Using matched observational data, we have previously shown that total knee arthroplasty (TKA) provides superior outcomes to an education and exercise (EduEx) program in patients with knee osteoarthritis (OA). In the present study, we conducted a secondary analysis of matched observational data from two Canadian prospective cohorts to determine if the treatment effect of EduEx versus TKA differs across patients with knee OA ranging from less severe to more severe symptoms. This was a secondary analysis of propensity-matched data from patients with knee OA receiving the GLA:D Canada EduEx program or TKA. Previously, TKA patients were matched to an EduEx patient on a 1:1 ratio based on a propensity score derived from 17 pre-treatment covariates. In the current analysis, matched patients were divided into three strata based on the propensity score (Between-group differences in improvement (the treatment effect) in pain, function, and quality of life (Knee injury and Osteoarthritis Outcome Score 12-item version; KOOS-12 subscales all scored 0 worst to 100 best) from baseline to 3- and 12-months were estimated using linear mixed models for repeated measures in each severity strata, adjusting for any unbalanced pre-treatment covariates after matching. A total of 522 patients (261 in each treatment group) were included in the matched analysis. The low propensity score strata (n=170) had the least medication use (77%) and best pain (mean 52.5), function (mean 56.2) and quality of life (mean 38.2) scores. The mid propensity score strata (n=245) had the next lowest medication use (94%) and next best pain (mean 44.0), function (mean 46.0), and quality of life (mean 29.9) scores. The high propensity score strata (n=107) used the most medication (97%) and had the worst pain (mean 35.7), function (mean 33.6), and quality of life (mean 20.2) scores. We labeled these strata as low, moderate, and high symptom severity. The analysis of treatment effects at 12-months revealed statistically significant differences in pain in favour of TKA over EduEx across all three severity strata (low severity: 21.6, 95% CI 16.3 to 26.9; moderate severity: 20.2, 95% CI 15.8 to 24.5; high severity: 29.8, 95% CI 22.2 to 37.3). The same pattern of improvement in favour of TKA over EduEx was found for function and quality of life outcomes. All estimated treatment effects at 12-months were considered clinically significant. From this study of well-matched EduEx and TKA patients, we report a clinically significant effect in favour of TKA over EduEx across all outcomes, regardless of knee OA symptom severity. TKA provided clinically superior outcomes of approximately the same magnitude in patients with low and moderate symptom severity, and of even greater magnitude in patients with the most severe symptoms. These findings challenge previous assumptions that only patients with severe knee OA symptoms should be prioritized for TKA and those with less severe symptoms should be prioritized for non-surgical treatments.
Young et al. (Wed,) studied this question.