CHAPTER 18 Mobile-Bearing Uni
Long-Term Outcomes
Introduction: Prosthesis Design
The implant most commonly used is the Oxford unicompartmental knee replacement.1 This design was originally used as a bicompartmental device but was first used as a unicompartmental device in 1982 (Phase 1). Since then two further design phases have been introduced, retaining the design features while improving the instrumentation and the method of implantation. In Phase 1 the femoral surface was prepared with cutting blocks to remove angular cuts of bone to fit the nonarticular surface of the femoral component. However, this did not allow accurate balance of the flexion-extension gap at implantation, and the Phase 2 prosthesis was introduced in 1987 and sought to improve this. The femoral component was changed to a spherically concave inner surface and the femoral bone surface was prepared with a mill rotating around a spigot. Incremental 1-mm milling of the distal femoral surface allows careful balancing of the knee ligaments throughout flexion (Fig. 18–1). The Phase 3 design, introduced in 1998, provided new instruments and an increased range of components to facilitate the implantation through a short incision. The aim was to gain the short-term advantages associated with this less invasive incision2 while maintaining the longer term survival, by maintaining the original design features of the fully congruent articulation (Fig. 18–2). Although the Oxford mobile-bearing unicompartmental knee was the first of its kind, there are several other mobile-bearing knees based on a similar philosophy. The AMC knee (Uniglide) utilizes a fully mobile bearing restoring the natural tension within the knee; however, there is a difference in the shape of the femoral component. The radius of the femoral component is constant up to 45° of flexion but decreases toward the posterior portion of the condyle.
We have reviewed the long-term results of the prostheses mentioned above. There is substantial literature regarding the Oxford Knee and more limited literature about the other prosthesis. Long-term outcome is an important factor since impressive early results of an implant may dramatically change with longer term follow-up. For example, in the PCA fixed-bearing unicompartmental prosthesis, 90% of patients were reported as having a satisfactory functional result at 2 years,3 yet the cumulative survival at 3.7 years was only 77%.4 Our aim is to review the long-term results, at 10 years and greater, of mobile-bearing unicompartmental replacements.
Measuring Outcome: Explanation of Study Types
Assessing Survivorship
The measurement of success in arthroplasty surgery is important to provide information not only for the surgeon selecting the type of surgery and individual prosthesis to use but also when counseling patients regarding expected outcome from an intervention. The most commonly used measure of success is the survival of the implant (time to revision), but it is also important to assess the functional outcome of the arthroplasty in surviving patients, since this provides more complete information regarding the success of the patients’ treatment than pure survival analysis.5 Modern survival analysis will define a failure end point and provide an assessment of how many patients failed and how long after the operation this failure occurred. It is therefore important to understand the exact nature of the defined end point in each study as some authors use removal of the implants as the failure end point, whereas others use further surgery of any type. The difference between the two end points may affect the reported survival. Survival analysis may be performed using the life table method or Kaplan-Meier method. Whichever is employed, there are a number of features that must be understood about the reported data. Survival figures are cumulative, which allows a prediction of the expected failure rate in the long term, reducing the need for large numbers of prostheses to have reached the long-term follow-up point. However, the number at risk at each time point must be known since, if this number reduces to less than 15, it becomes difficult to interpret the data. Loss to follow-up is also important, and studies should present worst- and best-case scenarios.
Study Types
Joint Registers
The prime function of registers is to assess the success of treatment in large, often nationwide, population-based cohorts. They provide data that represent clinical practice without the inherent bias in cohort studies and can be used to compare outcome of implant designs. For this reason, registers have always taken “revision” as the marker of failure and the “cumulative revision rate” as the comparator between implants. While revision is an event that will invariably bring a patient back into contact with a surgeon and is therefore easily measured, there are some difficulties with the interpretation of the revision rates. Unfortunately, arthroplasty registers collect few data and exert no control over patient selection, surgical expertise, and indication for revision. The term surgical routine was defined by Robertsson et al.6 in 2001 to describe these factors, and it was shown that the revision rate in centers that performed fewer than 23 unicompartmental knee arthroplasties (UKAs) annually had a risk of revision 1.63 times higher than centers with higher volumes of work. Over the last 10 years, the New Zealand Joint Registry has introduced patient-reported outcome data (PROMS) into the joint register. This allows additional information to be gained regarding true clinical outcome and patients’ functional score at the time of revision surgery, which offers an improved assessment of the true success of arthroplasty. Analysis of these data shows that, in every category of clinical outcome, a UKA is between four and six times as likely to be revised as a total knee arthroplasty (TKA). This highlights the problem of using registry data to compare between UKA and TKA, whereas a comparison between UKA designs is more relevant.
