Background
Tibial bone defects in total knee arthroplasty (TKA) often are managed with various types of bone grafts, such as solid bone, morselized bone, and combined grafts [
1‐
4]. For optimal results, patients with bone grafts require long-term postoperative therapy. In particular, during weight-bearing gait training, the mechanical condition of the grafted bone and micromotion-preventing fixation of the tibial component influence the degree of micromotion [
3]. Whether the patient is able to walk safely immediately postoperatively is an important concern after cementless TKA. The finite element method is being used increasingly for biomechanical analysis of tibial component micromotion. Several studies have analyzed the relative micromotion of implanted prostheses using this method [
5‐
8]. We have previously reported that tibial bone graft combined with cortical bone can be expected to reduce micromotion and stress [
8]. However, the simulation was not enough to include the characteristics of implant and bone graft materials in the models or to examine the detailed status of contact and stress distribution while considering micromotion. To our knowledge, no analysis has been conducted considering the actual bone graft size and material properties. We used the finite element method to develop a new precise model of the proximal tibia and tibial implant and meticulously analyzed the biomechanical effects of various types of bone grafts on relative micromotion and stress distribution at the tibial tray–bone interface in the immediate postoperative period.
Discussion
Medial loading resulted in subsidence of the loaded medial tibial plate and liftoff at the periphery on the opposite, unloaded side. The tilting motion of the tibial components observed in this study implies instability of the initial fixation, which could possibly compromise bony ingrowth. The success of cementless implants depends on the ingrowth of bone into the porous coating, which is inhibited by relative micromotion at the implant–bone interface during patient activity [
19,
20]. Excess implant micromotion prevents bone formation within porous-surfaced implants. According to Jasty et al. [
21], 150 μm of motion results in unstable ingrowth of bone into porous-surfaced implants. Although the 150-μm limit often is quoted in the literature, a range of ± 50 μm was used to assess the sensitivity of the micromotion limit [
22]. Our findings showed that soft bone grafting leads to considerable implant micromotion. The relative micromotion observed in the L–S model exceeded 100 μm. This degree of micromotion, if occurring in a clinical situation, may inhibit bony ingrowth and lead to implant loosening. Conversely, use of hard bone decreases micromotion. In all hard bone graft models, the relative micromotion was less than 100 μm. These results suggested that hard bone grafting is necessary in patients requiring a large bone graft.
Loads were transferred via contact between the tibial baseplate and tibial bone. This contact resulted in an increase in the loading of the medial plateau of the tibia. Analysis of the soft bone graft models revealed considerable unloading of the bone graft area concentrated around the bone graft compared with the I model. Tibial host cancellous bone stresses were reduced by hard bone grafting. The L–S model caused high stress in host cancellous bone. High stress levels, particularly acting on weak cancellous bone, have been implicated as the predominant cause of tibial component loosening [
23,
24], and the abnormal stress pattern may cause stress shielding in long-term follow-up [
25,
26]. The stress pattern produced by load sharing between the implant and bone suggests that large soft bone grafting should be avoided in favor of grafting with hard bone to eliminate excessive stress in host cancellous bone.
In the previous report, the maximum liftoff and relative micromotion were approximately 2–10 times higher than those in this study and suggested that bone grafting causes excessive instability [
8]. However, from a modeling viewpoint, the shape of the keel of the tibial baseplate was simple, and the roughness of the keel surface to improve fixation was ignored, treating the interface between the bone and implant as a frictionless contact in the previous report. Furthermore, the graft size and material properties were impractical. However, in this study, the shape of the keel designs was precise, treating the interface between the implant and bone as a frictional contact to provide additional stability, in addition to considering realistic graft size and material properties. We believe that these distinct differences are the main reasons that affected the results.
With regard to the elastic modulus of bone graft materials, several studies have demonstrated that material stiffness and strength vary substantially according to age, diagnosis, and composition of cancellous and corticocancellous bone and that improved fixation can be achieved by impaction during bone grafting [
13,
14,
27‐
29]. In this study, hard bone models showed lower relative micromotion than soft bone grafting models, suggesting that increased stiffness of the bone graft material by impaction led to decreased micromotion and improved the initial fixation of the tibial baseplate. In clinical cases where the quality of host bone varies substantially between patients, the appropriate bone, including impacted bone, should be grafted into bone defects to achieve optimal stability and a good clinical outcome [
4,
30,
31].
With respect to eccentric loading, the finite element analysis of bone grafts undertaken in this study represented the worst-case scenario. Our results suggested that grafting defects with hard bone is more likely to yield better fixation especially for larger, posteromedial defects. During surgery, mechanically effective fixation can be achieved by impaction according to the bone graft material used and type of tibial defects. However, this study did not consider changes in mechanical conditions because of bone remodeling, ingrowth, or stress relaxation. Thus, the qualitative trends outlined here offer an approximation of what occurs during the immediate postoperative phase. This study was designed to provide a foundation for a biomechanical rationale that will support selection of treatment. Because the severity of tibial bone defects and range of bone grafts vary in the clinical setting, care should be taken when interpreting these results. Considering these limitations, combining our results with those of previous cadaveric studies may enhance the clinical use of this technique.