Hip fracture rates and bone mineral density (BMD) values differ significantly between the sexes and among ethnic groups. The value of BMD is due to its strong statistical links with fracture incidence but the measurement lacks theoretical underpinnings explaining the observed associations. The interpretation of sex and ethnic differences is confounded by the fact that exactly what a BMD measurement indicates about bone fragility is poorly understood. It should be possible to approach this problem by the application of engineering principles to suitable data. A number of large population studies have acquired dual energy x-ray absorptiometric (DEXA) scans of the hip to study rates of bone mass loss with age as well as sex and ethnic differences in those rates. It is possible to derive limited structural information from such data sets for the application of engineering models. The process involves the use of specialized interactive computer programs which yield the (cortical equivalent) cross-sectional areas, cross-sectional moments of inertia and bone widths of the entire proximal femur as a continuum including the proximal shaft, intertrochanteric region and femoral neck. Dimensions such as the femoral neck length, hip axis length and neck-shaft angle are also computed as well as bending moments for specific loading conditions. The method has been used to extract structural information from approximately 15,000 hip scans acquired in the third National Health and Nutrition Examination Survey (NHANES III) as well as a subsample of the Study of Osteoporotic Fracture (SOF). Application of the method to the entire SOF data set as well as data from the European Prospective Osteoporosis Study (EPOS) are planned. To supplement the two dimensional analysis a pseudo-3D method has been developed which is also derived from DEXA data but can employ sophisticated finite element analysis (FEA) methods for structural analysis. The pseudo-3D FEA method will be validated by comparison with true 3D FEA data derived by computed tomography on cadaver femora. If the pseudo-3D method proves feasible, it should provide a rapid method for employing sophisticated FEA methods to the study of osteoporosis. These 2D and 3D analyses should permit improved understanding of specific patterns of structural change associated with aging as well as the differences in those patterns in ethnic groups with disparate hip fracture rates. It should also provide information permitting a biomechanical explanation of why certain dimensions such as hip axis length, are associated with increased hip fracture risk.
