Abstract
Identification of causative risk factors amenable for modification is essential for the prevention and treatment of osteoporosis. Observational studies have identified associations between several potentially modifiable risk factors and osteoporosis. However, observational studies are susceptible to confounding, reverse causation bias, and measurement error, all of which limit their ability to provide causal estimates of the effect of exposures on outcomes, thereby reducing their ability to inform prevention and treatment strategies against bone loss and fractures. In addition, not all risk factors are suitable for an analysis in a randomized clinical trial. Mendelian randomization is a genetic epidemiological method that exploits genetic variants as unbiased proxies for modifiable exposures (e.g., biomarkers, adiposity measures, dietary factors, and behaviors) to determine the causal relationships between exposures and health outcomes. This technique has been used to provide evidence of causal associations of serum estradiol concentrations, smoking, body mass index, and type 2 diabetes with bone mineral density and the lack of associations of serum thyroid stimulating hormone, urate, C-reactive protein, and 25‑hydroxyvitamin D concentrations with bone mineral density in generally healthy populations. This review will briefly explain the concept of Mendelian randomization, the advantages and potential limitations of this study design, and give examples of how Mendelian randomization has been used to investigate questions relevant to osteoporosis.