Association between nonalcoholic fatty liver disease and bone mineral density: Mendelian randomization and mediation analysis.

Abstract

Observational studies have reported significant association between non-alcoholic fatty liver disease (NAFLD) and bone mineral density (BMD), a critical indicator of bone health. We aimed to investigate whether NAFLD is a cause for changes in BMD. We selected 29 independent SNPs as instrumental variables for NAFLD. A range of Mendelian randomization (MR) methods, namely the inverse variance-weighted (IVW) method, weighted-median, weighted-mode, and MR-Egger regression, were utilized to determine the causal effects of NAFLD on BMD. Two-step MR analysis was conducted to determine the mediating effect of fasting glucose, insulin, glycosylated hemoglobin, low-density cholesterol, and body-mass index on the association between NAFLD and BMD. False-discovery-rate (FDR) was used to correct for multiple testing bias. The IVW-method indicated a significantly inverse association between genetically predicted NAFLD and total body BMD (β = -0.04, 95 % CI -0.07 to -0.02, FDR = 0.010). Notably, the relationship was more pronounced in participants over 60 years of age (β = -0.06, 95 % CI -0.11 to -0.02, FDR = 0.030). Inverse associations were observed in other subpopulations and in site-specific BMD, though they were not statistically significant after correcting for multiple testing. We observed a significantly positive association between NAFLD and the risk of osteoporosis. Consistency in results was observed across multiple MR methods and in the repeated analysis. Fasting glucose, insulin, and glycosylated hemoglobin mediated 25.4 % (95 % CI 17.6-31.5 %), 18.9 % (12.0-24.9 %), and 27.9 % (19.9-36.7 %) of the effect of NAFLD on BMD, respectively. Our findings underscore a probable causal negative link between NAFLD and BMD, indicating that NAFLD might detrimentally affect bone health, especially in older individuals.