Effect of variability in anatomical landmark location on knee kinematic description

Small variability associated with identifying and locating anatomical landmarks on the knee has the potential to affect the joint coordinate systems and reported kinematic descriptions. The objectives of this study were to develop an approach to quantify the effect of landmark location variability on both tibiofemoral and patellofemoral kinematics and to identify the critical landmarks and associated degrees of freedom that most affected the kinematic measures. The commonly used three‐cylindric open‐chain kinematic description utilized measured rigid body kinematics from a cadaveric specimen during simulated gait. A probabilistic analysis was performed with 11 anatomical landmarks to predict the variability in each kinematic. The model predicted the absolute kinematic bounds and offset kinematic bounds, emphasizing profile shape, for each kinematic over the gait cycle, as well as the range of motion. Standard deviations of up to 2 mm were assumed for the anatomical landmark locations and resulted in significant variability in clinically relevant absolute kinematic parameters of up to 6.5° and 4.4 mm for tibiofemoral and 7.6° and 6.5 mm for patellofemoral kinematics. The location of the femoral epicondylar prominences had the greatest effect on both the tibiofemoral and patellofemoral kinematic descriptions. A quantitative understanding of the potential changes in kinematic description caused by anatomical landmark variability is important not only to the accuracy of kinematic gait studies and the evaluation of total knee arthroplasty implant performance, but also may impact component placement decision‐making in computer‐assisted surgery. © 2007 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 25:1221–1230, 2007