FC094: Frax is Useful Tool for Predicting All-Cause Mortality in Addition to the Risk of Fragile Fracture in Hemodialysis Patients

BACKGROUND AND AIMS The fracture risk assessment tool (FRAX®) was developed to predict the probability of fragile fractures during the next 10 years for the general population, and it is calculated by inputting risk factors of fractures such as age, gender, height, weight, fracture history, presence or absence of rheumatoid arthritis and steroid therapy. Although contents of FRAX do not include CKD as a risk factor, previous studies have shown that FRAX was useful in predicting fracture risk in maintenance dialysis patients. However, many reports are short-term observational studies targeting fracture risk. In the present study, we examined whether FRAX can be useful to predict the risk of fragile fractures and all-cause mortality in hemodialysis patients through the long-term observation. METHOD A retrospective cohort study was conducted in 273 patients aged 40 to under 90 who started hemodialysis at our hospital between March 2004 and the present and who could be followed for at least 3 years or had a fracture within 3 years. In order to assess the association between FRAX score and fragile fracture or all-cause mortality, a multivariate analysis was performed using the Cox proportional hazard model with factors related to fractures and mortality as covariates. The 10-year probability (%) of major osteoporotic fracture without bone mineral density was adopted as the FRAX score. RESULTS The average follow-up period was 6.7 years, and the average age of patients was 66 years, with women accounting for 29%. Of all 273 patients, 55 developed fragile fractures and 97 died. FRAX scores were divided into two groups (lower FRAX score and higher FRAX score) in accordance with a cut-off value of 8.2%, determined by ROC curve for fracture. Univariate analysis of factors associated with fractures revealed that higher FRAX score (8.2 or more) was a significant factor [hazard ratio (HR): 4.49 (2.59–7.85)]. Multivariate analysis conducted with significant factors in univariate analysis (mean serum albumin level, serum phosphorus level, magnesium level, C-reactive protein level and use or non-use of phosphate binder during the 3 years) as the covariates revealed that higher FRAX score was significantly associated with fractures [HR: 2.40 (1.27–4.5)]. Regarding the association between FRAX score and all-cause mortality, in univariate analysis, higher FRAX score was a significant predictive factor [HR: 4.34 (2.84–6. 64)]. In multivariate analysis adjusted with mean serum albumin level, creatinine level, magnesium level, C-reactive protein level, hemoglobin level, use or non-use of phosphate binder and calcimimetics as the covariates, higher FRAX score was found to be a significant factor of all-cause mortality [HR: 3.22 (1.97–5.22)]. CONCLUSION FRAX score was also a useful tool for predicting fractures and all-cause mortality in hemodialysis patients even in long-term analysis.