The long-term revision risk of RSA-tested and non-RSA-tested TKA implants in The Netherlands: a Dutch arthroplasty register study

Background and purpose: Radiostereometric analysis (RSA) of total knee arthroplasty (TKA) is used as an early safeguard during the phased evidence-based introduction of new implants. The goal of our study was to compare the long-term revision risk between RSA-tested implants and non-RSA-tested implants in the Netherlands using patient-level data.Methods: All primary TKAs between 2007 and 2016 from the Dutch Arthroplasty Register were included, and procedures with an RSA-tested implant were identified. Both all-cause major revision risk and revision risk because of loosening were calculated at 5 and 10 years postoperatively using Kaplan–Meier analyses. Sensitivity analyses were performed with stricter definitions of implant characteristics to classify procedures as RSA-tested, to avoid camouflage of different subdesigns within the same brand implant portfolio.Results: 83,638 RSA-tested and 104,105 non-RSA-tested TKAs were included. Cumulative all-cause major revision percentages for the RSA-tested group at 5 and 10 years were 2.2% (95% confidence interval CI 2.1–2.3) and 3.6% (CI 3.4–3.7), respectively, compared with 2.5% (CI 2.4–2.6) and 3.3% (CI 3.2–3.4) for the non-RSA-tested group. RSA-tested TKAs showed higher 10-year revision risks because of loosening than non-RSA-tested procedures (1.8%, CI 1.7–1.9 vs 1.4%, CI 1.3–1.4, respectively). Comparable results were found after stratification by various patient characteristics and with stricter classification approaches.Conclusion: Regarding all-cause revision risk, RSA-tested TKAs had a slightly lower risk at 5 years. However, at 10 years the RSA-tested TKAs had a higher all-cause revision risk and higher revision risk because of loosening compared with the non-RSA-tested TKAs.