exhaled breath condensate cytokine patterns in chronic obstructive pulmonary disease

ARTICLE IN PRESS

Respiratory Medicine (2005) 99, 1229–1240

KEYWORDCytokines;Chronic obpulmonaryExhaled brcondensatFlow cytom

0954-6111/$ - sdoi:10.1016/j.r

Abbreviationchronic obstrucpulmonary diseobstructive pulcapacity; FEV1/volunteers; ICSInterleukin 10;metalloproteinResidual volumcapacity; TNF-a�CorrespondiE-mail addr

Exhaled breath condensate cytokine patterns inchronic obstructive pulmonary disease

Christian Gessnera,�, Robert Scheibeb, Michael Wotzelb,Stefan Hammerschmidta, Hartmut Kuhna, Lothar Engelmanna,Gerhard Hoheisela, Adrian Gillissenc, Ulrich Sackb, Hubert Wirtza

aDepartment of Respiratory Medicine, University of Leipzig, Johannisallee 32, 04103 Leipzig, GermanybInstitutes of Clinical Immunology and Transfusion Medicine, University of Leipzig, Germany, Johannisallee 30, 04103 Leipzig, GermanycPneumology, Robert-Koch-Hospital, Nikolai-Rumjanzew-Strasse, 100, 04207 Leipzig, Germany

Received 31 December 2004

S

structivedisease;eathe;etry

ee front matter & 2005med.2005.02.041

s: AECOPD, Acute exactive pulmonary diseasease with intensive caremonary disease; EBC, EFVC, Ratio of the force, Inhaled corticosteroidsIL-12p70, Interleukin 1ase 9; PaCO2, Carbon de; SLIP, Antiproteases se, Tumor necrosis factong author. Tel.: +49 341ess: [email protected]

Summary Differences in cytokine patterns in stable chronic obstructive pulmonarydisease (COPD), exacerbated COPD, smokers without apparent COPD, and healthyvolunteers should be of interest for pathophysiological and therapeutic reasons.Methods including lavage, biopsy and sputum have been employed to investigatecytokines in the lung. For asystematic comparison, exhaled breath condensate (EBC)appears to be well suited.

We investigated healthy volunteers, smokers without apparent COPD,stable and exacerbated COPD patients (7 inhalative steroids) and finallythose whose exacerbation made mechanical ventilation inevitable, for amore complete picture of inflammatory cytokines in COPD. We chose EBCbecause it is non-invasive and can be used repeatedly in spontaneous breathingindividuals and during mechanical ventilation. EBC cytokines (IL-1b; IL-6, IL-8, IL-10,IL-12p70, TNF-a) were assayed from a single sample using a multiplex arraytest kit.

Elsevier Ltd. All rights reserved.

erbation of a chronic obstructive pulmonary disease; AECOPD-GW, Acute exacerbation of awith general ward treatment; AECOPD-ICU, Acute exacerbation of a chronic obstructiveunit treatment; BAL, Bronchoalveolar lavage; BALF, Bronchoalveolar lavage fluid; COPD, Chronicxhaled breath condensate; FEV1, Forced expiratory volume in the first 1 s; FVC, Forced vitald expiratory volume in the first 1 s to the forced vital capacity of the lung; HS, Healthy smoking; ICU, Intensive care unit; IL-1b, Interleukin 1 beta; IL-6, Interleukin 6; IL-8, Interleukin 8; IL-10,2p70; LABA, Long acting beta agonists; MMP-1, Matrix metalloproteinase 1; MMP-9, Matrixioxide partial pressure in arterial blood; PaO2, Oxygen partial pressure in arterial blood; RV,cretory leukoprotease inhibitor; TIMP-1, Tissue inhibitor of metalloproteinase 1; TLC, Total lungr alpha; VOL, Healthy non-smoking volunteers.971 2674; fax: +49 341 971 2609.

e (C. Gessner).

ARTICLE IN PRESS

C. Gessner et al.1230

We observed a significant increase of all cytokines in acute exacerbationcompared to stable COPD, smokers, and volunteers. Stable COPD and volunteersexhibited only small differences in cytokine pattern with respect to IL-1b and IL-12ðPo0:01Þ: Smokers had increased levels of all investigated cytokines ðPo0:01Þcompared to non-smokers and, with the exception of IL-1b; to stable COPD. Inhaledsteroids resulted in reduced levels of IL-1b; IL-6, IL-8, IL-10, and IL-12 (all: Po0:01)in stable COPD (all: ex-smokers) with dose dependency for IL-8, IL-1b and IL-12.

EBC analysis successfully characterized important differences in stable COPDcompared to exacerbation or smoking and non-smoking healthy individuals.& 2005 Elsevier Ltd. All rights reserved.

Introduction

Chronic obstructive pulmonary disease (COPD) ischaracterized by airflow obstruction due to chronicbronchitis and/or emphysema1 which is progressiveand associated with an abnormal inflammatoryresponse of the lung. The final result of this processis destruction and loss of lung parenchyma. Variousmethods have been employed to investigate andpossibly quantify inflammation in COPD. Amongthem invasive methods such as biopsy of thebronchial walls2,3 as well as bronchoscopy andbronchoalveolar lavage (BAL) with subsequentexamination of cells and soluble markers.4–8 In-duced sputum, a less invasive method, has been thesource for many investigations aimed at describinginflammation.9–13 Exhaled breath condensate (EBC)constitutes another, entirely non-invasive methodwhich has been used for the same purpose.14–18

Different stages and variations of the process calledCOPD were investigated, with a majority of thestudies investigating stable disease9,10,12,16

although in some cases acute exacerbations ofCOPD (AECOPD) were also studied.19–22 The mostsevere and acute form of COPD, AECOPD that leadsto ICU treatment and the need for invasivemechanical ventilation has also been studied bythis technique with respect to pH, cytokines andnitrite.18 The different situations of exacerbationof COPD as well as stable disease were variablycompared to either healthy volunteers and/orsmokers without apparent signs of COPD.6–8,17 Inthis study we wanted to characterize the cytokinepattern of inflammation in stable and exacerbatedCOPD (stable COPD, AECOPD with general wardtreatment [AECOPD-GW], AECOPD with ICU treat-ment [AECOPD-ICU]) and compare it to healthyvolunteers as well as smokers without apparentCOPD. EBC was used to investigate the cytokinepatterns, because this material is non-invasive andsampling does not lead to any alteration of theinflammatory situation of the airways. In this studyEBC cytokine measurement proofed to be useful inestimating the inflammatory status of the different

patient groups. Volunteers, smokers, stable COPDpatients and exacerbated COPD patients all ex-hibited significant differences. The understandingof these patterns may facilitate therapeutic deci-sions.

Methods

Study subjects and clinical scores

EBC was collected from:

(a)
ICU patients with life-threatening exacerbationof COPD on mechanical ventilation (AECOPD-ICU),23 n ¼ 11:
(b)
Patients with severe and very severe exacer-bated COPD with the need for hospital admis-sion in a general ward (AECOPD-GW),23 n ¼ 34:
(c)
Patients with known and stable COPD (stableCOPD), n ¼ 40: (Stable COPD was defined by thelack of symptoms typical for an acute exacer-bation and no need for a change in medicationfor at least 8 weeks prior to presentation.)
(d)
Healthy smoking volunteers (HS), n ¼ 21: (e) Healthy non-smoking volunteers (VOL), n ¼ 24:
Patient’s characteristics are depicted in Table 1.Patients on mechanical ventilation were included inthis investigation at 8–48 h of ventilation. Mechan-ical ventilation was performed according to guidelines for protective ventilation with a tidal volumeat 6ml/kg ideal body weight.24 Ventilated patientsfulfilled criteria described by Burge and co-work-ers.23 EBC of patients with AECOPD was collectedbefore therapy at the time of diagnosis. AllAECOPD-GW patients exhibited criteria for admis-sion to a hospital as has been suggested23:respiratory rate 425/min; pulse rate 4110/min;PaO2o8 kPa; abnormal chest radiograph; seriousconcomitant disease; altered mental state; livingalone. All AECOPD patients (ICU and GW) weretreated with i.v. antibiotics. EBC collection was

ARTICLE IN PRESS

Table 1 Patient’s characteristics.

Patients VOL HS Stable COPD AECB-GW AECB-ICU

Total number 24 21 40 34 11Age (mean7SD)� 54.7710.5 59.0712.2 60.3715.8 62.1712.3 62.779.6SexMale 13 12 27 24 6Female 11 9 13 10 5

Smoking statusy Non-smoker Smoker Ex-smoker Ex-smoker Ex-smokerPack years 0 2675 3079 2877 2775FEV1 (l)z 3.7970.46 3.3870.58 1.3370.60 1.3370.59 1.2170.38FVC (l)z 4.1470.49 3.8970.58 2.5370.96 2.2870.82 2.3670.76FEV1/FVC (%)z 91.472.50 86.774.46 51.977.61 57.279.36 52.678.11PaO2 (kPa)y 10.0970.40 9.4470.40 7.3770.46 6.6370.28 5.3670.61PaCO2 (kPa)y 5.1570.29 5.4070.31 5.8470.31 6.5870.24 8.5870.63Protein in EBC (mg/ml) (mean7SD)� 13.978.9 13.475.5 11.574.4 10.775.3 14.479.1ICSz therapyYes 0 0 21 20 0No 24 21 19 14 11

All data are shown as mean 7SD.�P40.05 (no significant difference between investigated groups) in ANOVA test.ySmoking was defined as current smokers or ex-smokers that discontinued smoking no longer than 12 month, non-smoking as

no smoking longer than 1 year.zResults were measured after AECB.yResults were measured at time of hospitalization.zICS: inhalative corticosteroid.

Exhaled breath condensate cytokine in COPD 1231

performed between 8 and 36 h following hospitali-zation. At the time of EBC collection corticoster-oids had not yet been given. AECOPD diagnosis wasbased on criteria described by Anthonisen25: pre-sence of at least one of the following three majorsymptoms: increase in dyspnoe, sputum volumeincrease, sputum change to purulence and at leastone of the following minor symptoms: cough,wheeze, sore throat, nasal discharge, fever. Allpatients with stable COPD, AECOPD-GW, andAECOPD-ICU were nonsmokers for at least 1 yearaccording to their own saying.

Lung function was performed on the day of EBCcollection in all stable patients, within 1 week inAECOPD-GW patients and within 10 days followingextubation in AECOPD-ICU patients. Capillary bloodgas analysis was performed on all patients withinthe first few hours after admission.

All COPD patients were staged according tocurrent guidelines (GOLD), (1 update 2003) as stageIII COPD or stage IV COPD. None of the patients inthis series was treated with an oral steroid. Allpatients were on oral (theophylline) and inhaledtherapy (LABA and/or long acting anticholinergicand/or inhaled corticosteroids (ICs)) according tothe guidelines. All patients were regularly seen bypulmonary specialists. Approval for this investiga-tion was obtained from the ethics committee of theUniversity of Leipzig.

EBC collection and markers

EBC was collected by inserting a special conduit(FILT lung and Chest Diagnostics Ltd., Berlin,Germany) for the EcoScreens breath condensatecollecting device (ViaSys, Hoechberg, Germany)into the expiratory limb of the ventilator tubingdirectly after the Y-shaped connecting piece for a20min time period. Humidification of inspiratorygas was achieved using heat humidifiers.

EBC from spontaneously breathings was similarlycollected for 20min with the EcoScreens system aspreviously described.26

All EBC samples were examined for amylaseactivity (alpha-Amylase ESP1491300 kit; detectionlimit 0.05 mmol/l s; Boehringer Mannheim, Ger-many) in order to exclude contamination by saliva.

Protein concentration in EBC was measured usingthe Micro BCA Protein Assay (Pierce, Rockford USA).

Bronchoalveolar lavage (BAL)

When BAL was scheduled for microbiological reasons inAECOPD patients, an aliquot (2ml) was taken for thecomparison of cytokines in BAL fluid (BALF) and EBC. Acomparison of BALF and EBC was thereby achieved in14 cases. In all cases, EBC collection was performedprior to bronchoscopy. BAL was done according to

ARTICLE IN PRESS


published guidelines27 (5� 20ml aliquots). The cellfree supernant was used for further analysis.

Cytometric bead array

A multiplex fluorescent bead immunoassay (cyto-metric bead array [CBA] Becton Dickinson, SanJose, CA, USA) has been adapted to analysis inbreath condensate to detect cytokine concentra-tions. A mixture of six bead populations withdistinct fluorescence intensities and coated withcapture antibodies specific for IL-8, IL-1b; IL-6, IL-10, TNF-a; and IL-12p70 proteins were incubatedwith 2ml of lyophilized breath condensate recon-stituted with 50 ml of ddH2O (duplicate sampleswere prepared). Cytokines in EBC samples andrecombinant standards bound to capture beadswere detected by PE-conjugated detection anti-bodies in a flow cytometer (FACS CaliburTM, BectonDickinson). Effects of cryoconservation, lyophiliza-tion, and reduction of reagent amounts as well asrequirements to buffer composition were investi-gated in a group of 9 samples in order to ensurecorrect detection of cytokines by this commerciallyavailable system: defined cytokine concentrationswere dissolved in water and measured before andafter cryoconservation and lyophilization. A loss of10.6% of activity was observed. By adding 1% of BSAthis loss was reduced to 1.8%.

Mean intra-assay reproducibility and inter-assayreproducibility were 92.8% (IL-8: 89.5%; IL-1b:94.5%; IL-6: 90.7%; IL-10: 96.5%; TNF-a: 91.6%; IL-12: 93.9%) and 87% (IL-8: 80.6%; IL-1b: 85.7%; IL-6:90.4%; IL-10: 92.7%; TNF-a: 88.9%; IL-12: 83.8%),respectively.

Statistical analysis

Statistical analysis was performed with the SPSSsoftware package (SPSS Inc., Chicago, USA). Linear

Table 2 Detection rates of cytokines in EBC.

Marker Detection rates for cytokines

VOL HS Stable COPD

d/nd (%) d/nd (%) d/nd (%)

IL-8 14 (58)/10 (42) 19 (90)/2 (10) 31 (77)/9 (23IL-1b 19 (79)/5 (21) 21 (100)/0 (0) 39 (97)/1 (3)IL-6 11 (46)/13 (54) 20 (95)/1 (5) 15 (37)/25 (6IL-10 13 (54)/11 (46) 19 (90)/2 (10) 31 (77)/9 (23TNF-a 16 (67)/8 (33) 20 (95)/1 (5) 33 (82)/7 (18IL-12p70 19 (79)/5 (21) 21 (100)/0 (0) 37 (92)/3 (8)

d, cytokine detectable; nd, cytokine not detectable.

regression analysis was applied to investigate thecorrelation of cytokine levels in EBC and BALF.Comparison of patient groups (three or more) wasperformed by Kruskal–Wallis test and Mann–Whit-ney test. Data in box blots are median 725/75%(box) and 5/95% confidence intervals, with outlyingdata points indicated as closed circles. Statisticalsignificance was accepted at the 5% level.

Results

Patients’ characteristics

Patient characteristics are listed in Table 1.

General exhaled breath condensatecharacteristics

None of the condensate samples exhibited amylaseconcentrations measurable with the assay used.Therefore a relevant saliva contamination can beexcluded because amylase concentration in salivais proximally 10,000 times higher than those in EBCwith respect to detection limit of assays.28,29

Total EBC protein concentration was regularlymeasured. Results are shown in Table 1. There wasno significant difference in any of the subgroupsanalyzed ðP ¼ 0:27Þ:

Inflammatory cytokines in EBC

The rates of detection of the various cytokines inall of the patient subgroups are listed in Table 2. InAECOPD-ICU and AECOPD-GW patients all cytokinelevels were considerably higher and cytokines weremore often detectable than in patients with stableCOPD, healthy smokers or healthy volunteers (Fig.1, Table 2). Cytokines in relation to protein in EBCshowed comparable results shown in Fig. 2.

AECOPD-GW AECOPD-ICU All patients

d/nd (%) d/nd (%) d/nd (%)

) 34 (100)/0 (0) 11 (100)/0 (0) 109 (84)/21 (16)34 (100)/0 (0) 11 (100)/0 (0) 124 (95)/6 (5)

3) 33 (97)/1 (3) 10 (91)/1 (9) 89 (68)/41 (32)) 34 (100)/0 (0) 11 (100)/0 (0) 108 (83)/22 (17)) 34 (100)/0 (0) 11 (100)/0 (0) 114 (88)/16 (12)

34 (100)/0 (0) 11 (100)/0 (0) 122 (94)/8 (6)

ARTICLE IN PRESS

Figure 1 IL-8, IL-1b; IL-6, IL-10, TNF-a; and IL-12p70 levels in EBC of patients at ICU with life-threatening exacerbationof COPD on mechanical ventilation (AECOPD-ICU: n ¼ 11), patients with severe and very severe exacerbated COPD withthe need for hospital admission in a general ward (AECOPD-GW: n ¼ 34), patients with known and stable COPD (stableCOPD: n ¼ 40), healthy smoking volunteers (HS: n ¼ 21), and healthy non-smoking volunteers (VOL: n ¼ 24). Results aremedian 725/75% (box) and 5/95% confidence intervals, with outliers indicated as closed circles.


Differences between stable COPD and healthysmokers were observed in all cytokines investigatedwith the exception of IL-1b: Healthy smokersshowed significantly increased levels for all cyto-kines compared with healthy volunteers. For stableCOPD and healthy volunteers differences were onlyobserved for IL-1b and IL-12p70 (Figs. 1 and 2).

Subgroup analysis of COPD stage III versus COPDstage IV in both the AECOPD as well as in the stableCOPD patient groups did not reveal significant

differences with respect to cytokines analyzed(data not shown).

Correlation of EBC cytokines with lungfunction or ventilatory parameters

Results of lung function as well as blood gas analysisare shown in Table 1 for the following parameters:FEV1, FVC, FEV1/FVC, PaO2, and PaCO2. There was

ARTICLE IN PRESS

Figure 2 IL-8, IL-1b; IL-6, IL-10, TNF-a; and IL-12p70 levels in EBC in relation to protein in EBC of patients with life-threatening exacerbation of COPD on mechanical ventilation (AECOPD-ICU: n ¼ 11), patients with severe and verysevere exacerbated COPD with the need for hospital admission in a general ward (AECOPD-GW: n ¼ 34), patients withknown and stable COPD (stable COPD: n ¼ 40), healthy smoking volunteers (HS: n ¼ 21), and healthy non-smokingvolunteers (VOL: n ¼ 24). Results are median 725/75% (box) and 5/95% confidence intervals, with outliers indicated asclosed circles.


no strong, significant correlation observed betweenlung function parameters VC, Rtot, FEV1, FVC, FEV1/FVC, TLC, RV, PaO2, and PaCO2 and the level ofcytokines in EBC of either all patients or any of thesubgroups. This same result was also true for all the

parameters of ventilation examined in this study:positive end expiratory pressure (6.7mbar73.4),peak inspiratory pressure (20.6mbar73.3), tidalvolume adapted to ideal body weight (6.3ml/kgbody weight71.0), expiratory minute volume

ARTICLE IN PRESS


(9.8 l/min72.5) and breathing frequency(21.574.5) when correlated to cytokine levels inventilated patients.

Influence of inhaled corticosteroid in COPDpatients at cytokines in EBC

Stable COPD patients ðn ¼ 40Þ and patients withAECOPD-GW ðn ¼ 34Þ were evaluated for theinfluence of ICS on the level of cytokines in EBC.

Figure 3 EBC cytokines in patients with severe and very severin a general ward (AECOPD) with ðn ¼ 18Þ and without ðn ¼ 1(stable COPD) with ðn ¼ 21Þ and without ðn ¼ 19Þ ICS therapywith outliers indicated as closed circles).

We found lower levels for IL-1b; IL-6, IL-8, IL-12,and IL-10 (all: Po0:01) but not for TNF-a in patientswith stable COPD on ICS ðn ¼ 21Þ compared withstable COPD patients without ICS (n ¼ 19; Fig. 3).ICS-dose dependence (lower dose p1000 mg; higherdose 41000 mg/day) was recognized for IL-8, IL-1band IL-12 (Fig. 4). In contrast to these results,patients with AECOPD exhibited no difference incytokine levels depending on the use of ICS (Fig. 3).Although a lower level of EBC cytokines wasprominent in stable COPD patients on ICS, no

e exacerbated COPD with the need for hospital admission1Þ ICS therapy and patients with known and stable COPD(median 725/75% [box] and 5/95% confidence intervals,

ARTICLE IN PRESS

Figure 4 Cytokine levels in EBC of stable COPD patients on p1000mg ðn ¼ 10Þ; and 41000 ðn ¼ 11Þ of dailybeclomethasone equivalent of inhaled steroid (median 725/75% [box] and 5/95% confidence intervals, with outliersindicated as closed circles).


change in lung function parameters was apparent incomparison with stable COPD patients without ICS.

Correlation of cytokine levels in BALF andEBC

The correlation of BALF ingredients with the samemolecules in EBC is largely unknown. We found nosignificant correlation between cytokine levels inBALF and those in EBC (Table 3). However theresults were not equal for all of the cytokinesinvestigated and clearly the correlation was closer

for IL-12p70 (r ¼ 0:52; P ¼ 0:06; Fig. 5) than for anyof the remaining cytokines investigated. A largernumber of samples might well have demonstrated asignificant correlation in this case and possibly alsofor IL-6 and IL-10.

Discussion

This study was aimed at characterizing infla-mmation by analyzing cytokine profiles inhealthy individuals, smokers, stable COPD, acute

ARTICLE IN PRESS

Table 3 Correlation of EBC cytokines and BALFcytokines in a subgroup of 14 patients.

Marker Correlation of EBC and BALF

R P-value

IL-8 0.03 0.93IL-1b 0.06 0.83IL-6 0.40 0.15IL-10 0.41 0.15TNF-a 0.21 0.46IL-12p70 0.52 0.06

18

BA

L IL

-12p

70 (

pg/m

l)

16

14

12

10

8

6

4

2

00 100 200 300 400 500

EBC IL-12p70 (pg/ml)

Figure 5 Correlation of IL-12p70 in EBC with concentra-tion in BAL (r ¼ 0:52; P40:06) in a subgroup of 14patients.


exacerbation without and with the need formechanical ventilation. We chose to measurecytokine profiles in EBC because this material isacquired non-invasively. Investigating a potentialinfluence of ICS on the cytokine profiles was asecondary aim of our study.

Chronic inflammation, although different to thatin asthma is the predominant feature of COPD. Thenature of this inflammatory process has beenthoroughly investigated by, e.g. invasive methodssuch as biopsies and bronchoscopy.6,8 Severalinvestigators have relied upon induced sputum asa source for monitoring of inflammation.4,19,30

Detection of cytokines in EBC has only been usedin one study of COPD patients.16 This is the firststudy comparing a systematic profile of cytokines ina spectrum of healthy volunteers, smokers withoutCOPD, stable COPD patients and patients withAECOPD using EBC.

A prominent finding of this study is the order ofmagnitude in differences of cytokine levels of COPDpatients with acute exacerbations in comparisonwith either stable COPD patients, healthy smokersor non-smoking volunteers. This difference inmagnitude relates to all of the cytokines examined.A similarly strong increase of cytokine levels,approx. 15 fold in this study, was only paralleledin an analysis of Nys et al. in which a 5–25 foldincrease in IL-1b and IL-8 was observed in BALF ofpatients with acute lung injury with pneumoniaversus patients with acute lung injury withoutpneumonia.31 A much smaller difference wasobserved in investigations using induced sputumwhere AECOPD-GW versus stable COPD resulted inan IL-6 increase of approx. 1.5 fold32 or an IL-8increase of 2 fold and a TNF-a increase of 4 fold.22

EBC cytokine analysis using the bead array usedhere appears to be very sensitive in picking up largedifferences in cytokine levels in acute inflammationcompared to more chronic inflammation. This ofcourse would be a prerequisite of any standard toolfor monitoring inflammation.

Cytokines expressed in relation to protein con-centration in EBC as a means of standardizationshowed absolutely comparable results with regardto the differences between subgroups. Proteinconcentration in EBC has been suggested to beone of the potential ways to standardize concen-trations in EBC. However, no reliable way ofstandardization has yet been agreed upon. Wetherefore decided to present the raw cytokineconcentrations as measured in EBC in this paper.

Much smaller magnitudes of cytokine levels wereobserved in clinically stable and clinically healthysituations. However, volunteers were differentfrom smokers with respect to all cytokines demon-strating that smoking without any sign of COPD doesinduce an inflammatory lesion. In BALF a similarsubset of cytokines was analyzed before and did notresult in significant differences between stableCOPD patients, smokers and non-smokers at com-parable group sizes.33 Although not significant, therelative levels observed in this study fit well to ourdata. An IL-6 increase (2 fold) due to smokingcomparable to the one described in our study waspreviously reported in EBC.34 We also observed anincrease in IL-8 in smokers compared with non-smokers. In contrast IL-8 was not different betweenthese two groups in a study using BALF to analyzeneutrophil chemokines in smokers and non-smo-kers.7

Cytokine levels in stable COPD patients who wereall seen by pulmonary specialists and who weretreated according to current guidelines were ratherlow in comparison with those of smokers and

ARTICLE IN PRESS


volunteers. A significant difference between stableCOPD and healthy volunteers was only seen for IL-1b and IL-12. In fact, in stable COPD the level ofinflammatory cytokines appeared to be lower thanin smokers possibly due to the fact that all patientswith stable COPD had quit smoking at least 1 yearago. A study that has found elevated levels of, e.g.IL-8 in BALF of smokers5 is not really comparable toany of our groups, because all COPD patients weresmokers whereas in this study all COPD patientswere ex-smokers and smokers were clinicallyapparent. Another explanation for the small differ-ence in cytokine levels between stable COPDpatients and healthy volunteers might have beentherapeutic intervention. Because steroids have aninfluence on cytokine levels in many systems weevaluated the effect of ICS on cytokine levels instable COPD and AECOPD-GW patient subgroups. Instable COPD ICS therapy in this study was corre-lated with decreased levels of IL-1b; IL-8, IL-6, IL-10 and IL-12 but not TNF-a: Other authors have notobserved an ICS induced effect on IL-8, MMP-1,MMP-9, SLIP, and TIMP-1 in induced sputum35

indicating again, that EBC analysis of cytokinesmay be a method with increased sensitivity. In factthe effect of ICS in COPD is a matter of ongoingdebate. From this study it can be concluded, thatthere is definitely an effect of ICS on cytokines inEBC and this effect can even be underscored bydemonstrating an ICS-dose dependency for IL-1b;IL-8 and IL-12. Patel and coworkers similarly foundlower IL-6 and IL-8 concentrations in supernatant ofhuman bronchial epithelial cells harvested frombiopsies of stable COPD patients and a negativecorrelation of at least IL-6 with the ICS dose.2 Theeffect of dexamethasone on basal cytokine releaseby alveolar macrophages varied in another studybetween smokers and COPD patients.4 Dexametha-sone 10�5 M resulted in a significant reduction inboth smoking COPD patients (approx. 45% reductionof basal GM-CSF release in macrophages) andsmokers without COPD (approx. 57% reduction). Incontrast, dexamethasone 10�5 M did not result in asignificant reduction of the strongly increased IL-8release in smoking COPD patients, but did reducethe much lower basal release of IL-8 in smokerswithout COPD by approx. 30%.4 Further work of Itoand coworkers demonstrated a significant differ-ence in the effect of dexamethasone 10�6 M on IL-1b induced release of IL-8 and TNF-a betweensmokfers (without COPD) and non-smokers36: dex-amethasone failed to reduce the stimulated releaseof both cytokines in smokers. This and experimentsaiming at the activity and expression of the enzymehistone deacetylase (HDAC) has generated thehypothesis, that smoking inhibits HDAC expression

and promotes cytokine expression and the lack ofsteroid responsiveness.37 It was suggested thateven in COPD patients that have stopped smoking,this mechanism might continue. Our findings withCOPD patients that all were ex-smokers appear toargue against that assumption. Our findings in factseem to fit well with the clinical impression thatinhaled steroids do reduce the frequency of COPDexacerbations38 and that lower levels of cytokinesare related to less frequent COPD exacerbations.19

However, there was no relation of IL-8 level ininduced sputum and exacerbation frequency inanother investigation39 The significance of thisfinding will have to be further investigated. InAECOPD-GW patients, no impact of ICS on cytokinelevels was observed, but cytokine levels were muchhigher as mentioned before and the small doses ofinhaled steroids might not be enough to influencethis level of inflammation. Oral steroids haveproven use in this situation.

A methodological question we wanted to answerin this study concerned the correlation of cytokinelevels of EBC with BAL fluid of the same patient. Wedid not find any significant correlation of cytokinelevels in EBC compared with BALF although thenumbers varied for each cytokine and significanceseemed within reach for IL-12p70. This is in somecontrast to another marker, nitrite, for which wepreviously described a significant but not verystrong correlation between EBC and BALF.40 Wecannot offer a good explanation for the dissimilarbehavior of different cytokines. Instead we postu-late that a correlation of ingredients of EBC andBALF will have to be demonstrated for eachingredient separately.

An increasing number of papers indicate theusefulness of EBC in respiratory medicine. An ERSstatement on this subject is currently in prepara-tion. For the purpose of estimating the extent ofinflammation of the airways/lung in COPD, EBCseemed to perform well and might enable physi-cians to monitor the disease and rationally followtherapy.

We conclude that AECOPD-GW and AECOPD-ICUexhibit similar strong increases in cytokine levels inCOPD patients. Stable COPD (ex-smokers) andvolunteers were similar with respect to cytokinepatterns with the exception of elevated IL-1b andIL-12. Smokers have increased levels of all cyto-kines when compared to non-smokers. Inhaledsteroids resulted in reduced levels of IL-1b; IL-6,IL-8, IL-10 and IL-12 in stable COPD (ex-smokers)with dose dependency demonstrated for IL-1b; IL-8and IL-12. EBC therefore appears to be able tocharacterize important differences in various COPDpatient groups.

ARTICLE IN PRESS


Acknowledgments

This work was supported by the InterdisciplinaryCentre for Clinical Research at the UniversityLeipzig (IZKF Project Z10).

References

1. Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS.Global strategy for the diagnosis, management, and pre-vention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease(GOLD) Workshop summary. Am J Respir Crit Care Med2001;163(5):1256–76.

2. Patel IS, Roberts NJ, Lloyd-Owen SJ, Sapsford RJ, WedzichaJA. Airway epithelial inflammatory responses and clinicalparameters in COPD. Eur Respir J 2003;22(1):94–9.

3. Gizycki MJ, Hattotuwa KL, Barnes N, Jeffery PK. Effects offluticasone propionate on inflammatory cells in COPD: anultrastructural examination of endobronchial biopsy tissue.Thorax 2002;57(9):799–803.

4. Culpitt SV, Rogers DF, Shah P, et al. Impaired inhibition bydexamethasone of cytokine release by alveolar macrophagesfrom patients with chronic obstructive pulmonary disease.Am J Respir Crit Care Med 2003;167(1):24–31.

5. Pesci A, Balbi B, Majori M, et al. Inflammatory cells andmediators in bronchial lavage of patients with chronicobstructive pulmonary disease. Eur Respir J1998;12(2):380–6.

6. Kuschner WG, D’Alessandro A, Wong H, Blanc PD. Dose-dependent cigarette smoking-related inflammatory re-sponses in healthy adults. Eur Respir J 1996;9(10):1989–94.

7. Morrison D, Strieter RM, Donnelly SC, Burdick MD, Kunkel SL,MacNee W. Neutrophil chemokines in bronchoalveolar lavagefluid and leukocyte-conditioned medium from nonsmokersand smokers. Eur Respir J 1998;12(5):1067–72.

8. Lim S, Roche N, Oliver BG, Mattos W, Barnes PJ, Fan ChungK. Balance of matrix metalloprotease-9 and tissue inhibitorof metalloprotease-1 from alveolar macrophages in cigar-ette smokers. Regulation by interleukin-10. Am J Respir CritCare Med 2000;162(4 Pt 1):1355–60.

9. Keatings VM, Collins PD, Scott DM, Barnes PJ. Differences ininterleukin-8 and tumor necrosis factor-alpha in inducedsputum from patients with chronic obstructive pulmonarydisease or asthma. Am J Respir Crit Care Med1996;153(2):530–4.

10. Hill AT, Bayley D, Stockley RA. The interrelationship ofsputum inflammatory markers in patients with chronicbronchitis. Am J Respir Crit Care Med 1999;160(3):893–8.

11. Yamamoto C, Yoneda T, Yoshikawa M, et al. Airwayinflammation in COPD assessed by sputum levels of inter-leukin-8. Chest 1997;112(2):505–10.

12. Beeh KM, Beier J, Kornmann O, Mander A, Buhl R. Long-termrepeatability of induced sputum cells and inflammatorymarkers in stable, moderately severe COPD. Chest2003;123(3):778–83.

13. Vernooy JH, Kucukaycan M, Jacobs JA, et al. Local andsystemic inflammation in patients with chronic obstructivepulmonary disease: soluble tumor necrosis factor receptorsare increased in sputum. Am J Respir Crit Care Med2002;166(9):1218–24.

14. Montuschi P, Collins JV, Ciabattoni G, et al. Exhaled 8-isoprostane as an in vivo biomarker of lung oxidative stress

in patients with COPD and healthy smokers. Am J Respir CritCare Med 2000;162(3 Pt 1):1175–7.

15. Kharitonov SA, Barnes PJ. Exhaled markers of inflammation.Curr Opin Allergy Clin Immunol 2001;1(3):217–24.

16. Bucchioni E, Kharitonov SA, Allegra L, Barnes PJ. High levelsof interleukin-6 in the exhaled breath condensate ofpatients with COPD. Respir Med 2003;97(12):1299–302.

17. Garey KW, Neuhauser MM, Robbins RA, Danziger LH,Rubinstein I. Markers of inflammation in exhaled breathcondensate of young healthy smokers. Chest2004;125(1):22–6.

18. Gessner C, Hammerschmidt S, Kuhn H, et al. Exhaled breathcondensate acidification in acute lung injury. Respir Med2003;97(11):1188–94.

19. Bhowmik A, Seemungal TA, Sapsford RJ, Wedzicha JA.Relation of sputum inflammatory markers to symptoms andlung function changes in COPD exacerbations. Thorax2000;55(2):114–20.

20. Wedzicha JA, Seemungal TA, MacCallum PK, et al. Acuteexacerbations of chronic obstructive pulmonary disease areaccompanied by elevations of plasma fibrinogen and serumIL-6 levels. Thromb Haemost 2000;84(2):210–5.

21. Bresser P, Out TA, van Alphen L, Jansen HM, Lutter R. Airwayinflammation in nonobstructive and obstructive chronicbronchitis with chronic haemophilus influenzae airwayinfection. Comparison with noninfected patients withchronic obstructive pulmonary disease. Am J Respir CritCare Med 2000;162(3 Pt 1):947–52.

22. Aaron SD, Angel JB, Lunau M, et al. Granulocyte inflamma-tory markers and airway infection during acute exacerbationof chronic obstructive pulmonary disease. Am J Respir CritCare Med 2001;163(2):349–55.

23. Burge S, Wedzicha JA. COPD exacerbations: definitions andclassifications. Eur Respir J 2003;41(Suppl):46s–53s.

24. Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratorydistress syndrome. N Engl J Med 1998;338(6):347–54.

25. Anthonisen NR, Manfreda J, Warren CP, Hershfield ES,Harding GK, Nelson NA. Antibiotic therapy in exacerbationsof chronic obstructive pulmonary disease. Ann Intern Med1987;106(2):196–204.

26. Gessner C, Kuhn H, Seyfarth HJ, et al. Factors influencingbreath condensate volume. Pneumologie 2001;55(9):414–9.

27. Hunninghake GW, Gadek JE, Kawanami O, Ferrans VJ,Crystal RG. Inflammatory and immune processes in thehuman lung in health and disease: evaluation by bronch-oalveolar lavage. Am J Pathol 1979;97(1):149–206.

28. Huszar E, Vass G, Vizi E, et al. Adenosine in exhaled breathcondensate in healthy volunteers and in patients withasthma. Eur Respir J 2002;20(6):1393–8.

29. Effros RM, Hoagland KW, Bosbous M, et al. Dilution ofrespiratory solutes in exhaled condensates. Am J Respir CritCare Med 2002;165(5):663–9.

30. Beeh KM, Kornmann O, Buhl R, Culpitt SV, Giembycz MA,Barnes PJ. Neutrophil chemotactic activity of sputum frompatients with COPD: role of interleukin 8 and leukotriene B4.Chest 2003;123(4):1240–7.

31. Nys M, Deby-Dupont G, Habraken Y, et al. Bronchoalveolarlavage fluids of ventilated patients with acute lung injuryactivate NF-kappaB in alveolar epithelial cell line: role ofreactive oxygen/nitrogen species and cytokines. NitricOxide 2003;9(1):33–43.

32. Seemungal TA, Harper-Owen R, Bhowmik A, Jeffries DJ,Wedzicha JA. Detection of rhinovirus in induced sputum atexacerbation of chronic obstructive pulmonary disease. EurRespir J 2000;16(4):677–83.

ARTICLE IN PRESS


33. Soler N, Ewig S, Torres A, Filella X, Gonzalez J, Zaubet A.Airway inflammation and bronchial microbial patterns inpatients with stable chronic obstructive pulmonary disease.Eur Respir J 1999;14(5):1015–22.

34. Carpagnano GE, Kharitonov SA, Foschino-Barbaro MP,Resta O, Gramiccioni E, Barnes PJ. Increased inflam-matory markers in the exhaled breath condensateof cigarette smokers. Eur Respir J 2003;21(4):589–93.

35. Culpitt SV, Maziak W, Loukidis S, Nightingale JA, MatthewsJL, Barnes PJ. Effect of high dose inhaled steroid on cells,cytokines, and proteases in induced sputum in chronicobstructive pulmonary disease. Am J Respir Crit Care Med1999;160(5 Pt 1):1635–9.

36. Ito K, Lim S, Caramori G, Chung KF, Barnes PJ, Adcock IM.Cigarette smoking reduces histone deacetylase 2 expression,enhances cytokine expression, and inhibits glucocorticoid

actions in alveolar macrophages. Faseb J 2001;15(6):1110–2.

37. Barnes PJ, Ito K, Adcock IM. Corticosteroid resistance inchronic obstructive pulmonary disease: inactivation ofhistone deacetylase. Lancet 2004;363(9410):731–3.

38. van der Valk P, Monninkhof E, van der Palen J, Zielhuis G, vanHerwaarden C. Effect of discontinuation of inhaled corti-costeroids in patients with chronic obstructive pulmonarydisease: the COPE study. Am J Respir Crit Care Med2002;166(10):1358–63.

39. Gompertz S, Bayley DL, Hill SL, Stockley RA. Relationshipbetween airway inflammation and the frequency of exacer-bations in patients with smoking related COPD. Thorax2001;56(1):36–41.

40. Gessner C, Hammerschmidt S, Kuhn H, et al. Exhaled breathcondensate nitrite and its relation to tidal volume in acutelung injury. Chest 2003;124(3):1046–52.

exhaled breath condensate cytokine patterns in chronic obstructive pulmonary disease

Documents