Hip axis length is a FRAX-and bone density-independent risk factor for hip fracture in women

Context: Bone mineral density ( BMD ) measurement from dual-energy X-ray absorptiometry ( DXA ) is widely used to assess skeletal strength in clinical practice, but DXA instruments can also measure biomechanical parameters related to skeletal shape. Objective: The objective of the study was to determine whether DXA-derived hip geometry measures provide information on fracture prediction that is independent of hip fracture probability determined from the fracture risk assessment tool ( FRAX ) algorithm. Design and Setting: This was a retrospective registry study using BMD results for Manitoba, Canada. Patients: Women aged 40 years and older with baseline hip DXA, derived hip geometry measures, and FRAX scores ( n = 50 420 ) participated in the study. Main Outcome Measures: Hospitalized hip fracture ( n = 1020 ) diagnosed during 319 137 person-years of follow-up ( median 6.4 y ) was measured. Results: Among the hip geometry measures, hip axis length ( HAL ) showed a consistent association with hip fracture risk when adjusted for age [hazard ratio ( HR ) 1.30 per SD increase, 95% confidence interval ( CI ) 1.22–1.38], and this was unaffected by further adjustment for BMD or FRAX score. Adjusted for FRAX score with BMD, there was a significant effect of increasing HAL quintile on hip fracture risk ( linear trend P < .001 ); relative to quintile 1 ( referent ), the HR increased from 1.43 ( 95% CI 1.12–1.82 ) for quintile 2, 1.61 ( 95% CI 1.27–2.04 ) for quintile 3, 1.85 ( 95% CI 1.47–2.32 ) for quintile 4, and 2.45 ( 95% CI 1.96–3.05 ) for quintile 5. There was a modest but significant improvement in net reclassification improvement ( 1.5% ) and integrated discrimination improvement ( 0.7% ) indices. The effect of HAL was particularly strong among younger, nonosteoporotic women ( FRAX adjusted HR 1.70 per SD increase, 95% CI 1.48–1.94 ). Conclusions: DXA-derived hip geometry measurements are associated with incident hip fracture risk, but many do not confer significant independent predictive information. HAL was found to predict hip fractures when adjusted for BMD or FRAX score and may be of clinical value in refining hip fracture risk.