The Relationship Between FEV 1 and Peak Expiratory Flow in Patients With Airways Obstruction Is Poor*

Transcription

1 Original Research PULMONARY FUNCTION TESTING The Relationship Between FEV 1 and Peak Expiratory Flow in Patients With Airways Obstruction Is Poor* Ashutosh N. Aggarwal, MD, FCCP; Dheeraj Gupta, MD, FCCP; and Surinder K. Jindal, MD, FCCP Study objectives: To evaluate the correlation between FEV 1 and peak expiratory flow (PEF) values expressed as a percentage of their predicted value, and to assess factors influencing differences between the two measurements. Design: Cross-sectional. Setting: Pulmonary function laboratory at a tertiary-level teaching hospital in northern India. Participants: A total of 6,167 adult patients showing obstructive pattern on spirometry over a 6-year period. Interventions: None. Measurements and results: There was considerable variability between percentage of predicted FEV 1 (FEV 1 %) and percentage of predicted PEF (PEF%). Locally weighted least-square modeling revealed that PEF% overestimated FEV 1 % in patients with less severe obstruction and underestimated it in those with more severe obstruction. Using Bland-Altman analysis, PEF% underestimated FEV 1 % by a mean of only 0.7%; however, limits of agreement were wide ( 27.4 to 28.8%), indicating that these two measurements cannot be used interchangeably. PEF% and FEV 1 % were > 5% apart in approximately three fourths and differed by > 10% in approximately one half of the patients. On multivariate analysis, discordance > 5% was significantly influenced by female gender (odds ratio, 1.26; 95% confidence interval [CI], 1.01 to 1.58) and increasing FEV 1 % (odds ratio, 1.09 for every 10% increase; 95% CI, 1.04 to 1.14) but not by height or age. Conclusions: FEV 1 % and PEF% are not equivalent in many patients, especially women and those with less severe airflow limitation. Assumptions of parity between PEF% and FEV 1 % must be avoided. (CHEST 2006; 130: ) Key words: agreement; airflow limitation; respiratory function tests Abbreviations: CI confidence interval; FEV 1 % percentage of predicted FEV 1 ; PEF peak expiratory flow; PEF% percentage of predicted peak expiratory flow; VC vital capacity *From the Department of Pulmonary Medicine, Postgraduate Institute of Medical Education and Research, Chandigarh, India. The authors have no conflicts of interest to disclose. Manuscript received November 16, 2005; revision accepted April 13, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Surinder K. Jindal, MD, FCCP, Professor and Head, Department of Pulmonary Medicine, Postgraduate Institute of Medical Education and Research, Chandigarh , India; skjindal@indiachest.org DOI: /chest Spirometry is the recommended investigation for diagnosis and categorization of severity of airflow limitation. 1 Spirometry is a well-standardized technique, and elaborate guidelines already exist regarding procedure performance, evaluation of test quality, and interpretation of measured parameters. 1 3 However, spirometry is not widely available, and the pitfalls of spirometry frequently limit use of this test at the primary care level. 4,5 Peak expiratory flow (PEF) recording is proposed as an alternative to spirometry for this purpose. 6 8 The PEF instrument is cheap, portable, and easy to operate and maintain. Guidelines 9 11 on asthma management focus heavily on categorizing patients based on severity of airflow limitation measured on formal pulmonary function testing. It is suggested that either FEV 1 or PEF can be expressed as a percentage of predicted 1454 Original Research

2 values and used for this purpose. Similarly, definition and severity assessment of COPD is now based on measurement of percentage of predicted FEV 1 (FEV 1 %) and FEV 1 /vital capacity (VC), although a need for evaluating the role of PEF in situations and areas where spirometry is not routinely available is recognized. 12,13 There is, however, no consensus on whether or not FEV 1 % and percentage of predicted PEF (PEF%) can be used interchangeably in patients with obstructive lung diseases. Most clinicians assume a general parity between these measurements, and some guidelines 9,11 on asthma management also recommend the same. However, other guidelines 11,12 also suggest that PEF% may underestimate the degree of airways obstruction assessed by FEV 1 %. Previous studies 8,14 19 addressing comparisons between FEV 1 % and PEF% have been performed in highly selected patients and have been limited to some extent by inclusion of small number of subjects and inability to examine relationships in different subgroups of patients. We therefore studied adult patients with obstructive ventilatory defects from a large spirometry database to evaluate the correlation between FEV 1 % and PEF%, and to assess the factors influencing differences between the two measurements. North Indian adults. 20,21 The regression equations for spirometric indexes were generated from studies performed on 962 healthy nonsmoking North Indian adults, aged 15 to 74 years, using a water seal spirometer. The regression equations for PEF were derived from measurements performed on 3,166 healthy nonsmoking North Indian adults, aged 20 years, using a Wright peak flowmeter. Both these sets of regression equations predict lung function parameters using formulae containing gender, height, and age as variables (Table 1). A spirometry record with FEV 1 /VC ratio less than its predicted lower limit of normal was categorized as having an obstructive defect. The lower limit of normal for FEV 1 /VC ratio was computed by subtracting times the SE of estimate of the corresponding equation from the predicted value. To minimize analysis bias related to multiple records showing obstructive defects in the same individual because of repeated pulmonary function testing, only the first record was retained for further analysis in such situations. Wherever available, postbronchodilator values were chosen for interpretation of spirometry. Observed values for both FEV 1 and PEF were expressed as percentages of predicted values. A simple scatterplot was constructed to assess how closely PEF% approximated FEV 1 %. Iterative locally weighted smoothing techniques were applied to this distribution to obtain a regression curve fitting this data; such a method avoids the undue influence of outlying values on curve estimation by assigning a relatively lesser weight to these values. 22 In order to evaluate if PEF and FEV 1 could be used interchangeably across different categories of age, height, gender, and severity of airway obstruction, we calculated the limits of agreement between the two estimates using Bland-Altman analysis. 23 Multiple logistic regression analysis was also performed to study factors responsible for FEV 1 % and PEF% values being 5% apart from each other. Materials and Methods Our laboratory provides facilities for spirometry and other detailed investigations (such as estimation of static lung volumes, diffusion testing, and study of pulmonary and airway mechanics). Patients are referred for lung function testing not only from our department but also from other medical and surgical specialties of our hospital. Since our institute is a tertiary academic center in northern India, patients are referred for diagnosis and management from several northern Indian states. Records of all consecutive adult patients (aged 15 years) undergoing spirometry during a 6-year period were retrieved. Sources of referral, reasons for performing spirometry, as well as other clinical details were, however, not analyzed further. All subjects performed spirometry on a dry rolling seal spirometer (Spiroflow; P K Morgan Ltd; Kent, UK), followed by PEF estimation using a Wright peak flowmeter. VC, FEV 1, and PEF were measured using American Thoracic Society guidelines, and the highest measurements from among three technically acceptable and reproducible maneuvers were expressed at body temperature and pressure saturated with water vapor. 3 All technicians performing spirometry were experienced in pulmonary function testing and strictly followed standard procedures while analyzing the kymograph output. The spirometer was frequently calibrated to ensure optimum performance. Since the study was a retrospective analysis of already available laboratory records, and patient confidentiality was not breached in any fashion, prior approval from our Hospital Ethics Committee was not required. Age, gender, height, and spirometric data were recorded for all patients using software previously developed by us. 20 The predicted values for VC, FEV 1, FEV 1 /VC ratio, and PEF were generated using previously defined prediction equations for Results During the study period, 25,914 spirometry procedures were performed. After excluding 775 individuals 15 years old and 302 incomplete records, 24,837 records were retained. Of these, 7,395 records (29.8%) were interpreted as having obstructive defect. After excluding 1,228 repeat procedures, 6,167 records formed the database for further analysis. There were 3,213 men (52.1%) aged 16 to 95 years (median, 53 years; interquartile range, 41 to 64 years) and 136 to 189 cm in height (median height, Table 1 Prediction Equations for Lung Function in Healthy North Indian Adults* Parameters Regression Equation Formula SEE Men FEV 1,L A A H FEV 1 /FVC A A H 6.6 PEF A 0.08 A H 55.0 Women FEV A A H FEV 1 /FVC A A H 5.8 PEF A 0.04 A H 51.0 *A age in years; H height in centimeters; SEE standard error of estimate. CHEST / 130 / 5/ NOVEMBER,

3 165 cm; interquartile range, 161 to 170 cm); and 2,954 women (47.9%) aged 16 to 94 years (median, 47 years; interquartile range, 36 to 60 years) and 136 to 189 cm in height (median height, 153 cm; interquartile range, 149 to 157 cm). PEF was reduced in 4,935 patients (80.0%). In general, there was a moderate correlation between FEV 1 % and PEF%, with overall Pearson correlation coefficients of (p 0.001) and (p 0.001) among men and women, respectively. However, the scatter was wide, more so in patients with mild obstruction (Fig 1). The regression curve generated using iterative locally weighted smoothing techniques showed two zones. In patients with FEV 1 % 40%, PEF% tended to underestimate FEV 1 %; in patients with more severe obstruction, PEF% tended to overestimate FEV 1 % (Fig 1). Agreement between FEV 1 % and PEF% was only fair, with weighted estimate of using 10% categories for both FEV 1 % and PEF%. Using arbitrary severity categories based on 20% FEV 1 %intervals, PEF% and FEV 1 % severity categories were concordant in only 3,013 instances (48.9%), with better concordance as severity of obstruction (based on FEV 1 %) became more severe (Table 2). The differences between FEV 1 % and PEF% followed a normal distribution (Fig 2); for the entire study population, PEF% underestimated FEV 1 % by a mean of only 0.7%. However, limits of agreement were wide and exceeded 25% (Fig 3). Overall, differences were more marked in women and in patients at extremes of height distribution (Table 3). PEF% and FEV 1 % were 5% apart in 4,574 patients (74.2%) and 10% apart in 3,161 patients (51.3%). Although there is no consensus on the issue, a discordance 5% could be considered a clinically important error for estimation of severity of airway obstruction. On univariate analysis, the proportion of such discordant results was significantly more in women, and increased significantly with worsening FEV 1 %, and decreasing height (Table 4). On multivariate analysis, gender and FEV 1 % independently influenced discordance (Table 4). Figure 1. Scatterplot showing relationship between FEV 1 and PEF readings referenced to their respective predicted values. The regression curve is fitted using iterative locally weighted smoothing techniques Original Research

4 Table 2 Concordance Between Categorization of Severity of Airway Obstruction Based on FEV 1 % and PEF%* FEV 1 % PEF% 0 to 40% 40 to 60% 60 to 80% 80 to 100% 100% 0 to 40% 897 (75.6) 521 (29.6) 117 (6.7) 13 (1.2) 40 to 60% 254 (21.4) 881 (50.1) 618 (35.2) 164 (14.9) 16 (4.4) 60 to 80% 33 (2.8) 296 (16.8) 744 (42.4) 415 (37.7) 95 (25.9) 80 to 100% 3 (0.3) 48 (2.7) 246 (14.0) 394 (35.8) 159 (43.3) 100% 12 (0.7) 30 (1.7) 114 (10.4) 97 (26.4) *Data are presented as No. (%). Discussion In patients with obstructive lung diseases, both FEV 1 % and PEF% are widely used to estimate the degree of pulmonary impairment. In general, FEV 1 measurements are preferred as these are much more reproducible. However, spirometry is not widely available in developing countries such as India, and there is a need to assess if similar information could be acquired using PEF measurements, which are cheaper and much more widely available. Since clinical decisions are often based on the results of these measurements, we attempted to compare their utility in defining severity of airflow obstruction. A few investigators 14 17,19 have looked at correlation between PEF and FEV 1 in cross-sectional studies. The correlation between absolute values of PEF and FEV 1 values was rather low in one study. 14 In general, the correlation between PEF% and FEV 1 % has been moderate, with correlation coefficients ranging from 0.5 to ,17,19 In our own database, there was a moderate correlation between FEV 1 % and PEF%. In one follow-up study, 15 individual correlation coefficients were found to range from 0.68 to However, a numerical summary of information does not imply that PEF can be used as a surrogate Figure 2. Histogram showing the near-normal distribution of differences between FEV 1 and PEF values. Both are described as a percentage of their respective predicted values, and the latter is subtracted from the former to compute the difference. A discrepancy of 10% between FEV 1 % and PEF% occurred in more than one half of instances. CHEST / 130 / 5/ NOVEMBER,

5 Figure 3. Bland-Altman plot highlighting magnitude of difference between FEV 1 and PEF values expressed as a percentage of their respective predicted values. Horizontal dashed lines represent the mean bias and its 95% confidence limits. for FEV 1. Scatterplots from various reports show considerable difference in FEV 1 % and PEF% values in individual patients, although most coordinates lie close to the line of identity. To avoid undue influence from outlying values, we used iterative locally weighted smoothing to obtain a regression curve fitting our data; and we found that in patients with severe airway obstruction (FEV 1 40% of predicted), PEF% overestimated FEV 1 %, whereas exactly the opposite happened in patients with less Table 3 Mean Bias and Limits of Agreement Between PEF% and FEV 1 % in the Study Population Men Women Total Variables Bias Limits of Agreement Bias Limits of Agreement Bias Limits of Agreement Age, yr to to to to to to to to to to to to to to to to to to 28.5 Height, cm to to to to to to to to to to to to to to 17.2 Severity of obstruction FEV 1 % 60% to to to 35.0 FEV 1 %40to60% to to to 28.7 FEV 1 % 40% to to to 19.4 Total to to to 28.8 *Bias mean of (FEV 1 % PEF%); Limits of agreement bias (1.96 SD of bias). For details on method of calculation of bias and limits of agreement, see Bland and Altman Original Research

6 Table 4 Results of Logistic Regression Analysis To Study Factors Responsible for Patients Having FEV 1 % and PEF% Values > 5% Apart From Each Other Univariate Analysis Multivariate Analysis Variables Odds Ratio 95% CI Odds Ratio 95% CI Women 1.4* Height (10-cm intervals) 0.9* Age (10-yr intervals) FEV 1 % (10% intervals) 1.1* * *p p severe airway obstruction. Previous investigators 14 have also noted similar trends. In one study, 18 PEF% values were higher than corresponding FEV 1 % values, particularly in patients with moderate-to-severe asthma. In another study, 16 most patients could generate higher PEF% values than FEV 1 % values, although very few patients had severe airway obstruction. Overall, PEF% and FEV 1 % were 5% apart in 74.2% and 10% apart in 51.3% of patients in our data set. Thus, only a minority of patients had PEF% and FEV 1 % values close to each other (Fig 2). Data from previous studies 8,17,18,24,25 also show that limits of agreement are wide and point toward absence of parity between FEV 1 % and PEF% values. In our own data set, limits of agreement were 27.4 to This means that for a given value of PEF%, corresponding FEV 1 % could be 28.8% lower or 27.4% higher. These values render substitution of PEF% for FEV 1 % useless in routine clinical practice. Our results are at slight variance with observations in other reports. The mean difference between FEV 1 % and PEF% in this study was only 0.7%. Previous studies 8,17,18 have shown a much wider mean difference, with FEV 1 % being lower than PEF% by 9.1 to 17.2%. This possibly is related to the selection and size of study populations enrolled in these studies. Most of these studies 8,17,18 included small number of patients (25 to 101 patients). Some studies 8,15,18 recorded multiple paired observations on each subject and analyzed each pair as a separate unit. Both factors preclude generalization of results. Some studies 8,15,16,18 also included patients who, although suffering from an obstructive lung disease, did not have airflow limitation at time of evaluation. As discussed earlier, PEF% in these subjects is likely to be much greater than FEV 1 % in comparison to patients having more severe airway obstruction at time of evaluation, and it is difficult to generalize these results while categorizing severity of airway obstruction. There could be several reasons for lack of equivalence between FEV 1 % and PEF%. 26 For one, measured PEF values depend heavily on lung volumes. Any disease process leading to reduced lung volumes will effect a corresponding reduction in measured PEF. This implies that in addition to patients with airway obstruction, those with restrictive lung defects are also likely to have a reduced PEF. Secondly, normal population variability of PEF is quite large. Hence calculation of lower limits of predicted normal based on regression equations leads to values that are much lower than corresponding values for other spirometric indexes like FEV 1. Thirdly, while PEF is measured on the first effortdependent portion of the forced expiratory maneuver and predominantly reflects large airway function, FEV 1 is determined both by the effort-dependent and effort-independent portions of this maneuver and reflects both large and peripheral airway function. 27 Thus differential changes in FEV 1 and PEF may be observed, depending on the amount and predominant site of airways narrowing. These factors are likely to lead to a greater discrepancy in patients with COPD and airway collapsibility secondary to the loss of elastic tissue. In these patients, the initial rapid rise in expiratory flow is similar but, as intrathoracic pressure increases, that pressure is transmitted to the segmental and other large airways, which collapse and obstruct passage of air through those airways. This results in the rapid reduction in flow after a relatively normal peak has been attained, leading to significantly lower values of FEV 1 compared to PEF. These issues could lead to a significant discordance if FEV 1 % values are replaced by PEF% values for purpose of severity classification. As it is, there is no clearly defined objective strategy to categorize severity of airflow limitation based on FEV 1 % values. Various guidelines on management of asthma and COPD use arbitrary standards for this purpose, adding to the confusion. The FEV 1 % cutoffs used to categorize mild, moderate, and severe obstruction are variable for Global Initiative for Chronic Obstructive Lung Disease (80% and 30%), British CHEST / 130 / 5/ NOVEMBER,

7 Thoracic Society (60% and 40%), American Thoracic Society (50% and 35%), European Respiratory Society (70% and 50%) and new American Thoracic Society/European Respiratory Society (80% and 50%) guidelines on COPD. 12,13,28 30 Both Global Initiative for Asthma and National Institutes of Health guidelines on asthma use FEV 1 % cutoffs of 80% and 60%. 9,11 In the absence of any standard severity classification, we used arbitrary severity groups based on FEV 1 % values and have shown that less than one half the subjects were correctly classified if PEF% replaces FEV 1 % values (Table 2). In addition, there are other technical issues related to equipment. Several PEF meters do not show linear responses, with different proportional error at different flow levels. 31 Significant decrease in accuracy and precision has also been reported after regular peak flowmeter use. 32 Submaximal effort during PEF maneuver, supramaximal flow transients occurring early during a forced expiration, and PEF maneuver-induced bronchospasm are phenomena that may account for some of the discrepancies between PEF and FEV 1 values. It is clear from these results that if international guidelines are followed and PEF% is used as a surrogate for FEV 1 %, then severity of obstruction may be wrongly categorized in a large proportion of patients. The impact of use of PEF on categorization of severity of stable asthma was reported in an earlier study, 18 in which the categorization using PEF% and FEV 1 % was concordant in only one half of the patients. Misclassification was particularly evident in patients with severe asthma. Significant differences were also reported in a study 24 of patients with exacerbation of asthma. These differences are likely to be more pronounced in women and in patients with less severe airways obstruction (defined by FEV 1 %), as is evident from the results of logistic regression analysis conducted in this study. This has far-reaching implications for developing and resource-poor countries where facilities for conducting spirometry are not freely available, and physicians rely mainly on clinical features and/or PEF estimation to assess severity of airflow limitation. While our data set comprising of a large number of spirometry records does provide results that can perhaps be interpreted with greater confidence than previous studies, we are aware of certain inadequacies. The major limitation of our work is its retrospective nature, because of which we are unable to provide results separately for patients with asthma and COPD, or provide exact data regarding acceptability/ repeatability goals of individual spirometry maneuvers. Moreover, our results remain strictly applicable to only the kind of patient population investigated at our pulmonary function laboratory and may not be generalized to other sets of patients. We therefore suggest that until such time that more data are available, assumptions of parity between PEF% and FEV 1 % must be avoided in assessment of patients with obstructive airway disease. In case spirometry is not available, PEF measurements can be performed for broad assessment, but patients should not be categorized as having mild, moderate, or severe obstruction based on these results. References 1 Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J 2005; 26: Miller MR, Crapo R, Hankinson J, et al. General considerations for lung function testing. Eur Respir J 2005; 26: American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995; 152: Eaton T, Withy S, Garrett JE, et al. Spirometry in primary care practice: the importance of quality assurance and the impact of spirometry workshops. Chest 1999; 116: Walters JA, Hansen E, Mudge P, et al. Barriers to the use of spirometry in general practice. Aust Fam Physician 2005; 34: Hansen EF, Vestbo J, Phanareth K, et al. Peak flow as predictor of overall mortality in asthma and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001; 163: Jackson H, Hubbard R. Detecting chronic obstructive pulmonary disease using peak flow rate: cross sectional survey. BMJ 2003; 327: Llewellin P, Sawyer G, Lewis S, et al. The relationship between FEV 1 and PEF in the assessment of the severity of airways obstruction. Respirology 2002; 7: National Heart Lung and Blood Institute. Guidelines for the diagnosis and management of asthma: expert panel report 2. Bethesda, MD: National Institute of Health, British guideline on the management of asthma. Thorax 2003; 58(suppl 1):i1 i94 11 National Institutes of Health, National Heart, Lung, and Blood Institute. Update of the NHLBI/WHO Workshop Report: global strategy for asthma management and prevention, issued January Bethesda, MD: National Institutes of Health, National Heart, Lung, and Blood Institute, 2002; document No BTS guidelines for the management of chronic obstructive pulmonary disease. The COPD Guidelines Group of the Standards of Care Committee of the BTS. Thorax 1997; 52(suppl 5):S1 S28 13 Global Initiative for Chronic Obstructive Lung Disease. Global strategy for the diagnosis, management and prevention of chronic obstructive lung disease: NHLBI/WHO workshop report. Bethesda, MD: National Heart, Lung, and Blood Institute, Harrison BD, Swarbrick ET. Peak-flow percentage in asthma [letter]. Lancet 1971; 2: Kelly CA, Gibson GJ. Relation between FEV 1 and peak expiratory flow in patients with chronic airflow obstruction. Thorax 1988; 43: Vaughan TR, Weber RW, Tipton WR, et al. Comparison of PEFR and FEV 1 in patients with varying degrees of airway 1460 Original Research

8 obstruction: effect of modest altitude. Chest 1989; 95: Teeter JG, Bleecker ER. Relationship between airway obstruction and respiratory symptoms in adult asthmatics. Chest 1998; 113: Sawyer G, Miles J, Lewis S, et al. Classification of asthma severity: should the international guidelines be changed? Clin Exp Allergy 1998; 28: Thiadens HA, De Bock GH, Van Houwelingen JC, et al. Can peak expiratory flow measurements reliably identify the presence of airway obstruction and bronchodilator response as assessed by FEV(1) in primary care patients presenting with a persistent cough? Thorax 1999; 54: Aggarwal AN, Gupta D, Jindal SK. Development of a simple computer program for spirometry interpretation. J Assoc Phys India 2002; 50: Jindal SK, Wahi PL. Pulmonary function laboratory in the tropics: needs, problems and solutions. In: Sharma OP, ed. Lung disease in the tropics. New York, NY: Marcel Dekker, 1991; Cleveland WS. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 1979; 74: Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: Choi IS, Koh YI, Lim H. Peak expiratory flow rate underestimates severity of airflow obstruction in acute asthma. Korean J Intern Med 2002; 17: Emerman CL, Cydulka RK. Use of peak expiratory flow rate in emergency department evaluation of acute exacerbation of chronic obstructive pulmonary disease. Ann Emerg Med 1996; 27: Paggiaro PL, Moscato G, Giannini D, et al. Relationship between peak expiratory flow (PEF) and FEV 1. Eur Respir J Suppl 1997; 24:39S 41S 27 Robinson DR, Chaudhary BA, Speir WA Jr. Expiratory flow limitation in large and small airways. Arch Intern Med 1984; 144: Siafakas NM, Vermeire P, Pride NB, et al. Optimal assessment and management of chronic obstructive pulmonary disease (COPD): The European Respiratory Society Task Force. Eur Respir J 1995; 8: American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995; 152:S77 S American Thoracic Society/European Respiratory Society Task Force. Standards for the diagnosis and management of patients with COPD, version 1.2. New York, NY: American Thoracic Society, 2004; updated Paggiaro PL, Moscato G, Giannini D. The Italian Working Group on the use of peak expiratory flow rate (PEFR) in asthma. Eur Respir Rev 1993; 314: Irvin CG, Martin RJ, Chinchilli VM, et al. Quality control of peak flow meters for multicenter clinical trials: the Asthma Clinical Research Network (ACRN). Am J Respir Crit Care Med 1997; 156: CHEST / 130 / 5/ NOVEMBER,