Abstract
The aim of total knee arthroplasty (TKA) is to align both the femoral and tibial components perpendicular to the mechanical axis of the leg. Most instrument systems cut the femur and tibia independently. Accurate alignment of the femoral component is hampered by our inability to define precisely the centre of the hip in three-dimensional space. Femoral resection is therefore based on a number of assumptions, which unfortunately do not hold true all of the time.
First, it assumes that an intramedullary rod follows a predictable path in the femur; secondly, that there is a fixed relationship between the rod and the mechanical axis of the leg, and thirdly that the shape of the distal femur is constant. Fourthly, even if the resection is correct, it assumes that the femoral component sits perfectly on cut surfaces. Further, there are inherent inaccuracies in the assessment of femoral component position, in that rotation of the limb with a 10° fixed-flexion deformity greatly affects apparent component position.
The exact entry point into the femur also influences alignment in that an intramedullary rod placed through an entry point 10 mm anterior to the intercondylar notch of the femur gives a mean valgus angle of 8°. When the tibia is cut perpendicular to its long axis in the coronal plane, assuming 3° of tibial varus, the femur needs to be cut with the corresponding degree of valgus, i.e., 5°. Even this argument is based on a small number of cadavers and does not take account of variations in the anatomy of the distal femur. In particular, a valgus bow can result in valgus malposition of the component. Extramedullary alignment carries the problem of using only a surface representation of the centre of the hip in a single plane, which becomes inaccurate as the femoral jig is rotated.
Malalignment of the tibial component increases the stress on the ultra-high molecular weight polyethylene insert, predisposing it to increased wear and subsidence. Studies comparing intramedullary and extramedullary guidance systems for cutting the proximal tibia have shown that 71% to 94% of prostheses inserted with an intramedullary guide, and 82% to 88% inserted with an extramedullary guide, are within 2° of being perpendicular to the long axis of the tibia.
To set a benchmark for comparison with computer assisted and robotic techniques currently being developed, we felt that it was important to assess the accuracy of placement of both the tibial base plate and femoral component in the coronal plane using current guidance systems.
We developed a series of radiographs allowing accurate independent assessment of femoral and tibial components. A long anteroposterior view of the distal femur with the patient prone was used to assess femoral placement. Coned views of the proximal and distal femur on the same plate were used to assess tibial placement. Correct rotational alignment of the radiograph was confirmed by the profile of the components.
Using this technique, we radiologically assessed the varus/valgus alignment of the tibial components of 350 TKAs. All the tibial components were implanted using an extramedullary guide with no posterior slope. We implanted 96.3% of components within 2° of the perpendicular to the longitudinal axis of the tibia. In order to validate our radiological assessment, a subgroup of 40 knees was re-assessed on CT scan. Analysis of this subgroup showed a close correlation between the results using the two different methods (mean difference 0.88°, SD 0.75).
We also assessed the position of the femoral component in 362 TKAs. A subgroup of 32 knees, 18 with perfect alignment and 14 with imperfect alignment, underwent CT scout scan of the femur from which the mechanical axis of the femur could be measured. Radiologically, 92% of all components were implanted within 3° of the target value and 83% were within 2° of target. There was close correlation between the CT and radiological measurements in the subgroup. Deviation from the mechanical axis was 1.16° (− 2.5° to +2°) in the perfectly aligned knees, validating both surgical technique and radiological assessment.
Although the findings for the femoral components compared favourably with other studies, there was still room for improvement. We set out to achieve this through direct measurement of the mechanical axis of the femur. In a series of 80 TKAs, patients were subjected to a preoperative CT scout scan of the femur. We took care to eliminate rotational error. The angle between the slope of the distal femur and the mechanical axis of the femur was calculated. During surgery the distal cutting block (Wright Medical Medial Pivot Arthroplasty System) was applied directly to the distal femur without use of an intramedullary alignment rod and the angle corrected so as to be perpendicular to the mechanical axis. A right-angled jig resting on the anterior femoral cortex was used to assess the flexion/extension of the cut. Patients were scanned again postoperatively.
In 76 knees (95%) the femoral component was within 2° of the mechanical axis. The remaining three were within 3°. We continue to evaluate the technique with the use of a new jig, which allows incremental 1°-correction of the distal femoral cut.
In conclusion, accurate cutting of the tibia during knee arthroplasty is possible with careful use of extra-medullary instrumentation. The use of a simple pre-operative CT scan eliminates the errors inherent in intramedullary femoral systems and takes into account the femoral anatomy of each individual patient.
Robotic-assisted surgery may offer the opportunity of accurate placement of components. It is, however, likely to be both time consuming and expensive. We should not yet abandon thoughts of improving the use of our current mechanical instruments. Robots have yet to prove their superiority.
The abstracts were prepared by Professor M.B.E. Sweet. Correspondence should be addressed to him at PO Box 47363, Parklands, Johannesburg 2121, South Africa.
