Résumé : Total Knee Arthroplasty (TKA) is a very successful surgical procedure but clinical outcomes were reported to be affected by implant design, ligament balancing, alignment or patient-related anatomical factors. It was recently demonstrated that malpositioning of the TKA components and patient related anatomical factors can considerably alter tibio-femoral (TF) and patellofemoral maximum contact forces. However, up to now, how a component malpositioning and different soft-tissue anatomy changes TF knee kinematics was not yet fully investigated. The goal of this study was to evaluate how sensitive TF kinematics are to these factors during a simulated loaded squat for different TKA designs. Four TKA types (a fixed bearing, posterior stabilized prosthesis; a high flexion fixed bearing guided motion prosthesis; a mobile bearing prosthesis and a hinge prosthesis) were virtually implanted on the same virtual cadaver leg model which underwent a loaded squat between 0° and 120°. The reference models were then modified to simulate either component malpositioning in several directions or changes in ligaments geometry by change in the collateral ligament insertions. The results showed that, for all implant designs, TF kinematics were affected by changes in implant positioning and anatomical factors. While the ranges of motion predicted for all tested configurations were generally similar to the reference configuration for each type of TKA, the modifications resulted in shifts in the maximum and minimum values for the TF rotations and translations.