The medial patellofemoral ligament (MPFL) has been recognised as the most important medial structure preventing lateral dislocation or subluxation of the patella (LeGrand 2007). After MPFL rupture the patella deviates from the optimal path resulting in an altered retropatellar pressure distribution. This may lead to an early degeneration with loss of function and need for endoprosthetic joint replacement. The goal of this study was to develop a dynamic knee-simulator to test the influence of ligament instabilities to patella-tracking under simulation of physiological quadriceps muscle loading. On 10 fresh-frozen cadaveric knees the quadriceps muscle was divided into five parts along their anatomic fibre orientation analogous to Farahmand 1998. The muscular loading was achieved by applying weights to each of the fife components in proportion to the cross sectional muscle area. A total of 175 N was connected to the muscles using modified industrial cable connecting systems [Lancier Calbe, Drensteinfurt/Germany]. A novel light digital patellar reference base (DRB) was developed and attached to the patella with four bone screws. On addition a femoral and tibial digital reference base was constructed and secured to these two bones. Position data of the patella, the femur and tibia was tracked by a conventional tracking system [Optotrak, NDI Europe]. The relative movement between femur and tibia (“flexion path”) and patella and femur (“patella tracking”) was recorded. For retropatellar pressure measurement a custom made sensor was introduced between the patella and femur [Pliance, Novel/Germany]. The sensor consists of 85 single pressure measuring cells. The robot-control-unit is liked to a force-torque sensor (hybrid method). The force free knee-flexion-path from 0° to 90° was calculated during three “passive path” measurements using this hybrid robotic method. The actual measurements followed with identical parameters. The 3D-patella position was recorded (six degrees of freedom) along with the corresponding retropatellar pressure distribution according to knee-flexion and medial forces (intact vs. cut MPFL). Measurements were performed for the intact knee (“native”), with muscular quadriceps loading, after opening the joint capsule and with introduced pressure sensor to differentiate each of their influences. The load free knee-flexion-path (“passive path”) could be calculated for all of the ten knees and was utilised as the basis for all dynamic measurements. There was no alteration of the “flexion-path”. Thus the measurements were only influenced by the variables “capsular joint opening,” “muscular quadriceps loading” and “MPFL-tension”. The custom made connections between the five quadriceps components and weights proved to be a secure way to prevent rupture due to the applied forces of up to 70 N during the average measuring time of 7.5 h/knee. Only on one knee the Vastus lateralis obliquus muscle ruptured proximally. All reference bases were 100% visible despite the knee flexion form 0°–90°. No loosening of the reference base screws occurred. Overall the combination of a robotic-assisted, force free dynamic knee-flexion under quadriceps simulation and 3D-patella-tracking seems to be a promising method to evaluate the biomechanical influences of ligaments on the human knee.
