The effects of body composition and body configuration on impact dynamics during lateral falls: insights from in-vivo, in-vitro, and in-silico approaches

Abstract

Fall-related injuries are the current leading accidental cause of emergency room and hospital visits, hospitalizations, and injury-related deaths in Ontario. Hip fractures in particular are associated with poor functional and survival outcomes, and high medical and rehabilitation costs. While nearly a third of older adults fall in Canada each year, only 2% of falls result in a hip fracture, and only 10-37% of falls result in any injury requiring medical attention. The Factor of Risk model (i.e. applied loads/fracture tolerance) has been proposed as a conceptual model to explain this discrepancy, however, current approaches used to screen for hip fracture risk focus primarily on bone strength (i.e. only fracture tolerance) and population-level clinical risk factors for which the mechanistic link to fracture risk is unclear. Individual faller and falling characteristics have been proposed to influence the application and distribution of loads during a fall which could add predictive value to the Factor of Risk approach. However, the magnitude and interaction of these factors has not been quantified.

Therefore, the focus of this thesis was to examine the influence and interaction of individual anthropometry and falling configuration on impact dynamics during simulated falls, and development of a computational model to predict the magnitude and distribution of loads in the pelvis during a fall. The overarching theme was supported through five studies, with objectives to i) define the relationship between elements of body size (e.g. height) and composition (e.g. percent fat mass, trochanteric soft tissue thickness) and impact dynamic outcomes (i.e. peak vertical force, pressure, contact area and deflection) during a simplified simulated fall protocol; ii) determine the relationship between individual characteristics and model parameters for one- and three-dimensional contact models with elastic and viscoelastic components; iii) assess model performance (namely, Mass Spring, Voigt, Hertz, Hunt-Crossley and Volumetric) for the prediction of applied loads during simulated falls; iv) determine how trochanteric soft tissue influences deflection of skeletal structures during a controlled impact; and iv) examine how the relationships between body size and composition are affected when more complex fall simulation protocols are implemented. Studies 1 and 5 employed in vivo fall simulation protocols, Studies 2-3 were performed in silico¬ based on parameters and outcomes drawn from in vivo fall simulations, with comparisons based on both peak and time-varying force outcomes, and in Study 4 an in vitro drop tower protocol was used to apply loads directly to the greater trochanter.

In Study 1, pelvis impact dynamics were strongly related to individual characteristics, providing support for the development of a subject-specific hip fracture model. Peak force was strongly linearly related to mass, while peak pressure, contact area, and deflection were more strongly related to the quantity of adipose tissue overlying the hip. In Study 2, elastic parameters for the Voigt and Hertz models were not linked to any individual characteristic, while the Mass-Spring, Hunt-Crossley and Volumetric elastic parameters were related to body fat, sex and trochanteric soft tissue thickness, respectively. Damping parameters for the Voigt model differed between males and females; for the Hunt-Crossley and Volumetric models varied based on pelvis width. In Study 3, model performance was strongest for the Hunt-Crossley model compared to all other models tested, and improved for three- vs one-dimensional models and models including dampers compared to elastic-only models. In Study 4, when cadaveric greater trochanters were laterally impacted using a drop tower protocol, greater and more consistent deflection was found at the anterior superior iliac spine than the greater trochanter, and low, but substantial, deflection occurred medially at the lateral apex of the pelvic ring and medial border of the ilium. Deflections distributed between structures were different during conditions where trochanteric soft tissues were present vs. conditions where soft tissues were removed. In Study 5, while impact characteristics continued to link closely with individual faller characteristics, they were more strongly linked to fall simulation method. Though vertical impact velocity was similar between protocols, shear forces and pressure were greater when participants initiated a simulated fall from a squat position compared to initiation from a kneeling position or a passive “pelvis release” fall simulation.

In sum, the results of these studies provide evidence of the importance of faller characteristics, particularly trochanteric soft tissue thickness, and falling configuration, in predicting the magnitude and distribution of loads during a fall impacting the hip. Additionally, the modeling components point towards the ease of developing and implementing an individualized and mechanistic method of predicting fracture risk in older adults. The results of these studies help to illuminate why some fallers in some configurations experience different risk of injury than would be predicted based on clinical risk factors. These results can be used to improve screening of individuals who might be at greater risk of injury due to poor absorption or distribution of energy by the trochanteric soft tissues. Further, these results may be used to identify which type of hip protector may be appropriate based on individual anthropometry. Finally, risky falling configurations have been identified and can be linked to falling patterns within epidemiological and balance literature—these can be used to develop exercise- and environment-based interventions. Future work should focus on determining how faller strategy influences falling configuration and impact dynamics. Additionally, further model expansion and validation is required to improve the external validity of the models proposed and tested within this thesis, particularly with regards to non-vertical impact dynamics and load distribution within the femur and pelvis.