U-shaped relationship was observed between serum phosphate and mortality in dialysis patients. The connection between high serum phosphate and mortality can be explained partially by the contribution of phosphate to vascular calcification and oxidative stress in endothelial cells. Epidemiological study about the prior history of hip fracture in Japan revealed that high serum phosphate was associated with lower prevalence of prior fracture by univariate analysis. However, this association was eliminated by including many nutritional variables in multivariate analysis. This analysis also showed that severe hypophosphatemia less than 3 mg/dL was independently associated with high prevalence, implying the connection of malnutrition with bone health. Epidemiological data regarding incident fracture is also needed to know the factors really contributing to bone fragility. There is no way other than to determine target ranges of serum phosphate using observational studies, since interventional trial with hard outcome was practically impossible. The focus of observational studies will move on, in future, to the association between mortality and the methods to decrease serum phosphate level.