Robotic-assisted TKA is Not Associated With Decreased Odds of Early Revision: An Analysis of the American Joint Replacement Registry

Abstract

Abstract Background Robotic-assisted TKA continues to see wider clinical use, despite limited knowledge of its impact on patient satisfaction and implant survival. Most studies to date have presented small cohorts and came from single-surgeon or single-center experiences. Therefore, a population-level comparison of revision rates between robotic-assisted and conventional TKA in the registry setting may help arthroplasty surgeons better define whether robotic assistance provides a meaningful advantage compared with the conventional technique. Questions/purposes (1) After controlling for confounding variables, such as surgeon, location of surgery, and patient comorbidity profile, were robotic-assisted TKAs less likely than conventional TKAs to result in revision for any reason at 2 years? (2) After again controlling for confounding variables, were robotic-assisted TKAs less likely to result in any specific reasons for revision than the conventional technique at 2 years? Methods The American Joint Replacement Registry was used to identify patients 65 years or older who underwent TKA between January 2017 and March 2020 with a minimum of 2 years of follow-up. Patients were limited to age 65 yeas or older to link TKAs to Medicare claims data. Two retrospective cohorts were created: robotic-assisted TKA and conventional TKA. Patient demographic variables included in the analysis were age, gender, BMI, and race. Additional characteristics included the Charlson comorbidity index, anesthesia type, year of the index procedure, and length of stay. A total of 10% (14,216 of 142,550) of TKAs performed during this study period used robotics. Patients with robotic-assisted TKA and those with conventional TKA were similar regarding age (73 ± 6 years versus 73 ± 6 years; p = 0.31) and gender (62% [8736 of 14,126] versus 62% [79,399 of 128,334] women; p = 0.34). A multivariable, mixed-effects logistic regression model was created to analyze the odds of all-cause revision as a factor of robot use, and a logistic regression model was created to investigate specific revision diagnoses. Results After controlling for potentially confounding variables, such as surgeon, location of surgery, and Charlson comorbidity index, we found no difference between the robotic-assisted and conventional TKAs in terms of the odds of revision at 2 years (OR of robotic-assisted versus conventional TKA 1.0 [95% CI 0.8 to 1.3]; p = 0.92). The reasons for revision of robotic-assisted TKA did not differ from those of conventional TKA, except for an increased odds of instability (OR 1.6 [95% CI 1.0 to 2.4]; p = 0.04) and pain (OR 2.1 [95% CI 1.4 to 3.0]; p < 0.001) in the robotic-assisted cohort. Conclusion In light of these findings, surgeons should not assume that robotic assistance in TKA will lead to improved early implant survival. Our findings do not support an improvement over conventional TKA with robotic assistance with regards to common causes of early revisions such malalignment, malposition, stiffness, pain, and instability, and in some cases, suggest there is a benefit to conventional TKA. Differences in the mid-term and long-term revision risk with conventional versus robotic-assisted TKA remain unknown. Appropriate informed consent around the use of robotic assistance should not imply decreased early revision risk. Level of Evidence Level III, therapeutic study.