Article Text

Download PDFPDF

PEF versus FEV1
  1. Eynsham
  2. Witney
  3. Oxford, UK
  1. D S POSTMA,
  1. Department of Pulmonary Diseases
  2. University of Groningen
  3. The Netherlands
  4. Department of General Practice
  5. Leiden University Medical Center
  6. 2301 CB Leiden
  7. The Netherlands
  1. Dr H A Thiadens

Statistics from

The assertion by Dr Thiadens and colleagues1that identification of airflow limitation and estimation of its reversibility by a bronchodilator is less reliable when measured by peak expiratory flow (PEF) than by forced expiratory volume in one second (FEV1) cannot be allowed to go unchallenged. They measured both values with a Microlab 3300 turbine spirometer, disregarding the fact that, in primary care, PEF is almost always measured by peak flow meters of variable orifice type which employ an entirely different principle and give considerably higher values. Jones and Mullee,2 who compared a similar Microlab turbine spirometer with a mini-Wright meter, found that values of PEF measured by the latter were, on average, 87 l/min higher. Hence, the values reported by Thiadens et al would have been much higher if they had been measured with a peak flow meter.

To compare the reliability of PEF and FEV1 for estimating magnitude of airflow limitation, Thiadens et al expressed observed values of each as percentage predicted, using the reference values for each sex recommended by the European Respiratory Society (ERS).3 Those for predicting PEF were derived from regression equations which describe a linear fall with age and give predicted values much lower than curvilinear regressions such as those of Nunn and Gregg,4 which an ERS Working Party on PEF5 subsequently judged to be the most satisfactory reference values for prediction. The difference in l/min between predicted values derived from the latter and those from the ERS regressions was roughly equal in each sex (fig 1) to the difference between the turbine measured observed values and the values which would have been obtained with a peak flow meter. Since they are of opposite direction, they obscure the spuriously low absolute values measured by the turbine spirometer. Nevertheless, Thiadens et al considered that, in 19.2% of their patients, low values of PEF were associated with normal values of FEV1 whereas in only 3.3% of patients was a normal value of PEF associated with an abnormally low value of FEV1.

Figure 1

Linear regressions of peak expiratory flow with age in men and women of height 175 cm and 160 cm, respectively, of the ERS3 compared with curvilinear regressions of Nunn and Gregg.4

To evaluate the relative merits of FEV1 and PEF as indices of bronchodilator reversibility Thiadens et al compared changes in FEV1, expressed as percentage differences in predicted values, with changes in PEF expressed as percentage differences in absolute values. The dissimilar manner in which the values were expressed invalidates their comparison and, hence, any conclusions drawn from it; it also makes the authors' prolix discussion of the sensitivity and specificity of their findings wholly irrelevant.

There is no justification for stating that a cut off value for a rise in FEV1 after a bronchodilator is “useful and valid . . . in separating asthma from COPD” since a bronchodilator reveals only immediate reversibility. Moreover, true irreversibility does not necessarily signify COPD since it may be present in patients with longstanding asthma in whom structural damage of the bronchi has occurred. The most convenient and reliable test in primary care for distinguishing between potentially reversible and truly irreversible airflow limitation is still twice or thrice daily monitoring of PEF during a course of corticosteroid treatment.6

The conclusion by Thiadens et al that FEV1 is more reliable than PEF for assessment of airflow limitation and its reversibility is not supported by their findings. Although I am very reluctant to criticise their study, attention needs to be drawn to its faults because the prominence given to their study by Thorax is likely to persuade general practitioners that its findings are valid and its conclusions are authoritative.


authors' reply Dr Gregg's remarks on our paper concern four major points: (1) the differences between the Micro medical spirometer and the mini-Wright peak flow meter; (2) the choice of predicted values; (3) the “dissimilar manner” in which changes in PEF were compared with changes in FEV1; and (4) the use of changes in FEV1 rather than changes in the FEV1/FVC ratio.

(1) It is true that there are differences between the two devices based upon the different principles—turbine flow measurement and variable orifice peak flow measurement. As we stated in the discussion part of the paper, the turbine flow meter yields slightly lower values than a pneumotachograph1-1 1-2 and the variable orifice peak flow meter shows significantly higher values (200–300 l/min) than a pneumotachograph in the mid region.1-3 In both cases the pneumotachograph value is considered the reference value. Without pertaining to the brand of the portable spirometer used, it appears that the turbine values generally meet the criteria for monitoring devices set by the ATS.1-4In any case, the differences between the devices are systemic and should not interfere with the results of the study, provided that the same device is used throughout it.

(2) In the discussion this issue has also been questioned. The choice of the predicted values depends in part on the equipment used. The ERS predicted values for PEF are obtained from a mixture of pneumotachograph data and Wright peak flow data whereas the values proposed by Dr Gregg are obtained from mini-Wright peak flow meters. The values produced by the turbine spirometer come closest to the pneumotachograph values. In view of this, the ERS values are probably the best choice.

(3) This issue was referred to in the discussion of our paper. We did not only compare the changes in FEV1 expressed as percentage differences of predicted values (with changes in PEF expressed as percentage differences in absolute values); we also compared changes in FEV1 as percentages of the initial values (including absolute improvement of 200 ml) with changes in PEF, both absolute and percentage, to the initial values (see table 3). We agree that measuring longitudinal reversibility with a corticosteroid is the best method, although we prefer to use the FEV1value at the start and the FEV1 after some weeks of corticosteroid treatment as outcome parameters. This issue was also discussed in the editorial by Professor Jones.1-5

(4) Generally, the FEV1/FVC ratio is a reliable indicator of bronchial obstruction provided the manoeuvre is carried out correctly. This is a problem with hand held spirometers, and the recommendations are that the expiratory curve be followed in real time to ensure that a true beginning and end of the forced expiration is detected. This is not possible with hand held spirometers and inevitably leads to falsely low FVC values. In our opinion, therefore, it is wise to exclude this parameter from analysis.

Although Dr Gregg is very definitive in his opinion about the value of peak flow measurements, especially using the mini-Wright meter, he has not been able to convince us, nor has he referred to validity studies about the accuracy of this device in demonstrating (reversible) airflow limitation.


  1. 1-1.
  2. 1-2.
  3. 1-3.
  4. 1-4.
  5. 1-5.

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.