -0.084 to -0.036; P = 4x10 -6 ; standard deviation change in BMD per standard deviation change in LDL-C, with 95% CI), but not HDL or triglycerides. Multivariable MR analysis suggested that the effect of LDL-C on BMD was independent of HDL-C and triglycerides, and sensitivity analyses involving MR Egger and weighted median MR approaches suggested that the LDL-C results were robust to pleiotropy. MR analyses of LDL-C restricted to SNPs in the HMGCR region showed similar effects on BMD (β = -0.083; -0.132 to -0.034; P = 0.001) to those excluding these SNPs (β= -0.063; -0.090 to -0.036; P = 8x10 -6 ). Bidirectional MR analyses provided some evidence for a causal effect of BMD on plasma LDL-C levels. Our results suggest that effects of statins on BMD are at least partly due to their LDL-C lowering effect. Further studies are required to examine the potential role of modifying plasma lipid levels in treating osteoporosis.

Use of Mendelian Randomization to Examine Causal Inference in Osteoporosis

“…For example, Kemp et al used a one-sample multivariable method to show that BMI SNPs acted via both lean and fat mass to increase BMD (34). In an MR analysis of relationships between plasma lipids and BMD, we observed a strong inverse association between LDL cholesterol and forearm BMD in multivariate MR analyses adjusting for HDL cholesterol and triglycerides, which was not evident in univariate analyses involving only LDL cholesterol (59). This indicates that complex relationships may exist between the causal effects of different lipids and BMD, which MR analyses need to account for, and may help to explain the conflicting results from different MR analyses examining relationships between lipid levels and eBMD in the UK Biobank (53, 59, 60).…”

“…This analysis suggested that 394 of 404 SNPs exerted their primary effect on BMD as opposed to LDL-C levels. IVW MR, weighted median MR and MR-Egger regression results showed some evidence that eBMD might influence LDL-C, and the association remained unchanged after Steiger filtering to remove those SNPs that primary affected LDL-C levels (59).…”

The causal effect of serum vitamin D concentration on COVID-19 susceptibility, severity and hospitalization traits: a Mendelian randomization study

“…Trait Cases Controls Total Proportion of cases r 2 = 0.01 r 2 = 0.02 r 2 = 0.03 r 2 = 0.04 r 2 = 0.05 25 variance is explained by the genetic variants used in the current study.…”

“…When selecting SNPs from the GWAS, especially in very large GWAS, it can be difficult to determine whether a SNP has its primary association with the exposure being studied or the outcome [25]. For and outcome (COVID-19 phenotypes).…”