The control group matters: Pain, physical function and strength improvements relative to the comparator intervention in knee and hip osteoarthritis

Background: In knee and hip osteoarthritis (OA), the mechanism for resistance exercise improving clinical outcomes and the dose-response between strength and clinical outcomes are unknown; in part due to inconsistent trial designs across studies. Purpose: To determine whether the effects of resistance exercise interventions on pain and function differ based on comparator group; and whether there is an association between improvements in lower extremity strength with improvements in pain and function in knee and hip OA. Methods: We searched 6 databases (inception to January 28 2023,) for randomized controlled trials (RCTs) comparing land-based, resistance exercise-only interventions with no intervention or any other intervention. There were four subgroups for comparator intervention: NONE (none/placebo/sham/usual care), EXE (other exercise interventions alone), NONEXE (non-exercise interventions alone), COMBO (combined exercise + nonexercise interventions). The between-group effect (ES) was calculated for immediate post-intervention pain and function (activities of daily living (ADL) and sports/recreation (SPORT)). Meta-regression analyses were completed to evaluate the association between improvements in lower extremity strength (independent variable) and improvements in pain, ADL and SPORT (dependent variables), irrespective of comparator intervention. Results: For knee OA (257 studies), there were large benefits for pain [ES (95 % CI) = -0.92 (-1.15, -0.69)], ADL [-0.79 (-1.01, -0.56)] and SPORT [-0.79 (-1.02, -0.56)] favouring resistance exercise interventions compared to NONE. For knee pain, there was also a moderate benefit favouring COMBO interventions compared to resistance exercise interventions [0.44 (0.23, 0.65)]. For hip OA (15 studies), there were moderate benefits for pain [-0.51 (-0.68, -0.33)], ADL [-0.57 (-0.78, -0.36)] and SPORT [-0.52 (-0.70, -0.35)] favouring exercise interventions compared to NONE. For hip pain, there was also a moderate benefit favouring NONEXE interventions compared to resistance exercise interventions [0.57 (0.17, 0.97)]. For knee OA, greater strength gains were associated with larger improvements in pain [beta (95 % CI) = -0.24 (-0.38, -0.09)], ADL [-0.43 (-0.73, -0.12)] and SPORT [-0.37 (-0.73, -0.00)]. Conclusion: In knee and hip OA, the effects of resistance exercise on pain and function improvements depend on the comparator intervention. For knee OA, a dose-response relationship was observed between lower extremity strength gains with pain and function improvements.