Article Text


Peak flow meters still useful but require consistency rather than accuracy
  1. C K Connolly1
  1. 1
  1. V Brusasco2
  1. 2Professor of Respiratory Medicine, Medical School, University of Genoa, Italy;

Statistics from

I read with interest the editorial by Dr Brusasco on the usefulness of peak expiratory flow measurements1 in which he suggests that they may become obsolete. As a pioneer of the use of regular peak flow measurements in hospital patients in the 1970s,2,3 it might seem natural that I would be reluctant to see them go, but I do think I can rationally defend their place for two purposes—one dating from the start and the other from the end of my career.

The characteristic of asthma is variable airway obstruction. Despite the potential difficulties, variation in peak flow is of value in patients in whom asthma has been diagnosed as an adjunct to establishing the pattern of disease in the contexts of causation and management. Much of Dr Brusasco’s argument is dichotomous, which is inappropriate in developing an overall strategy when there are limitations to all approaches—for example, under-perception and over-perception of symptoms. Cheating does occur, but blinded readings may be used to obtain useful results. Symptoms may precede deterioration in peak flow but, in practice, a relevantly lower reading on the second attempt is one of the best early indications of onset of a deterioration in a consistent performer. Guidelines suggest that serial recordings may have a place in the diagnosis of occupational asthma4—for example, in aluminium workers.5 We have shown that, despite the difficulties relating to the patient’s best reading, there is an association between peak flow and symptoms,6 and overall mortality independent of and in addition to spirometric measurements.7

The second use of peak flow is as a useful check of quality in the diagnosis of minimal COPD when measured at the same time as and by the same instrument as forced expiratory volume in 1 second (FEV1). A normal peak flow should be of the same order as FEV1 and, in subjects with mild COPD, the geometry of the curve demands that peak flow is less affected than FEV1.

In the first case, accuracy of calibration is irrelevant as the absolute value of peak flow is not used in the assessment. Consistency is vital and that is achieved by the simple instruments. In the second case, again it is not the absolute value that matters but consistency against the definitive measurement, FEV1. Provided that it is accepted that the absolute value of peak flow rarely has any value, measurement of peak flow remains very useful in these two particular circumstances and it follows that, when restricted to these uses, elaborate calibration of the absolute value of peak flow is unnecessary.


Author’s reply

The letter from Dr Connolly gives me the opportunity to clarify some points not specifically addressed in my editorial1 leading the paper by Miller et al2 on the assessment of portable peak flow meters.

The first argument raised by Dr Connolly in defence of peak expiratory flow (PEF) measurement in asthma is that it may give additional information on the pattern of disease in patients in whom the asthma is already diagnosed. I respect this opinion, which was the basis for proposing the use of PEF for asthma monitoring and management, but this is exactly what has not received experimental support by controlled studied in two decades.3 Dr Connolly also argues that PEF measurements may be useful in the diagnosis of occupational asthma and that cheating can be detected by blinded recording. I am aware that electronic recording may help to identify fabricated or wrongly recorded data,4 but this would require more expensive devices and it is not known whether or not it increases adherence by itself. Furthermore, a smart cheater asking for compensation may easily realise that blows started from submaximal lung inflation will result in low PEF values, and I do not see how this can be detected even electronically in unsupervised measurements without recordings of forced vital capacity. Apart from these considerations, the assumption that accurate validation of flow meters is unnecessary for monitoring purposes is patently wrong for simple physical reasons. Even if the ability to measure true values in serial measurements is not crucial, the assessment of the dynamic response of the flow meter needs to be accurate. What Dr Connolly calls consistency is, according to the International Standards Organization,5 a combination of repeatability and reproducibility. Testing of repeatability is based on percentage and absolute differences between sequential measurements of known flows generated by suitable waveforms. Therefore, accuracy is necessary. In this context, it must be remembered that an inadequate dynamic response of a flow meter may affect measurements depending on the frequency content of the input signal, thus making any comparison of serial measurements impossible. Reproducibility is the ability to measure the same flow with different devices. Accuracy of measurements is therefore necessary if a given patient does not use the same peak flow meter for his/her whole life, which is not unlikely to be the case.

The second argument put forward by Dr Connolly is that comparison of PEF and FEV1 may be useful to confirm the diagnosis of minimal COPD, but no reference is given. I do not question the usefulness of looking at PEF on flow-volume curves to evaluate the quality of manoeuvres (my editorial was on the use of peak flow meters and not on spirometry) but, even in this case, accurate assessment of the dynamic response of the measuring device is imperative. If the system does not measure PEF and flows at lower lung volumes with the same accuracy, any inference from the shape of the flow-volume curve or derived parameters would be wrong and totally useless.

Finally, although Dr Connolly asserts that it is accepted that absolute values of PEF are rarely useful, the most recent guidelines on asthma management6 maintain that the severity of the disease can be classified based on PEF as percentage predicted. I agree that this may be inadequate, but we should acknowledge that a patient with PEF constantly below 200 l/min should be approached differently from a patient with PEF constantly above 500 l/min, even if the percentage variability of serial measurements is the same. As pointed out by Miller et al,2 an underdamped flow meter would give readings much greater than the true ones.

In conclusion, an accurate test of the dynamic characteristics of peak flow meters is imperative if their use is to be recommended. Even so, the usefulness of PEF measurements in asthma may be limited for reasons not related to instrument accuracy, which is what I tried to explain in my editorial.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.