Joint-line vs. pain-site–guided radial pressure wave therapy for knee osteoarthritis: a three-arm randomized trial

Background It remains unclear whether joint-line application of radial pressure wave therapy (RPWT) provides better outcomes than pain-site–guided application in knee osteoarthritis. We compared these strategies within a common exercise program. Methods In a single-center, open-label, three-arm randomized trial, adults aged 50–80 years received three weekly RPWT sessions (2,000 impulses, 2.0 bar, 10 Hz) targeting the two most painful points, the medial joint line, or both medial and lateral joint lines. All groups received supervised exercise and a subsequent home program. Pain intensity (NRS, 0–10) was primary and WOMAC total (0–96) secondary; four months was the primary assessment visit. We report the protocol-specified repeated-measures ANOVA and an exploratory post hoc reanalysis using baseline-adjusted mixed models for repeated measures. Four-month comparisons with pain-site–guided application were adjusted using the Holm method. Results Of 94 randomized participants, 82 provided post-baseline outcomes and 71 were assessed at four months. Mean pain and WOMAC scores improved from baseline in all three groups; ANOVA in 70 complete cases showed overall time effects (both p < 0.001 ), with no evidence of group-by-time interactions. Exploratory four-month MMRM differences vs. pain-site–guided application were − 0.01 NRS (95% CI − 1.33 to 1.31) and − 0.52 WOMAC ( − 10.78 to 9.74) for medial application, and 1.01 ( − 0.32 to 2.35) and 7.07 ( − 3.35 to 17.49) for combined application; none was significant after Holm adjustment. Imputation estimates were similar under a missing-at-random assumption, but findings remained sensitive to assumptions about unobserved outcomes. Six participants per group reported application pain (18.2%–20.0%), including three with concurrent erythema per group; the pain did not persist afterward, and no other adverse events were recorded. Conclusions Mean pain and WOMAC scores improved from baseline in all three groups. The primary ANOVA did not detect different trajectories; exploratory MMRM contrasts did not demonstrate four-month superiority. The estimates do not establish equivalence or support a preferred application site. Because all groups received RPWT and exercise, improvements cannot be attributed specifically to RPWT. Clinical Trial Registration https://clinicaltrials.gov/study/NCT06622512 , identifier NCT06622512.