Fall-related hip fractures are a major concern facing the older adult population, especially as rates of these injuries increasing worldwide. Based on the strong positive relationship between femoral areal bone mineral density (BMD) and femoral bone strength, various methods and tools have been developed to predict hip fracture risk and ultimately prevent these injuries. However, these BMD-based methods are currently limited in their sensitivity, with commonly used methods like the T-Score failing to predict approximately 70% of hip fracture case. This begs the question: What is currently limiting our ability to accurately predict hip fracture risk? One potential explanation is that currently established methods to predict bone strength may be limited as they are based on non-physiological, low loading rate experiments measuring bone strength. Additionally, current predictive methods are solely focused on the inorganic phase of bone which may limit their accuracy. Despite understanding the role of collagen in bone structure and on the mechanical properties of bone at the microscale, such as its influence on fracture toughness, the organic phase of bone has not been directly considered in the context of femoral bone strength and fall-related hip fractures.
Therefore, the overarching purpose of this thesis was to investigate the contributions of loading rate, fracture toughness, and bone composition (including both the inorganic and organic phases of bone) on the strength of the proximal femur in the context of fall-related hip fractures. The secondary purpose of this thesis was to investigate the effect of bone-affecting inflammatory disease states on these factors. These purposes were achieved through four studies which aimed to answer the following questions: 1) Does a biofidelic experimental paradigm, a vertical drop tower hip impact simulator (HIS), produce different bone strength measures than a traditional low displacement rate material testing system (MTS) approach? 2) Does a predictive bone strength model developed using a biofidelic test paradigm result in a model that is more accurate than previously developed models? 3) Does collagen network integrity and connectivity affect the fracture toughness of inferior femoral neck cortical bone under impact-like loading? 4) Are metrics of collagen quality significant predictors of proximal femur bone strength?
Matched pairs of fresh frozen cadaveric femurs were used to compare measures of femoral bone strength extracted from both inertially driven HIS experiments and constant d isplacement vrate MTS experiments, with the femurs of each pair being split between experiments (Study 1). Using a separate sample of matched pairs of femurs, measures of proximal femur bone strength (Study 2) were linked to measures of fracture toughness and collagen connectivity of cortical bone samples from the inferior femoral neck of the paired contralateral femur (Study 3). This allowed for direct consideration of site-specific measures of toughness and collagen in predictions of femoral bone strength (Study 4).
Although loading rate was significantly higher for specimens tested using the HIS in Study 1, there was no significant difference in the femoral bone strength measured between experimental paradigms. Within each experimental paradigm, there was a significant positive relationship between loading rate and bone strength which was likely mediated by individual specimen stiffness. In Study 2, the combination of femoral neck areal BMD, sex, and their interaction was identified as the strongest overall model for predicting femoral bone strength (adj. R² = 0.688, p <0.001). Model predictions of bone strength from four published materials testing system-based models revealed that each model produced significantly lower predictions of bone strength compared to the model developed in this study. When split by sex, each of the published models predicted significantly lower bone strength for males when compared to measured values. Study 3 revealed a significant relationship between collagen network connectivity (Max Slope) and both elastic and elastic-plastic fracture toughness (KMax and JMax ), of inferior femoral neck cortical bone samples when evaluated at, or up to, the point of peak load (adj. R² = 0.229, p = 0.002, and adj. R² = 0.163, p = 0.021, respectively). Collagen network thermal stability (Td ), however, was not associated with fracture toughness. Towards addressing the secondary purpose of this thesis, presence of bone affecting disease states was considered as a categorical variable, where it was found to be significantly associated to fracture toughness metrics alongside Max Slope and led to a marked improvement of model strength for JMax (adj. R² = 0.324, p = 0.003). After aggregating results for matched pairs of femurs that were split between Studies 2 and 3, regression analyses in Study 4 revealed that while fracture toughness was not associated with bone strength, Td was significantly associated with bone strength (adjusted R² = 0.395, p = 0.017). Td explained an additional 3.2% of the variance in the prediction of femoral bone strength when included alongside BMD and sex. The combination of these three variables resulted in the strongest overall model predicting bone strength (adj. R² = 0.942, p < 0.001). Considering that the Td is known to be associated with the connectivity of the collagen network, crosslinking content, and organization of vithe collagen network, this finding suggests that these molecular level characteristics of bone collagen are important contributors to femoral bone strength.
Through multiscale investigations of whole bone strength at the macroscale and characterization of aspects of the collagen network at the micro-scale, this thesis represents the first investigations to directly measure the influence of collagen on whole bone strength, specifically in the context of fall-related hip fractures. The aspects of collagen connectivity captured by Td were found to significantly contribute to femoral bone strength in simulated lateral hip impacts. However, the inclusion of Td as a predictor alongside BMD and sex only resulted in an additional 3.2% of the variance being explained, which brings into question the clinical significance of these findings. The difference between the bone strength regression model generated in Study 2 compared to other published models based on materials testing systems experiments suggests that it may be important to use more biofidelic test paradigms for the development of models to predict bone strength. These findings may lead to improvements in our ability to accurately predict femoral bone strength and consequently result in better estimations of injury risk. By improving the accuracy of our estimates of injury risk, we can ultimately aid in the prevention of fall-related hip fractures. Further research into the molecular-level aspects of collagen that relate to Td is needed to properly understand the mechanism through which collagen contributes to bone strength. This would then allow for the identification of potential biomarkers that could be used clinically to estimate bone strength.