Abstract
Abstract
Introduction
Component mal-positioning in total hip replacement (THR) and total knee replacement (TKR) can increase the risk of revision for various reasons. Compared to conventional surgery, relatively improved accuracy of implant positioning can be achieved using computer assisted technologies including navigation, patient-specific jigs, and robotic systems. However, it is not known whether application of these technologies has improved prosthesis survival in the real-world. This study aimed to compare risk of revision for all-causes following primary THR and TKR, and revision for dislocation following primary THR performed using computer assisted technologies compared to conventional technique.
Methods
We performed an observational study using National Joint Registry data. All adult patients undergoing primary THR and TKR for osteoarthritis between 01/04/2003 to 31/12/2020 were eligible. Patients who received metal-on-metal bearing THR were excluded. We generated propensity score weights, using Sturmer weight trimming, based on: age, gender, ASA grade, side, operation funding, year of surgery, approach, and fixation. Specific additional variables included position and bearing for THR and patellar resurfacing for TKR. For THR, effective sample sizes and duration of follow up for conventional versus computer-guided and robotic-assisted analyses were 9,379 and 10,600 procedures, and approximately 18 and 4 years, respectively. For TKR, effective sample sizes and durations of follow up for conventional versus computer-guided, patient-specific jigs, and robotic-assisted groups were 92,579 procedures over 18 years, 11,665 procedures over 8 years, and 644 procedures over 3 years, respectively. Outcomes were assessed using Kaplan-Meier analysis and expressed using hazard ratios (HR) and 95% confidence intervals (CI).
Results
For THR, analysis comparing computer-guided versus conventional technique demonstrated HR of 0.771 (95%CI 0.573–1.036) p=0.085, and 0.594 (95%CI 0.297–1.190) p=0.142, for revision for all-causes and dislocation, respectively. When comparing robotic-assisted versus conventional technique, HR for revision for all-causes was 0.480 (95%CI 0.067 –3.452) p=0.466. For TKR, compared to conventional surgery, HR for all-cause revision for procedures performed using computer guidance and patient-specific jigs were 0.967 (95% CI 0.888–1.052) p=0.430, and 0.937 (95% CI 0.708–1.241) p=0.65, respectively. HR for analysis comparing robotic-assisted versus conventional technique was 2.0940 (0.2423, 18.0995) p = 0.50.
Conclusions
This is the largest study investigating this topic utilising propensity score analysis methods. We did not find a statistically significant difference in revision for all-causes and dislocation although these analyses are underpowered to detect smaller differences in effect size between groups. Additional comparison for revision for dislocation between robotic-assisted versus conventionally performed THR was not performed as this is a subset of revision for all-causes and wide confidence intervals were already observed for that analysis. It is also important to mention this NJR analysis study is of an observational study design which has inherent limitations. Nonetheless, this is the most feasible study design to answer this research question requiring use of a large data set due to revision being a rare outcome.
Declaration of Interest
(b) declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported:I declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research project.
