The paper by Malmberg et al in the current issue of Thorax 1 deals with the difficult subject of the hyperventilation syndrome and finds that these patients have a disproportionately high ventilatory response to change of body position from supine to standing. The authors suggest that this can be used as a diagnostic criterion for hyperventilation syndrome. Hyperventilation is a confused and poorly documented subject and the publication of this paper provides an opportunity to review some of the particularly controversial aspects of this subject.

The first issue concerns the basis for the labelling of these patients as “hyperventilation syndrome”. Some of the controversies about the use of this term have recently been reviewed by Folgering2and by Gardner.3 The physiological definition of hyperventilation is alveolar ventilation that is inappropriately high for the metabolic production of carbon dioxide, leading to reduction of arterial Pco₂ (Paco ₂) below the normal range (hypocapnia) and respiratory alkalosis. The combination can lead both to vasoconstriction in selected vascular beds and to neuronal hyperexcitability producing symptoms involving most systems of the body. Many psychosomatic syndromes have been described in the past in which hyperventilation has a variable and uncertain role but the term “hyperventilation syndrome” was first used in 1938 to describe patients with the somatic symptoms of both hypocapnia and anxiety.4 This theme was extended by subsequent authors5-7 and the definition arrived at by Lewis and Howell in 1986 on the basis of a questionnaire of delegates at a psychophysiology meeting8 was “a syndrome induced by physiologically inappropriate hyperventilation and usually reproduced in whole or in part by voluntary hyperventilation”. However, the term “hyperventilation syndrome” is now used in so many different contexts that it could be argued that it has ceased to have any universal meaning. Some physicians diagnose it in the presence of the somatic symptoms of hypocapnia either at rest or induced by voluntary overbreathing without assumptions about aetiology,9 or regard it primarily as an abnormality of respiratory control10 or as a variant of disproportionate breathlessness.11 ,12 Folgering accepts that anxiety may be absent and has recently suggested a new definition as “a dysregulation of ventilation causing hypocapnia in the absence of organic causes for hyperventilation, with symptoms and complaints not exclusively associated with hypocapnia”. Many refuse to recognise it as a separate entity13 or regard it as secondary to organic disease and especially to asthma.14 ,15 Many would not use the term in the presence of any organic cause of hyperventilation, yet organic and psychiatric factors are usually difficult to separate. Lum16 regards hyperventilation as a form of conditioned response and avoids use of the term “hyperventilation syndrome”. Gardner3 believes that it is not useful in the clinical context to label a patient with hypocapnia as “hyperventilation syndrome” and that the term should be abandoned. He believes that hyperventilation is often due to a complex interaction between a range of organic, psychogenic, and physiological factors and that, in all cases, the initiating and sustaining cause or causes of the increased respiratory drive causing the hyperventilation should be sought and documented. Use of a label such as “hyperventilation syndrome” tends to preclude further search for underlying aetiological factors and can be dangerous in the context of the emergency room. It is difficult to assess which of the current definitions applies to the patients in the study by Malmberget al. There was no clear evidence of anxiety or organic causes of hyperventilation and the end tidalPco₂ was recorded as being no different from the control value at rest. However, the history suggested hypocapnia at other times although the presence of a low Pco₂ at that time was not documented. About the only certain statement that can be made is that these patients do not have chronic hyperventilation.17

Hyperventilation occurs in many different specialties including neurology, cardiology, chest medicine, and psychiatry. The patient population to which the term is applied will vary greatly between specialties. Without a clear understanding of the patient population being studied it is difficult to understand the significance of any findings concerning this subject, and the source of the patients and the clinical context from which they were recruited into the study requires particular emphasis in any study about hyperventilation. There is considerable ignorance among physicians about psychiatry and the precise criteria that are used for psychiatric diagnoses. Many physicians in medical specialties assume that the demonstration of a low arterial Pco₂ automatically diagnoses hyperventilation syndrome with an assumption that the patient has some unspecified psychiatric condition. Such patients are then considered to be no longer of interest to the physician. This is often unfair to the patient, and ignores the wide range of organic, behavioural and physiological causes of hyperventilation, many of which can and usually do coexist. These uncertainties reflect the complexity of this subject which falls between psychiatry, clinical medicine, and physiology.

It is difficult to determine the significance of the suggestion in the paper by Malmberg et al that their findings should be used as the basis for a diagnostic test for hyperventilation syndrome. At a physiological level this finding requires a more detailed physiological study to evaluate mechanisms before its significance can be ascertained. At a clinical level ventilation is difficult to measure and most routine lung function laboratories do not have tilt tables or the facilities for measurement of respiratory control variables, especially while the patient is moving from one position to another. The methodology for diagnosing hyperventilation is even more controversial than the issues surrounding the definition, and it could be argued that diagnostic criteria cannot adequately be defined unless the issue of definition has been clarified. Strict adherence to the physiological definition would require documentation of hypocapnia, but hyperventilation may be sporadic and there are few current techniques for measuring Pco₂ over prolonged periods of time outside the laboratory.18 ,19 As in the present study, hyperventilation is often diagnosed with, not only no evidence of a low arterial, end tidal or transcutaneousPco₂ , but even with evidence of a normalPco₂ on spot sampling. While it is possible to provide convincing evidence of hyperventilation on behavioural grounds, using the term hyperventilation in the presence of a normalPco₂ puts an onus on the authors to be more meticulous than usual in documenting the criteria by which hyperventilation was diagnosed. Many such descriptions are sadly unconvincing.

Hypocapnia induces a range of symptoms, and symptom checklists such as the Nijmegen questionnaire20 have often been used for diagnosing the hyperventilation syndrome. However, many would argue that most of these symptoms are non-specific and do not provide an adequate basis for diagnosis when used alone. It could be argued that the only symptoms specific to hypocapnia are paraesthesiae and tetany, possibly combined with symptoms due to cerebral vasoconstriction and hypoxia. In clinical practice the clinical finding of a low arterial or end tidal Pco₂ is of little relevance in itself if there are no associated symptoms of hypocapnia, or if the symptoms are of minor importance compared with the symptoms of the disorder causing the hyperventilation. In other situations the symptoms of hyperventilation are pivotal to the clinical presentation of the patients. The reporting of familiar symptoms during voluntary hyperventilation is often used as a basis for diagnosis, but this has been criticised by Hornsveld et al 21 in that similar symptoms are also reported when the Paco ₂ is artificially maintained at normal levels during voluntary overbreathing.

The role of anxiety disorders in the paper by Malmberg and colleagues is unclear. No psychiatric disorders were reported but their patients nevertheless had symptoms of “episodic dyspnoea or air hunger” and “palpitations, sweating, trembling, dryness of the mouth or other symptoms of overactivity of the autonomic nervous system . . . suggestive of panic disorder”. This definition of panic is imprecise and no formal psychiatric assessment was apparently performed. Contrary to the statement in the introduction to the paper, the association between anxiety and panic is still controversial.22 ,23Although anxiety was a core component in the original description of hyperventilation syndrome, the relation of hyperventilation to anxiety is not simple and hyperventilation can occur without anxiety17 ,24 ,25 or anxiety may be induced by hyperventilation.26 Anxiety can be associated with both mild hyperventilation and abnormalities of breathing pattern.27 Endogenous non-retarded depression can be associated with hyperventilation28 and phobic patients have a high prevalence of breathing difficulties.29 ,30The predisposition to overbreathe in response to stress may be dependent on biological vulnerability, personality, and cognitive variables22 as well as individual interpretation of the hyperventilation induced somatic symptoms,31 and may become a conditioned response.22 Because of the complexities of this subject, it could be argued that any paper about hyperventilation requires a collaborative input from a psychiatrist or psychologist. At a clinical level such an input is often required for no reason other than to counter the assumptions of the referring clinician that a patient with hyperventilation must automatically have an anxiety state. Often there is a complex interaction between organic and psychiatric factors such as depression and phobic states which require a combined input from a physician and a psychiatrist.

The authors describe their patients as having a history of recurrent or episodic dyspnoea. Howell11 ,12 regards disproportionate dyspnoea as being synonymous with hyperventilation syndrome, and he has described the characteristics of these patients. Gardner3has argued that the two are not synonymous and that, if a patient is breathless for reasons that are not clear, it is the breathlessness that is the primary condition for which a cause should be sought and that, if hyperventilation is also present, it is usually secondary to the breathlessness and of little clinical importance. Patients are often referred with a label of psychogenic dyspnoea when the degree of distress seems disproportionate to the clinical findings and lung function or blood gas data, but there is almost no literature on this subject.32 ,33 Dyspnoea is what the patient reports and it is therefore difficult to dispute. Studies of breathlessness are impeded by uncertainty about the basic mechanisms34 and there are probably many different forms. Air hunger, or a sensation of inability to take a satisfying breath or to fill the lungs, was reported by patients in the Malmberg study and may indicate a psychogenic component to breathlessness. It was a universal feature of the patients with chronic hyperventilation studied by Gardner and Bass,17 ,25 and more recently Gardner has reported patients in whom the primary presentation is air hunger which leads to variable degrees of panic and hyperventilation.3

Asthma is a common underlying basis for hyperventilation15 ,35 ,36 and Gardner has argued that many patients with a primary presentation of hyperventilation have very mild and often previously undiagnosed asthma.37 Malmberget al describe the exclusion of “physician diagnosed” asthma but give no details about how this was achieved, although patients had “histamine provocation tests and pulmonary diffusion capacity if clinically indicated”. Asthma is a surprisingly difficult and controversial diagnosis and many physicians would argue that these indications should be clarified.

In summary, it would seem that the patients in the study by Malmberg and colleagues might fit more accurately into a classification of dyspnoea and air hunger with secondary intermittent hyperventilation rather than receiving an automatic label of “hyperventilation syndrome”. Such a reclassification may lead to a different interpretation of these clearly interesting results and may suggest lines of enquiry for possible mechanisms of air hunger. The physiological basis for these responses requires investigation and may provide useful insights into mechanisms by which postural changes can influence control of breathing and respiratory sensations. It is probably unhelpful to suggest this as the basis for yet another indirect diagnostic test for hyperventilation syndrome.