factors · factors as atopy and smoking. these associa-tions are important since the value of pef...

Thorax 1993;48:899-905

Factors affecting peak expiratory flow variabilityand bronchial reactivity in a random populationsample

Bernard G Higgins, John R Britton, Susan Chinn, Kam Kwong Lai, Peter G J Burney,Anne E Tattersfield

Respiratory MedicineUnit, City Hospital,NottinghamB G HigginsJ R BrittonA E TattersfieldDepartment ofPublicHealth Medicine,UMDS, St Thomas'Campus, LondonS ChinnKKLaiP G J BurneyReprint requests to:Dr B G Higgins,Freeman Hospital,Newcastle upon TyneNE7 7DN, UKReceived 19 October 1992Returned to authors6 January 1993Revised version received13 May 1993Accepted 17 June 1993

AbstractBackground-Bronchial reactivity mea-surements are widely used in epidemio-logical studies to provide an objectivemarker of asthma. There are, however,several potential advantages of measur-

ing peak expiratory flow (PEF) variabil-ity instead, particularly in large studies.PEF variability and bronchial reactivitywere compared in a population sample toassess the relationships of the two mea-

surements to factors known to be associ-ated with airways disease, and tocompare their response rates.Methods-Subjects aged 18-65 were ran-

domly selected from the electoral regis-ter of an administrative area in easternEngland and randomised to attend eitherfor a bronchial challenge test measuringthe provocative dose of methacholineproducing a 20% fall in FEV, (PD20), or tomeasure PEF at two hourly intervalsduring waking hours for one week. Skintests with common allergens were per-formed and a smoking history obtained.PEF variability was expressed as theamplitude % mean (highest - lowest x100/mean).Results-A total of 273 subjects (69%)collected a PEF meter but a completedrecord sheet was returned by only 247(62%); this was still significantly more

than the 202 subjects (54%) who attendedfor and successfully completed a chal-lenge test. Amplitude % mean was higherin women than in men (9.7% v 8.5%). Inmultiple regression analysis amplitude %mean increased significantly with age,mean skin weal diameter, and with cur-rent smoking. The odds of having a PD20below 24 5,umol increased with mean

skin weal diameter and were greater incurrent smokers. Neither age nor sex hada significant effect on bronchial reac-itvity but there were significant inter-actions between age and the effects ofboth smoking and atopy.Conclusions-The higher response rateassociated with the use ofPEF variabilitymeasurement, and the association withfactors implicated in the pathogenesis ofairways disease, suggest that PEF vari-ability would be a useful measurement toemploy in epidemiological studies.

(Thorax 1993;48:899-905)

The inclusion of an objective measurement ofbronchial reactivity in epidemiological investi-gations of asthma overcomes some of theproblems of identifying asthma in the com-munity and enables comparisons to be madebetween studies performed in different coun-tries. It has become clear, however, thatbronchial reactivity does not have as precise arelationship to the clinical diagnosis ofasthma as was originally thought.' Thereare, in addition, several practical problemsassociated with the measurement of bronchialreactivity in the community, including theinvasive nature of challenge tests, the neces-sity of having a doctor present during the test,and the difficulties of obtaining a conven-tional measurement of reactivity such as theprovocative dose of methacholine producing a20% fall in FEV, (PD20) in most subjects in arandom population sample when dosage ofbronchoconstrictor agents is limited by sideeffects.We have explored the use of serial peak

expiratory flow (PEF) recordings to try to cir-cumvent some of these problems. In additionto the practical benefits, this method has theadvantage of measuring directly the diurnalvariability of airway calibre which is a cardinalfeature of asthma. We have shown that col-lection of PEF recordings from previouslyuntrained subjects is feasible, and that thedata can be analysed to provide a numericalindex of PEF variability which shows theexpected relationship to the diagnosis ofasthma.6 Although these findings suggest thatthe use of PEF variability might be an alter-native to measurement of bronchial reactivityfor community surveys of asthma prevalence,there are several unanswered questions. Ourearlier study was performed in a small ran-dom sample together with a group of subjectsselected because of respiratory symptoms,and did not permit an adequate considerationof the association of PEF variability with suchfactors as atopy and smoking. These associa-tions are important since the value of PEFvariability as an epidemiological tool wouldbe questionable if it showed no relationshipto factors implicated in the development andclinical diagnosis of airways disease.Furthermore, the possibility that measure-ment of PEF variability would be moreacceptable to subjects in a community set-ting, and thereby achieve higher responserates, has not been explored.We have therefore measured PEF

899 on O

ctober 22, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.48.9.899 on 1 Septem

ber 1993. Dow

nloaded from

http://thorax.bmj.com/

Higgins, Britton, Chinn, Lai, Burney, Tattersfield

variability and bronchial reactivity in a ran-

dom sample of the population aged 18-65from a semirural area in eastern England.Our aims were to compare the response ratesof the two methods -in conditions whichimposed the kind of logistic difficulties thatwould be encountered in a major epidemio-logical study, to define the relationshipbetween PEF variability and atopy and smok-ing in a random population sample, and tocompare this with the relationship betweenbronchial reactivity and the same factors.

MethodsSUBJECTSThe study was conducted in South Kesteven,an electoral district in Lincolnshire coveringan area of 364 square miles. The populationof 99 653 (established 1984) is dividedbetween three towns and numerous villages.A random start systematic sample of 1100names was drawn from the electoral registerof the area and randomised into two groups,one to undergo bronchial challenge testingwith methacholine and the other to keep a

record of serial PEF measurements.Subjects were sent a letter explaining the

purpose of the study and asking them to keepan appointment for a challenge test or to col-lect a peak flow meter. A prepaid envelopeand a reply slip were enclosed with the initialletter and the subjects were asked to informus whether or not they would participate. Theletter explained that we only wished to testthose aged 65 or under, and older subjectswere asked to indicate on the reply slip if theywould not attend for reasons of age. A secondletter containing the same information was

sent to all non-responders three weeks afterthe first approach. Tests were conducted atseveral health centres throughout the area toreduce the travel required and, if necessary,

subjects were visited at home. Attempts weremade to contact all subjects who did notreply to either letter by visiting their addressand, if possible, the test was rearranged atthis time.

Approval for the study was obtained fromthe ethics committees of Nottingham CityHospital, South Lincolnshire HealthAuthority, and the South LincolnshireGeneral Practitioner Committee.

QUESTIONNAIREInformation on respiratory symptoms, smok-ing history, age, and sex was obtained fromall subjects using the bronchial symptoms

Table 1 Response in the sample drawn from electoralregister

Challenge test group PEF group

Ineligible 170 158Refused 145 103Moved within area: 9 6not tracedNot accounted for 11 12Attended for test 212 273Total 547 552

Table 2 Distribution of sex, age, atopy, and smokinghistory among the subjects with complete data for allparameters

Challenge test group PEF group

SexMen 85 (50%) 104 (48%)Women 86 (50%) 114 (52%)

Age median (range) 39 (18-64) 41 (18-64)Atopy

Atopic* 60 (35%) 80 (37%)Non-atopic* 111 (65%) 138 (63%)

SmokingNon-smoker 75 (44%) 107 (49%)Ex-smoker 33 (19%) 42 (19%)Current smoker 63 (37%) 69 (32%)

*Atopic any skin weal > 2 mm greater than saline weal;non-atopic-no skin weal >2 mm greater than salineweal.

questionnaire of the International UnionAgainst Tuberculosis (IUAT). This was com-pleted before performing bronchial challengetests or starting peak flow recordings.

BRONCHIAL CHALLENGE TESTSThe initial letter to the subjects included arequest to abstain from cigarettes and bron-chodilators for six hours before attending thetest centre. On arrival the subjects restedwhile informed consent was obtained, thequestionnaire completed, and skin tests per-formed. Recordings of forced expiratoryvolume in one second (FEVy) and forced vitalcapacity (FVC) were then made with adry bellows spirometer (Vitalograph,Buckingham, UK) until two successive FEVymeasurements were within 5% of each other,the highest of these being taken as the base-line reading. Subjects were excluded fromchallenge testing if their baseline FEVy wasless than 60% predicted, if the FEVI/FVCratio was less than 50%, if they had recentlyexperienced a serious illness, or if they werepregnant.

Bronchial challenge tests were performedby the method of Yan et all using metha-choline inhaled from De Vilbiss No 40 nebu-lisers which had all been shown by priortesting to have an output in the range00025-0-0035 ml per activation. After com-pleting baseline spirometry subjects inhalednormal saline followed by increasing concen-trations of methacholine with measurementof FEVy one minute after each dose, alwaysrecording the highest of two readings within5% of each other. Quadrupling increments ofmethacholine from a starting dose of 0-096,umol were then given until the FEV1 fell by10-19% when the test continued with dou-bling increments. For subjects with a historysuggestive of asthma the starting dose was0-048 ,pmol and doubling increments wereused throughout. The test ended when theFEV1 had fallen by 20% or more, or a cumu-lative dose of 12-25 ,umol methacholine hadbeen given.

PEAK FLOW RECORDINGSOn arrival subjects completed the question-naire and skin tests were performed. Subjectswere then shown how to use a mini Wright

900

on October 22, 2020 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.48.9.899 on 1 S

eptember 1993. D

ownloaded from


Factors affecting peak expiratoryflow variability and bronchial reactivity in a random population sample

Table 3 Relationship between PEF variability and skin weal diameter, age, sex, and smoking history. Regressioncoefficients for amplitude % mean and absolute amplitude are for logO transformed values; coefficients for mean PEFvalues are untransformed

Mean (SE) multple regression coefficients

Amplitude % mean Absolute amplitude Mean PEF

Non-atopic non-smoking male aged 18 0-781 (0 058) 1-616 (0 050) 657-0 (19-7)Addition per mm increase in mean skin weal diameter 0-026 (0-01 1)** 0-019 (0009)* -8-5 (3.7)*Addition per year of age 0-002 (0-001)* t - 2-2 (0Q4)***Difference if:

Female 0-084 (0-033)** -0 039 (0 027) - 137-4 (10-8)***Ex-smoker -0-052 (0 041)1* -0-035 (0-035) 18-6 (13-9)Current smoker 0-059 (0-036)f 0-054 (0-031) -10-3 (12-3)

*p < 0 05; **p< 0-01; ***p< 0-001; tcoefficient negligible.

peak flow meter and how to fill in a purposedesigned record sheet with spaces for record-ings at two hourly intervals throughout theday, commencing at 02-00 hours. They wereasked to record their PEF within 15 minutesof the times shown every day for a week dur-ing waking hours. It was emphasised thatspaces should be left blank rather than insertinaccurate data if a measurement was missed.

Subjects were asked to return the meterand completed record sheet by post and weregiven a strong stamped addressed envelopefor this purpose. If the meter was notreturned by post subjects were visited athome.

SK1N TESTSSkin prick tests were performed on the ven-tral aspect of the forearm using Dermato-phagoides pteronnysinus, grass pollen, and catdander as test reagents with histamine andsaline as controls. Skin weal size was mea-sured after 10 minutes as the mean of thelargest diameter and that at right angles to-it,excluding any pseudopodial outgrowths.

ANALYSIS OF RESULTSThe results of the challenge tests wereanalysed by a curve fitting method8 to deter-mine the methacholine PD20. Extrapolation ofthe curve to one doubling dose above themaximum dose administered-that is, to 24-5umol-was carried out if the FEV, had not

Table 4 Effects ofskin weal size, age, sex, and smoking history on the odds ofhaving aPD,0 methacholine <24 5,umol. Coefficients are for log, (odds). Significance levels basedon difference in deviance compared with x2 distribution. Age coefficients for current andex-smokers are for difference compared with non-smokers

Factor Estimate (SE)

Non-atopic non-smoking male -3-46 (1-68)aged 18

Addition per mm increase in mean 1-48 (0.52)**skin weal diameter

Difference if:Female -0J13 (0 43)Ex-smoker 5-67 (3 00),**Current smoker 4-53 (1-70)f

Addition with age (per year): 0-06 (0-03)in ex-smokers -0-15 (0 08)

, Smoking/age interactionin current smokers -0-08 (0-04)J

Interaction between age and skin weal -0403 (0-01)*diameter

*p < 005; **p< 0-01.

fallen by 20% after the maximum dose ofmethacholine.

Peak flow records were first scrutinised todetect possible falsification of results. Thevariability of PEF in each subject was thendetermined as amplitude % mean6:

highest PEF reading - lowest 100mean

To calculate the response rates, subjectsaged over 65, those who had left the electoraldistrict, and those who had died since theelectoral register was compiled were excludedfrom analysis. Subjects who were ill or couldnot be traced were counted as non-respon-ders. Numbers responding for each surveymethod were compared by the x2 test.

Smoking history was determined from twoquestions, one asking if subjects had eversmoked on a daily basis for at least a year,and the second if they had smoked at all inthe last month. Subjects answering bothquestions affirmatively were classified as cur-rent smokers, those answering both negativelyas non-smokers, and those who had had nocigarettes in the last month but had smokedfor a year or more in the past were classifiedas ex-smokers.The mean skin weal diameter of each sub-

ject was calculated by subtracting the salinecontrol value from each of the house dust,grass, and cat weal diameters, and taking themean of these three values.The relationship between amplitude %

mean and age, sex, mean skin weal diameter,and smoking history was determined by mul-tiple linear regression with amplitude % meanas the dependent variable. Sex and smokingwere entered as categorical factors using threelevels of smoking representing current, ex-smokers, and non-smokers. Identical analyseswere performed with the two components ofamplitude % mean, amplitude and meanPEF as dependent variables. A similar analy-sis was performed for the results of bronchialchallenge tests but in this case multiplelogistic regression was used with reactor sta-tus as the categorical dependent variable, areactor being defined as a subject with a PD20value below 24-55umol. The interactionsbetween age and both atopy and smokingwere tested for significance. The effect ofcontrolling for baseline FEV1 on the relation-ship between bronchial reactivity and

901



j.com/


eptember 1993. D

ownloaded from



16r

14K

12k

101

8

6

<25 25-34 35-44 45-54 >54

Age group (years)

Figure 1 Mean (95% confidence intervals) amplitude % mean by mean skin weal sizeand age group. Atopic subjects (O) are those with one or more skin weals > 2 mm; non-

atopic subjects (U) are those with all skin weals < 2 mm.

1V

only 247 (62 3%) could be analysed.Challenge tests could not be performed in 10cases because of low FEV, (n = 8) or poorspirometric technique (n = 2) leaving 202(53 6%) completed tests. The differencebetween response rates (62-3% v 53 6%) issignificant (difference = 8-7%; 95% confi-dence interval = 8-1%; p < 0 05).

Incomplete questionnaires, skin testrefusal, or unsatisfactory skin tests excludeddata from a further 29 subjects from thefull regression analysis for PEF variabilityand 31 subjects from the full analysis ofbronchial reactivity. The age, sex, and smok-ing characteristics of the subjects included inthe analysis are given with their atopic statusin table 2.

COMPLETENESS OF PEF DATAFor each available two hourly time point atotal of 1729 PEF readings was possible-that is, seven days in each of 247 subjects. Atall times from 08-00 to 22-00 hours inclusivethe number of measurements recorded wasbetween 72% and 80% of the possible total.The percentage fell to 21% at 06-00 and 15%at midnight. Few measurements were madeat 02-00 or 04 00 hours.

C)

o

U)

0

a)C..

U)

(a

<25 25-34 35-44 45-54

Age group (years)

Figure 2 Percentage of reactors (PD20 <24-5 jmol) with 95% confidence limits b)and atopic status. Atopic subjects (a) are those with one or more skin weals > 2 mnon-atopic subjects (U) are those with all skin weals < 2 mm. The atopic subject gi>54 years is not shown since it contains only three subjects.

smoking was determined. Log,0 transfivalues of amplitude % mean and PD,1used throughout. The regression anwere performed using the statistical prGLIM.

ResultsRESPONSE RATES

After exclusion of subjects aged over 6those who had moved outside the ardied since the electoral register wasstructed, 377 subjects were eligiblebronchial challenge test and 394 for mement of PEF variability. A total of 21,jects (56 2%) attended for challengeand 273 subjects (69 3%) accepted ameter (table 1). However, 26 subjectsto return an adequate PEF record s(

RELATIONSHIP OF PEF VARIABILITY TO AGE,SEX, ATOPY, AND SMOKING HISTORYAmplitude % mean was higher in women

than in men (9 7% v 8-5%, p < 0 05) andthis was the only factor which had a signifi-cant effect on PEF variability when consid-ered alone. In multiple regression analysis,after controlling for sex, amplitude % mean

T showed a significant increase with increasingage, was lower in ex-smokers and higher incurrent smokers than in non-smokers, andincreased with increasing mean skin wealdiameter (table 3). There was no interaction

>54 between age and smoking nor between age

and atopy (fig 1).The differences in amplitude % mean seen

y age with smoking and atopy were produced by

zmOup changes in both absolute amplitude and mean'roupPEF. Absolute amplitude increased with skinweal diameter and was higher in currentsmokers: mean PEF showed the conversechanges (table 3). In contrast, both absolute

ormed amplitude and mean PEF were lower inO were women and mean PEF decreased with age;

ialyses the changes in mean PEF with age and sexogram were greater than those in absolute ampli-

tude, accounting for the increase in ampli-tude % mean with age and female sex.

RELATIONSHIP OF BRONCHIAL REACTIVITY TOAGE, SEX, ATOPY, AND SMOKING HISTORY

i5 and A total of 45 subjects (22 3%) had a PD20rea or methacholine < 24-5 ,umol. The odds werecon- significantly greater in current smokers and

for a with increasing skin weal diameter (table 4).asure- Allowance for baseline FEV, made little dif-2 sub- ference to the effect of smoking. Age and sextests, did not significantly alter the odds of having aPEF PD20 below 24-5 ,umol. There was, however,

failed a significant interaction between age ando that atopy (as defined by their product) with a

aEo0-

a)E._IQ)

T

902

------....

............

...... ............

...... ............

...... ............

...... ............

..... ........... ......



j.com/


eptember 1993. D

ownloaded from



100k

ClO

o

a)

01)

0'

0)

40~

T

80k

60

40

2

25-34 35-44 45-54 >54Age group (years)

Figure 3 Percentage of reactors (PD20 <24 5 mol), with 95% confidence limits, by ageand smoking history. Current smokers (X) and non-smokers (X) are shown; there werefew ex-smokers in the younger age groups and they are omitted.

decline in the effects of atopy with increasingage (table 4, fig 2), and between smoking andage, the increased reactivity in current smok-ers being less pronounced with age (table 4,fig 3).

DiscussionThe associations of bronchial reactivity withsmoking and atopy have been reported previ-ously, but this is the first time that the rela-tions of these factors to PEF variability havebeen studied in randomly selected subjects.Although bronchial reactivity and PEF vari-ability are both measures of airway lability,they are obtained in different ways and arelikely to reflect different aspects of airwaybehaviour. The results from this study in ran-

domly selected subjects, however, show thatboth measurements vary with factors knownto be associated with airways disease. Thestrength of the associations of the two mea-surements with these factors cannot be com-pared since the measurements were not madein identical subject groups.

Amplitude % mean was positively associ-ated with mean skin weal diameter, theincrease with increasing weal diameter beingproduced by complementary changes inabsolute amplitude (increase) and mean PEF(decrease). The difference is present in all agegroups except those aged > 54 years (fig 2).Atopic status is one of the major risk factorsfor the development of asthma and any pro-posed measure of airway lability would beexpected to show a relationship to atopy.Such an association might also be expectedon the basis of the changes which occur whenallergen challenge is carried out under con-

trolled laboratory conditions9-"1: airway cali-bre falls as an immediate response to allergen

challenge followed several hours later by anincrease in bronchial reactivity. Natural expo-sure of atopic asthmatic subjects to allergenwould thus be expected to lower PEF read-ings directly and to increase the susceptibilityto bronchoconstrict when exposed to otherstimuli, leading to increased PEF variability.The odds of having a methacholine PD20

below 24-5 ,imol also increased as skin wealdiameter increased. Several previous surveyshave shown a relationship between atopicstatus and bronchial hyperreactivity.12-'7Seasonal variation in bronchial reactivity hasbeen found in wheat workers18 and in a com-munity population selected because of occa-sional wheeze,'9 and these changes areassumed to be related to seasonal changes inallergen levels. In the laboratory allergenchallenge can produce an increase inbronchial reactivity9-1 and the airwayresponse to allergen can be predicted from aprior knowledge of the histamine PD20.20-23We also found, as in a previous study,'3 aninteraction between age and atopy with a sig-nificant reduction in the proportion of atopicsubjects with measurable PD20 values withincreasing age.

Current smokers had higher levels ofamplitude % mean than non-smokers or ex-smokers. The difference between smokingcategories was slightly more pronounced inthe younger age groups but this trend was notstatistically significant. The increased valuesin young current smokers suggest that theeffect of cigarettes on the airways can occurafter relatively short exposure. As with atopythe changes in amplitude % mean with smok-ing history were contributed to by an increasein the absolute amplitude and a decrease inmean PEF.Most studies which have considered the

effects of cigarette smoking on bronchialreactivity in community samples have shown,as in the present study, an increased pro-portion of smokers among subjects with ameasurable PD20 value."32428 In some suchstudies the effects of smoking have been seenmainly in subjects with low baseline airwaycalibre,2728 a finding consistent with thetheory that long term smoking altersbronchial reactivity through an effect on air-way calibre. There is also evidence of shortterm effects of smoking on airway calibre andbronchial reactivity,2930 and in our study theincreased proportion of current smokers witha measurable PD20 was present, as withamplitude % mean, in all age groups and was,in fact, more obvious in the younger agegroups as shown in fig 3. The figure mayexaggerate the extent of the increase in theyoungest age group which contained onlyseven smokers, but the results suggest thatsmoking can alter bronchial reactivity afterrelatively short exposure when changes in air-way calibre would be expected to be minimal.The effect of smoking on bronchial reactivityremained after controlling for baseline FEV1,again favouring the suggestion that smokingmay have a more direct influence on reactivity.

Previous reports which have considered the

903 on O



/T


ber 1993. Dow

nloaded from



effects of age on bronchial reactivity havegenerally, like the present study, been basedon cross sectional surveys. Studies whichhave included both children and adults haveshown a fall in the proportion of reactors tomethacholine and cold air with increasingage," 32 the lowest levels being seen in sub-jects in their forties. In adults there appears tobe an increase in the number of reactors asage increases above 40-50 years,33-36 althoughone recent study has shown the opposite.37 Incombination these results tend to suggest aU-shaped distribution with reactivity greatestin early and late adult life. In contrast toPD20, amplitude % mean showed a smallsteady increase with age (fig 2). This appearsto be due principally to the well recogniseddecline in mean PEF with age, since absoluteamplitude was relatively constant across theage range studied. Although the age relatedvariation in amplitude % mean was signifi-cant it was not large, the expected value foramplitude % mean in a non-smoking non-atopic man changing from 6&6% at age 20 to8-5% at age 60.

There was no difference in PD20 valuesbetween men and women, and no consistentassociation with sex has been reported inadults.5'37 Both mean PEF and absoluteamplitude were lower in women than in men,but the difference was proportionately largerfor mean PEF so that amplitude % mean wassignificantly greater in women. In our samplemore women than men had a history ofasthma (9-7% v 6&1%) which may accountfor some of the difference in PEF variabilitybetween the sexes, but amplitude % meanwas also higher in women in the subjects whoreported no asthma. In practice the sex differ-ence could be allowed for in studies usingamplitude % mean measurements as is cur-rently done for other indices of lung functionsuch as FEVI.One of the main reasons for performing

this study was to compare response rates forthe two survey methods. In terms of the num-ber of people attending for a challenge test orto collect a PEF meter there was a distinctlybetter response with PEF recordings. Somesubjects failed to return their PEF record,however, and some were unable to completechallenge tests, so that the final differencewas 8-7%. Of the 26 subjects who accepted aPEF meter but failed to return a completedrecord 23 claimed to have posted it, and it isdifficult to disbelieve them all. Some pack-ages were returned in a damaged state and itis possible that some were lost in the postalsystem. Even allowing for these losses PEFrecordings were associated with a higherresponse rate than challenge tests and this islikely to be the case in future studies, at leastwhen conducted in adults in countries withhigh literacy rates.

In addition to the difference in responserates the use of PEF recordings was muchsimpler logistically than challenge tests,although this is difficult to quantify formally.PEF meters can be distributed more quicklythan challenge tests can be performed, even

by experienced personnel. Appointment sys-tems for PEF recordings are easier to organ-ise and several subjects can be instructedtogether. If subjects need to be visited athome it is much easier to deliver a PEF meterand give instructions than to perform a chal-lenge test, and the PEF method can be usedby non-medically trained personnel.

There are limitations to PEF measure-ments. Some researchers may not feel com-fortable with the lack of supervision themethod entails. The need to correct for ageand sex adds a minor complication. Perhapsmost important is the finding from a previousstudy, in subjects selected because of a his-tory of wheeze, that subjects who had beengiven a diagnosis of asthma were separatedmore clearly from non-asthmatic subjects byuse of bronchial reactivity than by PEF vari-ability measurements.'8 However, there waslittle difference between the measurementswhen the relationships to respiratory symp-toms were compared.

In summary, we have confirmed in a ran-dom population sample that bronchial reac-tivity measurements are associated with atopyand smoking, and have shown that these fac-tors are also related to increased PEF vari-ability. The effects of atopy and smoking onbronchial reactivity show interactions withage which we did not find for PEF variabilitybut, despite this difference, our findings sug-gest that measurement of PEF variability is asuitable alternative to bronchial challengetests for epidemiological studies. In particu-lar, it may be more suitable for studies involv-ing repeated measurements in which itsgreater acceptability and absence of censoreddata would be important. A measure of PEFvariability can be obtained in all subjects in apopulation sample who accept a peak flowmeter; it appears to produce a higherresponse rate than bronchial challenge testsand is associated with factors known to berelated to airways disease.

We would like to thank Mrs S Cooper and Ms R Millard forassistance with bronchial challenge testing and the distribu-tion of PEF meters; Dr S Walsh and Mrs Y Himsworth ofSouth Lincolnshire Health Authority for help in setting up thestudy; the many general practitioners in South Kesteven whosupported the work; and the British Lung Foundation forfinancial support (grant 86/18). Miss S Chinn and Mr KWLai were funded by the Department of Health.

1 Britton J, Tattersfield AE. Does measurement ofbronchial hyperreactivity help in the clinical diagnosis ofasthma? Eury Respir Dis 1986;68:233-8.

2 Sears MR, Jones DJ, Holdaway MA, Hewitt CJ, FlanneryEM, Herbison GP, et al. Prevalence of bronchial reac-tivity to inhaled methacholine in New Zealand school-children. Thorax 1986;41:283-9.

3 Adelroth E, Hargreave FE, Ramsdale EH. Do physiciansneed objective measurements to diagnose asthma? AmRev Respir Dis 1986;134:704-7.

4 Salome CM, Peat JK, Britton WJ, Woolcock AJ.Bronchial hyperresponsiveness in two populations ofAustralian schoolchildren. 1. Relations to respiratorysymptoms and diagnosed asthma. Clin Allergy 1987;17:271-81.

5 Pattemore PK, Asher MI, Harrison AC, Mitchell EA, ReaHH, Stewart A. Ethnic differences in prevalence ofasthma symptoms and bronchial hyperresponsiveness inNew Zealand schoolchildren. 7horax 1989;44:168-76.

6 Higgins BG, Britton JR, Chinn S, Jones T, Jenkinson D,Burney PGJ, et al. The distribution of peak expiratoryflow variability in a population sample. Am Rev RespirDis 1989;140:1368-72.

904



j.com/


eptember 1993. D

ownloaded from



7 Yan K, Salome C, Woolcock AJ. Rapid method formeasurement of bronchial responsiveness. Thorax 1983;38:760-5.

8 Chinn S, Britton JR, Burney PGJ, Tattersfield AE,Papacosta AO. Estimation and repeatability of theresponse to inhaled histamine in a community survey.Thorax 1987;42:45-52.

9 Cockcroft DW, Ruffin RE, Dolovich J, Hargreave FE.Allergen-induced increase in non-allergic bronchialreactivity. Clin Allergy 1977;7:503-13.

10 Cartier A, Thomson NE, Frith PA, Roberts R, HargreaveFE. Allergen-induced increase in bronchial hyperre-sponsiveness to histamine: relationship to the late asth-matic response and change in airway calibre. J AllergyClin Immunol 1982;70:170-7.

11 Durham SR, Craddock CF, Cookson WO, Benson MK.Increases in airway responsiveness to histamine precedeallergen-induced late asthmatic response. Jf Allergy ClinImmunol 1988;82:764-70.

12 Cockcroft DW, Murdock KY, Berscheid BA. Relationshipbetween atopy and bronchial responsiveness to hista-mine in a random population. Ann Allergy 1984;53:26-9.

13 Burney PGJ, Britton JR, Chinn S, Tattersfield AE,Papacosta AO, Kelsen MC, et al. Descriptive epidemiol-ogy of bronchial reactivity in an adult population:results from a community survey. Thorax 1987;42:38-44.

14 Peat JK, Britton WJ, Salome CM, Woolcock AJ.Bronchial hyperresponsiveness in two populations ofAustralian schoolchildren. II. Relative importance ofassociated factors. Clin Allergy 1987;17:283-90.

15 Trigg CJ, Bennett JB, Tooley M, Sibbald B, De SouzaMF, Davies RJ. A general practice based survey ofbronchial hyperresponsiveness and its relation to symp-toms, sex, age, atopy and smoking. Thorax 1990;45:866-72.

16 Peat JK, Woolcock AJ. Sensitivity to common allergens:relation to respiratory symptoms and bronchialhyperresponsiveness in children from three differentclimatic areas of Australia. Clin Exp Allergy1991;21:573-81.

17 Clough JB, Williams JD, Holgate ST. Effect of atopy onthe natural history symptoms, peak expiratory flow, andbronchial responsiveness in 7 and 8-year-old childrenwith cough and wheeze. A 12 month longitudinal study.Am Rev Respir Dis 1991;143:755-60.

18 Hensley MJ, Scicchitano R, Saunders NA, Cripps AW,Ruhno J, Sutherland D, et al. Seasonal variation in non-specific bronchial reactivity: a study in wheat workerswith a history of wheat associated asthma. Thorax 1988;42:103-7.

19 Britton J, Chinn S, Burney P, Papacosta AO, TattersfieldAE. Seasonal variation in bronchial reactivity in a com-munity population. J Allergy Clin Immunol 1988;82:134-9.

20 Kreukniet J, Pijper MM. Response to inhaled histamineand to inhaled allergen in atopic subjects. Respiration1973;30:345-59.

21 Killian D, Cockcroft DW, Hargreave FE, Dolovich J.Factors in allergen induced asthma: relevance of theintensity of the airways allergic reaction and non-specific bronchial reactivity. Clin Allergy 1976;6:219-25.

22 Bryant DH, Burns MW. Bronchial histamine reactivity: itsrelationship to the reactivity of the bronchi to allergens.Clin Allergy 1976;6:523-32.

23 Cockcroft DW, Murdock KY, Kirby J, Hargreave FE.Prediction of airway responsiveness to allergen fromskin sensitivity to allergen and airway responsivness tohistamine. Am Rev Respir Dis 1987;135:264-7.

24 Welty C, Weiss ST, Tager IB, Munoz A, Becker C,Speizer FE, et al. The relationship of airways respon-siveness to cold air, cigarette smoking, and atopy to res-piratory symptoms and puimonary function in adults.Am Rev Respir Dis 1984;130:198-203.

25 Taylor RG, Joyce H, Gross E, Holland F, Pride NP.Bronchial reactivity to inhaled histamine and annualrate of decline in FEV, in male smokers and ex-smokers. Thorax 1985;40:9-16.

26 Woolcock AJ, Peat JK, Salome CM, Yan K, AndersonSD, Schoeffel RE, et al. Prevalence of bronchial respon-siveness and asthma in a rural adult population. Thorax1987;42:361-8.

27 Rijcken B, Schouten JP, Weiss ST, Speizer FE, Van derLende R. The relationship between airway responsive-ness to histamine and pulmonary function level in arandom population sample. Am Rev Respir Dis 1988;137:826-32.

28 Lim TK, Taylor RG, Watson A, Joyce H, Pride NP.Changes in bronchial responsiveness to inhaled hista-mine over four years in middle-aged male smokers andex-smokers. 77wrax 1988;43:599-604.

29 Gerrard J, Cockcroft D, Mink J, Cotton D, Poonwala R,Dosman J. Increased non-specific bronchial reactivity incigarette smokers with normal lung function. Am RevRespir Dis 1980;122:577-81.

30 Cerveri I, Bruschi C, Zoia MC, Maccarini L, Grassi M,Lebowitz MD, et al. Smoking habit and bronchial reac-tivity in normal subjects: a population based study. AmRev Respir Dis 1989;140:191-6.

31 Weiss ST, Tager IB, Weiss JW, Munoz A, Speizer FE,Ingram RH. Airway responsiveness in a population sam-ple of adults and children. Am Rev Respir Dis 1984;129:898-902.

32 Townley RG, Bewtra A, Wilson AF, Hopp RJ, ElsteinRC, Nair N, et al. Segregation analysis of bronchialresponse to methacholine inhalation challenge in fami-lies with and without asthma. J Allergy Clin Immunol1986;77: 101-7.

33 Alanko K. Prevalence of asthma in a Finnish rural popula-tion. ScandJRespir Dis 1970;76(Suppl):1-64.

34 Van der Lende R, Visser BF, Wever-Hesse J, De Vries K,Orie NGM. Distribution of histamine threshold valuesin a random population. Rev Inst Hyg Mines 1973;28:186-90.

35 Lang DM, Hopp RJ, Bewtra AK, Nair NM, Watt GD,Townley RG. Distribution of methacholine inhalationchallenge responses in a selected adult population.JAlkergy Clin Immunol 1987;79:533-40.

36 Rijcken B, Schouten JP, Weiss ST, Speizer FE, Vander Lende R. The relationship of non-specificbronchial responsiveness to respiratory symptoms in arandom population sample. Am Rev Respir Dis 1987;136:62-8.

37 Bakke PS, Baste V, Gulsvik A. Bronchial responsivenessin a Norwegian community. Am Rev Respir Dis 1991;143:317-22.

38 Higgins BG, Britton JR, Chinn S, Jones TE, Burney PGJ,Tattersfield AE. A comparison of bronchial reactivityand peak expiratory flow variability measurements forepidemiological studies. Am Rev Respir Dis 1992;145:588-93.

905 on O



/T


ber 1993. Dow

nloaded from


factors · factors as atopy and smoking. these associa-tions are important since the value of pef...

Documents