Diaphragmatic Training for Non-Specific Low Back Pain: A Systematic Review and Meta-Analysis.

Abstract

To investigate the effectiveness of diaphragmatic training interventions in adults with non-specific low back pain. PubMed (MEDLINE), Embase, Web of Science, Cochrane Library, Cumulative Index to Nursing and Allied Health Literature, and China National Knowledge Infrastructure databases were systematically searched from inception to December 2024, with an updated search conducted through December 2025. Inclusion criteria included the following: (1) adults (≥18 years) with non-specific low back pain; (2) interventions involving diaphragmatic training, either as a standalone treatment or in combination with other therapies; (3) any comparator; (4) pain intensity and functional disability as primary outcomes, with diaphragmatic thickness as a secondary outcome; and (5) randomized controlled trial designs. Two reviewers independently screened the studies. Data were extracted independently by two reviewers. The Cochrane Risk of Bias (RoB 1) assessment and Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) certainty ratings were conducted by four reviewers. Thirteen studies (626 participants) were included. Diaphragmatic training likely reduced pain intensity compared with control interventions (SMD = -0.60; 95% CI = -0.86 to -0.34; 11 trials, n = 465; moderate-certainty evidence). It also likely improved functional disability (SMD = -0.71; 95% CI = -0.90 to -0.53; 9 trials, n = 469; moderate-certainty evidence). Evidence for changes in diaphragmatic thickness was of very low certainty. Subgroup analyses suggested that better pain outcomes typically occurred with 2-3 sessions per week. Moderate-certainty evidence suggests that diaphragmatic training may be considered as an adjunct to rehabilitation to reduce pain intensity and improve functional disability in adults with non-specific low back pain. Prespecified trials with longer follow-up are needed to clarify effects on diaphragm thickness and the influence of training parameters.