A decade ago several research studies highlighted the underdiagnosis of asthma, particularly in children1-3 but also in the elderly,4 ,5 and most general practitioners felt an increased pressure on them to diagnose this common chronic respiratory disorder. It was implied that asthma should be considered whenever a patient presented with a persistent cough and that far more people “deserved” to be on effective treatment with inhaled steroids. Undoubtedly, as a consequence of this message, many with asthma gained treatment which otherwise would not have been provided and, presumably, improved their morbidity and quality of life.

However, there has emerged a down side to this campaign in that some subjects with other respiratory conditions and some with no lung disease at all have been labelled as having asthma, leading not only to years of receiving unnecessary medication but also to the development of psychological dependence on the asthma label and its associated quest for improvements in symptom control. Most of such subjects have chronic obstructive pulmonary disease (COPD), but in my clinical practice I have also found cases of hyperventilation, recrudescent tuberculosis, and severe sleep apnoea syndrome. The records of the asthma diagnosis in one subject (now correctly labelled as having COPD) read merely: “Cough. Wheeze. Ventolin. Becotide”. It is now clear that, whenever possible, diagnoses of obstructive airways diseases—whether asthma, COPD, or other rarer conditions—should be supported by objective evidence from measures of lung function.

Within the last decade in the UK, and similarly elsewhere, there have been consensus publications on the management of asthma6 ,7 and COPD.8 Much is written in these documents about the management of these conditions but less on diagnosis. The 1995 review and position statement on asthma management6 said only that “If the variable nature of airway narrowing which is characteristic of asthma cannot be demonstrated by any other means, then in adults and older children a trial of high dose oral steroids with peak flow monitoring for a minimum of two weeks is essential”. The COPD guidelines are more specific, grading the condition into mild, moderate, and severe categories on the basis of forced expiratory volume in one second (FEV₁) as a percentage of predicted as well as symptoms and signs.8

General practitioners in the UK have been able to prescribe peak flow meters on the National Health Service since 1990 and are thus most likely to use these devices for both the diagnosis and management of asthma and COPD in the community. The benefits of peak flow monitoring in the management of established asthma have become clear since Beasleyet al first published their initial study of self-management.9 However, it is much less clear that measurements of peak flow alone can be considered sufficient in the diagnosis of asthma and COPD. A primary and secondary care respiratory specialists working group, with representation from the UK and the Netherlands, clearly stated that formal spirometric testing was necessary.10

In this issue of Thorax Thiadenset al,11 another group from the Netherlands, have carefully compared the usefulness of peak flow recordings and their changes in response to bronchodilators with the presumed gold standard of changes in FEV₁. They gathered data from 240 adults not previously labelled as having either COPD or asthma who presented with cough persisting for at least two weeks to one primary health care centre over a 15 month period. Low peak flow (found in 86 subjects) had a positive predictive value (PPV) of only 47% for low FEV₁ (found in 48 subjects) but a negative predictive value (NPV) of 95%. Several different definitions of increase in peak flow with a bronchodilator challenge were compared with a 9% or more change in FEV₁. The PPVs for both absolute and percentage changes were all less than 66%, though the NPVs were 88–93%.

With the recent focus on COPD as well as asthma and the resultant increased clarity in how these conditions should be managed, it is clearly imperative that patients with obstructive pulmonary diseases are correctly diagnosed from the outset. Thiadenset al did not examine reversibility testing with courses of oral prednisolone nor the usefulness of diurnal variation in peak flow, but doubt has correctly been expressed about the importance of the finding of a low peak flow rate in patients who present with cough. The message appears to be that the finding of normal peak flow at presentation makes COPD unlikely (though not excluding asthma), but that low peak flow should be taken as a trigger for spirometric testing. This presents a considerable challenge for primary care practitioners.

Spirometric measurements can be conducted “in house” in general practices or in the lung function laboratories of local hospitals, either by consultant referral or by open access arrangements. The former route has the advantage of convenience for patients but a major disadvantage in relation to quality control, with difficulties in equipment selection, staff training, and interpretation of results. The local hospital service, though less convenient, obviates these disadvantages but there is a considerable risk of it being overwhelmed if anything like all the potential patients are referred to it. There may well be scope for locally based services, set up by primary care groups, whereby one primary care team provides spirometric services for others in its vicinity with quality control achieved by links to hospital staff. If these issues are not tackled in a coordinated fashion, there is a risk that future patients will continue to be diagnosed with the wrong condition in primary care and not receive the best treatment for their symptoms.