Thorax 1993;48:722-729

Variability of bronchodilator response and effectsof inhaled corticosteroid treatment in obstructiveairways disease

Huib AM Kerstjens, Paul L P Brand, Philip H Quanjer,Brigitte A H A van der Bruggen-Bogaarts, Gerard H Koeter, Dirkje S Postma on behalf ofthe Dutch CNSLD Study Group

AbstractBackground-In the day to day care ofobstructive airways diseases (asthma andchronic obstructive pulmonary disease)inportant decisions such as diseaseclassification and choice of therapy arebased on assessment of the broncho-dilator response. However, surprisinglylittle is known of the long term course ofthe bronchodilator response in patientswith obstructive airways disease.Methods-Data from a multicentre trialwere used in which 274 patients aged18-60 years with airways obstructionwere selected with PC20 < 8 mglml andFEVy < 95% CI of predicted. FEV, wasmeasured before and 20 minutes after1000 ug terbutaline and 40 minutes afteran additonal 80 pg ipratropium bromide.Data were analysed from 185 patientswho were followed up for 21 months (fivemeasurements). Four different expres-sions of bronchodilator response (BDR)were examined for change under longterm therapy, long term variability, andprognostic value in predicting responseto inhaled corticosteroids.Results-There was a significant reduc-tion in BDR of 117 ml after three monthsof treatment with a /i agonist plus a cor-ticosteroid (BA + CS), but not afterbronchodilators only. Significant reduc-tions with BA + CS were also found inBDR as a percentage of initial FEV,, andin BDR as a percentage of predictedFEV,. Bronchodilator tests were quitevariable (SD 186 ml or 11% of initialvalue) and less than half of the patientscould consistently be classified as "irre-versible" with recommended cutofflevels. The bronchodilator response atthe start of the study proved to be a poorpredictor of improvement in FEVy underBA + CS treatment (correct prediction60%).Conclusions-Bronchodilator responsesdecrease substantially with inhaled corti-costeroid therapy, and within subjectvariability is considerable both in asthmaand chronic obstructive pulmonary dis-ease. Dichotomous decisions on whetherpatients are "irreversible" according toany single bronchodilator measurementshould therefore be made with great cau-tion. The bronchodilator response can-

not be used accurately as a predictor ofresponse to inhaled corticosteroids inobstructive airways disease.

(Thorax 1993;48:722-729)

Assessment of bronchodilator response(BDR) in patients with obstructive airwaysdisease is a routine procedure both in clinicalpractice and research. The results of the testare used to separate patients with asthmaboth from normal subjects and from thosewith chronic obstructive pulmonary disease(COPD) and to make therapeutic decisions.'There are, however, numerous problemsassociated with bronchodilator testing.The interpretation of any test requires

knowledge about its variability. Althoughdata are available for patients with COPD,'no such data exist on the variability ofbronchodilator testing in patients withasthma. In addition there is no consensus onhow to express the BDR. The most com-monly used expression-the response as apercentage of the initial value-has the disad-vantage that the response is highly dependenton the initial (baseline) value.3-5 There is alsono consensus on what constitutes a "signifi-cant" BDR.2-7 The American ThoracicSociety has recently issued new guidelinesrecommending that a "significant" BDRshould be both larger than a 12% increaseover baseline and larger than 200 ml.' To ourknowledge, no published data exist to sup-port this recommendation. Finally, the use ofthe BDR for therapeutic decisions has beenstudied in patients with COPD where alarger BDR predicted, to some extent, alarger improvement with oral cortico-steroids,9-1' but whether a higher BDR pre-dicts a more favourable reaction to inhaledcorticosteroid therapy in obstructive airwaysdisease is unknown.We have studied the BDR prospectively in

a large group of patients with obstructive air-ways disease (asthma and COPD) and abroad range of clinical presentations and lungfunction. They were randomly allocated toone of three different treatment regimens(inhalation of a single bronchodilator, twobronchodilators, or a bronchodilator and acorticosteroid), and followed up for 2 5years.'2 13 In this report we specifically addressthe following questions:

Department ofPulmonary Medicine,University HospitalGroningen,Oostersingel 59, 9713EZ Groningen,The NetherlandsH A M KerstjensP L B BrandG H KoeterD S PostmaDepartment ofPulmonary Medicine,University HospitalUtrecht,The NetherlandsB A H A van derBruggen-BogaartsDepartment ofPhysiology, Universityof Leiden,The NetherlandsP H QuanjerReprint requests to:Dr H A M KerstiensReceived 9 November 1992Returned to authors3 March 1993Revised version received14 April 1993Accepted 20 April 1993

722 group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Vanrability of bronchodilator response and effects of inhaled corticosteroid treatment in obstructive airways disease

1. Is the degree of BDR affected by mainte-nance treatment with inhaled broncho-dilators, or corticosteroids, or both? Howlarge is the long term variability of bron-chodilator tests? Does a significant treat-ment effect or a high variability, or both,have consequences for classification ofdisease as irreversible on the basis of asingle measurement?

2. Does a greater response to bronchodila-tors predict an improvement in airwaysobstruction during long term mainte-nance therapy with inhaled cortico-steroids (immediate v long termreversibility)?

MethodsPatients aged 18-60 years with respiratorysymptoms and no other major illnesses'2 wereselected from the respiratory outpatientdepartments according to the following crite-ria: (1) Forced expired volume in one second(FEVy) ranging between 4-5 and 1-64 resid-ual standard deviations (RSD) below thepredicted value-that is, 2-30-0-84 1 belowpredicted FEV1 for men or 1-71-0-62 1 belowpredicted FEV1 for women-or FEV1/IVC(inspiratory vital capacity) more than 1 64RSD (men 11-76%, women 10-68%) belowthe predicted value, provided that total lungcapacity was normal (higher than 1 64 RSD(men 1-15 1, woman 0-98 1) below the pre-dicted level). The rationale of using RSDsand predicted values has been detailed else-where.'415 FEV1 had to be larger than 1-2 1.(2) Concentration of histamine causing a20% fall in FEV1 (PC20) < 8 mg/ml.

Inhaled corticosteroids were tapered offand discontinued completely four weeksbefore a prerandomisation visit. Other main-tenance medication was withheld for at leastsix weeks (ketotifen, antihistamines), fourweeks (cromolyn sodium), or 48 hours (theo-phyllines) before the start of the study.Maintenance medication with oral steroidswas not allowed. Patients were randomly allo-cated to one of three double blind regimensusing identical metered dose inhalers: allpatients received an inhaled /2 agonist (terbu-taline 250 ug, two puffs four times daily)combined with either an inhaled cortico-steroid (beclomethasone 100 ,ug, two puffsfour times daily (BA + CS)), an inhaled anti-cholinergic (ipratropium bromide 20,ug, twopuffs four times daily (BA + AC)), or aninhaled placebo, two puffs four times daily(BA + PL). Additional bronchodilator medi-cation was supplied as salbutamol dry powderinhalations (400 ,ug) on demand. Duringexacerbations a 12 day course of oral pred-nisolone was administered.'2 The primary cri-terion for physician initiated withdrawal wasthe need for more than two courses of pred-nisolone over a three month period or morethan four courses over a year. 13

At alternate three month visits BDR andPC20 histamine were assessed. FEV, and PC20were measured only during clinically stableperiods and not within four weeks of termina-

tion of a course of prednisolone. Eight hoursbefore these tests all study medication wasdiscontinued. FEV1 was measured with cali-brated water sealed spirometers according tostandardisation guidelines.'4 At least threereproducible (less than 5% difference)recordings were obtained and the highestvalue was then used for analyses. Histamineprovocation tests were performed with a twominute tidal breathing method.'2 16

ASSESSMENT OF BDRBDRs were measured as the change in FEV1before and 20 minutes after four separateinhalations of 250,g of terbutaline sulphatefrom a metered dose inhaler administeredthrough a 750 ml spacer device (Nebuhaler,Astra Pharmaceuticals, Rijswijk, The Nether-lands). Patients rested at least 15 minutesbefore the first measurement and refrainedfrom coffee, tea, and smoking between mea-surements. Results are expressed in fourdifferent ways: (1) absolute response in milli-litres (BDR[ml]); (2) percentage of initial(prebronchodilator) FEV1 (BDR%init); (3)percentage of the predicted normal FEVY(BDR%pred); (4) difference between thestandardised residuals (SR) of the FEV1 val-ues before and after use of a bronchodilator(BDR-SR). An SR is obtained by subtractinga patient's FEV1 from the predicted FEV1 anddividing this difference by the RSD of thepredicted FEVI.14 An SR therefore indicateshow many RSDs a patient's FEVY is awayfrom the predicted FEVY.'5

Immediately after the BDR measurementwith terbutaline, four single inhalations of20 fg ipratropium were administered and 40minutes later the FEVY was assessed again.

DIAGNOSTIC CLASSIFICATIONUsing data from a standardised questionnairewe identified different syndromes, adheringto the criteria of the American ThoracicSociety.' Patients reporting attacks of breath-lessness and wheeze (asthmatic attacks) with-out chronic (that is, for more than threemonths per year) cough and sputum produc-tion were identified as having asthma.Current or former smokers without a historyof asthmatic attacks, reporting either chroniccough, with or without sputum production,or dyspnoea when walking quietly on levelground, or both, were included in the COPDgroup. Patients with both asthmatic attacks orrecurrent wheeze and chronic cough and spu-tum production were labelled asthmatic bron-chitics. Subjects with insufficient data toestablish a diagnosis from history taking wereincluded in an undefined diagnosis group.12

DATA ANALYSISAnalyses of PC20 values were performed withthe base-2 logarithm as this reflects doublingconcentrations. In the analyses patientsalready responding to saline or to the lowestconcentration of histamine (0-03 mg/ml) wereassigned a PC20 value of 0-015 mg/ml, beinghalf the lowest concentration applied.'2Means (SE) are presented unless otherwise

723 group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Kers6yens, Brand, Quanjer, Bruggen-Bogaarts, Koiter, Postma on behalf of the Dutch CNSLD Study Group

Figure 1 Prebronchodlawrand postbronchodilawrpercentage predicted FEV,values for patients with afollow up ofat least 21months. FEV, values arecorrectedfor sex, height,and age at every timepoint. 14

0 3

stated. Statistical mtests and one wa(ANOVA) to compafor differences in pr4signed rank test for ldistribution.'7 To ssubject standard denas well as intraclas;(ICC: between sutance); ICC ranges biof 1 signifies the idea

Table I Mean (SD) baseline characteistics by treatmeni

Treatment groi

BA + PL

n 51No. men 36 (71%)Age (y) 37 9 (11-2)FEVy prebronchodilation

litres 2-40 (0 77'% predicted 64-3 (17-5)number ofRSD below predicted 2-84 (1-42,

FEV, postbronchodilation (terbutaline)litres 2-93 (0-89'% predicted 78-1 (17-7)

Bronchodilator response (terbutaline)BDR[ml] 535 (357)BDR%init 24-2 (17-0)BDR%pred 13-9 (8-5)BDR-SR 1- 15 (0-78'

FEV, postbronchodilation (terbutaline + ipratropium)% predicted 83-6 (16-9)

Histamine challengelog,PC20 histamine (mg/ml) -2-06 (2-33,Geometric mean PC20 (mg/ml) 0-24

Allergy 33 (65%)Smoking

never 18 (35%)ex-smoker 15 (29%)current 18 (35%)

Symptom based diagnosis groupsasthma 21 (41%)asthmatic bronchitis 16 (31%)COPD 14 (28%)undefined diagnosis 0 (0%)

1.2 agonist + placebo (n = 51)--132 agonist + anticholinergic (n = 51)-12 agonist + corticosteroid (n = 83)o PrebronchodilatorA Postbronchodilator

ation in results is due to differences betweensubjects and a value of 0 denotes that all vari-ation is due to within subject variation alone(biological + measurement error).18 Linearregression analysis was employed to study therelationship between variables over time forthe group as a whole, and within subjectSpearman rank correlation coefficients (p) forwithin subject relations over time. The rate ofchange of FEVy and BDR over 21 monthsand the associated standard deviations werecalculated for each patient by least squareslinear regression. Analyses were performedusing the SPSS/PC + package.'9

Results9 15 21 The results of the intervention study in termsMonths of differences in withdrawal rate, FEV1, and

PC20 between patients treated with and with-out inhaled corticosteroids have been pub-

iethods used included t lished elsewhere."3 Briefly, 12 of 91 patientsLy analysis of variance on BA + CS were withdrawn, comparedire group means, X2 tests with 44 of 91 on BA + PL and 45 of 92 onoportions, and Wilcoxon BA + AC, p < 0x0001); 70% of withdrawalsnon-parametric testing of were due to increases in pulmonary symp-study variability, within toms."3 Mean (SE) prebronchodilator FEV1jiations were calculated'8 improved by 374 (51) ml in patients receivings correlation coefficients BA + CS in the first three months and)ject variance/total vari- remained stable afterwards, while noetween 0 and 1. An ICC increases in FEV1 were detected in the BA +d situation where all vari- AC and BA + PL groups (fig 1). These

group differences remained significant at allfollow up visits (p < 0 001). Over the 2 5years of follow up, PC20 improved on average

tgroup by 2-0 doubling concentrations in the BA +up CS group compared with no change in the

other two groups (p < 0-001). No significantBA + AC BA + CS differences were detected between the BA +51 83 AC and BA + PL group in terms of changes34 (67%) 55 (66%) in prebronchodilator and postbronchodilator41-2 (13-5) 40-1 (12-4) FEVy, PC20, and exacerbation rates."3

As the current analyses were focused on2)37 (0(78) 2-45(0(77) long term variability in BDR and effects of

2-70 (1-15) 2-70 (1-22) inhaled corticosteroid treatment we onlyanalysed data of those patients who had a fol-

2-82 (0 85) 2-86 (0-86) low up of at least 21 months (five measure-77*3 (16.1) 77-2 (16-1) ments). Withdrawals (those not reaching 21

months of follow up) had slightly lower20-7 (14-9) 1876 (1543) baseline prebronchodilator FEV1%pred than12-4 (7-6) 11-4 (8 2) non-withdrawals (60-8% (1-5%) v 65-1%0-99 (0 67) 0-92 (0 69) (1-2%), p < 0 05), lower postbronchodilator

FEV,%pred (72.1% (1-7%) v 77-5% (1.2%),81-4 (16-0) 82-8 (15-4) p < 0-01), and lower log2 PC20 (-2-44 (2.3) v

)-1.57 (2 33) -1.60 (220) -1-71 (2.3), p < 0 05). There were no signifi-0-34 0 233 cant differences at baseline between with-

26 (51%) 50 (60%) drawals and non-withdrawals in BDR (p >0 15 for all four expressions). Baseline

16 (31%) 27 (33%) characteristics for the 185 patients with a fol-15 (29%) 27 (33%) low up of at least 21 months are shown in20 (39%) 29 (35%) table 1.

22 (43%) 32 (39%)21 (41%) 28 (34%)8 (16%) 18 (22%)0 (0%) 5 (6%)

BA + PL,-2 agonist plus placebo; BA + AC-f/2 agonist plus anticholinergic; BA + CS-,82agonist plus corticosteroid; RSD-residual standard deviation: a larger figure denotes moreabnormality; BDR[ml]-bronchodilator response expressed as absolute response in ml;BDR%init-bronchodilator response expressed as a percentage of initial (prebronchodilator)FEV,; BDR%pred-bronchodilator response expressed as a percentage of the predicted normalFEV,; BDR-SR-bronchodilator response expressed as the difference between the standardisedresiduals of the postbronchodilator and prebronchodilator FEV,.

CHANGE IN BDR DURING TREATMENTIn fig 2A-D time courses are shown of theBDR expressed in four different ways. Afterthree months the BDR[ml] had decreased inthe BA + CS group by 117 (38) ml com-pared with an increase of 11 (39) ml in theBA + PL group (p < 0 05). This mean dif-ference of 127 ml remained approximately

I

724 group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Variability of bronchodilator response and effects ofinhaled corticosteroid treatment in obstructive ainvays disease

Figure 2A-D Mean(SE) bronchodilatorresponse expressed in fourdifferent ways (see text).

X agonist + placebo (n = 51)---P agonist + anticholinergic (n = 51)

A -_ agonist + corticosteroid (n = 83)620,

E0

a)C00.

m

0

o

m

'a0.C0

565-

510-

455-

400'

345-

2

;a0)a-

CL00-

U)c

0

CoU)

CG

o

mT

I..... ....

N.

N" --1 1

6 i 16ilC .1..

Months

stable afterwards (fig 2A). In the BA + PLand BA + AC groups there were no signifi-cant changes in BDR[ml] from baseline toany of the follow up visits; the slopes of theBDR in the BA + AC and BA + PL groupswere not significantly different from zero(mean slope expressed as ABDR[ml]/year of+12 (19) and +13 (19) in the BA + PL andBA + AC groups respectively, fig 2A).

T

BA-CS

I

BA--AC

_7 T

._fBA- PL

BA-PL 132 agonist placeboBA-AC 12 agonist + anticholinergicBA-CS 02 agonist + corticosteroidL112 post-terbutaline (at start of study)t-.14 post-terbutaline + ipratropium bromide (at start of study)_ pre-bronchodilator (at 3 months)

E,/2 post-terbuta line (at 3 months)post-terbutaline - ipratropium bromide (at 3 months)

Figure 3 Mean (SE) change in FEV, (ml) from baseline prebronchodilator FEV,.

Similarly, when expressed as BDR%init,BDR%pred, and BDR-SR, the BDR droppedmarkedly in the first three months of cortico-steroid treatment to remain stable afterwards(decrease in BDR%init: -7d1 (1 8) p < 0-005compared with BA + PL, fig 2B),BDR%pred: -3-2 (1-0), p < 0-01, fig 2C,BDR-SR: -0-27 (0 08), p < 0 01, fig 2D). Inthe BA + AC and BA + PL groups therewere no significant changes from baselineover time with any of the BDR expressions,nor were the slopes of the changes over timesignificantly different from zero.The decrease in BDR in three months in

the BA + CS group was correlated to theincrease in prebronchodilator FEV,%pred inthree months, more strongly so whenexpressed as BDR%init (r = -0-65) than asBDR[ml] (-0 52), BDR%pred (-0-51), orBDR-SR (-0-51), all p < 0-001.To determine to what degree the decrease

in BDR in the BA + CS treated group wascaused by the FEV, reaching normal levelsthe data were recalculated for a subgroup of49 patients whose FEV, value after terbu-taline remained more than 1-64 RSD belowthe predicted value after three months oftreatment-that is, whose post-terbutalineFEV, remained abnormal. The decrease inBDR[ml] in the first three months diminishedto -61 (39) ml (p > 01), the decrease inBDR%init to -5 0% (2d1%) (p < 0-05),BDR%pred to -2-0% (1-1%) (p > 0-05), andBDR-SRto -0-16 (0 1) (p > 0 05).

Additional administration of 80 pg of ipra-tropium bromide immediately after terbu-taline (1000 jug) gave a mean additionaldilatation of 197 (30) ml before and 200 (24)

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

ITI

6 3 9Months

15 21

1000-

800-

600 -

400 -

200-

Lu

C:cJ

0-

-200 -

P90 i

L

_

u

.....

.smes< 44S4

zr5

725

ILI.I%e

ickA'JV-

17C 1

group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Kerstjens, Brand, Quanjer, Bruggen-Bogaarts, Koeter, Postma on behalfof the Dutch CNSLD Study Group

Table 2 Intraclass correlation coefficients (ICC) * and within subject standard deviations (SD) * according to treatmentand diagnosis groups

Total groupBA + PL BA + AC BA + CS Asthmat ABt COPDt

ICC SD (SD) (SD) (SD) (SD) (SD) (SD)

BDR[ml] 0-76 186 195 184 182 193 169 195BDR%init 0-76 10-7 13-4 11-7 7 9 11-8 10-3 9.1BDR%pred 0 74 5-1 5-2 5 0 5-3 5-1 4-6 5 9BDR-SR 0 77 0-41 0 43 0 40 0 39 0 43 0 37 0-41

*ICC and SD are calculated for measurements under double blind treatment, i.e. without the baseline visit (see methodssection); tfive of 185 patients had an undefined diagnosis as specified in the methods section and are left out of the comparisonof diagnosis groups; AB-asthmatic bronchitis; COPD-chronic obstructive pulmonary disease. Other abbreviations as intable 1.

ml after three months of treatment withinhaled corticosteroids (fig 3). In the sub-group of 49 patients whose post-terbutalineFEVy value remained more than 1 64 RSDbelow predicted as described above, the addi-tional increase with ipratropium bromideafter terbutaline amounted to 227 ml (morethan 100 ml in 39 of 49 patients).

LONG TERM VARIABILITY OF BRONCHODILATORRESPONSESSince it is clear from fig 2A-D that there wasa considerable decrease in BDR in the BA +CS group in the first three months, the longterm variability of the various BDRs was cal-culated for measurements under double blindtreatment only-that is, without the baselinevisit. The within subject standard deviationswere quite high: 10-7%init, 186 ml,5 1%pred, and 0-41 RSD (table 2). The vari-ability in BDR was comparable between thesymptom based diagnosis groups of asthmaand COPD (table 2). It tended to diminishduring treatment with inhaled corticosteroids,most notably so when expressed as BDR%init(table 2). The decrease in variability withinhaled corticosteroids existed only in thesymptom based asthma and asthmatic bron-chitis groups.

Within subject change in BDR%init over21 months was more dependent on parallelchanges in prebronchodilator FEVy [ml] thanthe other expressions of BDR: median withinsubject p for FEVy and BDR%init was -0 70,BDR[ml] -056, BDR%pred -0-60, andBDR-SR -0-60 (all p < 0-001). These withinsubject correlation coefficients were slightlystronger for the group with a symptom baseddiagnosis of asthma than for COPD (medianp for FEVy and BDR%init, -0 70 and -0-60,respectively).

Table 3 Frequency of "significant" BDRs in 97 patients with five measurements during21 months of bronchodilator treatment

Frequency of a "significant" bronchodilator responseCut off level(reference) x 0 x I x 2 x 3 x 4 x 5

BDR%init 15%init' 12 18 8 12 17 30BDR%pred 9%pred6 8 15 7 16 16 35BDR[ATS] 12%init and 200 ml" 6 14 9 13 16 39

BDR[ATS]-significant bronchodilator response expressed as percentage of initial(prebronchodilator) FEV, > 12%init if absolute response > 200 ml;" other abbreviations as intable 1.

SHIFTS IN DIAGNOSTIC CLASSIFICATION AS"REVERSIBLE") SPONTANEOUSLY AND WITHCORTICOSTEROID TREATMENTPublished criteria for "significant" reversibil-ity 820 were applied to the data in order todetermine how many patients measured onlyin a stable condition would shift from classifi-cation category ("reversible" or not) duringfive measurements in 21 months (table 3).Five patients with one missing value havebeen left out of the table. High proportions ofpatients shifted from "reversible" to "irre-versible" and back several times. Consistentclassification-for example, for BDR%init>15%-occurred in only 42 of 97 (43%)patients (table 3).The proportion of patients that would be

labelled as having "reversible" airways diseaseby the various criteria fell significantly duringmaintenance treatment with BA + CS, butnot in both non-steroid treated groups(table 4).

PROGNOSTIC VALUE OF BDR FOR CHANGE INFEVy WITH INHALED CORTICOSTEROIDSBDRs at the start of the study were tosome extent predictive of improvement inFEVy during BA + CS treatment: correlationcoefficient for change in prebronchodilatorFEVI[ml] in the first three months with pre-study BDR[ml] r = 0-46, BDR%init r = 0.52,BDR%pred r = 0 40, and BDR-SR r = 0-42(all p < 0-001). Thus the explained variancein change of FEVy [ml] by the pre-study BDRis still low: r2 = 21%, 27%, 16%, and 17%respectively.As there is no agreement on what consti-

tutes a "significant" long term effect of corti-costeroid treatment, analogous to the ATSstatement on BDR, an improvement in FEVywithin three months of >12%init and>200 ml was labelled "significant".8 A "sig-nificant" long-term corticosteroid responsedefined in this way was correctly predicted bya baseline BDR%init > 15% in only 50 of 83patients (60%), by a BDR%pred >9% in60%, and by the new ATS criterion for BDR(>12%init and >200 ml) in 57% of cases. Inthe COPD subgroup the treatment responseto inhaled corticosteroids was somewhat bet-ter predicted by the baseline BDR expressedaccording to the ATS than in the asthma sub-group (12 of 18 (67%) and 17 of 32 (53%),respectively).

726

group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Variability of bronchodilator response and effects of inhaled corticosteroid treatment in obstructive ainvays disease

Table 4 Number ofpatients with a "significant" BDR per treatment group before and after treatment

Percentage with "significant" response (no/total)Cut off level(reference) Months BA + PL BA + AC BA + CS P* Pt

BDR%init 15%init' 0 63 (32/51) 63 (32/51) 52 (43/83) 0-33 0 003 60 (30/50) 61 (31/51) 29 (24/83) 000

21 67 (33/49) 57 (29/51) 26 (21/82) 0 00BDR%pred 9%pred6 0 65 (33/51) 73 (37/51) 54 (45/83) 0 10 0-08

3 68 (34/50) 65 (33/51) 46 (38/83) 0-0221 71 (35/49) 65 (33/51) 39 (32/82) 000

BDR[ATS] 12%init 0 65 (33/51) 73 (37/51) 58 (48/83) 0-26 0 05and 200 ml8 3 68 (34/50) 71 (36/51) 46 (38/83) 0-04

21 69 (34/49) 65 (33/51) 42 (34/82) 0-06

*p value for x2 comparison of proportions between the three treatment groups; tp value for X2 comparison of trend with time inproportions within BA + CS group; all p values for trends in BA + AC and BA + PL >0 50. Abbreviations as in tables 1 and 3.

DiscussionBronchodilator responses in our study aresubstantially reduced by treatment withinhaled corticosteroids and /32 agonists (BA +CS), but not by treatment with,2 agonistsalone (BA + PL), or Ih agonists with an anti-cholinergic drug (BA + AC). Both prebron-chodilator and postbronchodilator FEVyimproved within three months of treatmentwith corticosteroids, but the prebroncho-dilator FEV, improved to a greater extentcausing the reduction in BDR.We hypothesised that part of the decrease

in BDR in patients treated with BA + CScould be explained by the fact that patientsreach normal levels of FEVI: in those patientswhose post-terbutaline FEV, remainedabnormal after three months of treatment asmaller decrease in BDR was found (61 ml v117 ml, or 5%init v 7%init). However, inboth the complete BA + CS group and thesubgroup mentioned above the FEV, mea-sured after 1000 pg of terbutaline was not theindividual's personal maximum FEV1, sincethe administration of an additional 80,pg ofipratropium bromide caused a further 200 mlbronchodilatation (both at the start of thestudy and after three months of treatment).Since FEV, was measured before, 20 minutesafter terbutaline, and 40 minutes after ipra-tropium bromide, it could be argued that partof this further improvement with ipratropiumbromide could in fact have been due to a fur-ther improvement with terbutaline after thefirst 20 minutes. An additional effect of ipra-tropium bromide added to a betamimetic,however, has been documented by severalothers, though usually at lower dosages of /32agonist than the 1000 pg of terbutaline givenin our study.21-24

Theoretically airway calibre can bedecreased by at least five mechanisms25:decreased preload, increased smooth muscletone, increased smooth muscle strength(hypertrophy, hyperplasia), airway wall thick-ening due to hypertrophy and hyperplasia ofmuscles and glands, oedema, influx of inflam-matory cells, and secretions within the airwaylumen. In this study prebronchodilator FEV,after three months of treatment with cortico-steroids rose to approximately the post-terbu-taline value at the start of the study. Thiscould indicate that the improvement in air-ways obstruction with inhaled corticosteroids

is due to a decrease in bronchomotor toneonly. However, after three months the meanBDR to terbutaline was still 300 ml, com-pared with 417 ml at the initial visit. Thissuggests that inhaled corticosteroids alsocause an alteration in airway geometry apartfrom the change in bronchomotor tone,which is well in line with the proposed anti-inflammatory effects of inhaled cortico-steroids (reduction of number ofinflammatory cells and decreases in release ofcytokines and mediators thereby reducing air-way wall thickness, airway secretions andmuscle tone).26 Although in our study FEV1and BDR showed no further change afterthree months of corticosteroid treatment andhence the maximum improvement in smoothmuscle tone and airway geometry had beenachieved, airway hyperresponsiveness to hist-amine nevertheless continued to improveuntil at least one year of treatment in thesame patient population."3 It is possible thatmore sensitive measurements than FEV1would still detect improvements in airwayresistance in parallel with further improve-ments in the extent of inflammation as mea-sured by PC20. Other explanations forcontinuing improvement in PC20 withoutconcomitant change in airway calibre seemmore appealing, however, such as improve-ments in epithelial healing and decreases inpermeability which would not by itself affectairway geometry.As indicated above, only part of the

improvement in the level of airways obstruc-tion is caused by a decrease in bronchomotortone. This by itself already suggests that aBDR measured before treatment with inhaledcorticosteroids will not accurately predict thechange in FEV, caused by their use. Althoughsignificant regression coefficients were foundbetween baseline BDR and long termimprovement in FEV1, the maximum correla-tion coefficient was 0-52 signifying 27%explained variance. The correct prediction ofa favourable long term response was calcu-lated using the baseline BDR as the test: onlyapproximately 60% of patients would havebeen correctly given or withheld inhaledcorticosteroids. This correct prediction of an"inhaled corticosteroids responder" by a pre-treatment BDR was slightly better in COPDthan in asthma, but because of the smallnumbers it is uncertain whether any firm

727

group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Kerstjens, Brand, Quanjer, Bruggen-Bogaarts, Koeter, Postma on behalf of the Dutch CNSLD Study Group

conclusions should be attached to the differ-ence-that is, whether a single bronchodilatortest is more useful as a prognostic test forinhaled corticosteroids in COPD than it is inasthma.

Surprisingly few data have hitherto beenpublished on long term variability of BDR inasthma and COPD. Considerable within sub-ject variability of BDR was found in thisstudy, which was performed under rigorouslycontrolled trial conditions, all patients beingon maintenance medication and withoutmeasurements during or shortly after exacer-bations (table 2). The standard deviations ofBDR found are comparable to those reportedby Anthonisen et al in patients with COPDparticipating in the IPPB trial.2 Even thoughwe found significant differences (but with alarge overlap) in BDR between the differentdiagnosis groups,4 within subject standarddeviations were quite comparable for patientswith a symptom based diagnosis of asthma orCOPD (table 2).The American Thoracic Society has

recently recommended a new criterion for thedefinition of a "significant" bronchodilatorresponse, again from one test (a change inFEV, larger than 12%init and larger than 200ml). Table 3 shows that in this large group ofpatients with a broad range of clinical charac-teristics, variability in the result of a singletest is not reduced by the new ATS criterion.

Although the decrease in BDR withinhaled corticosteroids is probably not sur-prising, to our knowledge it has not beendocumented to date. In many guidelines forthe management of obstructive airways dis-ease it is recommended that healthy subjectsshould be distinguished from those withobstructive airways disease, and reversible air-ways obstruction should be distinguishedfrom irreversible airways obstruction on thebasis of a (single) bronchodilator test.' 27 Theresults of our study suggest that this is haz-ardous, both because of a high within subjectvariability and because of a decrease in BDRduring treatment with inhaled corticosteroids.We feel that one single bronchodilator test isclearly insufficient to make a diagnosis of"(ir)reversible" airways obstruction, andhence to classify a patient as having asthma orCOPD on the basis of the test.From cross sectional data several authors

have stressed the dependence on baselineFEV, of the BDR when expressed as apercentage of the initial value.2 4 From ourlongitudinal data we document further mani-festations of this baseline dependency:changes in BDR%init were more dependenton improvements in prebronchodilator FEV,with corticosteroid treatment than the otherexpressions employed. Similarly, in the non-steroid treated groups the within subject vari-ability of BDR was most closely linked tovariability in prebronchodilator FEV, whenexpressed as %init.5 From the data presentedin this report no clearcut argument can begiven in favour of any one of the expressionsof BDR. Expressing the BDR relative to thepredicted FEVI, however, makes some cor-

rection for gender, height, and age whichmakes comparison between patients andbetween studies easier.4 Expressing the FEVy,and hence the BDR, relative to the standard-ised residual offers theoretical advantages'415but no advantages in the sense of decreasedbaseline dependency or decreased variabilityin this study.The study population was selected on the

basis of symptomatic respiratory disease, age,and objective functional criteria of airwaysobstruction (FEVI, FEV,IVC) and airwayresponsiveness (PC2o histamine). Selectionbias of a certain level of reversibility to a Ragonist was therefore not introduced.Because considerable dropout occurred inboth treatment arms without inhaled corti-costeroids, however, the current selectionfrom our original population introduced biasto some extent. Withdrawals had slightlylower baseline FEV,%pred and lower PC20than non-withdrawals, but no significant dif-ferences were found in BDR. Since morethan 70% of withdrawals resulted fromincreases in respiratory symptoms it seemslikely that, if anything, variability of the BDRin the currently selected group is a conserva-tive estimate of the true variability in ouroriginal population.

In conclusion, BDR decreases substantiallywith inhaled corticosteroid therapy. Thisimprovement is related to the improvement inFEV, which is caused by both altered airwaygeometry and a decrease in bronchomotortone as measured by a decreased BDR.Moreover, measurements of BDR are veryvariable over 21 months. These results sug-gest that a single BDR should not be used asthe basis for therapeutic decisions nor as aselection criterion for clinical trials. Theseproblems are greatest with a BDR expressedas a percentage of the initial value because ofits dependency on baseline FEV,. Finally,because of the imperfect relation between thebronchodilator test before inhaled cortico-steroids and the future response to them, wethink that therapeutic decisions on whether ornot to prescribe corticosteroids should not bemade on the basis of bronchodilator tests, butprobably on combined information fromsymptoms, FEV1, PC20, peak flow, atopy, andBDR.

The Dutch CNSLD study group consists of a steering com-mittee (KF Kerrebijn, PH Quanjer and HJ Sluiter), and ofmembers from the departments of Pulmonology of theUniversity Hospital of Amsterdam (EM Pouw, DFMESchoonbrood, CM Roos, HM Jansen), Groningen (PLPBrand, A de Gooyer, HAM Kerstjens, DS Postma, ThW vander Mark, HJ Sluiter, GH Koeter), Leiden (PM de Jong, PJSterk, AMJ Wever, JH Dijkman), Nijmegen (PNRDekhuijzen, HTM Folgering, CLA van Herwaarden),Rotterdam (SE Overbeek, JM Bogaard, C Hilvering) andUtrecht (SJ Gans, HJJ Mengelers, BAHA van der Bruggen-Bogaarts, J Kreukniet), from the departments of PaediatricPulmonology at the Sophia Children's Hospital, Rotterdam(EEM van Essen-Zandvliet, KF Kerrebijn), the JulianaChildren's Hospital, the Hague (EJ Duiverman, JMKouwenberg, JE Prinsen), University Hospital of Groningen(HJ Waalkens, J Gerritsen, K Knol), from the department ofAllergology, University Hospital of Groningen (JGR deMonchy), from the department of General Practice,University of Leiden (FW Dekker), from the department ofPsychiatry, University Hospital Leiden (AA Kaptein), andfrom the department of Physiology, University of Leiden

728

group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

Variability of bronchodilator response and effects of inhaled corticosteroid treatment in obstructive airways disease

(PJFM Merkus, PH Quanjer). Scientific counsel: SJ Pocock,MD Hughes, (London, UK), ER Bleecker, DA Meyers(Baltimore, USA).

We gratefully acknowledge the support from AstraPharmaceuticals, Boehringer Ingelheim, and Glaxo who sup-plied the study medication in identical metered dose inhalers.

This study was supported by a government grant from theNetherlands Health Research Promotion Program (SGO).

1 American Thoracic Society. Standards for the diagnosisand care of patients with chronic obstructive pulmonarydisease (COPD) and asthma. Am Rev Respir Dis1987;136:225-44.

2 Anthonisen NR, Wright EC, IPPB Trial Group.Bronchodilator response in chronic obstructive pul-monary disease. Am Rev Respir Dis 1986;133:814-9.

3 Weir DC, Burge PS. Measures of reversibility in responseto bronchodilators in chronic airflow obstruction: rela-tion to airway calibre. Thorax 1991;46:43-5.

4 Brand PLP, Quanjer PH, Postma DS, Kerstjens HAM,Koeter GH, Dekhuijzen PNR, et al. Interpretation ofbronchodilator response in patients with obstructive air-ways disease. Thorax 1992;47:429-36.

5 Dompeling E, van Schayck CP, Molema J, Akkermans R,Folgering H, van Grunsven PM, et al. A comparison ofsix different ways of expressing the bronchodilatingresponse in asthma and COPD; reproducibility anddependence of prebronchodilator FEV,. Eur Respir J1992;5:975-81.

6 Dales RE, Spitzer WO, Tousignant P, Schechter M,Suissa S. Clinical interpretation of airway response to abronchodilator. Epidemiologic considerations. Am RevRespir Dis 1988;138:317-20.

7 Eliasson 0, Degraff C Jr. The use of criteria for reversibil-ity and obstruction to define patient groups for bron-chodilator trials. Influence of clinical diagnosis,spirometric, and anthropometric variables. Am RevRespir Dis 1985;132:858-64.

8 American Thoracic Society. Lung function testing: selec-tion of reference values and interpretative strategies. AmRev Respir Dis 1991;144:1202-18.

9 Harding SM, Freedman S. A comparison of oral andinhaled steroids in patients with chronic airwaysobstruction: features determining response. Thorax1978;33:214-8.

10 Mendella LA, Manfreda J, Warren CPW, AnthonisenNR. Steroid response in stable chronic obstructive pul-monary disease. Ann Intern Med 1982;96:17-21.

11 Nisar M, Walshaw M, Pearson MG, Calverley PMA.Assessment of reversibility of airway obstruction inpatients with chronic obstructive airways disease.Thorax 1990;45:190-4.

12 Brand PLP, Kerstjens HAM, Postima DS, Sterk PJ,Quanjer PH, Sluiter HJ, et al. Long-term multicentretrial in chronic non-specific lung disease: methodologyand baseline assessment in adult patients. Eur Respir 71992;5:21-31.

13 Kerstjens HAM, Brand PLP, Hughes MD, Robinson NJ,Postma DS, Sluiter HJ, et al. A comparison of bron-chodilator therapy with or without inhaled cortico-steroid therapy in obstructive airways disease. N Eng1 JMed 1992;327:1413-9.

14 Quanjer PH. Standardized lung function testing. BuUl EurPhysiopathol Respir 1983;19(Suppl 5):1-95.

15 Miller MR, Pincock AC. Predicted values: how should weuse them? Thorax 1988;43:265-7.

16 Cockcroft DW, Killian DN, Mellon JJA, Hargreave FE.Bronchial reactivity to inhaled histamine: a method andclinical survey. Clin Allergy 1977;7:235-43.

17 Armitage P, Berry G. Statistical methods in medical research.Oxford: Blackwell, 1987.

18 Chinn S. Statistics in respiratory medicine. 2 Repeatabilityand method comparison. Thorax 1991;46:454-6.

19 Norusis MJ. Statistical package for the social sciences, PCversion 4-0. Chicago: SPSS, 1990.

20 Tweeddale PM, Merchant S, Leslie M, Alexander F,McHardy GJR. Short term variability in FEV,: relationto pretest activity, level of FEVy, and smoking habits.Thorax 1984;39:928-32.

21 Petrie GR, Palmer KNV. Comparison of aerosol iprat-ropium bromide and salbutamol in chronic bronchitisand asthma. BMJ 1975;i:430-2.

22 Grandordy BM, Thomas V, Lauture D, Marsac J.Cumulative dose-response curves for assessing com-bined effects of salbutamol and ipratropium bromide inchronic asthma. EurRespirJ 1988;1:531-5.

23 Elwood RK, Abboud RT. The short-term bronchodilatoreffects of fenoterol and ipratropium in asthma.JAllergy Clin Immunol 1982;69:467-73.

24 Higgins RM, Stradling JR, Lane DJ. Should ipratropiumbromide be added to beta-agonists in treatment of acutesevere asthma? Chest 1988;94:718-22.

25 Moreno RH, Hogg JC, Pare PD. Mechanics of airwaynarrowing. Am Rev Respir Dis 1986;133:1171-80.

26 Morris HG. Mechanisms of action and therapeutic role ofcorticosteroids in asthma. JAllergy Clin Immunol 1985;75: 1-13.

27 International Asthma Management Project. Internationalconsensus report on the diagnosis and treatment of asthma.Publication no. 92-3091. Bethesda: NHLBI, NIH,1992.

729

group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from

doi: 10.1136/thx.48.7.722 1993 48: 722-729Thorax

 H A Kerstjens, P L Brand, P H Quanjer, et al. disease. Dutch CNSLD Study Group.treatment in obstructive airwaysand effects of inhaled corticosteroid Variability of bronchodilator response

http://thorax.bmj.com/content/48/7/722Updated information and services can be found at:

These include:

References http://thorax.bmj.com/content/48/7/722#related-urls

Article cited in:

serviceEmail alerting

up in the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on July 15, 2011 - Published by thorax.bmj.comDownloaded from