| The ability to non-invasively predict vertebral fracture risk is important in the treatment of osteoporosis. At the macroscopic level, vertebral fracture is fundamentally a structural phenomenon that occurs when the loading conditions exceed the capacity of the bone. However, unlike a tensile test on a steel rod, vertebral fracture in vivo is not well defined due to a high degree of variability in bone geometry, material properties, and loading conditions. Because is unclear whether the factors complicating structural assessment of vertebral strength limit its applicability in osteoporotic fracture risk prediction, the goal of this thesis is to assess the need to develop more structurally sophisticated measures of vertebral body strength.; The results of two biomechanical experiments on isolated vertebral bodies demonstrated that sophisticated structural models performed no better than simple metrics in predicting ex vivo strength. Finite element (FE) and mechanics of solids (MOS) axial rigidity measures derived from quantitative computed tomography (QCT) scans were strongly correlated with vertebral strength in axial compression (R2 = 0.80 for FE, R2 = 0.81 for MOS). In anterior bending FE-derived metrics were better correlated with ex vivo strength in anterior bending. However, neither FE nor MOS metrics were good strength predictors (R2 = 0.22 and 0.43, respectively, and their performances were statistically equivalent (p > 0.05 for R2 comparisons). Incorporation of specimen geometry and material property distribution information did yield some improvements in basic structural metrics. For example, integral bone mineral density (iBMD) was better correlated with strength in compression than trabecular bone mineral density (tBMD) (adj. R2 = 0.62 vs. 0.16); and MOS axial rigidity was a better predictor of compressive strength than iBMD (adj. R2 = 0.81 vs. 0.62).; This thesis also presents evidence that vertebral strength is affected by several factors unrelated to radiographically assessable features of the vertebral body. Destructive biomechanical tests of donor-matched vertebrae showed that vertebral strength in compression and anterior bending were only modestly correlated (adj. R2 = 0.61, p < 0.001). Adjacent level QCT-based metrics displayed approximately 50% of the capacity of same-level metrics in predicting ex vivo compressive strength. Lastly, preliminary data from in situ biomechanical tests in neutral and forward flexed spinal positions indicated that: (1) load transfer through the facet joints may be significant in the elderly spine, and (2) adjacent disc degeneration increases vertebral strength in compression, a conclusion that was also reached using a theoretical, finite element-based analysis. This latter finding suggests that clinical fracture assessment protocols should include patient-specific information regarding disc health.; The results of this thesis suggest that future research efforts in developing more accurate non-invasive methods of vertebral fracture risk-prediction should be shifted from creating more structurally sophisticated modeling strategies, e.g., the finite element technique, to integrating non-structural patient-specific and epidemiological information with the existing clinical technique. For instance, improvements may be made to the current fracture risk assessment techniques by incorporating the results presented in this thesis concerning the site-specificity of radiographic measurements in the spine and the effects of intervertebral disc health on vertebral strength. |
