Benchmarking total knee replacement constructs using noninferiority analysis: the New Zealand joint registry study.

<bold>Background: </bold>The aim of this study was to compare the relative performance of total knee replacement constructs and discern if there is variability in performance in currently commonly used prostheses in the New Zealand Joint Registry (NZJR) using a noninferiority analysis.<bold>Methods: </bold>All patients who underwent a primary total knee replacement (TKR) registered in the NZJR between 1st January 1999 to June 2020 were identified. Using a noninferiority analysis, the performance of total knee replacement prostheses were compared with the best performing contemporary construct. Construct all-cause revision rate was estimated using the 1-Kaplan Meier survival function method to estimate net failure. The difference in all-cause revision rates between the contemporary benchmark and all other constructs was tested.<bold>Results: </bold>In total 110 183 TKR were recorded and 25 constructs (102 717 procedures) had > 500 procedures at risk at 3 years post-primary of which 5 were inferior by at least 20 % relative risk of which, one was inferior by at least 100 % relative risk. 14 constructs were identified with > 500 procedures at risk at 10 years with 5 inferior by at least 20 %, of which 2 were inferior by > 100 % relative risk.<bold>Conclusions: </bold>We discerned that there is great variability in construct performance and at all time points, greater than 25 % of constructs are inferior to the best performing construct by at least 20 %. These results can help inform patients, clinicians and health care funders when considering TKR surgery. [ABSTRACT FROM AUTHOR]