Total Hip Bone Area Affects Fracture Prediction With FRAX® in Canadian White Women

Context: Areal bone mineral density (BMD) measurements are confounded by skeletal size. Hip BMD is an input to the FRAX® tool (Centre for Metabolic Bone Diseases, University of Sheffield, United Kingdom), but it is unknown whether performance is affected by hip area. Objective: To examine whether fracture prediction by FRAX® is affected by hip area. Design and Setting: Cohort study using a population-based BMD registry. Patients: A total of 58,108 white women aged ≥40 years. Main Outcome Measures: Incident major osteoporotic fracture (MOF; n = 4913) and hip fracture (n = 1369), stratified by total hip area quintile, before and after adjustment for hip axis length (HAL). Results: Smaller hip area was associated with younger age and lower FRAX® scores, whereas incident fractures were greater in those with larger hip area (P for trend < 0.001). Larger hip area quintile increased risk for MOF and hip fracture when adjusted for FRAX® score with BMD (P for trend < 0.001). Each standard deviation increase in hip area was associated with greater risk for incident MOF [adjusted hazard ratio (HR), 1.08; 95% confidence interval (CI), 1.05 to 1.11] and hip fracture (HR, 1.16; 95% CI, 1.11 to 1.21), but not after adjustment for HAL. FRAX® with BMD underestimated MOF risk in the largest hip area quintile and underestimated hip fracture risk in the three largest hip area quintiles. Conclusions: In Canadian white women, skeletal size based on hip area affects fracture risk assessment based on FRAX® score with BMD, with risk underestimated in those with larger hip areas. Including HAL in the risk assessment compensates for this confounding by skeletal size and provides for more accurate assessment of fracture risk.