Résumé : Introduction: Unicompartmental knee arthroplasty (UKA) and total knee arthroplasty (TKA) are standard procedures for treating knee joint arthritis. Neither UKA nor TKA seems to be optimally suited for patients with bicompartmental osteoarthritis that affects only the medial and patellofemoral compartments. A bicompartmental knee arthroplasty (BKA) was designed for this patient group. This study aimed to compare the effectiveness of a BKA and TKA in restoring the kinematics of the knee joint. Materials and methods: In this in vitro study, three types of knee arthroplasties (BKA, posterior cruciate ligament-retaining, and posterior cruciate ligament-resecting TKA) were biomechanically tested in six freshly frozen human cadaveric specimens. Complete three-dimensional kinematics was analyzed for each knee arthroplasty during both passive and loaded conditions in a validated knee kinematics rig. Infrared motion capture cameras and retroreflective markers were used for recording data. Results: No significant differences could be found between the three types of arthroplasties. However, similar kinematic changes between BKA and a native knee joint were documented under passive conditions. However, in a weight-bearing mode, a significant decrease in femoral rotation during the range of motion was found in arthroplasties compared to the native knee, probably caused by contraction of the quadriceps femoris muscle, which leads to a decrease in the anterior translation of the tibia. Conclusions: Kinematics similar to that of the natural knee can be achieved by BKA under passive conditions. However, no functional advantage of BKA over TKA was detected, which suggests that natural knee kinematics cannot be fully imitated by an arthroplasty yet. Further prospective studies are required to determine the anatomic and design factors that might affect the physiologic kinematics.