Illi et al. 2012, respiratory muscle training and exercise performance (meta-analysis)

Sports Medicine systematic review and meta-analysis of respiratory muscle training (RMT) in healthy people, pooling eight controlled trials and running a regression across 46 studies. RMT improved endurance performance on constant-load tests (about +16%) and on intermittent incremental tests (about +18.5%), but produced no significant improvement on conventional incremental (VO₂max-style ramp) tests. The benefit was larger in less-fit participants and shrank as fitness rose, by roughly 6% for every 10 mL/kg/min of VO₂max, so trained endurance athletes should expect small effects at best. Longer test durations showed larger gains, and there was no difference between running, cycling, rowing and swimming. The dependence of the effect on the test protocol is the key caveat: it appears in some tests and vanishes in others, pointing to a modest and possibly partly non-specific effect. The underlying trials were small, unblinded, and likely subject to publication bias.