Abstract
In observational studies, serum urate concentrations are positively associated with bone mineral density (BMD) and reduced risk of fragility fractures, raising the possibility that urate is a direct mediator of bone density. We used Mendelian randomization analysis to examine whether urate has a causal effect on BMD. We analyzed data from the Generation 3 cohort in the Framingham Heart Study (FHS) (N=2501 total; 1265 male, 1236 female). A weighted genetic urate score was calculated using the SLC2A9, ABCG2, SLC17A1, SLC22A11, and SLC22A12 single-nucleotide polymorphisms (SNPs) that explains 3.4% of the variance in serum urate. Mendelian randomization analysis was performed using the two-stage least squares method with >80% power at =0.05 to detect an effect size equivalent to that observed in the ordinary least squares analysis between serum urate and total femur BMD. A strong association between serum urate and BMD was observed in the crude ordinary least squares analysis (total femur crude beta=0.47, p=1.7E-51). In the two-stage least squares analysis using the weighted genetic urate score as the instrumental variable, no significant relationship was observed between serum urate and BMD (total femur crude beta=-0.36, p=0.06). Similar findings were observed in both the male and female subgroups, and there was no evidence for causality when individual SNPs were analyzed. Serum urate is strongly associated with BMD. However, controlling for confounders by Mendelian randomization analysis does not provide evidence that increased urate has a causal effect on increasing BMD. (c) 2015 American Society for Bone and Mineral Research.