Bone microarchitecture differences at the tibia and radius, measurable by high-resolution peripheral quantitative computed tomography (HR-pQCT) exist between individuals with a prevalent vertebral fracture and healthy controls. As fracture is a result of mechanical failure, FE models of these sites in physiologically relevant loading conditions (compression, torsion, bending) may reveal insight into fracture pathogenesis uncaptured by microarchitecture. In this study, HR-pQCT scans at the tibia and radius were collected from post-menopausal women with prevalent vertebral fracture and healthy controls, from which microarchitecture and finite element outcomes (Pistoia failure load, apparent modulus) were calculated. At the tibia, a one standard deviation increase in compressive failure load was most strongly associated with fracture risk reduction (OR= 0.369, p=0.002), followed by compressive apparent modulus (OR= 0.473, p=0.010), torsion apparent modulus (OR= 0.493, p=0.023), and bending apparent modulus (OR=0.550, p=0.010). At the radius, failure load was the only FE outcome associated with fracture risk reduction (OR=0.534, p=0.023). None of the FE outcomes improved fracture discrimination compared to measures of bone microarchitecture. This study demonstrates the applicability of using FE under multiple loading conditions to evaluate bone structure changes related to osteoporotic fracture.