Purpose: Most series on revision total knee arthroplasty (TKA) have cited femorotibial instability as a frequent cause of failure, after loosening and patellar complications. The purpose of this study was to analyse TKA failure due to femorotibial instability and to search whether an initial defect in technique or indication was the cause of stability and thus to draw therapeutic conclusions for revision surgery.

Material and methods: Between 1989 and 2000, 43 aseptic TKA required revision with implant replacement (tibial, femoral or both). During the same period 1013 first-intention TKA were implanted. Among the failures, implant loosening (femoral, tibial or both) was noted as the cause in 22 cases, isolated femorotibial instability in 15. Among the 22 loosenings, there were seven cases of femorotibial instability not related to implant migration or wear. We retained the 22 cases of femorol instability related to ligaments (15 cases of isolated instability and 7 cases associated with loosening) for study. Clinical data recorded were: initial diagnosis, patient age and sex, manifestations of instability, time to revision after first intervention. Radiological data recorded were: type of prosthesis implanted, implant position (alpha and beta angles), pre- and postoperative mechanical femorotibial alignment, tibial slope, tibial and femoral mechanical angles (searching for extra-articular deformation).

Results: The 22 revisions conserned 17 women and five men. Signs were pain and sensation of instability. Mean time to revision was two years eight months for isolated instability and six and one half years for instability associated with loosening. Prostheses were implanted in different units and thus varied: all were semi-constrained implants. Among the 22 instabilities leading to revision, we found 13 frontal instabilities, three sagittal instabilities, and six global instabilities. Analysis of the patient files demonstrated that failure could be explained in 19 cases by several defects, sometimes associated: insufficient release during initial intervention (medial or lateral release), excessive release (n=1), varus or valgus frontal or tibial cut, excessive tibial slope, internal rotation of the femoral or tibial implant, extra-articular deformation corrected intra-articularly (n=4), insufficient medial collateral ligament with major genu valgum (n=3). Certain failures were particular for certain implants, posterior laxity after implant with preservation and insufficiency of the posterior cruciate ligament, dislocaton of a posterior stabilised implant (n=1).

Discussion: This analysis of factors contributing to failure by femorotibial instability demonstrated that the majority of the cases have a technical explanation: 1) defective cuts and ligament imbalance are frequent; the cut or ligament release should be revised when changing the prosthesis. 2) Ligament insufficiency generally involving the medial collateral ligament in knees with major genu valgum; a more constrained prosthesis should be used or, as advocated by some, ligamentoplasty. 3) Extra-articular deformations are generally observed in knees with major genu varum; osteotomy may be needed if the extra-articular deformation exceeds 10°.

Conclusion: Femorotibial instability is a frequent cause of early failure of TKA. Greater precision in prosthesis implantation and correct ligament balance as well as proper choice of the degree of constraint should allow reduction of this frequency.