Forty-one studies were included in the review.
Breathing modification technique: There were 12 RCTs and one before-and-after study. Follow-up ranged from one week to 28 weeks. Drop-out rates ranged from zero to 24%. The 12 RCTS (six Buteyko Breathing Technique and six employed respiratory physiotherapy) examined the effect of breathing modification techniques compared with control interventions. Double blinding was effective on four studies and there was a true randomisation in 11 studies.
Meta-analysis showed no significant effect of breathing retraining in peak expiratory flow, forced expiratory volume in one second (FEV1) percentage predictive, FEV1 (litre) and asthma-related quality-of-life score between the two groups. There was a statistically significant effect for end tidal CO2 from breathing retraining (WMD 1.65 mmHg, 95% CI 0.31 to 2.99; Ι²=41.3%; five trials, 313 participants). The statistical test result was not reported for the before-and-after study.
Yoga: There were nine RCTs and six before-and-after studies. Follow-up times ranged from immediate to 54 months. Drop-out rates ranged from zero to 43%. True randomisation was not clear in six RCTs and double blinding was either not clear or not reported in seven RCTs.
Meta-analysis showed no effect on peak expiratory flow (PEF) or absolute values of FEV1. There was a significant effect of yoga on FEV1 % predicted (WMD 8.79, 95% CI 0.37 to 17.22; heterogeneity present Ι²=65.2%; three trials, 221 participants) and asthma-related quality-of life (SMD -0.61, 95% CI -1.01 to -0.21; Ι²=0%; two trials, 102 participants). Five non-randomised studies reported improvement in lung function, asthma-related quality-of-life or airway hyper-responsiveness for the intervention compared to the control group.
Respiratory muscle training: There were four RCTS. Follow-up times ranged from four months to six months. Drop-out rates ranged from zero to 27%. All studies used muscle strengthening techniques.
Meta-analysis showed significant effect of respiratory training on B2-agonist (WMD -1.19, 95% CI -2.15 to -0.24; Ι²=0%; two trials, 41 participants).
Biofeedback training: There were seven RCTs and five before-and-after studies. Follow-up ranged from zero to 12 months. Drop-out rate ranged from zero to 19%.
Six RCTs and two before-and-after studies found significant improvement in lung function, medication use or asthma symptoms (test results not reported).