Abstract
Introduction
Total Knee Arthroplasty (TKA) is an established procedure for relieving patients of pain and functional degradation associated with end-stage osteoarthritis of the knee. Historically, alignment of components in TKA has focused on a ‘reconstructive’ approach neutral to the mechanical axes of the femur and tibia coupled with ligament balancing to achieve a balanced state. More recently, Howell et al. have proposed an alternate approach to TKA alignment, called kinematic alignment. (Howell, 2012) This approach seeks to position the implants to reproduce underlying, pre-disease state femoral condylar and tibial plateau morphology, and in doing is ‘restorative’ of the patients underlying knee kinematic behaviour rather than ‘reconstructive’. While some promising early clinical results have been reported at the RCT level (Dosset, 2014), in vivo comparisons of the kinematic outcome achieved at patient specific levels with the two alignment techniques remain an impossibility. The aim of this research is to develop and report preliminary findings of a means of simulating both alignment techniques on a number of patients.
Method
In 20 TKR subjects, 3D geometry of the patient was reconstructed from preoperative CT scans, which were then used to define a patient specific soft tissue attachment model. The knees were then modelled passing through a 0 to 140 degree flexion cycle post TKR under each alignment technique. A multi-radius CR knee design has been used to model the TKA under each alignment paradigm. Kinematic measurements of femoral rollback, internal to external rotation, coronal plane joint torque, patella shear force and varus-valgus angulation are reported at 5, 30, 60, 90 and 120 degrees of flexion. Student's paired 2 sample t-tests are used to determine significant differences in means of the kinematic variables.
Results
The mean femoral component alignment to the femoral mechanical axis was 3.3° ± 2.2° valgus and 2.3° ± 1.6° internal to the surgical transepicondylar axis in the kinematically aligned models. The mean tibial component alignment to the tibial mechanical axis was 3.5° ± 2.4° varus and 7.6° ± 6.5° internal to Insall's tibial anterior-posterior axis. The mechanically aligned model sims were all neutral to both axes.
As a result of the relative match in femoral valgus & tibial varus component angulation, mean long leg varus at 5° degrees through 60° is not significantly different from the mechanically aligned knees, though with much higher variance in the kinematically aligned group. Statistically significant differences were observed at 90 and 120 degrees, where the long leg angle is dictated by posterior condylar contact on the femur rather than distal. Other statistically significant differences in mean results were observed, notably for coronal plain joint torque (at 5° and 30°, mechanical alignment higher).
Discussion
Kinematic aligned TKR is conceptually a very different operation to mechanically aligned TKR, targeting different biomechanical goals. While evidence exists for improved clinical results in patients at a broad level, simulation tools at a patient specific level are a platform that, with development, could distinguish between patients benefiting most from a restorative or a reconstructive approach to their surgery.
