Ex Vivo Mechanical Testing and FEA Modelling of Bovine Trabecular Bone

Osteoporosis, a disease characterized by bone tissue deterioration, primarily occurs in trabecular bone regions. Trabecular bone remodels in response to mechanical loads and is of interest to researchers for developments in osteoporosis treatments and improved patient-specific implant designs. Established methods for investigating trabecular bone remodelling include ex vivo experimentation and in silico finite element analysis (FEA). The first objective of this research was to measure change in apparent elastic modulus (Eapp) in response to mechanical compressive strain during a long-term ex vivo bone organ culture. The second objective was to test if modelled local mechanical stimuli were correlated with the measured change in Eapp. To achieve these objectives, trabecular bone cores (n=23) were excised from a viable bovine sternum and placed in bioreactors to maintain bone core viability for 21 days. Half the bone cores were cyclically loaded to stimulate bone growth. The “load” group increased Eapp by 53.4% over 21 days compared to 20.9% for the no-load control group (p=0.059), demonstrating the ability of ex vivo bone organ culture for characterizing the remodelling response. Following ex vivo testing, the remaining surviving bone cores (n=17) were micro computed tomography (µCT) scanned and their morphology was characterized. µCT-base FEA of the loaded bone cores (n=9) was performed to calculate the distributions of strain energy density (SED), von Mises stress (σVM), and maximum compressive principal strain (εmax). The εmax best predicted the 21-day change in Eapp (R2=0.72, p=0.005), showing that local mechanical stimuli can predict remodelling. Limitations of the experimental procedure included a small number of samples relative to the inherent variance in bone structure and cell viability, a single mechanical test at each time point, and not adjusting for end-artifacts. Other limitations included assumed homogeneous mechanical properties in FEA and the smoothed bone core surfaces to help with computational efficiency. Future work should modify the bioreactor design to reduce leakage, increase the number of samples, and perform dynamic histological analyses to compare regions of high growth rate to local stress-strain FEA results. Models should include heterogeneous mechanical properties, reduced element size, higher order shape functions, and investigate other remodelling criteria.