Critical velocity metric derived from dynamic finite element analysis classifies hip fracture risk in a clinical cohort.

Abstract

Hip fractures are debilitating and life altering events that substantially reduce quality of life. Areal bone mineral density (aBMD) is currently used to quantify hip fracture risk, along with finite element (FE) based measures such as virtual femur strength (S), and the Fragility Ratio (FR) - which quantifies the nonlinearity of the predicted mechanical bone response. As more sophisticated dynamic FE models become prevalent and can model multiple bones and the soft tissue, these values are either not biomechanically relevant (aBMD), not directly measurable (S), or not easily interpretable (FR). Therefore, we sought to derive a new parameter, the Critical Velocity (V_c) which is the minimum initial impact velocity of a dynamic FE model that results in a simulated fracture. Simulated fracture is determined by the length of connected failed elements - we used 10 mm (V_c10) and 20 mm (V_c20) in this study. We then compared the discriminatory ability against existing measures in an existing case-control cohort of 55 subjects consisting of 22 fracture cases and 33 non-fracture controls. We found that FR had the highest area under the Receiver Operating Characteristic, AUC (0.835, 95% CI = 0.686 - 0.947); followed by: V_c10 (0.795, 95% CI = 0.614 - 0.934), V_c20 (0.785, 95% CI = 0.603 - 0.928); S (0.779, 95% CI = 0.625 - 0.899), and aBMD (0.689, 95CI = 0.500 - 0.856). V_C10 and V_C20 had similar discriminatory performance. Although only FR was significantly more discriminatory than aBMD (p = 0.027), the Critical Velocity offers biomechanically relevant measurement to better understand hip fracture and hip fracture risk.