Variation in KOOS JR improvement across total knee implant designs: a cohort study from Michigan Arthroplasty Registry Collaborative Quality Initiative

Abstract

Background and purpose: Arthroplasty registries report revision risk, but patient-reported outcomes may also measure implant performance. We aimed to evaluate (i) change in patient-reported outcome measures (PROMs) across multiple total knee arthroplasty (TKA) designs in a regional registry, (ii) the association of patellar resurfacing on the change in PROMs, and (iii) the variation in PROMs change within implants with or without patellar resurfacing.

Methods: This is a cohort of primary TKAs from Michigan Arthroplasty Registry Collaborative Quality Initiative (MARCQI) performed between January 1, 2017 and September 30, 2021. The dependent measure was change in KOOS JR. Independent variables were implant name and patellar resurfacing. Multivariate modeling adjusted for patient-level factors. A previous report suggests a change of 23 points in KOOS JR as clinically relevant in achieving acceptable pain/function levels. A clinically relevance ratio (CRR) of those achieving the threshold of 23 points to the overall group was calculated for each implant.

Results: 18 implant designs met the inclusion criteria. There were 51,606 cases with complete preoperative and postoperative KOOS JR matched pairs. There was variation in improvement from preoperative to postoperative unadjusted KOOS JR scores across implant designs (P < 0.001), ranging from 18.7 (95% confidence interval [CI] 16.8-20.6) to 27.0 (CI 24.9-29.2). Patellar resurfacing resulted in greater KOOS JR improvement 1.0 (CI 0.5-1.5, P < 0.001). Of the cases with resurfaced patellae, the CRR was 50.2% (CI 49.7-50.7). For cases without resurfaced patellae, the CRR was 47.2% (CI 45.9-48.5). The association of implant design persisted whether the patella was resurfaced or not, evident in the adjusted mean change in KOOS JR (P < 0.001), ranging from 20.1 (CI 17.6-22.6) to 25.5 (CI 24.3-26.7) for resurfaced and from 17.0 (CI 13.9-20.1) to 23.3 (CI 20.3-26.2) for not resurfaced, and the CRR difference (P < 0.001), ranging from 45.8% (CI 42.5-48.6) to 55.8% (CI 50.4-60.8) for resurfaced and from 37.9% (CI 27.4-44.7) to 51.4% (CI 43.9-56.6) for not resurfaced.

Conclusion: Implant design and patellar resurfacing both show an association with KOOS JR improvement. Variations in implant design persist whether the patella is resurfaced or not. Implant selection and patellar resurfacing may be associated with patient outcomes.

Figure 1

Flow diagram of cases illustrating how the final analytical dataset was generated from the MARCQI registry.

Figure 2

Adjusted mean change in KOOS JR. 5 implant designs had 95% confidence intervals for KOOS JR change that were below the MARCQI average of 23: Evolution MP/Evolution MP, iBalance/iBalance, Journey II/Journey, Legion/Genesis II, and Vanguard/Maxim Mono-Lock. 2 had 95% confidence intervals entirely above 23: NexGen LPS Option/NexGen TM and NK II GS/NK II. The remainder had confidence intervals that overlapped the MARCQI average.

Figure 3

Clinical relevance ratio of KOOS JR change. 3 implant designs had 95% confidence intervals entirely below the MARCQI average CRR of 50%: Evolution MP/Evolution MP, Journey II/Journey, and Legon/Genesis II. 3 had 95% confidence intervals above the average: NK II GS/NK II, Persona/Persona, and Triathlon/Triathlon TS. The remainder had confidence intervals that overlapped the MARCQI average.

Figure 4

Adjusted mean change in KOOS JR in cases with patella resurfaced in red. 2 implant designs with patella resurfaced had 95% confidence intervals entirely below 23 points: Journey II/Journey and Legion/Genesis II. 6 implant designs with patella resurfaced had 95% confidence intervals entirely above the MARCQI average of 23: Triathlon/Triathlon, Triathlon/Triathlon TS, Vanguard/Maxim, Persona/Persona, NK II GS/NK II, and NexGen LPS Option/NexGen TM. The other 9 implant designs with patella resurfaced had overlapping the MARCQI average. Adjusted mean change in KOOS JR in cases with patella not resurfaced in green. 2 implant designs with patella not resurfaced had 95% confidence intervals entirely below the MARCQI average: Evolution MP/Evolution MP and Legion/Genesis II. The remaining 5 implant designs with patella not resurfaced had 95% confidence intervals crossing the MARCQI average.

Figure 5

Clinical relevance ratio of KOOS JR change in cases with patella resurfaced in red. There were 2 implant designs that had CRR 95% confidence intervals entirely < 50%: Legion/Genesis II and Journey II/Journey. 2 implant designs had 95% confidence intervals entirely > 50%: Triathlon/Triathlon TS and Persona/Persona. All others had confidence intervals that overlapped 50%. Clinical relevance ratio of KOOS JR change in cases with patella not resurfaced in green. There were 2 implant designs that had CRR 95% confidence intervals entirely < 50%: Legion/Genesis II and Evolution MP/Evolution MP. All others had confidence intervals that overlapped 50%.