Full-field in vitro measurements and in silico predictions of strain shielding in the implanted femur after total hip arthroplasty.

Abstract

Alterations in bone strain as a result of implantation may contribute towards periprosthetic bone density changes after total hip arthroplasty. Computational models provide full-field strain predictions in implant-bone constructs; however, these predictions should be verified using experimental models wherever it is possible. In this work, finite element predictions of surface strains in intact and implanted composite femurs were verified using digital image correlation. Relationships were sought between post-implantation strain states across seven defined Gruen zones and clinically observed longer-term bone density changes. Computational predictions of strain distributions in intact and implanted femurs were compared to digital image correlation measurements in two regions of interest. Regression analyses indicated a strong linear correlation between measurements and predictions (R = 0.927 intact, 0.926 implanted) with low standard error (standard error = 38 µε intact, 26 µε implanted). Pre- to post-operative changes in measured and predicted surface strains were found to relate qualitatively to clinically observed volumetric bone density changes across seven Gruen zones: marked proximal bone density loss corresponded with a 50%-64% drop in surface strain, and slight distal density changes corresponded with 4%-14% strain increase. These results support the use of digital image correlation as a pre-clinical tool for predicting post-implantation strain shielding, indicative of long-term bone adaptations.