Peak Expiratory Flow Variability and Bronchial Responsiveness to Methacholine An Epidemiologic Study in 117 Workers1.2

FRANCOISE NEUKIRCH, RENATA LIARD, CLAIRE SEGALA, MYRIAM KOROBAEFF, CHRISTINE HENRY, and JACQUELINE COOREMAN

Introduction

Bronchial hyperresponsiveness, which is greater than normal variability of airwaycaliber in response to various stimuli, is a main physiologic feature of asthma. In population studies, it has been found to be associated not only with asthma but also with chronic bronchitis (1), smoking habits (2), and upper respiratory disorders (3). Moreover, it has been demonstrated to be a risk factor for the development of airway obstructive diseases (4-6). In epidemiologic studies, bronchial hyperresponsiveness has been investigated by measuring the reactivity to inhaled methacholine or histamine. Higgins and coworkers (7) recently suggested that studying airway lability by measuring the variability of peak expiratory flow (PEF) might provide an alternative to reactivity measurements in some surveys of respiratory morbidity. The limitations of bronchial provocation testing could thereby be avoided (8, 9). To date, serial measurements of PEF have been widely used in clinical management of asthma, and it has been suggested that excessive diurnal variation can be used to identify asthmatic subjects (10, 11). The relationship between reactivity of histamine or methacholine and airway lability have been investigated in clinical studies (12-16). Ryan and coworkers (12) observed a strong negative association between the levelof bronchial responsiveness to histamine and the diurnal variation of PEF in nine nonasthmatic and 32 asthmatic subjects. Bahous and coworkers (13)studied 27 subjects with mild symptoms of bronchial hyperexcitability: diurnal changes in PEF were significantly negatively correlated with the response to histamine. A recent longitudinal study of 20 asthmatic patients by Josephs and coworkers (14) confirmed these results. However, the relationship

SUMMARY We studied the relationships between peak expiratory flow (PEF) variability and bronchial responsiveness to methacholine In 117 workers attending the annual compulsory examination (mean age, 38.7 yr ± 9.5; men, 86.3%). SUbjects recorded their highest PEF out of three, every 3 waking hours (I.e., flva times s day) for 7 days, each using a neWly purchased Vltalograph peak flow meter, and underwent methacholine challenge tests with a maximal cumulatlva dose of 1,200 ~g. Those with a FEV, fall of 15% or more were considered as reactors. The variability of PEF was expressed as the amplitude percent mean, calculated from dally amplitude (highest-lowest readInglmean reading of the day x 100), avaraged over 6 days, from the second to the seventh. This Index had a continuous distribution, skewed towerds the greatest emplltudes, and correlated negatively with FEV, (r -0.25, P 0.01). SUbjects with asthma (n 8) had greater VIIrlations. In the 109 nonasthmatlcs, greater variability was observed In SUbjectswith whaeze apart from colds, breathlessness, or hay faver; the average amplitude was greater In reactors than In nonreactors to methachollna (16.9% versus 9.3%, p < 0.001). The subjects with excessive PEF variability were all methacholine reactors, but they were only a subgroup of the reactors. These results provide evidence thet excesslvaPEFvariability Is an Indlcetor of bronchial hyperresponslvaness to methacholine In a population sample. AM REV RESPIR DIS 199Z; 148:71-75

=

=

between responsiveness to challenge tests and PEF variability was based only on studies of small highly selected groups of patients attending hospital. The existence and significance of this relationship in general populations has not been rigorously investigated, although the distribution of PEF variability has been described in population samples (17, 18). This report describes a study of the relationship of PEF variability to methacholine reactivity in 117 workers, taking possible confounding factors into account. Methods Population The study was conducted in the industrial medical center of a detergent plant in June and July 1990.One hundred fifty blue-collar workers 22 to 58 yr of age, working in the production unit, were initially included in the study as part of the annual compulsory examination. Informed consent was obtained from all participants according to the regulations prevailing in France at that time, and subjects wereasked to avoid medications during data collection. Only 28 subjects were unwilling to perform the PEF protocol. With

=

regard to methacholine challenge, four subjects refused and one met exclusion criteria. The study population was, therefore, 117 subjects. The characteristics of the study group were similar to those of the 33 subjects for whom data were not available (table 1). A reference"normal" subgroup of the study population was selected.The inclusion criteria were: (1) negative answers to all questions asking whether the subject suffered from asthma, wheezeapart from colds, chronic wheeze, shortness of breath with wheeze, hay fever, chronic cough, chronic phlegm, history of chronic bronchitis, or emphysema; (2)absence of acute symptoms of upper and lower airways on the day of examination and of acute respiratory illness during the days of PEF measurements; (3) no use of bronchodilator inhalers. Heavy smokers (~ 20 cigarettes/day) werealso excluded from this "normal" population, which consisted of 53 subjects. (Received in original form July 12, 1991 and in revised form March 10, 1992) 1 From the National Institute of Health and Medical Research INSERM U226, Paris, France. 2 Correspondence and requests for reprints should be addressed to Dr. F. Neukirch, Epiderniologie, Faculte Xavier-Bichat, 16 rue Henri Huchard, 75018 Paris, France.

71

72

NEUKIRCH, LIARD, SEGALA, KOROBAEFF, HENRY, AND COOREMAN

TABLE 1 CHARACTERISTICS OF THE STUDY POPULATION'

Asthma Wheeze apart from colds Wheeze on most days Attacks of shortness of breath with wheeze Hay fever Breathlessness Chronic cough Chronic phlegm Sex = male Age, yr Mean SO Smoking Neversmokers Ex·smokers Current smokers" 20 cig/day Heavy smokers> 20 cig/day

With PEF Measurements and Methacholine Challenge Test (n = 117)

With Questionnaire Only (n = 33)

6.8 11.1

12.1 15.1

2.6

o

10.3 29.1 21.4 4.3 6.0 86.3

9.1 12.1t 33.3

o 3.0 90.9

38.7 9.5

35.2 9.4

38.5 21.4 34.2 6.0

39.4 18.2 30.3 12.1

• Values are percentages (except for ege). p _ 0.05, chi square test.

t

Data Collection Each subject was given a newly purchased peak flow meter (Vitalograph, Buckingham, UK) and was trained in its correct use by a trained technician. Each was instructed to record the highest value out of three measurements every 3 h (± 15 min) during waking hours, i.e., five times per day, for the 7 days after the examination. The first measurement had to be recorded on rising from bed. Measurements and their times were recorded on a special sheet that was given together with an instruction sheet. Subjects were asked to leave blank spaces rather than complete the sheets with false data. Subjects answered a modified British Medical Research Council/European Coal and Steel Community questionnaire on respiratory symptoms (19), with additional questions on symptoms on the day of examination (20). Asthma was defined as a positive answer to the question "Have you ever had asthma?" (cumulative prevalence). Breathlessness was defined as being troubled by shortness of breath when hurrying on levelground or walking up a slight hill. With regard to smoking habits, subjects wereclassified as neversmokers, ex-smokers, current smokers of less than 20 cigarettes/day, and heavy smokers (~ 20 cigarettes/day). Basic pulmonary function tests were conducted by trained spirometry technicians using a lung function analyzer (no. 3 Fleish pneumotachograph) with a built-in computer (Spiromatic). Subjects were tested in the sitting position and wearing a noseclip. Three technically acceptable measurements were made for each subject and converted to BTPS. The largest values werechosen in accordance with American Thoracic Society criteria (21).The

followingmeasurements wererecorded: FEV, and forced expiratory flow during the middle half of forced vital capacity (FEF2s- ,.) . Methacholine provocation testing was performed provided that a subject's baseline FEV, was ~ 70070 predicted. Aerosols were generated using a dosimeter nebulizer (Mediprom FDC88, Minato, Japan) giving an output of 0.04 ml in 0.8 s or of 0.08 ml in 1.6s. The concentration of methacholine was 1mg/ml. After a control inhalation of saline, doubling doses were administered in a series of eight steps, with a starting dose of 40 ug, or four steps, with a starting dose of 80 ug, according to whether or not subjects had a history of asthma and/or wheeze. The tests were continued until a total dose of 1,200 Ilg had been givenor the FEV, had fallen by 15%. Subjects with a FEV, fall of 15% or more were considered methacholine reactors. Skin prick tests wereperformed to extracts of common allergens, including cat, house dust, and a mixture of five grass pollens (Dome Hollister, Dome, Epernon, France), Dermatophagoides pteronyssinus (Stallergens), and a control saline solution. Atopy was defined as the presence of a wheal in response to at least one allergen,with a diameter both ~ 3 mm and 3 mm larger than the saline control wheal.

Data Analysis The SAS-PC statistical package was used for statistical analysis. The variability ofPEF was calculated as follows: daily amplitude = highest - lowest reading/mean reading of the day x 100. Amp % mean = dailyamplitude averaged over the 6 days from the second to the seventh, sincethe first day was considered as a learning day (see RESULTS).

This index could not be calculated for three subjects who had missed at least one of the five daily tests. The frequency distribution of the index was calculated for the study group as a whole and for the subgroup of "normal" subjects. When parametric statistical tests were used, the data were normalized by base 10 logarithmic conversion. Statistical analysis included chi-square tests, or Fisher's exact test when expected numbers were small, to evaluatethe associations of categoricalvariables. For continuous variables, t tests and linear regression analysis were used. Repeated measures analysis of variance, with days and hours as measurement factors, was used to test the consistency among the 6 days considered. Intraclass correlation coefficient (the ratio of between-subject variance to the sum of between-subject and within-subject variance) wascalculated for daily amplitude. Multiple regression analysis was also performed; p values ~ 0.05 were considered significant. Associations of borderline statistical significance (0.05 < P < 0.10)are indicated. For spirometric values, regressions on sex, age, and height were calculated on the basis of data from the entire study population. Normalized residuals were used for analysis. In RESULTS, the normalized residuals are reported as values for subjects of mean age and height.

Results

PEF Variability and Its Relationships with Other Factors The expected pattern of diurnal PEF was observed, with the lowest values corresponding to the first morning measurements, and the maximal values occurring at noon or in the afternoon. For example, in the subgroup of "normal" men, the mean value of the first morning readings averaged over the 7 days of the study was 541.7L/min (SD = 75.3), the mean value at noon was 565.0 (SD = 81.0), and the mean evening value was 563.3 (SD = 81.5).The same pattern was observed in those defined as asthmatics, with mean values lower than in "normal" subjects, especially in the morning. In the study population as a whole, the daily amplitude percent mean was the greatest on the first day (12.7%) and varied very little over the next 6 days (11.6, 11.4, 11.0, 10.5, 9.4, and 10.4070). Similar patterns were observed when different subgroups were considered: asthmatics and nonasthmatics, methacholine reactors and nonreactors, and "normal" subjects. The data from the first day were therefore discarded. The F-test testing for the variability between the 6 days from the second to the seventh was not significant, whereas this analysis confirmed the variability between hours (p = 0.0005). Intraclass correlation coefficient mea-

73

PEAK FLOW VARIABILITY AND BRONCHIAL RESPONSIVENESS TO METHACHOLINE

suring the day-to-day repeatability of PEF variability in the 6 days considered was 0.44. The distribution of Amp 11,10 mean, averaged over 6 days, in the whole study population and in "normal" subjects is shown in figure 1. Both distributions were positivelyskewed, although that for "normal" subjects was less skewed. The mean value for the study population as a whole was 10.7% (SD = 6.2070), and the median value was 9.9070. In the subgroup of "normal" subjects, the mean was 9.8070 (SD = 5.7070), and the median value was 8.6070.

Relationships with Other Factors PEF variability was greater in asthmaticsthan in nonasthmatics (Amp 070 mean = 14.4070 versus 10.4070), although the difference was of borderline statistical significance (p = 0.08). In the 109 nonasthmatics, no significant relationships wereobserved between Amp % mean and any of chronic cough, chronic phlegm, acute symptoms of upper or lower airways on the day of examination, wheezing, or atopy. Greater variability was observed in subjects with than without wheeze apart from colds (13.7070 versus 10.2070, p = 0.08), breathlessness (12.3070 versus 10.0070, p = 0.09), or hay fever (12.7070 versus 9.6070, p = 0.08). Amp 070 mean was not related to sex, but it correlated positively with age (r = 0.23, p = 0.02). Variability was greater in heavy smokers than in other subgroups, although the difference was not statistically significant (12.2070 versus 9.9% in nonsmokers, 10.6% in exsmokers, and 10.6070 in smokers of less than 20 cigarettes/day). Amp % mean was higher in symptomatic than in asymptomatic current smokers taken as a whole (12.8070 versus 9.2070, p = 0.006),

whereas this was not the case in nonsmokers. PEF variability correlated negatively with FEV dr = - 0.25, p = 0.01), and FEF zs- 7 S (r = -0.21, p = 0.03).

Reactivity to Methacholine and Its Relationships with Other Factors Nineteen ofthe 117 subjects tested reacted to methacholine (16.2070). The proportion was twice as great among asthmatics (37.5070) as among nonasthmatics (14.7%) (NS). Three of the eight asthmatics used medications regularly (beta2-agonists): only one of them could not avoid it prior to methacholine challenge, but reacted in spite of that. In the 109 nonasthmatics, no significant relationship was observed between reactivity to methacholine and any of the respiratory symptoms studied. Reactors were more frequent among women than among men (33.3070 versus 11.7070, p < 0.05). The mean age of reactors was older than that of nonreactors (44.3 ± 11.0 yr versus 37.9 ± 9.1 yr, p < 0.02). No relationship was observed between reactivity and smoking habits. Among atopic subjects, the proportion of reactors was 30.4070 versus 10.5070 in nonatopics (p = 0.01). FEV. and FEFzs- 7S , adjusted for sex, age, and height on the basis of the regression of the whole study population, were lowerin reactors than in nonreactors: 3.87 L/s versus 4.09, p < 0.10, for FEV.. and 3.63 Lis versus 4.08, p < 0.04, for FEFzs- 7S ' Relationships between PEF Variability and Reactivity to Methacholine Of the five subjects not tested for methacholine reactivity, only one had a FEV. < 70% predicted: his Amp 070 mean was 11.9070. Of the four that refused chal-

TABLE 2 MULTIPLE REGRESSION ANALYSIS WITH AMP % MEAN (Logl0) AS THE DEPENDENT VARIABLE (R' = 0.30) Variable Reactivity to methacholine FEV, Wheeze apart from colds Hay fever Age

Parameter Estimate 0.21 -0.05 0.15 0.09 0.004

p Value

< 0.001 0.03 0.03 0.07 0.10

lenge test, one was asthmatic, and his Amp 070 mean was 22.1 070, the percentages for the three others were 10.6, 12.0, and 14.2070, respectively. In the population studied, the PEF variability was greater in reactors than in nonreactors (17.6070 versus 9.3070, p < 0.001). Similar results were observed after exclusion of the eight asthmatics (16.9070 in reactors versus 9.3070 in nonreactors, p < 0.001). Multiple regression analysis with Amp 070 mean as the dependent variable, including age, methacholine reactivity, wheeze apart from colds, hay fever, asthma, breathlessness, and FEV., showedthat methacholine reactivity was a significant risk factor of PEF variability, independently of the other factors (table 2). In the study population as a whole, as well as in the nonasthmatics (figure 2), the distribution of the PEF variability of reactors and nonreactors to methacholine, showed that, despite the overlap between the two groups, those with the greatest PEF variability werereactors. Tocalculate the specificity and sensitivity of Amp 070 mean compared with methacholine reactivity, wefirst used the cutoff point for normal PEF variability identified by the 95% percentile in the

subjecls 40

%

'Y. aubj_cla 40 35

35

30

30

25

25

20

20

15

15

10

10

5

5 0 0·4

4·8

8·12

12·16 16·20 20·24 24·28 28·32 32·36 AMPLITUDE % MEAN

0·4

4·8

8·12

12·16 16·20 20·24 24-28 28·32 32·36 AMPLITUDE % MEAN

Fig. 1. Distribution of 6·day PEF variability expressed as Amp % mean in the stUdy population (left pane~ and in the subgroup of "normal" (right pane~ subjects.

74

NEUKIRCH, LIARD, SEGALA, KOROBAEFF, HENRY, AND COOREMAN

35 30 25

Fig. 2. Distribution of Amp % mean according to methacholine reactivity in nonasthmatics. Bars represent the percentage of the total nonasthmatics: the sum of all the bars is 100%. Open bars = PD15-; closed bars = PD15+.

20

15 10

0·4

4-8

8·12

12-16

16-20

20-24

24-28

28-32

32-36

AMPLITUDE % MEAN

afternoon, was consistent with the circadian rhythm observed by other investigators (10, 18). The PEF variability for each subject can be calculated and expressed using various indices (13, 14, 17, 18). We used Amp % mean, averaged over the days from the second to the seventh: the greater amplitude of the first day could be considered to be a learning effect since the amplitude of PEF variation wasvery consistent among the 6 other days. Other investigators discarded the first 2 days (18). The intraclass correlation coefficient we calculated (0.44) was slightly higher than the one observed by Higgins and coworkers (17) in their random sample, from the same index (0.40). The distribution of Amp % mean, which was positively skewed and was correctly normalized by base 10 logarithmic conversion, was also consistent with that in the population study of Higgins and coworkers. With regard to bronchoprovocation testing, only relatively low total dose of methacholine (1,200 IJ.g) could be administered because examinations were carried out in an office of occupational medicine and not in a hospital. For this

subgroup of "normal" subjects, which was 21.3070. Specificity was 100%, and sensitivity was 31.3070. In order to assess how sensitivity could be improved, we constructed a receiver-operating characteristic (ROC) curve; figure 3 shows that sensitivity increased sharply without heavy losses in specificity. Visual inspection of the ROC curve might suggest that 12070 variability is the best cutoff, considering both sensitivity and specificity. However, this value of PEF variability is in the range of mean ± SO, both in the study population and in "normal" subjects. Discussion In this population of 117 workers, the mean value of PEF variability expressed as Amp % mean was 10.7% (SO = 6.2%), and the median value was 9.9%. This variability was greater in asthmatics and in subjects with wheeze apart from colds or breathlessness. A strong association was observed between PEF variability and reactivity to methacholine. The diurnal pattern of PEF, with the lowest values occurring in the morning and the highest values at noon or in the

.",

100

•

••

10

12'4

•

•

10'1l0

• • • ••

~

t-

••

10

Fig. 3. Receiver-operatingcharacteristic (ROC) curve in nonasthmatics. The percentages (21, 12. 10, and 7%) are the cutoff values corresponding to the squares just below.

> tIII

Z III III

•

40

••

•

'" 20

O+----+----+---+--------'f-----+ 100

10 40 SPECIFICITY

20

o

reason we considered that a FEV I fall of 15% or more was satisfactory for classifying subjects as reactors. Although subjects are most often considered as reactors if they have a FEV I fall of at least 20%, some epidemiologists have used other indices such as PC to (22) or PC IS (23). Popa and Singleton (24) compared seven indices in 20 normal subjects and 20 asthmatic patients and concluded that the most discriminating were POlS and P0 20 • The results of a population study by Rijcken and coworkers (25) suggested that the difference between 10 and 20% decrease in FEV I is simply one doubling concentration. Bruschi and coworkers (26) studied different indices for bronchial reactivity in a general population and concluded that LnPO ISwas the best index to assess bronchial responsiveness in epidemiologic studies. The poor relationships of PEF variability and methacholine reactivity to asthma were probably due to our definition (cumulative prevalence). Moreover, the test of difference in Amp % mean between asthmatics and nonasthmatics had a statistical power of 65% only. The relationships of PEF variability and methacholine reactivity to other factors were consistent with those reported in the literature. Greater PEF variability was related to hay fever, breathlessness, and wheeze apart from colds, as previously observed by other investigators (7, 18). Methacholine reactivity was more frequent among women, as has been observed by ltigg and coworkers (27). The relationship between reactivity and age seems to be more complex: no association was observed in a population between 18 and 75 yr of age (27), but in another study the proportion of the population who were reactors had a Ll-shaped distribution by age, with higher proportions in the 18 to 24 and 55 to 64 yr age ranges than in the 35 to 44 yr age range (28). The subjects we studied were 22 to 58 yr of age (average, 38.7 yr; SO 9.5), and no relationship was observed. Respiratory symptoms were not related to reactivity in nonasthmatics, in agreement with Quackenboss and coworkers (18). In our study population, atopy among nonasthmatics was related to reactivity but not to PEF variability. The former result confirms previous studies with adults or children (27-29). The relationship between atopy and PEF in general populations is still under debate: in a longitudinal study of children, atopy and wheeze were both associated with a significantly greater PEF variation, and there was a significant interac-

PEAK FLOW VARIABILITY AND BRONCHIAL RESPONSIVENESS TO METHACHOLINE

tion (29). Other researchers found that the mean amplitude of subjects with some history of atopy without asthma was not significantly different from that of subjects with no history of atopy (10). Moreover, in a follow-up study of pollensensitized nonasthmatic subjects with seasonal rhinitis, carried out during and out of the pollen season, no significant variation in peak flow rates was observed throughout the study (30). The main result of the present study was the association between PEF variability and reactivity to methacholine, even after exclusion of asthmatics. This result confirms the relationships observed in clinical studies, mainly in asthmatic patients, and suggests that measures of PEF variability can be used in epidemiologic studies to identify subjects with airway lability. However, although the subjects with excessivePEF variability were all methacholine reactors, they wereonly a subgroup of the reactors. The cutoff point for excessive variability calculated from the 95th percentile in the "normal" subjects of our population (21.3010) identified subjects who werecertainly hyperresponsive. This cutoff was not very different either from the 20010 variation generally considered as the dividing line between asthmatics and nonasthmatics (10), or from the 19.1010 normal limit, based on the 95th percentile in the reference group, in a population study of subjects 35 to 65 yr of age (18). Such groups of people who are certainly hyperresponsive can be studied in epidemiology to identify risk factors. High sensitivity is obtained only with cutoff points that represent normal values, as shown in figure 3. This fact precludes the use of PEF variability as a clinical diagnostic test for bronchial hyperreactivity, but in epidemiology a cutoff corresponding to high sensitivity may be chosen to use PEF variability as a screening tool to limit the number of methacholine challenge tests. The association between response to methacholine and mean daily PEF variability has also been observed in a general practice-based survey (27), but the small number of subjects with known asthma were not excluded: the proportion of hyperresponsive subjects was23010, probably because of the tendency of symptomatic subjects to attend for tests in population studies. This bias was reduced in the present study, where the study group was taken from subjects attending the annual examination, which is compulsory for French workers: the prevalence of methacholine reactors observed

(16010) was similar to that observed in other general populations (28). Thus, this study showed that excessive PEF variability identified subjects who were certainly hyperresponsive. Further studies with larger populations are needed to compare such subjects to hyperresponsives without bronchial lability, who may have bronchial hyperreactivity related to factors other than asthmalike symptoms or allergy. If our results are confirmed, it appears that PEF variability, possibly measured over fewer days, is an objective and satisfactory measure of airway lability for epidemiologic surveys of asthma. References 1. Bahous J, Cartier A, Ouimet G, Pineau L, Malo JL. Non-allergic bronchial hyperexcitability in chronic bronchitis. Am Rev Respir Dis 1984; 129: 216-20. 2. Gerrard JW, Cockcroft DW, Mink JT, Cotton DJ, Poonawala R, Dosman JA. Increased nonspecific bronchial reactivity in cigarette smokers with normal lung function. Am Rev Respir Dis 1980; 122:577-81. 3. Empey DW, Laitinen LA, Jacob I, Gold WM, Nadel JA. Mechanisms of bronchial hyperreactivity in normal subjects after upper respiratory tract infections. Am Rev Respir Dis 1976; 113:131-9. 4. Sparrow D, O'Connor G, Colton T, Barry CL, WeissST. The relationship of nonspecific bronchial responsiveness to the occurrence of respiratory symptoms and decreased levelsof pulmonary function. The normative aging study. Am Rev Respir Dis 1987; 135:1255-60. 5. Taylor RG, Joyce H, Gross E, Holland F, Pride NB. Bronchial reactivity to inhaled histamine and annual rate of decline in FEV I in male smokers and ex-smokers. Thorax 1985; 40:9-16. 6. Annesi I, Neukirch F, Orvoen-Frija E, et at. The relevance of hyperresponsiveness but not atopy, to FEV. decline. Preliminary results in a working population. Bull Eur Physiopathol Respir 1987; 23:397-400. 7. Higgins BG, Britton JR, Chinn S, Jones TD, Burney PGJ, Tattersfield AE. Relationship of methacholine PD,. and peak flow variability measurements to respiratory symptoms in a community population. Am Rev Respir Dis 1988; 137:249. 8. Davies RJ, Blainey AD. Occupational asthma. In: Clark TJH, Godfred S, eds. Asthma. 2nd ed. London: Chapman & Hall, 1983; 205-13, 228-30. 9. Newman-Taylor AJ, Davies RJ. Inhalation challenge testing. In: Weill H, Thrner-Warwick M, eds. Oceupationallung diseases. New York: Marcel Dekker, 1981; 143-68. 10. Hetzel MR, Clark TJH. Comparison of normal and asthmatic circadian rhythms in peak expiratory flow rate. Thorax 1980; 35:732-8. 11. Venables KM, Burge PS, Davison AG, Newman-Taylor AJ. Peak flow rate records in surveys: reproducibility of observers' reports. Thorax 1984; 39:828-32. 12. Ryan G, Latimer KM, Dolovich J, Hargreave FE. Bronchial responsiveness to histamine: relationship to diurnal variation of peak flow rate, improvement after bronchodilator, and airway calibre. Thorax 1982; 37:423-9. 13. Bahous J, Cartier A, Malo JL. Monitoring of peak expiratory flow rates in subjects with mild airway hyperexcitability. Bull Eur Physiopathol

75 Respir 1985; 21:25-30. 14. Josephs LK, Gregg I, Mullee MA, Holgate ST. Nonspecific bronchial reactivity and its relationship to the clinical expression of asthma. A Iongitudinalstudy. Am Rev Respir Dis 1989;140:350-7. 15. Davies RJ, Morgan DJR, Blainey AD. Bronchial reactivity: its assessment and clinical value. In: Morley J, ed. Bronchial hyper reactivity. London: Academic Press, 1982; 187-207. 16. Ramsdale EH, Morris MM, Hargreave FE. Interpretation of the variability of peak flow rates in chronic bronchitis. Thorax 1986; 41:771-6. 17. Higgins BG, Britton JR, Chinn S, et at. The distribution of peak expiratory flow variability in a population sample. Am Rev Respir Dis 1989; 140:1368-72. 18. Quackenboss JJ, Lebowitz D, Krzyzanowski M. The normal range of diurnal changes in peak expiratory flow rates. Relationship to symptoms and respiratory disease. Am Rev Respir Dis 1991; 143:323-30. 19. Minette A, Brille D, Casula D, Van Der Lende R, Smidt U. Commentaires relatifs au questionnaire pour l'etude de la bronchite chronique et de l'emphyseme pulmonaire. Coli Med Hyg Travail CEE, 1972. 20. Neukirch F, Kauffmann F, Korobaeff M, Liard R. Common cold with cough on the day of examination: a factor that should be taken into account in epidemiological studies on pulmonary function. Int J Epidemiol 1985; 14:635-6. 21. Gardner RM. Snowbird workshop on standardization of spirometry. ATS statement. Am Rev Respir Dis 1979; 119:831-8. 22. Rijcken B, Schouten JP, Weiss ST, Speizer FE, Van Der Lende R. The relationship of nonspecific bronchial responsiveness to respiratory symptoms in a random population sample. Am Rev Respir Dis 1987; 135:62-8. 23. Kabiraj MU, Simonsson BG, Groth S, Bjorklund A, Bulow K, Lindell SE. Bronchial reactivity, smoking, and alpha-antitrypsin. A populationbased study of middle-aged men. Am Rev Respir Dis 1982; 126:864-9. 24. Popa V, Singleton J. Provocation dose and discriminant analysis in histamine bronchoprovocation. Are the current predictive data satisfactory? Chest 1988; 94:466-75. 25. Rijcken B, Schouten JP, Weiss ST, Meinesz AF, De Vries K, Van Der Lende R. The distribution of bronchial responsiveness to histamine in symptomatic and in asymptomatic subjects. A population-based analysis of various indices of responsiveness. Am Rev Respir Dis 1989;140:615-23. 26. Bruschi C, Cerveri I, Zola MC, Maccarini L, Grassi M, Rampulla C. Bronchial responsiveness to inhaled methacholine in epidemiological studies: comparison of different indices. Eur Respir J 1989; 2:630-6. 27. Thigg CJ, Bennett JB, Tooley M, Sibbald B, D'Souza MF, Davis RJ. A general practice based survey of bronchial hyperresponsiveness and its relation to symptoms, sex, age, atopy, and smoking. Thorax 1990; 45:866-72. 28. Burney PGJ, Britton JR, Chinn S, et at. Descriptive epidemiology of bronchial reactivity in an adult population: results from a community study. Thorax 1987; 42:38-44. 29. Clough JB, Williams JD, Holgate ST. Effect of atopy on the natural history of symptoms, peak expiratory flow, and bronchial responsiveness in 7- and 8-year-old children with cough and wheeze. Am Rev Respir Dis 1991; 143:755-60. 30. Boulet LP, Morin D, Milot J, Turcotte H. Bronchial responsiveness increases after seasonal antigen exposure in non-asthmatic subjects with pollen-induced rhinitis. Ann Allergy 1989;63:114-9.