Abstract
Computer assisted total knee arthroplasty may have advantages over conventional surgery with respect to component positioning. Femoral component mal-rotation has been shown to be associated with poor outcomes, and may be related to posterior referencing jigs. We aimed to determine the variation between the transepicondylar axis (TEA) and posterior condylar axis (PCA) in a series of knees undergoing navigated total knee arthroplasty, and to determine the correlation between final intra-operative coronal alignment and post-operative radiographic functional alignment.
A review of 170 consecutive patients undergoing primary total knee arthroplasty between June 2007 and August 2010, using Precision navigation and Triathlon implants (Stryker). The difference between the TEA and PCA was measured as was the initial coronal alignment. Referencing of the TEA had been previously validated against computerised tomography in a previous study. During arthroplasty, neutral alignment was aimed for, and the final alignment after implant insertion was recorded. Pre- and 1 year post-operative flexion was measured. A standing four foot alignment radiograph was obtained 6 weeks after surgery to determine the weight-bearing mechanical axis.
The mean difference between the TEA and PCA was 3.94 degrees (−2.80 to 11.59) and median difference was 3.6 degrees. (A positive value implies the PCA is internally rotated with respect to the TEA). The median pre-operative flexion was 120 degrees (80–130) and the median post-operative flexion was 125 (85–145). The mean change in flexion was −2.5 degrees (−40 to 40; p=0.001). The mean intra-operative alignment was 0.75 degrees (−3 to 6, SD 1.9) and the mean radiographic alignment was 1.24 degrees (−6.5 to 6.5, SD 1.6).
Taking −3 to +3 to be neutral, the outlier rate intra-operatively was 6.5% and radiographically was 16.5%. The intra-operative and radiographic alignment showed correlation (coefficient 0.289). There was poor correlation between pre-operative deformity and degree of difference between intra-operative and radiographic alignment (coefficient −0.1).
Conclusion: There is a wide variation in the difference between the TEA and PCA, and there is not a good relationship with coronal alignment. Although most valgus knees had a bigger difference, such a difference was also seen in many varus knees. This should alert the surgeon when using posterior referencing jigs when determining the femoral component size and rotation. Although these patients achieved good post-operative flexion, this was determined by the pre-operative range. There was reasonable correlation between the final intra-operative mechanical alignment and the weight-bearing alignment as determined by a standing radiograph.
