OP0188 CLINICAL EFFECTIVENESS OF STRATIFIED EXERCISE THERAPY COMPARED TO USUAL EXERCISE THERAPY IN PATIENTS WITH KNEE OSTEOARTHRITIS: A CLUSTER RANDOMIZED CONTROLLED TRIAL (OCTOPuS-STUDY)

Abstract

Background: There is strong, high-quality evidence for the effectiveness of exercise therapy in people with knee osteoarthritis (OA).1 However, although effective, the average effect size is only moderate (approximately 0.5).1 This may be attributed to the current 'one-size-fts-all' exercise approach, while a stratifed approach may yield superior clinical and economic outcomes. We are the frst to test a model of stratifed exercise therapy in patients in knee OA. This model was based on 3 previously identifed subgroups2 that are aligned with well-accepted OA phenotypes3, namely a 'low muscle strength subgroup' ('age-induced phenotype'), 'high muscle strength subgroup' ('post-traumatic phenotype') and 'obesity subgroup' ('metabolic phenotype'). For each subgroup, a subgroup-spe-cifc exercise therapy intervention was developed and pilot-tested4, which was supplemented by a dietary intervention for the 'obesity subgroup'. Objectives: The OCTOPuS-trial aimed to determine the effectiveness of strat-ifed exercise therapy in reducing knee pain and improving physical function, compared to usual, 'non-stratifed' exercise therapy, in patients with knee OA. Methods: We conducted a pragmatic cluster randomized controlled trial in a primary care setting in 335 people with knee osteoarthritis: 153 in the experimental arm and 182 in the control arm. Physiotherapy practices were randomized into the experimental arm providing the model of stratifed exercise therapy supplemented by a dietary intervention from a dietician for the 'obesity subgroup' or the control arm proving usual, 'non-stratifed' exercise therapy. Primary outcomes were knee pain severity (NRS pain, 0-10) and physical function (KOOS subscale daily living, 0-100). Measurements were performed at baseline, and 3-(primary endpoint), 6-, and 12-months follow-up. Intention-to-treat, multilevel, regression analysis was performed. Results: We found statistically non-signifcant differences in knee pain (mean difference (95% confdence interval): 0.19 (-0.31, 0.69)) and physical function (-0.40 (-3.91, 3.12)) at 3-months follow-up, with within-group effect sizes ranging between 0.5 and 0.7. Non-signifcant differences were also found for all other time points and for nearly all secondary outcome measures. Moreover, effects of experimental and control intervention were similar in each of the 3 subgroup separately. Conclusion: This trial demonstrated no added value with respect to clinical outcomes of our model of stratifed exercise therapy compared to usual exercise therapy. This could be attributed to the experimental arm therapists facing difficulty in effectively applying the model (especially in the 'obesity subgroup') and to elements of the model possibly being applied in the control arm.