Quantification of human bone microarchitecture damage in press-fit femoral knee implantation using HR-pQCT and digital volume correlation

Primary press-fit fixation of femoral knee prostheses is obtained thanks to the inside dimensions of the implant being undersized with respect to the bone cuts created intra-operatively, dictated by a press-fit specified by the implant design. However, during prostheses press-fit implantation, high compressive and shear stresses at the implant-bone interface are generated, which causes permanent bone damage. The extent of this damage is unknown, but it may influence the implant stability and be a contributing factor to aseptic loosening, a main cause of revisions for knee arthroplasty. The aim of this ex-vivo study was to quantify, using high-resolution peripheral quantitative computed tomography (HR-pQCT) imaging and Digital Volume Correlation (DVC), permanent bone deformation due to press-fit femoral knee implantation of a commonly used implant. Six human cadaveric distal femora were resected and imaged with HR-pQCT (60.7 mu m/voxel, isotropic). Femurs were fitted with cementless femoral knee implants (Sigma PFC) and rescanned after implant removal. For each femur, permanent deformation was examined in the anterior, posterior-medial and posterior-lateral condyles for volumes of interest (VOIs) of 10 mm depth. The bone volume fraction (BV/TV) for the VOIs in pre- and post-implantation images was calculated, at increasing depth from the bone surface. DVC was applied on the VOIs pre- and post-implantation, to assess trabecular bone displacements and plastically accumulated strains. The BV/TVpost/BV/TVpre ratio vs. depth showed, consistently among the six femurs, three consecutive points of interest at increasing bone depth, indicating: bone removal (ratio < 100%), compaction (ratio > 100%) and no changes (ratio = 100%). Accordingly, the trabecular bone displacement computed by DVC suggested bone compaction up to 2.6 +/- 0.8 mm in depth, with peak third principal strains of -162,100 +/- 55,000 mu epsilon (mean absolute error: 1,000-2,000 mu epsilon, SD: 200-500 mu epsilon), well above the yield strain of bone (7,000-10,000 mu epsilon). Combining 3D-imaging, at spatial resolutions obtainable with clinical HR-pQCT, and DVC, determines the extent of plastic deformation and accumulated compressive strains occurring within the bone due to femoral press-fit implantation. The methods and data presented can be used to compare different implants, implant surface coatings and press-fit values. These can also be used to advance and validate computational models by providing information about the bone-implant interface obtained experimentally. Future studies using these methods can assist in determining the influence of bone damage on implant stability and the subsequent osseointegration.