If one assumes that the observed hypocapnia in stable asthmatics is due to hyperventilation, it is an attractive proposition that retraining breathing patterns to reduce hyperventilation in such individuals should make it possible to raise carbon dioxide levels. This should theoretically reverse any bronchoconstrictive effects of hypocapnia, and possibly make use of the bronchodilatory effects of hypercapnia. There is some evidence that various forms of breathing retraining (eg, physiotherapy, yoga, and Buteyko) have had some beneficial effects on asthma patients. However, at present it is uncertain what mechanism produces these effects over and above any placebo effect. Apart from the carbon dioxide hypothesis, another potential physiologic mechanism behind breathing retraining is the effect of rhythmic stretching on the ASM.31–32 A number of psychological mechanisms are also possible such as increasing feelings of mastery or control, induction of relaxation/ meditation, or nonspecific attention effects. A recent randomized controlled trial of the Buteyko breathing technique by Cooper et al4 found that Buteyko breathing improved asthma symptoms and reduced bronchodilator use, but without affecting airway caliber or responsiveness. Thomas et al2 reported on a study of physiotherapy breathing retraining for asthma patients and found improved quality-of-life scores. However, neither of these studies shed any light on the mechanism producing the improvements generated. Buteyko, physiotherapy, and yoga therapies share very similar breathing techniques that encourage slowing the rate of breathing, and the use of breath holding. Very few studies of breathing therapy have included measures of carbon dioxide in their outcomes. One exception is the randomized controlled trial of Buteyko by Bowler et al,33who found no significant differences in means between the intervention and control groups after intervention, but unfortunately do not report any within subject changes. In a controlled trial currently only published in abstract form, Abramson et al34 reported no changes in lung function, but did find some increase in ETCO2 in one group and some reduction in response to carbon dioxide in another. They employed a factorial design involving four groups, one receiving “full” Buteyko (Buteyko practitioner plus Buteyko video), two receiving “partial” Buteyko (Buteyko practitioner plus placebo video, or Buteyko video plus placebo educator), and one receiving no Buteyko (placebo video plus placebo educator). Their findings are interesting, but the complexity of their design makes interpretation difficult.