Does Transitioning to Robotic-Assisted Total Knee Arthroplasty Change Surgeon Revision Risk or Failure Modes?: A Baseline Comparative Analysis Before and After Robotic Adoption.

Abstract

Robotic-assisted total knee arthroplasty (RA-TKA) may disproportionately affect revision rates based on a surgeon's revision profile. This national registry study (1) stratified surgeons by pre-RA-TKA revision rates as low, medium, or high and (2) compared the change in revision rates and modes of failure within each cohort after commencing RA-TKA. All TKA procedures performed for osteoarthritis and recorded by the Australian Orthopaedic Association National Joint Replacement Registry from September 1999 to December 2023 were screened for eligibility. A total of 178 surgeons were included, performing 85,075 TKAs since commencing RA-TKA. TKA with hinged, constrained or medial pivot bearings, and American Society of Anesthesiologists (ASA) Scores 4 and 5 were excluded. Surgeons were stratified into quartiles based on their 2-year cumulative-percent-revision (CPR) rates before commencing RA-TKA into low (2-year CPR ≤1%), middle (1%-2.5%), and high revision cohorts (≥2.5%). The primary outcome was CPR with comparisons between time-matched RA-TKA and non-RA-TKA for each surgeon cohort after commencing RA-TKA. Secondary outcomes were changes in failure modes. Cox proportional hazards modelling controlled for age, sex, body mass index, ASA, patellar resurfacing, surgical year, fixation, polyethylene, and bearing type. After commencing RA-TKA, there was no difference in CPR between RA-TKA and non-RA-TKA in low (hazard ratio [HR] 1.27; 95% CI 0.84-1.91; <i>p</i> = 0.26) and middle-revision rate surgeon groups (HR 0.99; 95% CI 0.83-1.19; <i>p</i> = 0.95). High-revision rate surgeons had decreased three-month CPR (HR 0.51; 95% CI 0.34-0.75; <i>p</i> < 0.001) but increased >9-month CPR (HR 2.28; 95% CI 1.74-2.98; <i>p</i> < 0.001). The mean follow-up was 2.4, 2.4 and 2.3 years, respectively, with maximum 7-year follow-up. Failure modes including infection, loosening, and instability did not change within any revision rate cohort after commencing RA-TKA (<i>p</i> > 0.05). At a grouped cohort level, commencing RA-TKA does not improve revision rates nor change failure modes. Surgeons with higher revision rates before commencing RA-TKA experienced further increases in revision rates. Level III. See Instructions for Authors for a complete description of levels of evidence.