OBJECTIVE: To investigate the effect of a 12-week exercise-based physical therapy on movement-evoked pain (MEP) in knee osteoarthritis (OA) and assess the relations between baseline psychological factors and treatment responses based on MEP. BACKGROUND: Pain during movement (i.e., MEP) is one of the most common complaints amongst people with knee OA, which induces greater functional challenges during daily activities. Although the worsening of pain during exercise appears to hinder activity-based treatment adherence, the assessments of MEP tend to be overlooked in the pain management in knee OA. METHODS: The data analyzed in this study was from the parent study, Wearable Sensors in Knee Osteoarthritis (WESENS-OA) study, a longitudinal, single-arm clinical trial of a 12-week exercise intervention in people with symptomatic knee OA. In the WESENS-OA (n = 60), participants self-reported pain intensity during the nominated activity as well as psychological symptoms (Center for Epidemiologic Studies Depression Scale), cognitive pattern (Pain Catastrophizing Scale), and central sensitization (Central Sensitization Inventory). Pain intensity during functional tasks and strength testing were collected during the laboratory visits. Three characteristics of MEP outcomes were assessed: (1) pain during the nominated activity, (2) functional MEP, and (3) exercise MEP. The participants reported the pain intensity during the nominated activity that is most troublesome due to their knee pain weekly using the 11-point numerical rating scale (NRS, range: 0-10). Pain intensity during or immediately after functional tasks (i.e., 6-minute walking, chair stand, climbing the staircase, range: 0-10) and strength testing (i.e., maximal voluntary isometric contractions of knee extensors and flexors, isokinetic knee flexion-extension at 60°/s and 120°/s each, range: 0-10) were evaluated to measure functional MEP or exercise MEP, respectively. MEP index score (i.e., maximum pain corrected for baseline pain) was reported as the outcome. Mixed model repeated measures (MMRM) analysis was utilized to evaluate the effect of an exercise intervention on the change in movement-evoked pain from baseline to 12 weeks, and least square means with the standard error were reported simultaneously. Each participant’s treatment response was determined by comparing the change in MEP from baseline to Week 12. A participant with a difference of ≥ 2 points in NRS for pain-nominated activity was classified as a responder, and for functional MEP and exercise MEP, treatment responder classification was done based on the hierarchical agglomerative clustering analysis. After dichotomizing participants, using binomial regression, the relative risk (RR) was reported to explore the predictive ability of CSI, CES-D, and PCS on treatment response based on MEP. RESULTS: For pain during nominated activity, the mean difference (MD) from baseline to Week 6 was -1.72 with 95% CI (-2.35 to -1.08; P <0.0001), and the difference at Week 12 was -2.41 with 95% CI (-2.92 to -1.89; P <0.0001). For functional MEP, MD with 95% CI at Week 6 was -0.34 (-0.52 to -0.16; P = 0.0003), and at Week 12 was -0.44 (-0.58 to -0.31; P<0.0001). Unlike the other MEP measures, for exercise MEP, MD was measured at one timepoint, at Week 12, which was -0.58 with 95% CI (-0.88 to -0.27; P = 0.003). Our secondary analysis did not provide evidence that baseline measures of CSI, CES-D, and PCS were associated with the treatment response based on the change in MEP measures after the exercise intervention. CONCLUSION: In conclusion, our study strongly supports the effect of exercise on reduction in three measures of MEP—pain during nominated activity, functional MEP, and exercise MEP. Our hypothesis to confirm the predictive ability of psychological factors on treatment response based on MEP was not reached, with insufficient evidence to substantiate such a notion. However, our findings hold strength to pioneer discovering the underlying mechanism of understudied MEP in knee OA. / 2026-02-28T00:00:00Z