The causal relationship between 25-hydroxyvitamin D and serum lipids levels: A bidirectional two-sample mendelian randomization study

Abstract

Serum vitamin D levels were linked to lipid metabolism in observational studies, but the exact mechanism was unclear. Several studies have attempted to decipher the relationship between 25(OH)D and lipid levels. Conventional observational studies are vulnerable to confounding. Mendelian randomization (MR) analysis can better control for confounding factors and reverse causality, allowing for the inference of causal association. We, therefore, sought to use MR to investigate the possible causal relationship between 25(OH)D and blood lipid levels (HDL cholesterol, LDL cholesterol, triglycerides, and total cholesterol). A bidirectional two-sample Mendelian randomization (MR) was performed on data primarily from European ancestors. In addition, the potential causal effect of lipids on 25(OH)D was assessed by regressor-based multivariate magnetic resonance (MVMR). The single-nucleotide polymorphisms (SNPs) related to 25(OH)D were selected from a large-scale genome-wide association study (GWAS) database named IEU GWAS, and the SNPs associated with the four blood lipids were chosen from UK Biobank (UKB) lipid GWAS. When blood lipids were the outcome, the results of bidirectional two-sample MR demonstrated that 25(OH)D exhibited a negative causal association with TG, TC, and LDL-C: β = - 0.23, 95% CI = -0.28 to -0.19, P<0.001; β = - 0.16, 95% CI: - 0.30 to—0.03, P < 0.05; β = - 0.11, 95% CI: - 0.23 to 0, P < 0.05. There was no causal relationship between 25(OH)D and HDL-C (β = 0.05, 95% CI: - 0.11 to 0.20, P = 0.56). When setting blood lipids as exposure, TG and 25(OH)D, β = -0.13, 95% CI: - 0.15 to -0.10, P < 0.05; TC and 25(OH)D, β = -0.11, 95% CI: - 0.15 to -0.07, P < 0.05; HDL-C and 25(OH)D, β = 0.02, 95% CI: 0 to 0.03, P = 0.07; LDL-C and 25(OH)D, β = -0.08, 95% CI: - 0.11 to -0.05, P < 0.05). Our MVMR study also showed a significant relationship between genetically determined lipid traits and 25(OH)D levels (TG and 25(OH)D, P < 0.05; TC and 25(OH)D, P < 0.05). In all MR analyses, there was no horizontal pleiotropy (all P > 0.05), or statistical heterogeneity. The "Leave-one-out" sensitivity analysis confirmed the stability of our results. MR Studies have shown a bidirectional causal relationship between genetically-determined 25(OH)D levels and serum TG and TC levels. The findings have potential implications for etiological understanding and disease prevention.