Background. The aim of this study was to evaluate causes and results of revision surgery in unstable total knee arthroplasties. Methods. We retrospectively reviewed 24 knees that underwent a revision arthroplasty for unstable total knee arthroplasty. The average follow-up period was 33.8 months. We classified the instability and analyzed the treatment according to its cases. Stress radiographs, postoperative component position and joint level were measured. Clinical outcomes were assessed using the score and range of motion of the Hospital for Special Surgery (HSS). Results. Causes of instability included coronal instability with posteromedial polyethylene wear and lateral laxity in 13 knees(Fig. 1) and coronal instability with posteromedial polyethylene wear in 6 knees(Fig. 2), coronal and sagittal instability in 3 knees including post breakage in 1 knee(Fig. 3), global instability in 1 knee(Fig. 4) and flexion instability in 1 knee. Mean preoperative/postoperative varus and valgus angles were 5.8°/3.2° (p = 0.713) and 22.5°/5.6° (p = 0.032). Mean postoperative α, β, γ, δ angle were 5.34°, 89.65°, 2.74°, 6.77°. Mean changes of joint levels were from 14.1 mm to 13.6 mm from fibular head (p = 0.82). The mean HSS score improved from 53.4 to 89.2 (p = 0.04). The average range of motion was changed from 123° to 122° (p = 0.82). Conclusions. Revision total knee arthroplasty with or without a more constrained prosthesis will be a definite solution for an unstable total knee arthroplasty. The solution according to the causes is very important and seems to have a chance of avoidance of unnecessary over-constrained implant selection in revision surgery for
