Comparative analysis of micro- and homogenized finite element simulations to predict the load-bearing capacity of trabecular bone screws.

The main function of fracture fixation implants is to ensure sufficient fracture stability that is a prerequisite for successful osteosynthesis. The failure of bone-screw constructs can occur when the biomechanical stability of their interface is insufficient. In complement to in vitro mechanical testing, these failure mechanisms can be investigated with finite element (FE) simulations. Two modeling approaches have been developed: micro-FE (μFE) discretizes trabecular bone at the microstructural scale while homogenized FE (hFE) assigns bone properties based on the local average bone density. μFE is presumably more accurate while hFE has lower computational costs and thus a higher potential for clinical translation. The elastic response of these approaches has previously been investigated and compared, but not their post-yield behavior. The present study aimed to assess and compare the ability of μFE and hFE to replicate the load-bearing capacity of bone-screw constructs determined experimentally. Twelve stainless steel screws were inserted in trabecular bone biopsies extracted from bovine proximal tibiae and scanned with μCT after pilot hole drilling. The samples were subsequently loaded monotonically at an inclination of 30° until failure. Previously developed methodologies were followed to build sample-specific μFE and hFE models based on μCT data. The elasto-plastic behavior of the bone and all features of the FE models, with the exception of meshing and material property assignment, were identical and aimed to replicate the experimental conditions. The experimental ultimate load correlated well with the ultimate load predicted by μFE (R2 = 0.89) and hFE (R2 = 0.95), although both simulations overestimated it systematically. The hFE ultimate load correlated well with the one of μFE (R2 = 0.84), while being closely 75 times faster to solve. These findings indicate that hFE is a suitable modeling approach for predicting the primary stability of bone screws in trabecular bone. In conjunction with its computational efficiency and its ability to use lower-resolution CT images, these results support the potential of hFE to be translated towards clinical applications.