Why knees move like they do: A simulation of the effect of varying distal femoral and proximal tibial anatomy on coronal kinematic curve morphology

OBJECTIVE: Computer assisted total knee arthroplasty (CA TKA) platforms can provide detailed kinematic data that is presented in various forms including a coronal plane graphic that maps the flexion arc from full extension to deep flexion. Graphics obtained from normal tibiofemoral articulations reveal varied and complex kinematic patterns that have yet to be explained. An understanding of what drives curve variation would allow prediction of how a preoperative curve would be altered by total knee arthroplasty. Implant position could then be tailored to maintain a desirable curve or avoid an undesirable one. METHODS: An articulated lower limb saw bone with a stable hip pivot was obtained. Adjustable osteotomies were created so that femoral torsion, femoral varus-valgus and tibial varus-valgus could be altered independently. The saw bone limb was registered with a CA TKA navigation system using the posterior condyles as a rotational axis. Axial and coronal plane morphology of the distal femur and coronal plane morphology of the proximal tibia were systematically altered and a kinematic curve obtained for each morphologic combination. Femoral rotational position was varied from 10(0) of internal torsion to 10(0) of external torsion in 2(0) increments. Similarly, femoral coronal position was varied from 2(0) of varus to 6(0) of valgus and tibial coronal position was varied from 5.5(0 of) varus to 1(0) of valgus. Curves were obtained by manually flexing the joint through a full range of motion with the femoral condyles in contact with proximal tibia at all times. RESULTS: Varying femoral rotation has no effect in full extension but drives the curve away from neutral as the knee flexes. Maximal deviation is seen at around 90(0) of flexion. Internal torsion drives the curve into valgus as the knee flexes and external torsion has a reciprocal effect. Varying femoral varus-valgus causes maximal deviation from neutral in full extension. Femoral varus drives the curve from varus in extension towards valgus as the knee flexes with the effect peaking in maximal flexion. Femoral valgus has a reciprocal effect. Varying tibial varus-valgus has no effect on curve shape but does move the curve either side of neutral. Complex (parabolic) curves are caused by large rotations or the opposing effects of femoral varus-valgus and femoral rotation. The modal human anatomy of slight femoral internal rotation, slight femoral valgus and slight tibial varus produces a straight neutral curve. CONCLUSION: Kinematic curve shape is driven by distal femoral anatomy. The typical changes made to distal femoral articular anatomy in TKA by externally rotating a neutrally orientated femoral component will bring many native curves towards neutral. Externally rotating when the preoperative curve begins neutral and drives into varus as the knee flexes will drive the curve harder into varus. Conversely, kinematic femoral placement will reconstitute the premorbid curve morphology. Which outcome is preferable has yet to be determined.