We recognise and applaud the efforts of Holloway and West1 and are encouraged by their findings. We agree that further investigation is warranted but that greater scientific methodology needs to be applied. Hyperventilation appears to be the only form of dysfunctional breathing that most physicians recognise yet even within this label, acute and chronic hyperventilation are physiologically distinct.2 Efforts must be made to further elucidate the nature of dysfunctional breathing in its many forms3 and direct appropriate therapy towards appropriate patients. It is a reasonable assumption that breathing retraining will be efficacious only in those patients where dysfunctional breathing exists. This lends support to the positive results of Holloway and West,1 given the unselected nature of the cohort of subjects recruited. It is hardly surprising that many trials of breathing retraining in unselected asthmatics prove equivocal. We liken this approach to a trial of thrombolytic therapy in unselected chest pain patients.

As a sub-note, we would guard against making assumptions on the basis of a reduction in Nijmegen scores. The Nijmegen questionnaire has only ever been validated in primary hyperventilation and even then the gold standard was physician interpreted typical symptoms.4 It was only described as a threshold value for the diagnosis of hyperventilation and there is no evidence that there is a graded correlation between Nijmegen scores and symptom severity. The domains within the questionnaire overlap with asthma symptoms. We have found it not to be predictive of capnography in patients with severe asthma. Since there is a trend towards improved spirometric values in the study group of Holloway and West,1 we would suggest caution in interpreting a reduction in Nijmegen scores as a positive signal of reduced dysfunctional breathing.