Magnitude of the placebo response across non-surgical treatment modalities used for knee osteoarthritis: A meta-analysis with meta-regression.

OBJECTIVE

To quantify pain-related placebo responses across treatment modalities in placebo-controlled clinical trials of non-surgical osteoarthritis interventions.

DESIGN

A systematic review and meta-analysis of randomized controlled trials (RCTs) published since January 2008 was conducted in accordance with PRISMA guidelines. Searches of the PubMed and Embase databases were conducted through May 2025. Hedges' g was calculated as the effect size for pre-to-post change in pain outcomes using a random-effects model. Heterogeneity was assessed using the I2 statistic, and sources of variance were analyzed with multivariate meta-regression.

RESULTS

One hundred and seventy-nine RCTs were included. Four trials also included a no-treatment arm. A Hedges' g of 0.5 or greater for improvement from baseline was observed in the placebo arm of 109 (60%) trials, while 15 (8%) reported worsening. Substantial heterogeneity was present (I2 = 88%, p < 0.0001), with a wide prediction interval (-0.31 to 1.60). Multivariate meta-regression identified four covariates that together accounted for 33% of the explained variance: baseline pain, funding source, effect size in the active treatment arm, and geographical region. In subgroup analysis, outcomes were not significantly affected by treatment modality. Trials conducted in North America reported the highest effect sizes.

CONCLUSION

Most trials reported moderate to large effect sizes for improvement in the placebo group. Treatment modality did not significantly affect outcomes and most of the variance in the placebo response remained unexplained. Important limitations include substantial statistical heterogeneity, the lack of data on psychological variables and few trials with a no-treatment group.