Stratified exercise therapy does not improve outcomes compared with usual exercise therapy in people with knee osteoarthritis (OCTOPuS study): a cluster randomised trial

Question: In people with knee osteoarthritis, how much more effective is stratified exercise therapy that distinguishes three subgroups (high muscle strength subgroup, low muscle strength subgroup, obesity subgroup) in reducing knee pain and improving physical function than usual exercise therapy? Design: Pragmatic cluster randomised controlled trial in a primary care setting. Participants: A total of 335 people with knee osteoarthritis: 153 in an experimental arm and 182 in a control arm. Intervention: Physiotherapy practices were randomised into an experimental arm providing stratified exercise therapy (supplemented by a dietary intervention from a dietician for the obesity subgroup) or a control arm providing usual, non-stratified exercise therapy. Outcome measures: Primary outcomes were knee pain severity (numerical rating scale for pain, 0 to 10) and physical function (Knee Injury and Osteoarthritis Outcome Score subscale activities of daily living, 0 to 100). Measurements were performed at baseline, 3 months (primary endpoint) and 6 and 12 months (follow-up). Intention-to-treat, multilevel, regression analysis was performed. Results: Negligible differences were found between the experimental and control groups in knee pain (mean adjusted difference 0.2, 95% CI –0.4 to 0.7) and physical function (–0.8, 95% CI –4.3 to 2.6) at 3 months. Similar effects between groups were also found for each subgroup separately, as well as at other time points and for nearly all secondary outcome measures. Conclusion: This pragmatic trial demonstrated no added value regarding clinical outcomes of the model of stratified exercise therapy compared with 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 stratified exercise therapy possibly being applied in the control arm. Registration: Netherlands National Trial Register NL7463.