Respiratory Medicine (2008) 102, 1473e1482

ava i lab le at www.sc ienced i rec t . com

j ourna l homepage : www.e lsev ier . com/ loca te / rmed

The course of persistent airflow limitation insubjects with and without asthma*

Stefano Guerra a,b,*,c, Duane L. Sherrill a,b, Margaret Kurzius-Spencer a,b,Claire Venker a,b, Marilyn Halonen a, Stuart F. Quan a, Fernando D. Martinez a

a Arizona Respiratory Center, University of Arizona, Tucson, AZ, USAb Mel and Enid Zuckerman College of Public Health, University of Arizona, Tucson, AZ, USA

Received 5 December 2007; accepted 4 April 2008Available online 6 August 2008

KEYWORDSAsthma;COPD;Eosinophilia;Airflow limitation

* This study was funded in part by athe National Heart, Lung, and Blood Inand Janet Lang Donor Advised Fund. TArizona.

* Corresponding author. Arizona Resp5030, USA. Tel.: þ1 520 626 7411; fax

E-mail address: sguerra@arc.arizonc Dr Guerra is the recipient of a Par

0954-6111/$ - see front matter ª 200doi:10.1016/j.rmed.2008.04.011

Summary

Rationale: Most patients who develop persistent airflow limitation do so either as a manifesta-tion of chronic obstructive pulmonary disease that is largely related to smoking or as a conse-quence of persistent asthma. We sought to compare the natural course of lung functionassociated with persistent airflow limitation in subjects with and without asthma from earlyto late adult life.Methods: We studied 2552 participants aged 25 or more who had multiple questionnaire andlung function data from the long-term prospective population-based Tucson EpidemiologicalStudy of Airway Obstructive Disease. Persistent airflow limitation was defined as FEV1/FVCratio consistently <70% in all completed surveys subsequent to the first survey with airflow lim-itation. Participants were divided into nine groups based on the combination of their physician-confirmed asthma status (never, onset �25 years, or onset >25 years) and the presence ofairflow limitation during the study follow-up (never, inconsistent, or persistent).Results: Among subjects with an asthma onset �25 years, blood eosinophilia increased signif-icantly the odds of developing persistent airflow limitation (adjusted ORs: 3.7, 1.4e9.5),whereas cigarette smoking was the strongest risk factor for persistent airflow limitation amongnon-asthmatics and among subjects with asthma onset after age 25 years. Among subjects withpersistent airflow limitation, the natural course of lung function differed between subjectswith asthma onset �25 years and non-asthmatics, with the former having lower FEV1 levelsat age 25 (predicted value for a 175-cm tall male of 3400 versus 4090 ml, respectively;

grant award by the American Thoracic Society/Alpha1 Foundation, grants HL14136 and HL085195 bystitute, grant 0660059Z by the American Heart Association, and an unrestricted grant from the Barryhis study was conducted with the support of the General Clinical Research Center of the University of

iratory Center, University of Arizona, 1501 N. Campbell Avenue, P.O. Box 245030, Tucson, AZ 85724-: þ1 520 626 6970.a.edu (S. Guerra).

ker B. Francis Fellowship.

8 Elsevier Ltd. All rights reserved.

1474 S. Guerra et al.

p< 0.001) and the latter having greater FEV1 loss between age 25 and 75 (1590 versus 2140 ml;p Z 0.003).Conclusion: In subjects who have asthma onset before 25 years of age and persistent airflowlimitation in adult life, the bulk of the FEV1 deficit is already established before age 25 years.ª 2008 Elsevier Ltd. All rights reserved.

Introduction

Airflow limitation that is variable and reversible eitherspontaneously or with treatment is a defining feature ofasthma.1,2 Yet, subgroups of patients with long-term persis-tent asthma may develop irreversible airflow limitation,3e5

which has been associated with markers of disease severity5

and overall mortality risk.6 The possible progression of per-sistent asthma into chronic airflow limitation is consistentwith observations showing that, at the population level,co-existing diagnoses of asthma, chronic bronchitis and em-physema are frequently reported by the same patient7,8

and that active asthma substantially increases the risk ofacquiring a subsequent diagnosis of chronic obstructivepulmonary disease (COPD).9 Indeed, in the presence of irre-versible airflow limitation, a differential diagnosis betweenasthma and COPD can be quite difficult using the phy-siologic tests that are routinely utilized in the clinicalsetting.10e12

Understanding the natural course of lung function and thepatterns of risk factors for development of persistent airflowlimitation among asthmatics, as contrasted with those ofsubjects with classical smoking-related COPD, is essential todetermine whether, how, and when prevention and treat-ment strategies can be implemented to reduce the long-term sequelae of this disease. However, population-basedlong-term prospective studies on persistent airflow limita-tion in subjects with asthma are scant. Seminal work byBurrows and colleagues1 assessed the course and progno-sis of chronic airway obstruction in 27 asthmatics whoalready had moderate to severe airflow limitation at theirenrollment in the Tucson Epidemiological Study of AirwayObstructive Disease (TESAOD). They found that subjectswith asthma had a slower FEV1 decline than did subjectswith smoking-related emphysema, but they did not studyhow the natural course of lung function from young to lateadult life differed in these two groups. The large cohort ofthe TESAOD study has now covered a sufficiently long periodwith prospective follow-up of well-characterized respiratoryphenotypes and repeated lung function tests over a substan-tial proportion of adult life for most of the participants, toaddress this issue. The goal of the present study was todetermine risk factors and natural course of lung functionassociated with persistent airflow limitation in subjectswith and without asthma in the TESAOD study.

Methods

Study population

TESAOD is a population-based prospective cohort studyinitiated in Tucson, AZ, in 1972. Details of the enrollment

process have been previously reported.13 During the follow-up, new enrollees were added by marriages and births fora final total of 5377 white participants.

TESAOD is composed of 13 surveys that took placeapproximately every two years between 1972 and 1996.All participants were eligible to take part in the first 12surveys, whereas survey 13 was conducted on selectedsubgroups based on previous reports of respiratory symp-toms/disease or residence in Arizona. During each survey,participants completed a standardized questionnaire and,with the exception of survey 4, performed spirometric lungfunction tests with a pneumotachygraph device, which wascalibrated daily, according to ATS guidelines.14 In survey 13,spirometric tests were repeated 15 min after inhalation of180 mg of albuterol via a pressured metered-dose inhalerand a valved holding chamber.

For the present study, we used data from 2552 partic-ipants who completed questionnaire and lung function testsin at least two TESAOD surveys in which they were 25 yearsor older. Only lung function data from surveys completed ator after age 25 were used to minimize potential confound-ing by individual variation in onset and length of theplateau phase of lung function growth. Of the 2552participants, 515 (20%) were 25 years or younger at thetime of enrollment into TESAOD, but they completed atleast two surveys in which they were 25 years or olderduring the study follow-up.

Definitions and measurements

Participants were considered to have physician-confirmedasthma if they reported to have ever had asthma and tohave been seen, diagnosed, or treated for asthma bya doctor at any completed survey. Participants with asthmawere classified into two mutually exclusive groups based onwhether they had disease onset at/before age 25 years orafter age 25 years. We selected this age cut-off to beconsistent with the above mentioned age cut-off that waschosen as initial point for analyses on lung function. Asthmaonset �25 years was defined as having asthma (as describedabove) at any survey before age 25 or, for subjects enrolledafter age 25, as having asthma at enrollment into the studyand reporting being �25 years at first asthma attack orhaving had ‘respiratory trouble’ before age 16.15 Asthmaonset >25 years was defined as either having no asthmaat enrollment and having new asthma at any subsequentsurvey after age 25 or as having asthma at enrollmentinto the study after age 25 and reporting age at firstasthma attack >25 years and no ‘respiratory trouble’before age 16.

Airflow limitation was defined as FEV1/FVC ratio <70%.Airflow limitation was defined as ‘‘persistent’’ if the subjecthad FEV1/FVC values <70% in all completed lung function

Course of persistent airflow limitation 1475

tests subsequent to the first survey with airflow limitation.Otherwise, it was coded as ‘‘inconsistent’’. The categoriesof no, inconsistent, and persistent airflow limitation werecreated based on available observations. In lung functionanalyses, missing observations were assumed to be at ran-dom. Both subjects with prevalent airflow limitation andsubjects with incident airflow limitation were included inthis study to avoid selective removal of subjects with earlyonset of airflow limitation. Because the FEV1/FVC ratiodeclines significantly with aging, analyses on lung functionwere also repeated after defining airflow limitation basedon lower limit of normal equations for FEV1/FVC.16

We categorized subjects as ever smokers if they hadsmoked at least 1 packyear by the time they completedtheir last survey.

A detailed description of assessment methods for skinprick tests, IgE, eosinophil measurements, and otherclinical variables is included in Table E1 in the onlineappendix.

Study design and statistical analyses

The 2552 study subjects were divided into nine mutuallyexclusive groups basedon the combination of their physician-confirmed asthma status (no asthma, asthma onset �25years, or asthma onset>25 years) and the presence of airflowlimitation during the study follow-up (never, inconsistent, orpersistent). The study design is summarized in Fig. 1.

Multinomial logistic regression was used to determinefactors associated with persistent and inconsistent airflowlimitation, as compared with no airflow limitation, in thetotal population and separately for subjects with noasthma, asthma onset �25 years, and asthma onset >25years. Robust variance estimates were used to minimize theeffects of correlation within households. In order tocompare FEV1 trends over time across the groups with

Figure 1 Summary of study design. The nine study groups were crstatus (no asthma, asthma onset �25 years, or asthma onset >25follow-up (never, inconsistent, or persistent). *Only participants wtwo surveys in which they were �25 years or older were eligible fo

persistent airflow limitation while adjusting for the intra-subject serial correlation of repeated observations andreducing the impact of missing observations,17 we usedrandom coefficients models that included as fixed covari-ates a categorical indicator of study groups, sex, height,length of follow-up, age, age squared, an interactionterm between study groups and age, and an interactionterm between study groups and age squared. BecauseFEV1 is part of the defining criteria of persistent airflow lim-itation, FEV1 trends were compared using linear contrastsonly between pairs of groups with persistent airflow limita-tion. In addition, the slope of FEV1 was computed for eachsubject with at least three lung function tests �25 years ofage by regressing FEV1 against age. Two multivariatelogistic regression models were then used to determine fac-tors associated with being in the fastest tertile and being inthe fastest quintile of FEV1 decline (i.e., having an FEV1

slope of at least 33 ml per year or at least 45 ml per year,respectively). Statistical analyses were completed usingthe statistical packages SPSS version 15.0 and Stataversion 9.0.

The study was approved by the Human Subjects Com-mittee at the University of Arizona and all participantsprovided informed consent.

Results

Of the 2552 subjects included in this study, 1778 (70%) hadno airflow limitation, 460 (18%) had inconsistent airflowlimitation, and 314 (12%) had persistent airflow limitation.Fig. 2 shows the frequency of persistent and inconsistentairflow limitation among subjects with no asthma, asthmaonset �25 years, and asthma onset >25 years. Subjects ineither asthma group were at increased risk of inconsistentand persistent airflow limitation as compared with subjectswith no asthma. When analyses were stratified by smoking,

eated based on the combination of physician-confirmed asthmayears) and the presence of airflow limitation during the studyho completed questionnaire and lung function tests in at leastr the current study.

0%

10%

20%

30%

40%

50%

60%

70%

NO Asthma

(n=21

09)

Asthma o

nset

<= 25

yrs

(n=19

3)Asth

ma ons

et

> 25y

rs

(n=25

0) NO Asthma

(n=97

0)

Asthma o

nset

<= 25

yrs

(n=88

)Asth

ma ons

et

> 25y

rs

(n=98

)NO Asth

ma

(n=11

39)

Asthma o

nset

<= 25

yrs

(n=10

5)Asth

ma ons

et

> 25y

rs

(n=15

2)

Percen

tag

e o

f p

articip

an

ts

with

airflo

w lim

itatio

n

NON-SMOKERS SMOKERSTOTAL POPULATION

p < 0.001 p < 0.001 p < 0.001

Inconsistent Airflow Limitation Persistent Airflow Limitation

Figure 2 Proportion of subjects with inconsistent and persistent airflow limitation in the groups with no asthma, asthma onset�25 years, and asthma onset >25 years. Data are presented for the total population as well as stratified by the groups of non-smokers and smokers. P values refer to the comparison of proportion of any airflow limitation among subjects with no asthma,asthma onset �25 years, and asthma onset >25 years.

1476 S. Guerra et al.

the associations between asthma and airflow limitationwere significant both among nonsmokers and smokers.

Factors associated with airflow limitation

Table 1 shows the characteristics of the subjects in the ninestudy groups (see methods). Factors potentially associatedwith inconsistent and persistent airflow limitation weretested in multinomial logistic regression models for thetotal population and separately for subjects with no asthma,asthma onset �25 years, and asthma onset >25 years(Table 2). Male sex, age, packyears, elevated IgE, eosino-philia, and asthma status were significantly associatedwith persistent airflow limitation in the total population.The risk for persistent airflow limitation was eight timeshigher for subjects with asthma onset �25 years than sub-jects with no asthma. The corresponding RR was lower(4.9) but still highly significant for subjects with asthmaonset >25 years. Age and smoking were significantly asso-ciated with persistent airflow limitation in all models strat-ified by asthma status. However, the RRs for the associationbetween packyears and persistent airflow limitation weresubstantially higher among subjects with no asthma orasthma onset >25 years than among subjects with asthmaonset �25 years. In contrast, eosinophilia was associatedwith an almost 4-fold increased risk of developing persis-tent airflow limitation only among subjects with asthmaonset �25 years. Of note, in this group the RR for persi-stent airflow limitation associated with eosinophilia wasparticularly high among never smokers (6.8; 1.6e28.9).Table E2 in the online data supplement shows multinomiallogistic regression models for inconsistent and persistentairflow limitation separately for males and females. Theincreased risk for persistent airflow limitation associatedwith eosinophilia appeared stronger among females thanmales, with the corresponding RRs being 2.1 (1.3e3.5)and 1.5 (0.9e2.5), respectively. This trend was particu-larly evident among subjects with asthma onset �25 years.In this group, eosinophilia was associated with an RRof 6.8 (1.6e28.4) among females and an RR of 1.9(0.5e7.3) among males, although the interaction term

between eosinophilia and sex was not significant (data notshown).

Lung function

FEV1 levels during adult age were assessed in the studygroups based on lung function tests completed over age25 (see Fig. E1 for distribution of available observationsby age). Fig. 3 shows that, inherent in the classification,the groups with persistent airflow limitation had lowerFEV1 values as compared with controls, but the natural his-tory of their lung function differed based on the presenceand time of onset of asthma. As compared with controlswith no asthma and no airflow limitation, subjects withno asthma and persistent airflow limitation had only moder-ate deficits of FEV1 at age 25 (mean deficit, 95% CI: 213,16e410 ml), but they had 874 (665e1084) ml excess lossof FEV1 between age 25 and 75. In contrast, subjects withasthma onset �25 years and persistent airflow limitationhad larger deficits of FEV1 at age 25 (904, 637e1171 ml)and a moderate excess loss of FEV1 between age 25 and75 (322, 15e630 ml). Subjects with asthma onset >25 yearsand persistent airflow limitation had a mean 524 (145e904)ml deficit of FEV1 at age 25 and a 763 (372e1154) ml excessloss of FEV1 between age 25 and 75. Linear contrasts indi-cated that, as compared with subjects with no asthmaand persistent airflow limitation, subjects with asthma on-set �25 years and persistent airflow limitation had signifi-cantly lower FEV1 levels at age 25 (predicted values fora 175-cm tall male: 4090 versus 3400 ml, respectively;p< 0.001) and a significantly smaller FEV1 loss betweenage 25 and 75 (2140 versus 1590 ml, respectively;p Z 0.003). Predicted FEV1 values and FEV1 loss are shownfor each of the study groups in Fig. 3. In the online data sup-plement, Fig. E2aed shows predicted FEV1 values from thebest-fitting random coefficients models stratified by sexand smoking.

In order to estimate slopes of FEV1 decline for each ofthe groups with persistent airflow limitation, random coef-ficients models were re-run after excluding the age squaredterm. The expected FEV1 decline in controls with no asthma

Table 1 Characteristics of the subjects in the nine study groups (N is 2552 unless otherwise specified)

No asthma (n Z 2109) Asthma onset Overall p value*

No airflowlimitation

Inconsistentairflowlimitation

Persistentairflowlimitation

�25 years (n Z 193) >25 years (n Z 250)

No airflowlimitation

Inconsistentairflowlimitation

Persistentairflowlimitation

Noairflowlimitation

Inconsistentairflowlimitation

Persistentairflowlimitation

N 1552 360 197 99 45 49 127 55 68 N/ASex % (female) 57.9 55.3 34.5 53.5 62.2 46.9 73.2 58.2 47.1 <0.001Age at first survey in years:

mean� SD45� 18 58� 15 59� 15 35� 14 43� 14 48� 16 43� 17 56� 14 60� 13 <0.001

Number of completed lung functiontests over age 25 years: mean� SD

5.4� 3 7.2� 3 5.9� 3 5.4� 3 7.4� 3 6.4� 3 5.3� 3 6.9� 3 5.7� 3 <0.001

Years of follow-up on lung functionover age 25 years: mean� SD

10.7� 7 13.2� 7 10.9� 7 10.5� 6 14.8� 6 11.6� 7 11.5� 6 13.2� 6 9.8� 7 <0.001

Years of formal education (n Z 2512):% with more than 12 years

55.0 39.9 33.8 69.1 54.5 58.3 59.7 41.8 23.5 <0.001

Chronic bronchitis (n Z 2550)% Never 84.6 67.8 49.7 76.8 53.3 28.6 67.7 56.4 29.4 <0.001% Sporadic 10.8 22.8 32.5 12.1 33.3 26.5 21.3 25.5 27.9% Frequent 4.5 9.4 17.8 11.1 13.3 44.9 11.0 18.2 42.6

Dyspnea (MMRCa score� 2) (n Z 2551)% Never 78.1 64.4 50.8 63.6 46.7 34.7 52.8 45.5 16.2 <0.001% Sporadic 13.8 22.2 21.8 22.2 35.6 22.4 22.8 27.3 22.1% Frequent 8.1 13.3 27.4 14.1 17.8 42.9 24.4 27.3 61.8Physician-confirmed emphysema

(n Z 2551) % ever2.3 9.7 31.0 4.0 13.3 42.9 8.7 25.5 51.5 <0.001

Packyears at last completed survey% <1 packyear 52.3 34.7 17.3 47.5 48.9 38.8 52.0 27.3 25.0 <0.001% 1e19.9 packyears 18.7 15.3 8.1 27.3 8.9 12.2 25.2 20.0 11.8% 20e49.9 packyears 21.6 28.6 26.9 22.2 24.4 26.5 17.3 38.2 22.1% �50 packyears 7.5 21.4 47.7 3.0 17.8 22.4 5.5 14.5 41.2Allergy skin tests (n Z 2390): % ever

positiveb49.6 42.8 45.4 80.2 87.8 68.1 70.1 65.5 42.4 <0.001

Elevated IgE (n Z 2238): % everc 12.4 15.0 21.2 27.2 45.0 46.5 24.5 51.0 27.9 <0.001Eosinophilia (n Z 2116): % everd 12.8 18.2 19.0 23.4 30.0 41.9 24.8 25.5 31.6 <0.001

*p value for the comparison across the nine groups.a Modified Medical Research Council dyspnea scale.33

b Wheal at least 2 mm larger than the control wheal for at least one tested allergen in at least one of the three test surveys.c IgE z score at least one standard deviation higher than the mean of the subject’s sex- and age-specific category in any of the measurements performed during the study.34 For example, this

would correspond to an IgE level >174 IU/ml for a woman between age 25 and 35 years.d Eosinophils >4% or eosinophil z scores at least one standard deviation higher than the sex-specific mean in any of the measurements performed during the study. For example, this would

correspond to eosinophils >274 per mm3 for a woman.

Course

of

persiste

nt

airfl

ow

limita

tion

1477

Table 2 Multinomial logistic regressions to predict inconsistent and persistent airflow limitation in the total population and s arately for subjects with no asthma, asthma onset�25 years, and asthma onset >25 years

Model 1: totalstudy population (n Z 2552)(reference group Z subjects withno airflow limitation)

Model 2: no asthma (n Z 2109)(reference group Z subjects withno asthma and no airflow limitation)

Model 3: asthmaonset �25 years (n Z 19(reference group Z subj ts withasthma onset �25 years nd noairflow limitation)

Model 4: Asthmaonset >25 years (n Z 250)(reference group Z subjectswith asthma onset >25 yearsand no airflow limitation)

InconsistentairflowlimitationRR (95% CI)

PersistentairflowlimitationRR (95% CI)

InconsistentairflowlimitationRR (95% CI)

PersistentairflowlimitationRR (95% CI)

InconsistentairflowlimitationRR (95% CI)

Persi entairflolimit ionRR (9 % CI)

InconsistentairflowlimitationRR (95% CI)

PersistentairflowlimitationRR (95% CI)

Sex (male) 1.14 (0.90e1.45) 1.79 (1.34e2.38) 1.10 (0.84e1.43) 1.82 (1.29e2.58) 0.79 (0.31e2.02) 1.17 .50e2.72) 2.21 (1.00e4.88) 2.49 (1.17e5.26)Age at first surveya 1.67 (1.55e1.81) 1.67 (1.52e1.84) 1.67 (1.53e1.82) 1.59 (1.42e1.78) 1.56 (1.17e2.09) 1.91 .43e2.55) 1.85 (1.45e2.35) 1.98 (1.54e2.54)Packyearsb 1.20 (1.15e1.26) 1.34 (1.27e1.42) 1.20 (1.14e1.26) 1.35 (1.27e1.44) 1.23 (1.03e1.46) 1.22 .05e1.43) 1.28 (1.06e1.53) 1.43 (1.19e1.73)Elevated IgEc 1.35 (0.99e1.82) 1.47 (1.02e2.12) 1.05 (0.72e1.53) 1.44 (0.90e2.32) 1.87 (0.77e4.56) 2.23 .92e5.41) 3.36 (1.52e7.46) 1.45 (0.61e3.45)Eosinophiliad 1.25 (0.92e1.70) 1.74 (1.21e2.51) 1.29 (0.92e1.83) 1.36 (0.84e2.20) 1.72 (0.62e4.72) 3.66 .41e9.54) 1.01 (0.40e2.53) 2.16 (0.84e5.51)

AsthmaOnset �25 years

versus Never3.64 (2.40e5.55) 8.10 (5.11e12.8) N/A N/A N/A N/A N/A N/A

Onset >25 yearsversus Never

2.01 (1.40e2.88) 4.86 (3.36e7.03) N/A N/A N/A N/A N/A N/A

Sex, age, packyears, skin tests, IgE, eosinophilia, and asthma status (in Model 1) were tested as independent variables; all models were lso adjusted for the duration of follow-up; factors thatwere significant in Model 1 were retained in all models.

a RR associated with 10-year increase.b RR associated with 10-packyear increase.c IgE z score at least 1 SD higher than the mean of the subject’s sex- and age-specific category in any of the measurements perfor ed during the study.34

d Eosinophils >4% or eosinophil z scores at least one SD higher than the sex-specific mean in any of the measurements performed d ring the study.

ep

3)eca

stwat5

(0(1(1(0(1

a

mu

1

2

3

4

5P

red

icted

F

EV

1 (L

iters)

fo

r a 175-cm

tall m

an

20 30 40 50 60 70 80

Age in years

No Asthma / No Airflow Limitation

No Asthma / Persistent Airflow LimitationAsthma <= 25 yrs / Persistent Airflow Limitation

Asthma > 25 yrs / Persistent Airflow Limitation

Figure 3 Levels and decline of FEV1 during adult age for thestudy groups of subjects with no asthma and persistent airflowlimitation (black line; number of subjects Z 197; number of ob-servations Z 1159), subjects with asthma onset �25 years andpersistent airflow limitation (red line; number of subjects Z 49;number of observations Z 313), and subjects with asthma onset>25 years and persistent airflow limitation (gold line; number ofsubjects Z 68; number of observations Z 385). Predictedvalues for subjects with no asthma and no airflow limitation(green line; number of subjects Z 1552; number of observa-tions Z 8433) are also reported for comparison. Depicted valuesrepresent predicted values for a 175-cm tall male from the best-fitting random coefficients model.1

Course of persistent airflow limitation 1479

and no airflow limitation was 26 ml per year. Subjects withno asthma and persistent airflow limitation had a signifi-cantly faster FEV1 decline than did subjects with asthmaonset �25 years and persistent airflow limitation or sub-jects with asthma onset >25 years and persistent airflowlimitation (43 [95% CI: 40e46] ml per year, 30 [24e36] mlper year, and 35 [29e40] ml per year, respectively).

We used multivariate logistic regression models to iden-tify factors associated with being in the fastest tertile of FEV1

decline (i.e., having an FEV1 slope of at least 33 ml per year).Only male sex and packyears, but not IgE levels or eosino-philia, were associated with increased odds of being in thefastest tertile of FEV1 decline (adjusted ORs: 1.98, 1.63e2.41, for male sex; and 1.09, 1.05e1.13, for a 10-packyearsincrease). Similarly, male sex and packyears were the onlysignificant risk factors for being in the fastest quintile ofFEV1 decline (i.e., having an FEV1 slope of at least 45 ml

1 Statistical comparisons were conducted only between pairs of groupsbelow

Subjects with noasthma/no airflowlimitation

Subjecasthmairflow

Predicted FEV1 values (95% CI) inliters at age 25

4.31 (4.26e4.35) 4.09 (3

Predicted FEV1 loss (95% CI) inliters between age 25 and 75

1.26 (1.21e1.32) 2.14 (1

*Significantly lower than subjects with no asthma/persistent airflow l**Significantly lower than subjects with no asthma/persistent airflowasthma onset >25 years/persistent airflow limitation (p Z .08).

per year), with the corresponding adjusted ORs being 2.24,1.77e2.84; and 1.10, 1.05e1.15, respectively.

Sensitivity analyses

To evaluate potential loss-to-follow-up bias, analyses wererepeated and results confirmed after selecting for the1149 subjects who had �6 lung function tests (Fig. E3)and after removing observations from survey 13, in whichonly subgroups of participants were eligible to participate(Fig. E4). Results were also unchanged when persistent air-flow limitation was defined as FEV1/FVC ratio consistently<70% plus FEV1 % predicted <80% during or after the firstsurvey with FEV1/FVC< 70% (Fig. E5). Random coefficientsmodels including only observations over age 40 years re-turned very similar trends across the study groups (Fig. E6).

Because the ratio FEV1/FVC declines significantly withaging in the general population, random coefficientsmodels for lung function were also repeated after definingthe study groups based on airflow limitation as assessedby an FEV1/FVC ratio below the lower limit of normal16

(Fig. E7). The main differences in the natural course oflung function associated with persistent airflow limitationbetween subjects with asthma onset �25 years and non-asthmatics were confirmed, with the former having lowerFEV1 levels at age 25 and the latter having greater FEV1

loss between age 25 and 75. In these models, the naturalcourse of lung function of subjects with persistent airflowlimitation was similar between the groups with asthma on-set �25 years and asthma onset >25 years.

Finally, using data from a subgroup of 418 adult subjectswho completed pulmonary function tests before and afteralbuterol in survey 13, we found that 70% (23/33) of subjectswith persistent airflow limitation had GOLD-defined COPD(FEV1/FVC< 70% after bronchodilator) in survey 13. This pro-portion was 60% (9/15) and 78% (14/18) among subjects withand without asthma, respectively.

Discussion

Using a population-based long-term prospective cohort,we identified subjects who had persistent airflow limita-tion during the study and compared risk factors andnatural course of their disease based on the presence andage at onset of asthma. Findings of this study providenovel evidence that (1) subjects with asthma account fora significant proportion of persistent airflow limitation in

with persistent airflow limitation (see methods) and are reported

ts with noa/persistent

limitation

Subjects with asthmaonset �25 years/persistent airflowlimitation

Subjects withasthma onset>25 years/persistentairflow limitation

.90e4.29) 3.40 (3.14e3.67)* 3.78 (3.40e4.16)

.94e2.34) 1.59 (1.29e1.89)** 2.03 (1.64e2.42)

imitation (p< .001).limitation (p Z .003) and borderline lower than subjects with

1480 S. Guerra et al.

the general population; (2) the profile of risk factors fordeveloping persistent airflow limitation is dependentupon the presence and the age at onset of asthma,with blood eosinophilia being the strongest risk factoramong subjects with asthma onset before age 25 years;(3) among patients who have asthma before age 25,development of persistent airflow limitation in adult ageis mainly associated with deficits that are alreadyestablished by the time these subjects reach the growthplateau of their lung function.

Factors associated with persistent airflowlimitation

Consistent with the well-known inverse relationship be-tween the ratio FEV1/FVC and age,18,19 aging was stronglyassociated with airflow limitation both among subjectswith and without asthma. However, lifetime exposure tocigarette smoking was associated with an increased riskfor persistent airflow limitation that was almost twice ashigh among subjects with no asthma or asthma onset >25years than among subjects with asthma onset �25 years.In contrast, among subjects with asthma onset �25 years,eosinophilia was the strongest risk factor for persistent air-flow limitation. Eosinophilia is a known risk factor for thedevelopment of respiratory symptoms20,21 and, among asth-matics, is associated with lung function impairment22,23

and mortality risk.24,25 In a cross-sectional study of 132 non-smoking outpatients with severe asthma, sputum eosino-phils �2% were associated with an almost 8-fold increasein the odds of having postbronchodilator FEV1 or FEV1/FVC< 75% predicted.22 Similarly, among children withasthma in the Childhood Asthma Management Program co-hort levels of sputum and circulating eosinophils werehigher in subjects with at least 1% per year loss in FEV1 %predicted than in subjects with non-significant FEV1

loss.23 However, the present study is the first population-based study showing that eosinophilia predicts develop-ment of persistent airflow limitation among adults whohad asthma onset in their first 25 years of life. Our findingssupport a major role for eosinophilia in the smoking-inde-pendent component of the phenotypic overlap betweenasthma and COPD, as suggested by the particularly strongassociation between eosinophilia and persistent airflow lim-itation among non-smoker asthmatics.

Natural history of lung function

Because FEV1 represents the numerator of the ratio ‘FEV1/FVC’, levels of FEV1 are expected to be reduced in most pa-tients with airflow limitation. However, in adults these FEV1

deficits can be the result of any combination betweenlower FEV1 levels at the beginning of adult age or an accel-erated FEV1 decline during adulthood.26 One of the majorfindings of our study is that, among subjects who hadasthma onset �25 years, persistent airflow limitation inadult age was strongly associated with FEV1 deficits thatwere already established by young adulthood. Whetherthese deficits are related to early airway remodeling orimpaired lung development, or both, remains to be deter-mined. These findings are consistent with those of several

other prospective studies e including the Dunedin Multidis-ciplinary Health and Development Study,18 the MelbourneAsthma Study,27 and the British 1958 Birth Cohort28 e inwhich persistent childhood asthma was associated withlung function deficits that are present before adult life be-gins and track over time from childhood to mid-adult life. Inthe European Community Respiratory Health Survey, per-cent predicted FEV1 values at age 20e44 years were strongpredictors of asthma severity and persistence during thefollow-up period.29 In addition, in the birth cohort of theChildren’s Respiratory Study our group has previously foundthat children who start life with low levels of lung functionhave lower expiratory flows throughout childhood30 as wellas lower FEV1/FVC levels at age 2231 as compared with theirpeers, suggesting that early exposures and/or genetic fac-tors also play a role in determining levels of lung functionthat will be attained before entering adult life. In ourstudy, the impact of asthma on early deficits of lung func-tion appeared stronger among males (Fig. E2a), who areknown to be at increased risk for severe and/or persistentchildhood asthma as compared with females, and amongnonsmokers (Fig. E2c).

Conversely, subjects with no asthma developed persis-tent airflow limitation mainly in response to cigarettesmoking and, thus, they showed only mild FEV1 deficits atage 25 but steeper FEV1 decline over adult life. The naturalcourse of lung function of subjects who had asthma onsetafter age 25 years included both moderate FEV1 deficitsin young adulthood and accelerated FEV1 decline thereaf-ter, although these results should be interpreted with cau-tion because of the possibility of reverse causality (i.e.,acquiring a diagnosis of adult-onset asthma might be theconsequence rather than the cause of the steep declineof lung function) and the small number of observations be-fore age 40 in this group.

The different natural history of lung function betweensubjects who developed persistent airflow limitation withor without asthma is consistent with the observation thateosinophilia, the strongest risk factor for persistent airflowlimitation among subjects with asthma onset �25 years,was not associated with FEV1 decline. This observation sug-gests that eosinophilia influences lung function among sub-jects with childhood asthma mainly by affecting FEV1 levelsthat are attained before young adult age.

Within our study design, we were unable to studypotential effects of anti-inflammatory treatments32 onasthma progression. In addition, the small sample size ofsome of our study groups suggests the importance of otherstudies to replicate these findings. As with most largeepidemiological studies, postbronchodilator lung functiontests were not available for the vast majority of TESAODparticipants. Thus, to what extent persistent airflow limita-tion can be used as a surrogate of GOLD-defined COPD is notknown, although sensitivity analyses from a subgroup of 418TESAOD participants suggested acceptable correlations be-tween these two phenotypes. We also acknowledge that,although a maximum 24-year follow-up was possible in TE-SAOD, subjects enrolled in this study were followed on av-erage for 11 years. Thus, we studied the natural course ofthe disease over adult life by combining information frommultiple sub-cohorts of subjects who differed in age atenrollment into the study and length of follow-up.

Course of persistent airflow limitation 1481

However, this is an unavoidable limitation of most largeprospective studies in adults. The relatively homogeneousdistribution of available observations over the entire spanof adult life (Fig. E1) and the use of statistical techniquesthat are specifically designed for analysis of unbalancedlongitudinal data17 should have minimized the impact ofthis limitation in our study.

Conclusions

In summary, development of persistent airflow limitation inadult patients with asthma is a common event and accountsfor a significant proportion of the public health burden ofobstructive lung disease. We showed that blood eosino-philia is the strongest risk factor for the development ofpersistent airflow limitation among patients with asthmaonset �25 years and that, among these patients, the bulk oflung function impairment is already established by youngadulthood. Therefore, future prevention programs willneed to identify and target these patients before theyenter adult life.

Conflict of interest statement

SG, DLS, MK-S, CV, MH, and SFQ have no conflicts of interestto disclose in relation to this manuscript. FDM has servedon the Merck Advisory Board and as a consultant forGlaxoSmithKline, Pfizer and Genentech. In the last threeyears, he has also received lecture fees from speaking atthe invitation of Merck and Genentech. In each of the lastthree years, he has been selected as the Pfizer VisitingScholar, a program meant to increase opportunities forscientific exchange and education at medical schools,teaching hospitals and other organizations. Values of eachservice are determined by quantity of time and effortrequired.

Acknowledgements

The TESAOD was initiated in 1972 as a result of the long-term vision and leadership of Benjamin Burrows, MD. Wegratefully acknowledge Dr Burrows’ enduring contribu-tions to our understanding of chronic obstructive lungdiseases. We are also thankful to Dr Michael D Lebowitzwho had an essential role in the design and implementa-tion of the TESAOD study and to Dr Ronald Knudson whodirected pulmonary function testing. Finally, we thankBobbe Boyer, RN who has caringly followed the cohortsubjects since 1972 and the several thousands of TESAODparticipants for their extraordinary dedication to thestudy.

Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.rmed.2008.04.011.

References

1. Burrows B, Bloom JW, Traver GA, Cline MG. The course andprognosis of different forms of chronic airways obstruction ina sample from the general population. N Engl J Med 1987;317:1309e14.

2. Global initiative for asthma. Global strategy for asthmamanagement and prevention 2006.

3. Vonk JM, Jongepier H, Panhuysen CI, Schouten JP, Bleecker ER,Postma DS. Risk factors associated with the presence of irre-versible airflow limitation and reduced transfer coefficient inpatients with asthma after 26 years of follow up. Thorax2003;58:322e7.

4. Kesten S, Rebuck AS. Is the short-term response to inhaledbeta-adrenergic agonist sensitive or specific for distinguishingbetween asthma and COPD? Chest 1994;105:1042e5.

5. Ulrik CS, Backer V. Nonreversible airflow obstruction in life-long nonsmokers with moderate to severe asthma. Eur RespirJ 1999;14:892e6.

6. Hansen EF, Phanareth K, Laursen LC, Kok-Jensen A, Dirksen A.Reversible and irreversible airflow obstruction as predictor ofoverall mortality in asthma and chronic obstructive pulmonarydisease. Am J Respir Crit Care Med 1999;159:1267e71.

7. Soriano JB, Davis KJ, Coleman B, Visick G, Mannino D, Pride NB.The proportional Venn diagram of obstructive lung disease:two approximations from the United States and the UnitedKingdom. Chest 2003;124:474e81.

8. Mannino DM. COPD: epidemiology, prevalence, morbidity andmortality, and disease heterogeneity. Chest 2002;121:121Se6S.

9. Silva GE, Sherrill DL, Guerra S, Barbee RA. Asthma as a risk factorfor COPD in a longitudinal study. Chest 2004;126:59e65.

10. Rabe KF, Hurd S, Anzueto A, Barnes PJ, Buist SA, Calverley P,et al. Global strategy for the diagnosis, management, and pre-vention of COPD e 2006 update. Am J Respir Crit Care Med2007;176(6):532e55.

11. Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS.Global strategy for the diagnosis, management, and preven-tion of chronic obstructive pulmonary disease. NHLBI/WHOGlobal Initiative for Chronic Obstructive Lung Disease(GOLD) workshop summary. Am J Respir Crit Care Med2001;163:1256e76.

12. Kraft M. Asthma and chronic obstructive pulmonary disease ex-hibit common origins in any country!. Am J Respir Crit CareMed 2006;174:238e40 [discussion 44].

13. Lebowitz MD, Knudson RJ, Burrows B. Tucson epidemiologicstudy of obstructive lung diseases. I: methodology and preva-lence of disease. Am J Epidemiol 1975;102:137e52.

14. ATS statement e snowbird workshop on standardization ofspirometry. Am Rev Respir Dis 1979;119:831e8.

15. Burrows B, Knudson RJ, Lebowitz MD. The relationship of child-hood respiratory illness to adult obstructive airway disease.Am Rev Respir Dis 1977;115:751e60.

16. Hankinson JL, Odencrantz JR, Fedan KB. Spirometric referencevalues from a sample of the general U.S. population. Am J Re-spir Crit Care Med 1999;159:179e87.

17. Brown H, Prescott R. Applied mixed models in medicine. Chi-chester, UK: John Wiley & Sons, Ltd; 2001.

18. Sears MR, Greene JM, Willan AR, Wiecek EM, Taylor DR,Flannery EM, et al. A longitudinal, population-based, cohortstudy of childhood asthma followed to adulthood. N Engl JMed 2003;349:1414e22.

19. Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B. Changes inthe normal maximal expiratory flowevolume curve withgrowth and aging. Am Rev Respir Dis 1983;127:725e34.

20. Mensinga TT, Schouten JP, Rijcken B, Weiss ST, Speizer FE, vander Lende R. The relationship of eosinophilia and positive skintest reactivity to respiratory symptom prevalence in

1482 S. Guerra et al.

a community-based population study. J Allergy Clin Immunol1990;86:99e107.

21. Jansen DF, Schouten JP, Vonk JM, Rijcken B, Timens W, Kraan J,et al. Smoking and airway hyperresponsiveness especially inthe presence of blood eosinophilia increase the risk to developrespiratory symptoms: a 25-year follow-up study in the generaladult population. Am J Respir Crit Care Med 1999;160:259e64.

22. ten Brinke A, Zwinderman AH, Sterk PJ, Rabe KF, Bel EH. Fac-tors associated with persistent airflow limitation in severeasthma. Am J Respir Crit Care Med 2001;164:744e8.

23. Covar RA, Spahn JD, Murphy JR, Szefler SJ. Progression ofasthma measured by lung function in the childhood asthmamanagement program. Am J Respir Crit Care Med 2004;170:234e41.

24. Ulrik CS, Frederiksen J. Mortality and markers of risk of asthmadeath among 1,075 outpatients with asthma. Chest 1995;108:10e5.

25. Hospers JJ, Schouten JP, Weiss ST, Rijcken B, Postma DS.Asthma attacks with eosinophilia predict mortality fromchronic obstructive pulmonary disease in a general populationsample. Am J Respir Crit Care Med 1999;160:1869e74.

26. Weiss ST, Ware JH. Overview of issues in the longitudinal anal-ysis of respiratory data. Am J Respir Crit Care Med 1996;154:S208e11.

27. Phelan PD, Robertson CF, Olinsky A. The Melbourne asthmastudy: 1964e1999. J Allergy Clin Immunol 2002;109:189e94.

28. Marossy AE, Strachan DP, Rudnicka AR, Anderson HR. Childhoodchest illness and the rate of decline of adult lung functionbetween ages 35 and 45 years. Am J Respir Crit Care Med2007;175(4):355e9.

29. de Marco R, Marcon A, Jarvis D, Accordini S, Almar E,Bugiani M, et al. Prognostic factors of asthma severity:a 9-year international prospective cohort study. J AllergyClin Immunol 2006;117:1249e56.

30. Morgan WJ, Stern DA, Sherrill DL, Guerra S, Holberg CJ,Guilbert TW, et al. Outcome of asthma and wheezing in thefirst 6 years of life: follow-up through adolescence. Am JRespir Crit Care Med 2005;172:1253e8.

31. Stern DA, Morgan WJ, Wright AL, Guerra S, Martinez FD. Poorairway function in early infancy and lung function by age 22years: a non-selective longitudinal cohort study. Lancet 2007;370:758e64.

32. Lange P, Scharling H, Ulrik CS, Vestbo J. Inhaled corticosteroidsand decline of lung function in community residents withasthma. Thorax 2006;61:100e4.

33. Celli BR, Cote CG, Marin JM, Casanova C, Montes de Oca M,Mendez RA, et al. The body-mass index, airflow obstruction,dyspnea, and exercise capacity index in chronic obstructivepulmonary disease. N Engl J Med 2004;350:1005e12.

34. Burrows B, Martinez FD, Halonen M, Barbee RA, Cline MG.Association of asthma with serum IgE levels andskin-test reactivity to allergens. N Engl J Med 1989;320:271e7.