In this study, Namba et al. review eight years of data from Kaiser Permanente’s arthroplasty registry and report on the relative risk of revision for fixed versus mobile-bearing implants. In brief, the study cohort included 47,339 arthroplasties, with 62% performed in women. The overwhelming majority of procedures utilized fixed-bearing systems, while just over 10% (4830) used rotating-platform mobile bearings. Of those, 3112 had a posterior-stabilized Rotating-Platform Press-Fit Condylar (RP-PFC) design, 1053 received a Low Contact Stress (LCS) implant, and 665 had a cruciate-retaining RP-PFC design. Of note, the patients receiving a mobile-bearing implant were significantly younger than those receiving a fixed-bearing implant.
Of this group, 881 knees were revised for any reason. Three hundred and sixty-six knees were revised for infection, while 515 were revised for other causes; these were fairly typical for a large cohort, including instability, pain, aseptic loosening, osteolysis, arthrofibrosis, failure of osseous ingrowth, and wear1. Of note, approximately 4% of the patients died within the study period, and 9% were lost to follow-up.
The major concern for the authors, and consequently for readers, is that the risk of revision for the LCS was two times higher than that of the fixed-bearing systems. The other mobile-bearing components did not show this, although a slight but not significant trend was present.
When I reviewed this study, two questions came to mind. First, is there something particularly odious about the LCS? Second, is there an advantage to the use of a mobile-bearing implant?
To address the first question, I looked to other registry data, since the LCS has been in clinical use for many years. In the most recent England and Wales review, the LCS appears comparable with other knee implants, with a revision rate of 1.76% (95% confidence interval [CI], 1.41% to 2.18%) at five years, with 13,270 cases in the registry2. This compares reasonably well with the five-year data for all implants (n = 373,954), which show a revision rate of 1.41% (95% CI, 1.36% to 1.47%). Similarly, the Australian registry shows acceptable but not spectacular results: it reports ten-year data on 441 revisions in 8247 primaries performed with the LCS, compared with 4585 of all 138,000 total knee arthroplasties3. The ten-year revision rate for the LCS is 6.5% (95% CI, 5.8% to 7.2%), which is good but not outstanding. My conclusion is that the LCS is a solidly performing implant choice, but it does not stand out in comparison with other systems. The current study appears to be an outlier in that regard. Several possible explanations come to mind: the LCS, as the authors note, demands a somewhat different surgical technique. If surgeons utilize the LCS on an occasional basis, their lack of routine use may result in an inferior outcome. In addition, only about 2% of the total cohort received the LCS. There may be substantial patient or surgeon variability in this relatively small group compared with the whole. For example, the patients receiving an LCS implant are younger, heavier, and more likely to have bilateral surgery. They are markedly more likely to receive an uncemented implant. Their surgeons are significantly less likely to have had fellowship training. In short, a number of confounding factors are present.
Thus, I come to the crux of this study. Is there an advantage to the use of a mobile-bearing total knee replacement? I freely confess that I routinely use a posterior-stabilized, fixed-bearing implant: I find it easier to achieve a reproducible result, I find it easier to teach to residents, and it removes my probably arbitrary judgment as a factor in deciding which patient should receive which design. However, there are some kinematic reasons to favor a mobile-bearing knee implant. By adding mobility between the tibial insert and the tray, a tibial surface that conforms to the femoral condyle can be used. Presumably, this would reduce localized contact stress in the polyethylene and thus reduce wear. Further, by allowing rotation, the design should allow the knee to align itself with flexion and extension.
That being said, there is a paucity of data to support any viewpoint. A recent Cochrane review showed no differences in knee motion, pain relief, or overall function between fixed-bearing and mobile-bearing implants4. With respect to wear, retrieval analyses demonstrate no differences5. The force across the tibiofemoral joint is similar for mobile-bearing and posterior-stabilized devices6.
The current study represents a valuable addition to a growing body of registry data, with the attendant strengths and limitations of any registry. The positive? It reflects the real world, with many surgeons performing operations on many patients. The negative? Registries can only provide a bird’s-eye view—they cannot analyze an individual patient’s needs, or a surgeon’s rationale for choosing a particular implant.