Although total knee replacement (TKR) has good long term reliability, some patients remain unhappy; this may relate to abnormal motion causing pain or instability. This study measured the effect of TKR femoral component internal-external rotation position upon knee kinematics.

The kinematics of eight fresh-frozen cadaveric legs were measured, with a range of loading and states of preparation. The stages of preparation included intact; TKR in standard navigated position aligned to mechanical and epicondylar axes, TKR with three and six degree internal and then external rotation of femoral component. The loads applied were 70N anterior and posterior draw; Five Nm internal and external rotation; Five Nm valgus and varus. All these were applied in every state of preparation with the knee moved passively in 0–120deg flexion-extension, then repeated with the quadriceps tensed to 400N by a pneumatic cylinder and cable. The TKR used was a Stryker Scorpio posterior cruciate retaining. The implant positions and tibio-femoral kinematics were measured continuously using a modified software Stryker knee navigation system, leading to ′envelopes of laxity′ for each degree of freedom across the range of flexion-extension. In order to vary the implant rotation, the ‘standard’ TKR was removed and then remounted on an adjustable intra-medullary rod-intube mechanism that was also linked to the navigation system. Adjustments in 6 degrees of freedom allowed the datum position to be regained within 1mm and 1deg, using a custom software module and a sensor located on the implant.

Internal rotation of the femoral component caused increasing tibial valgus with knee flexion, with the increase in valgus at 90deg matching the changed rotation. Similarly, external component rotation caused matching tibial varus with knee flexion. Varus and valgus laxities were not altered significantly from those in the datum condition by femoral component internal rotation, across the whole range of flexion. However, external rotation caused increased valgus laxity in flexion. Tibial rotational effects were complex. In the extended knee, femoral component rotation caused a matching tibial rotation. Thus, an externally rotated femoral implant magnified tibial external rotation (the screw-home) with terminal knee extension. The tibial internal rotation with knee flexion was then increased above normal, so that the tibia was internally rotated at 90deg flexion. Internal rotation of the component caused increased internal rotation laxity and decreased external rotation laxity; the opposite occurred after femoral component external rotation.

Changes in femoral component position had complex effects on the movement and posture of the tibia across the range of knee flexion. Some have easily-understood consequences, such as component internal rotation caused tibial valgus in flexion, thus increasing the lateral force vector acting on the patella. The changes in rotational laxity patterns are related to the differing structures of the medial and lateral collateral ligament complexes, the lateral collateral ligament allowing greater freedom of movement in response to the altered height of the ligament attachment above the joint line at that side of the knee, whereas the medial collateral ligament maintained greater control of rotational laxity. These effects explain loss of stability in flexion and the tendency of the knee to pivot about a medial axis.

Introduction Revision total knee arthroplasty is becoming a more frequent procedure throughout Europe. Painful patello-femoral problems, patellar dislocation, impingement pain as well as aseptic loosening and gross malalignment are among many causes. We investigated the routine use of CT scans in identifying alignment causes for failure as well as in the pre operative planning of the procedure.

Method Twenty poorly functioning total knee arthroplasties were analysed using the Perth CT protocol. All patients were awaiting revision total knee arthroplasty and were scanned using a GE multislice CT scanner. The measurements were performed using standard CT software.

Results The mean coronal position of the components was 3 degrees of valgus for the femoral component and 2.5 degrees of varus for the tibial component. Fourteen knees had errors of femoral component rotation, which ranged from 1 degree of external rotation to 9 degrees of internal rotation. Nine knees had errors of tibial base-plate rotation with all being internally rotated relative to the PCL/Tibial tuberosity axis from 3 to 12 degrees.

The cumulative error of implantation ranged from 6- 24 degrees in all 7 planes.

Discussion Revision total knee arthroplasty remains a difficult procedure but is increasing in frequency. The use of a CT protocol allows all coronal, sagittal and rotational errors in previous implantation to be accurately identified prior to surgery. We believe that all knee revision operations should have a CT scan as part of the pre operative planning.