RSA-tested TKA Implants on Average Have Lower Mean 10-year Revision Rates Than Non-RSA-tested Designs

The number of revisions after TKA is expected to rise because of aging populations in many countries and because patients are undergoing TKA at younger ages. Aseptic loosening is a major reason for late revision, which can be predicted by radiostereometric analysis (RSA) of small groups of patients at 2 years of follow-up. RSA is therefore an ideal tool to assess new TKA designs before they are introduced to the market, although not every TKA design has been studied with RSA. If RSA-tested TKA designs have lower 10-year revision rates in national registries than non-RSA-tested TKA designs, RSA testing of all new designs could be advocated.

In this study, we asked: Is there a difference in the all-cause revision rate between non-RSA-tested and RSA-tested TKA designs registered in national knee arthroplasty registries at 5 and 10 years of follow-up?

Knee arthroplasty registries were identified through the European Federation of National Associations of Orthopaedics and Traumatology webpage and through a manual internet search. Inclusion criteria were a minimum follow-up duration of 10 years and available revision or survival data per TKA design. Twenty-six registries were identified; seven were included comprising 339 TKA designs, of which 236 designs were classified as RSA-tested and 103 as non-RSA-tested. Six registries were excluded because no report was published. One registry was excluded because no fixation method was mentioned (79 TKA designs). Another registry was excluded because there was no 10-year data available (22 non-RSA-tested designs; 10 RSA-tested designs). Eleven registries were excluded because they did not provide revision rates per design and had not reached 10 years follow-up. The revision rates with their standard errors were extracted per design. We used the data from a recent meta-analysis to identify whether a TKA design was previously tested with RSA. This meta-analysis found 53 RSA studies comprising 70 different TKA designs. The prosthesis model, fixation method and insert type were extracted from these RSA-studies. The design characteristics of the TKA reported in the knee arthroplasty registries were also extracted, and if possible, matched to the TKA designs reported in the RSA-studies. At 5 years of follow-up, 191 TKA designs were identified as non-RSA-tested and 92 were identified as RSA-tested. At 10 years of follow-up, 154 TKA designs and 74 TKA designs were classified as non-RSA-tested and RSA-tested, respectively. A random-effects model using the Metafor Package in R statistics was used to estimate the pooled revision rate at 5 and 10 years of follow-up for both groups. The difference in revision rates between groups at 5 and 10 years of follow-up was estimated by including RSA as a factor in the random-effects model.

Mean all-cause revision rates at 5 years for non-RSA-tested and RSA-tested implants were 3.6% (95% CI 3.4 to 3.8) and 2.9% (95% CI 2.7 to 3.0), with a mean difference of 0.6% favoring RSA-tested implants (95% CI 0.4 to 0.8; p < 0.001). Mean all-cause revision rates at 10 years for non-RSA-tested and RSA-tested implants were 5.5% (95% CI 5.2 to 5.9) and 4.4% (95% CI 4.1 to 4.7), with a mean difference of 0.9% favoring RSA-tested implants (95% CI 0.4 to 1.3; p < 0.001).

Although there are exceptions, across registries, TKA designs that have been tested in an RSA setting have a slightly lower (about 1%) mean all-cause revision rate at 5-year and 10-year follow-up than those tested in a non-RSA setting. Acknowledging the inherent limitations of this observational study, a risk difference of 1% could potentially translate into an approximate 20% decrease in revision burden up to 10 years, which may have a profound impact on patient morbidity and health-related costs.

Level III, therapeutic study.