Abstract
Component position is critical to longevity of knee arthroplasties. Femoral component rotation is typically referenced from the transepicondylar axis (TEA), the anterior-posterior (AP) axis or the posterior condylar axis. Other studies have shown high variability in locating the TEA while proposing digitization of other landmarks such as the AP axis as a less-variable reference. This study uses a navigation system to compare the reproducibility of computing a kinematically-derived, navigated knee axis (NKA) to digitizing the TEA and AP axis. Twelve knees from unembalmed cadavers were tested. Four arthroplasty surgeons digitized the femoral epicondyles and the AP axis direction as well as flexed and extended the knee repeatedly to allow for NKA determination. The variance of the NKA axis determined under neutral loading conditions was smaller than the variance of the TEA axis when the kinematics were measured in the closed surgical condition (P<0.001). However, varus, valgus, and internal loading of the leg increased the variability of the NKA. Distraction of the leg during knee flexion and extension preserved the low variability of the NKA. In conclusion, a kinematically-derived NKA under neutral or distraction loading is more reproducible than the TEA and AP axis determined by digitization.