BMI mediates the effects of gut microbes on bone mineral density in the "intestinal bone axis": an observational study versus a mediation Mendelian randomization study

Abstract Background Several pieces of evidence suggest a strong association between gut microbiota (GM) and bone mineral density (BMD), but the intermediate factors between them are still unknown. While one study proposes that body mass index (BMI) might serve as an intermediary factor between gut microbiota and bone mineral density, there is insufficient evidence to substantiate this claim. We conducted an observational study and a mediation mendelian randomization analysis to investigate whether BMI could function as a mediator between gut microbiota and bone mineral density, thereby addressing this gap in knowledge. Methods We conducted an observational study using data from the National Health and Nutrition Examination Survey (NHANES) collected between 2013 and 2018. The independent effect of BMI on BMD was verified by using multiple linear regression analysis to exclude interference from confounders. We subsequently performed three-category Mendelian randomization analyses ((a) We used 211 categories of gut microbes as the exposure and eBMD as the outcome; (b) we used 211 categories of gut microbes as the exposure and BMI as the outcome; (c) we used BMI as the exposure and eBMD as the outcome). Ultimately, we identified one bacterium that could act as a mediator in multivariable and mediation Mendelian randomization analysis. For the primary analysis, we employed inverse variance weighting (IVW) and combined it with four other analysis methods and multiple sensitivity analyses, including heterogeneity analysis, horizontally pleiotropic analysis, "leave-one-out" analysis, MR-PRESSO, and MR-LASSO. Results Our multiple linear regression analysis showed that BMI had an independent influence on BMD (β = 0.011, t = 8.155, P < 0.05), after adjusting for other confounding factors. The results of our Mendelian randomization analysis revealed that eight bacterial genera were significantly associated with eBMD, while 15 bacterial genera were significantly associated with BMI. The results of the multivariate MR analysis of phylum-Actinobacteria showed that BMI acted as a mediator between phylum-Actinobacteria and eBMD, with BMI accounting for 84.9% of the intermediate effect. Conclusion Our study found that BMI fully mediated the association between phylum-Actinobacteria and eBMD. However, we cannot extrapolate this finding to suggest that BMI can mediate the association between other gut microbiota and eBMD. Our findings suggest that phylum-Actinobacteria could potentially serve as a biomarker or therapeutic target for osteoporotic patients who are obese. Modulating the relative abundance of phylum-Actinobacteria may be particularly effective in obese patients with osteoporosis. Further studies are required to confirm these findings.