Breathing exercises for asthma: panacea or placebo?

B reathing is easy—we all do it all of the time from the moment we are born. So why should it be necessary to teach people how to breathe? How is it possible to get something wrong when it appears so simple? And, if it is wrong, can it be corrected? In this issue of Thorax Holloway and West report a randomised trial of adding the Papworth technique—a combination of breathing and relaxation exercises first introduced for hyperventilation syndromes—to usual care in patients with asthma (see p 1039). They recruited 85 volunteers with asthma from a primary care asthma register, 72 of whom completed the parallel group study and were followed up for 12 months. The group had relatively mild asthma with mean forced expiratory volume in 1 s (FEV1) .90% predicted and thus, unsurprisingly, there were no significant improvements in the conventional physiological measures. However, the actively managed group who received five sessions of the Papworth technique, each lasting 1 h, had significant improvements in the St George’s Respiratory Questionnaire (SGRQ), symptoms and anxiety and depression scores. This study raises as many questions as it answers. Why might breathing be amenable to change? Most homeostatic mechanisms (such as body temperature, hormonal cycles, blood sugar levels) are automatic and impervious to voluntary control. But breathing is different. The body’s automatic ‘‘settings’’ can be overridden by voluntary control. It is possible to breath hold and, with training, to do so for prolonged periods (eg, several minutes as exemplified by the extreme diving sport). It is equally possible and easier to hyperventilate, so inducing hypocapnia and alkalosis. Alkalosis alters the levels of ionised calcium, nerve conduction speeds and smooth muscle control, especially of the vessels supplying the brain and heart. Cerebral blood flow may be reduced 40% by hyperventilation. Patients become aware of a myriad of symptoms that can include faintness, dizziness, chest tightness/discomfort and paraesthesia, especially of the hands. Some patients appear to hyperventilate on a regular basis leading to the description of ‘‘hyperventilation syndromes’’, although there is continuing controversy about their existence and some prefer terms such as ‘‘behavioural breathlessness’’. Lum in the 1960s recognised patients who had continuing symptoms despite normal cardiorespiratory physiology on standard testing and who appeared to be hyperventilating often with an irregular pattern of breathing. A programme of breathing and relaxation exercises was developed at Papworth Hospital and Lum reported observational data claiming that the Papworth technique could resolve symptoms in 70% or more of those he had defined as hyperventilating. His work—which included patients with asthma—was not well received by mainstream respiratory medicine. There were two major reasons for the doubt at the time: (1) a prevalent belief that the hyperventilation syndromes were more likely to be psychological than physiological which, coupled to less than exacting definitions of diagnostic criteria and treatment evaluation, was understandable; and (2) there was an asthma treatment revolution that was backed by solid science. The new inhaled drugs were able to improve dramatically the lives of many patients with asthma, and the inhalers also addressed the airway inflammation. The mantra for managing asthma in the 1980s became ‘‘asthma is an inflammatory condition—prescribe an inhaled steroid’’. While the emphasis on inflammation dominated research and treatment plans, it was easy to overlook other demonstrable means of inducing bronchospasm. Exercise-induced asthma and voluntary hyperventilation without exercise can both induce bronchospasm in sensitive individuals with relatively little inflammation. However, despite the armamentarium of effective drugs, not all patients with asthma are fully controlled. Surveys have shown that, despite treatment, a significant proportion of patients have continuing symptoms such as continuing to wake several nights per week. Every clinician is aware that increasing doses of inhaled treatments is not effective in all patients. It is common to blame patient factors such as emotional overlays, poor compliance or poor inhaler technique for the drug failures. But it is also worth remembering that many patients with asthma exhibit a degree of hyperventilation with hypocapnia in the early phases of acute asthma. Patients with asthma are relatively frequent users of alternative therapies, often without telling their physicians. The old medical adage advises ‘‘listen to the patient’’; perhaps they are telling us something by their pursuit of buteyko, hypnosis, yoga or breathing exercises. All these share a common theme of attention to the breathing pattern. A Cochrane review concluded that, while there are insufficient data for firm conclusions about efficacy, there is enough evidence to justify further trials. The study by Holloway and West is interesting, but first one must acknowledge the negatives. It was an open study, it accepted the GP diagnosis without further proof and it relied on patient volunteers (who may not necessarily be representative). Indeed, at randomisation, the ‘‘active’’ group had more symptoms at the outset (ie, more room for improvement). And the same person delivered the treatments and recorded the outcomes. It cannot therefore be considered definitive. But, on the positive side, the study accepted most of the volunteers and is probably a more typical asthma primary care population than is included in many asthma drug trials. Randomised controlled trials have tended to categorise lung function and symptoms in great detail to avoid ‘‘contamination’’ both from patients without asthma and from patients with asthma who could not benefit. The non-selective intake to this study could potentially have diluted the ability to achieve measurable effects. This cohort of patients from general practice felt and functioned better. This is an outcome measure increasingly used in studies of chronic illness and the effects were present at 12 months, long after any placebo effect ought to have worn off. Holloway and West describe an SGRQ benefit of 7 points compared with the control group. This is of similar magnitude to the randomised trials that led to long-acting b agonists being adopted enthusiastically for the treatment of chronic obstructive pulmonary disease (eg, 6.7 points on the SGRQ). The changes reported using other measures EDITORIALS 1033