Role of lifestyle factors in mediating the effect of educational attainment on bone mineral density: a Mendelian randomization study

We performed two-step multivariable Mendelian randomization analysis to explore the mediating role of lifestyle factors in educational attainment (EA) and bone mineral density (BMD). Summary statistics from genome-wide association studies of European lineages were used. Coffee intake and processed-meat intake mediated the association between EA and BMD.Purpose This study aimed to explore the causal relationship between educational attainment (EA) and bone mineral density (BMD), as well as the potential mediating roles of lifestyle factors in the expected EA-BMD relationship. By identifying modifiable lifestyle factors, we hope to provide relevant information to prevent osteoporosis or low BMD in the less educated population.Methods Using summary statistics from genome-wide association studies (GWAS) of major European lineages, one- and two-sample Mendelian randomization (MR) analyses were performed to estimate the association between EA (in the social sciences genetic association consortium (SSGAC) involving 766,345 individuals and in the UK Biobank (UKB) involving 293,723 individuals) and BMD (in the Genetic Factors for Osteoporosis Consortium involving 426,824 individuals selected from the UKB). The EA variable in both consortia were expressed by years of schooling completed. Two-step multivariable MR was used to assess the mediating roles of eight lifestyle-related factors (moderate-to-vigorous physical activity, watching television, computer using, smoking initiation, coffee intake, alcohol intake frequency, tea intake, and processed-meat intake) in the EA and BMD association, and the corresponding mediating proportion was calculated. Meta-analysis was used to present a pooled estimate.Results A total of 317 and 73 independent single-nucleotide polymorphisms (SNPs) of GWAS significance (P < 5.0 x 10-8) were selected as instrumental variables (IVs) for EA in the SSGAC and UKB, respectively. A total of 513 SNPs were selected as IVs for the BMD. The results of one- and two-sample MR revealed that the genetically predicted BMD increased by 0.094 and 0.047 g/cm2, respectively, in response to each SD increment of genetically predicted schooling years. Among the eight candidate mediators, coffee intake and processed-meat intake were potential mediators revealed by the two-step multivariable MR analysis, mediating 26.87% and 23.92% of EA's effect on BMD, respectively. Meta-analysis showed consistent findings. Results of sensitivity analysis indicated the robustness of our findings.Conclusion We elucidated the causal protective effect of EA on BMD and the mediating roles of coffee intake and processed-meat intake. Intervening with these factors can potentially reduce the burden of bone density loss or osteoporotic fractures among the less educated population.