Aim: To report our experience of revision knee arthroplasty with respect to surgical technique, joint line restoration and clinical outcome.

Methods: A clinical and radiological review was made of 45 knee revisions performed between 1996 to1998 using the Kinemax system. The reasons for revisions were infection (19), wear (11), loosening (seven), base-plate fracture (four) and instability (five). A primary prosthesis was used in 10 (22%), a posterior stabiliser in 24 (53%) and a superstabiliser in 11 (24%). WOMAC pain and function scores, arc of motion and measurements of joint line and patellar height were made.

Results: Three patients had died. Three knees had been revised a second time for instability and one had been arthrodesed for infection. The mean joint line elevation was 1.3mm compared with the primary and depressed 1.6mm when compared with the natural knee. Twenty-seven percent had patella infera and 13% had patella alta. A significant correlation was found between the change in patellar tendon length and the change in the joint line. The flexion arc was significantly improved from 83 degrees to 95 degrees. No relationship was found between the clinical outcomes and the changes in the joint-line. There were two deep infections (4.4%) and five knees were unstable. The three cases of flexion instability were due to failure to reconstruct adequately the antero-posterior (AP) diameter of the femur.

Conclusions: By restoring the level of the joint-line, a less constrained prostheses can be used in revision surgery with good functional results. The AP diameter of the femur must also be reconstructed to avoid flexion instability.