Abstract
Background/Objectives: Temporomandibular disorders (TMDs) are musculoskeletal conditions in which rehabilitation aims to reduce pain and restore mandibular mobility and function. We estimated the between-group effects of conservative and minimally invasive treatments versus inactive controls and examined important sources of variation. Methods: We included randomized controlled trials published in 2015–2025 in adults diagnosed with RDC/TMD or DC/TMD. PubMed/MEDLINE, Embase, Web of Science Core Collection, Scopus and CENTRAL were searched, with the final searches completed in January 2026. The review was prospectively registered in PROSPERO (CRD420251250251). We synthesized comparisons of conservative and minimally invasive treatments with inactive controls in adults with TMD. Random-effects meta-analyses used REML and modified Hartung–Knapp confidence intervals. Two human reviewers independently assessed risk of bias with RoB 2 and certainty with GRADE. Results: Thirty-five studies contributed to 69 effects. Compared with inactive controls, short-term pain was 1.31 points lower on a 0–10 scale (95% CI 0.62–2.00; 23 studies; very-low certainty) and 0.94 points lower (95% CI 0.15–1.74) after high-risk results were excluded. Mouth opening was 1.96 mm greater (95% CI 0.33–3.60; 26 effects from 25 studies; very-low certainty) and 0.75 mm greater (95% CI −0.62 to 2.12) after high-risk results were excluded; the latter confidence interval included no difference. Prediction intervals included no benefit. Evidence for jaw function, oral-health-related quality of life and longer-term effects was sparse. None of the 73 result-specific risk-of-bias assessments was low risk, and 25 were high risk. Conclusions: The average short-term pain benefit was modest and decreased after high-risk results were removed. The mouth-opening difference became inconclusive after high-risk results were excluded. These descriptive pooled estimates support transparent discussion of uncertainty in rehabilitation practice but should not be used to rank treatments or predict an individual response.