American Journal of Respiratory and Critical Care Medicine

Exacerbations of asthma are likely to be due to an increase in airway inflammation. We have studied noninvasive markers of airway inflammation in asthma exacerbations induced by reducing the dose of inhaled corticosteroids. Following a 2-wk run-in period, mild exacerbations were induced in subjects with stable asthma controlled with medium- to high-dose inhaled corticosteroids (beclomethasone dipropionate ⩾ 800 μ g or equivalent daily) by switching them to budesonide 200 μ g daily given from a dry-powder inhaler (Turbohaler). Fifteen subjects were enrolled and were seen twice weekly for 8 wk after steroid reduction. At each visit, exhaled nitric oxide (NO), and methacholine airway responsiveness were measured and spirometry and sputum induction were performed. Mild exacerbation was defined as: (1) a decrease in morning peak expiratory flow (PEF) of ⩾ 20% but < 30% on at least two consecutive days as compared with the mean for the last 7 d of the run-in period; (2) awakening on two consecutive nights because of asthma; or (3) increased use of a short-acting β2-agonist to eight or more puffs daily. Eight subjects did not develop exacerbations during the 8-wk study, whereas seven subjects developed mild exacerbations at Week 4 (n = 1), Week 6 (n = 1), and Week 8 (n = 5). The only significant difference between these two groups at baseline was a higher baseline sputum eosinophil count in subjects with subsequent exacerbations (p < 0.05). The increases in sputum eosinophils and exhaled NO were correlated with decreases in airway function, including decreases in morning PEF and FEV1. However, multiple regression analysis suggested that the change in sputum eosinophils is a potentially useful marker in predicting loss of asthma control reflected by loss of airway function. Jatakanon A, Lim S, Barnes PJ. Changes in sputum eosinophils predict loss of asthma control.

Asthma is a chronic inflammatory disorder of the airways, and long-term therapy for it is therefore directed at suppressing airway inflammation (1, 2). Ideally, adequate suppression of inflammation should be maintained throughout the course of therapy. Assessment of asthma severity and the effectiveness of treatment should therefore be guided directly by the degree of airway inflammation. This is inappropriate in clinical practice, since bronchoscopy is required to obtain samples of lower airway tissue. Because of the invasiveness of bronchoscopic procedures, the counting of inflammatory cells and measuring their activity in airway biopsies can be used to assess airway inflammation only for research purposes.

Assessment of asthma severity and the adequacy of asthma therapy is currently guided by indirect markers, such as lung function and existing asthma symptoms (1, 2). A major drawback to these markers is their uncertain relationship to the degree of airway inflammation and their ability to be temporarily improved by bronchodilators without controlling underlying inflammation. The use of short-acting β2-agonist inhalers for rescue or of long-acting β2-agonists may confound measurements of lung function and symptoms and therefore obscure the severity of asthma and the effectiveness of treatment for it. Persistent inflammation has been demonstrated in airway biopsies of asthmatic patients whose treatment was guided by asthma symptoms and lung function (3).

Recently, interest has been shown in developing relatively less invasive markers for monitoring airway inflammation more directly. It is hoped that asthma therapy guided by monitoring of these noninvasive markers may improve the outcome of treatment. These markers should be able to reflect airway inflammation and its changes in agreement with the findings in airway biopsies. They should be useful for monitoring treatment effectiveness and at the same time be sensitive enough to reflect any worsening of airway inflammation. Among various newer methods being developed for monitoring airway inflammation, sputum induction and measurement of exhaled nitric oxide (NO) seem to be the most promising, as they are reproducible and less invasive, and therefore applicable to sequential measurements. Both methods have been found useful in evaluating the antiinflammatory effects of drugs (4-8). However, their usefulness for monitoring loss of asthma control has not been investigated to the same extent. We have previously shown that exhaled NO increases when inhaled corticosteroid treatment is withdrawn at the time of increased symptoms, but the timing of these changes in relationship to exacerbations has not been studied (9). Similarly, an increase in sputum eosinophil numbers has also been reported after steroid withdrawal (10), but it is not certain how this relates to changes in exhaled NO.

The aim of our study was to evaluate whether parallel monitoring of eosinophils in induced sputum and of exhaled NO could be useful in reflecting loss of asthma control as indicated by loss of airway function. No prospective studies have compared the changes in both parameters in the same patients in relation to changes in airway function. We also attempted to find a marker that might be useful in identifying asthmatic patients who might be at risk of developing exacerbations of asthma following dose reduction of inhaled corticosteroid.


Patients 18 to 70 yr old with a previous history of asthma and who had been treated with an inhaled corticosteroid for at least 3 mo were enrolled in the study. Asthma was diagnosed from a history of recurrent wheezing and chest tightness and a previous physician diagnosis. This was subsequently confirmed by the presence of airway hyperresponsiveness (AHR) (a provocative concentration of methacholine causing a 20% decrease in FEV1 [PC20] < 4 mg/ml). Subjects with stable asthma requiring inhaled corticosteroids at medium to high doses (beclomethasone ⩾ 800 μg or equivalent daily by pressurized metered-dose inhaler [MDI]) to maintain asthma control were recruited. FEV1 at baseline had to be more than 70% of predicted. For stable asthma, all of the following were required: peak flow (PF) variability < 20% during a 2-wk run-in period, use of an inhaled β2-agonist at not more than four puffs per day, no nocturnal awakening due to asthma, and no changes in frequency or intensity of asthma symptoms or use of medications in the previous 3 mo. No subjects had a history of upper respiratory tract infection within the month preceding the study. Subjects gave written informed consent for their participation, and the study was approved by the Ethics Committee of the Royal Brompton Hospital.

Study Design

In order to determine the onset of asthma relapse after corticosteroid dose reduction, the study was conducted in a single-blind manner, including a 2-wk run-in and an 8-wk steroid reduction period. Sixteen subjects were recruited. This was based on the assumption that half of the subjects would have an exacerbation of asthma. Eight subjects with mild exacerbations were required for detection of a 1.5-fold increase in exhaled NO or sputum eosinophils within the group, for an α specification of 0.05 and a β specification of 0.20 (80% power).

At the screening visit, patients completed a questionnaire about their asthma history and current medications. Additionally, skin prick tests and spirometry were performed. Diary cards were issued. The patients' previous corticosteroid inhalers were replaced with budesonide dry-powder inhalers (Turbohalers; AstraZeneca, Lund, Sweden) at equivalent doses given twice daily. Subjects returned 2 wk later (Week 0) for baseline measurements. Subsequent visits were made twice weekly for a total of four visits (Week 2, Week 4, Week 6, and Week 8). At each visit, exhaled NO and methacholine airway responsiveness were measured, and spirometry and sputum induction were performed. At Week 0, all subjects were switched to budesonide delivered from a Turbohaler at a dose of 100 μg twice daily, given together with a matched placebo. The number of puffs taken daily was adjusted to match the number used during the run-in period. Subjects were informed that the budesonide dose could be the same or reduced to 200 μg daily. The nurses who performed all measurements were blind to the dose of budesonide used.

The study was suspended for individual patients whenever exacerbations occurred or at Week 8 if there had been no exacerbations. A mild exacerbation was defined as: (1) a decrease in morning PEF of ⩾ 20% but < 30% on at least two consecutive days as compared with the mean of the last 7 d of the run-in period; (2) awakening on two consecutive nights because of asthma; or (3) an increase in the use of short-acting inhaled β2-agonist to eight or more puffs daily. A severe exacerbation was defined as a decrease in morning PEF to ⩾ 30% below the baseline for two consecutive days. Subjects who developed an exacerbation went to the laboratory as soon as possible within the next 24 h for the same investigations as used regularly in the study. If there was a chance of severe bronchospasm (as judged by the investigator or from variability in PF > 30% or an FEV1 of less than 60% predicted), the methacholine challenge test was not to be performed.

Throughout the study, subjects recorded asthma symptoms including daytime, nighttime, and early morning chest tightness, with a score ranging from 0 to 3 for each symptom (0 = none, 1 = mild, 2 = moderate, 3 = severe). Morning and evening PEF, and the use of short-acting inhaled β2-agonist (puffs/day) for rescue, were also recorded. The levels of morning PEF values that indicated mild exacerbations were calculated and are given individually. At each visit, subjects were encouraged to contact the principal investigator, who was accessible on a 24-h basis, as soon as possible if there was a doubt about their asthma status or any criterion of mild exacerbation was reached. Treatment of an exacerbation included oral prednisolone at 40 mg/d for 1 wk and high-dose budesonide from a Turbohaler (⩾ 800 μg/d) as appropriate. All subjects were treated as outpatients, and telephone assessments were made over the following few days. Subjects returned to the laboratory 1 wk after exacerbations, and subsequent visits were arranged as suitable until asthma status was well controlled.

Exhaled NO Measurement

End-exhaled NO was measured with a chemiluminescence analyzer (Model LR2000; Logan Research, Rochester, UK), using a previously described method (11). In brief, subjects exhaled slowly through a mouthpiece with an exhalation flow of 5 to 6 L/min from TLC over a 30 to 40-s period. NO was sampled from a sidearm attached to the mouthpiece. The mean NO value was taken from the point corresponding to the plateau of the end-exhaled CO2 reading (5 to 6% CO2) and representing a lower respiratory tract sample. Results of the analyses were computed and graphically displayed on a plot of NO and CO2 concentrations, pressure, and flow against time. Measurements were made at the same time of day, and factors that might have influenced NO measurement were controlled throughout the study (12).

Spirometry and Methacholine Responsiveness

FEV1 and VC were measured at the same time of day through dry spirometry (Vitalograph Ltd., Buckingham, UK). The best value of three maneuvers was expressed as a percentage of the predicted value.

Inhalation methacholine challenge tests were performed with a dosimeter (Mefar, Bovezzo, Italy). Doubling concentrations of methacholine (0.06 to 32 mg/ml) were inhaled at tidal breathing while patients wore a noseclip. A total of five inhalations of each concentration of methacholine were administered (inhalation time: 1 s; breathholding time: 6 s). FEV1 was measured 2 min after the last inhalation, until there was a decrease in FEV1 of ⩾ 20% compared with the control inhalation (0.9% saline solution) or until the maximal concentration of methacholine was inhaled. The PC20 for methacholine was calculated by interpolation on a logarithmic dose–response curve; a value of 8 mg/ ml or less indicated AHR (13).

Morning and evening PEF values (best of three) were measured with a Mini-Wright peak flow meter (Clement Clarke International, Ltd., Harlow, UK).

Sputum Induction and Processing

Sputum was induced through the method described by Keatings and coworkers (6). Inhaled albuterol at 200 μg was given via an MDI at 15 min before sputum induction. After spirometric values were recorded, subjects were instructed to wash their mouths thoroughly with water. They then inhaled nebulized 3.5% saline at room temperature from an ultrasonic nebulizer (DeVilbiss Co., Heston, UK) at the maximum saline output (4 ml/min). The total period of sputum induction was 15 min. Subjects were encouraged to cough deeply at 3-min intervals until the 15-min induction time had been completed. Mouthwashing before each cough was encouraged in order to minimize salivary contamination. The initial sample from the first cough was discarded. Sputum was collected into a 50-ml polypropylene tube, kept at 4° C, and processed within 2 h.

Spirometry was repeated after sputum induction. If there was more than a 15% decrease in FEV1, the subject was required to stay for observation until FEV1 had returned to baseline.

For sputum processing, 1 ml Hanks' balanced salt solution (HBSS) containing 1% dithiothreitol (DTT) (Sigma Chemicals, Poole, UK) was added to the sputum. The mixture was vortexed and repeatedly aspirated at room temperature until the sputum was homogenized. Samples were left at room temperature for 5 min. Sputum volume was then recorded, the sputum was further diluted with HBSS to a volume of 5 ml, and the preparation was vortexed briefly and centrifuged at 400 × g for 10 min at 4° C. The final concentration of DTT in all specimens was 0.2%.

The cell pellets from the sputum centrifugation were resuspended for total cell counts recorded with on a hemocytometer, using Kimura stain. Slides were prepared by using a Cytospin instrument (Shandon, Runcorn, UK) and were stained with May–Grunwald–Giemsa stain for differential cell counts. All slides were blinded before counting, and at least 400 inflammatory cells were counted in each sample. An adequate sample was defined as having less than 50% squamous epithelial cells in a Cytospin preparation.

Examination of the reproducibility of differential cell counts performed on 18 pairs of samples obtained from the same asthmatic subjects within an interval of 2 wk showed intraclass correlation coefficients of 0.75 for eosinophils, 0.78 for neutrophils, 0.76 for macrophages, and 0.56 for lymphocytes.

Statistical Analysis

The values of morning PEF, total symptom scores, and rescue inhaler use (puffs/day) were averaged from the last 7 d before each visit. Variability in PF was calculated with the following formula: (maximum PEF − minimum PEF) × 100/maximum PEF. PC20 values were log transformed prior to analysis. FEV1, methacholine PC20, and exhaled NO measured at Week 0 were used as baseline values. To determine whether changes in sputum eosinophil counts and exhaled NO accompanied loss of asthma control, the differences of the last-visit measurements from baseline were used for analysis of correlation with Pearson's product-moment technique. The log-transformed values of the change in sputum eosinophil count were used for analysis. Multiple regression analysis (forward) was then applied to examine whether the changes in sputum eosinophil count and exhaled NO could predict loss of asthma control as reflected by the change in morning PEF and FEV1. To identify the parameters that might be useful in identifying asthmatic individuals who could be at risk of exacerbations after corticosteroid dose reduction, subjects were classified into two groups according to whether or not they developed an exacerbation, and the differences in these groups' parameters at baseline were compared through the unpaired t test or Mann–Whitney U test for parametric or nonparametric data, respectively.

One subject dropped out of the study after Visit 3 (Week 4) because he was unable to attend for all of the study visits, and was therefore excluded from the analysis. Fifteen subjects completed the study. All subjects were atopic and used albuterol as rescue medication. Patients' clinical characteristics at baseline are summarized in Table 1. Eight subjects did not develop exacerbations during the 8 wk of the study, whereas seven subjects subsequently developed mild exacerbations at Week 4 (n = 1), Week 6 (n = 1), and Week 8 (n = 5). Three patients (Patients 2, 6, and 7) had exacerbations reflected by a decrease in morning PEF of > 20%; four patients (Patients 2, 3, 5, and 6) had exacerbations reflected by nocturnal awakening, and one patient (Patient 1) had an exacerbation reflected by the need for eight or more puffs of rescue medication. Six patients with exacerbations (Patients 1 and 3 through 7) had a decrease in FEV1 ⩾ 20%. None had severe exacerbations that required hospital admission. Exacerbations were successfully treated with oral prednisolone. The changes in clinical parameters, lung function, airway responsiveness, exhaled NO, and sputum inflammatory cells over time in the groups with and without exacerbations are summarized in Table 2. Methacholine PC20 at the time of exacerbations was not determined in six of seven subjects for safety reasons.


Pt. No. (Gender)Age (yr)FEV1(% pred )MornPEF (L/min)PC20(mg/ml )PFvar (%)Morning Tightness Nocturnal Asthma Daytime Asthma β2-agonist (puffs/d )Inhaled CS (μg/d )
Subjects without subsequent exacerbations
 1 (F)60 76.93870.9512.10000.0Bud 800
 2 (F)32 77.64852.636000.10.6BDP 800
 3 (M)52 70.14650.0213.60.400.40Bud 800
 4 (F)27 96.84072.73 4.80.700.30.6Bud 800
 5 (M)28 93.75200.22 7.40000Bud 800
 6 (F)39 88.64210.1016.60.700.32.6BDP 800
 7 (F)24 87.73500.48 7.80.300.30.1BDP 800
 8 (M)21101.55202.42 7.60.600.31.7BDP 1000
Mean35.4 86.64440.47*  9.20.300.20.7
SEM 4.9  3.8 221.85*
Subjects with subsequent exacerbations
 1 (F)50 86.9427.00.5912.5101.13.1Bud 800
 2 (M) 32 81.76100.3915.10.4001BDP 800
 3 (F)36 79.83800.4113.10.3000BDP 1000
 4 (M)46 70.54581.17 6.20004FP 1000
 5 (F)25 86.44421.05 4.50001.4Bud 800
 6 (F)60 793700.0812.50000.9Bud 600
 7 (M)55 885100.2320000Bud 400
Mean43.4 81.84560.45*
SEM 4.8  2.3 311.42*

Definition of abbreviations: BDP = beclomethasone dipropionate; Bud = budesonide Turbohaler; CS = corticosteroid; FP = fluticasone propionate; PC20 = provocative dose of methacholine needed to reduce FEV by 20%; mornPEF = morking peak expiratory flow; PFvar = peak flow variability.

*Geometric value.

Daily mean score (minimum = 0, maximum = 3).


WeekSymptom Scoreβ2-Agonist (puffs/d )FEV1(L/min)MornPEF (L/min)PFvar (%)PC20(mg/ml )Exhaled NO (ppb)Spt TCC (× 106/ml )Spt Eos(%)Spt Neut(%)
Subjects without subsequent exacerbations 9.20.46
(0.3, 1.1)(0.0, 1.5)(2.3, 3.8)(391, 502) (5.6, 12.8)(0.11, 2.00) (6.0, 12.2)(0.81, 4.10)(0, 1.4)(39.2, 61.6)
(0.5, 1.7)(0.3, 1.3)(2.2, 3.5)(384, 476) (7.2, 20.1)(0.16, 2.42) (6.4, 15.6)(0.98, 2.14)(0.3, 2.2)(32.5, 70.1)
(0.3, 13)(0.3, 1.0)(2.2, 3.5)(382, 493) (7.6, 26.8)(0.14, 2.81) (4.5, 21.2)(1.32, 3.71)(0, 2.7)(24.3, 72.0)
(0.0, 1.4)(0.0, 1.0)(2.3, 3.7)(370, 477) (9.1, 20.1)(0.15, 1.58) (7.3, 21.7)(0.90, 1.82)(0, 2.7)(42.7, 61.0)
(0.0, 1.1)(0.0, 1.1)(2.6, 3.4)(379, 479) (6.4, 18.1)(0.09, 3.72) (6.1, 20.1)(1.22, 3.20)(0.2, 2.0)(50.9, 63.0)
Subjects with subsequent exacerbations 9.50.4511.30.713.653.7
(0.0, 1.4)(0, 3.5)(2.4, 3.0)(380, 533) (4.7, 14.3)(0.19, 1.05) (5.6, 17.1)(0.61, 1.35)(0.6, 15.0)(30.8, 57.3)
(0.2, 2.5)(0.0, 4.0)(2.2, 2.8)(353, 505) (7.8, 17.3)(0.30, 2.58) (6.8, 25.6)(1.17, 1.86)(2.2, 10.0)(30.2, 46.0)
(0.0, 3.2)(0.0, 6.7)(2.0, 2.7)(340, 488)(11.1, 24.3)(0.10, 2.69)(12.2, 33.2)(0.88, 1.73)(4.4, 22.6)(35.0, 65.3)
(1.4, 3.6)(0.0, 5.9)(1.9, 2.8)(343, 472) (5.1, 29.6)(0.18, 2.34)(12.7, 38.1)(0.74, 1.37)(8.4, 17.5)(34.0, 56.6)
(1.3, 4.8)(1.8, 5.3)(1.2, 2.8)(318, 453)(11.9, 35.3)(20.7, 47.8)(0.82, 1.84)(5.1, 45.5)(27, 76.2)

Definition of abbreviations: CI = confidence interval; Eos = eosinophils; MornPEF = morning peak expiratory flow; neut = neutrophils; PC20 = provocative dose of methacholine needed to reduce FEV by 20%; PFvar = peak flow variability; Spt = sputum; Symptom Score = total daily symptom scores (minimum = 0, maximum = 9); TCC = total inflammatory cell count; ND = not done.

*  Significant linear trend (p < 0.05),

differential cell count,

 significant linear trend (p < 0.001).

§Six patients,

five patients. Data are shown as mean (95% CI), except sputum data, which are presented as median (25th and 75th percentile).

Although there were no significant differences in patient clinical characteristics at baseline between the groups that did and did not develop exacerbations (Table 1), a significantly higher baseline sputum eosinophil count (p < 0.05) was found in subjects who developed exacerbations (Figure 1A). The median (interquartile range) values were 13.6 (13.9)% and 0.2 (1.4)% for the subjects with and without subsequent exacerbations, respectively. Exhaled NO levels at baseline, however, were not different between the two groups, with values of 11.3 ± 2.3 ppb and 9.1 ± 1.3 ppb (mean ± SEM), respectively, in patients with and without exacerbations (Figure 1B). The changes in exhaled NO, sputum eosinophil count, and PF variability after steroid reduction in individual subjects are summarized in Figures 2 and 3.

The changes in parameter levels measured on the last visit are summarized in Table 3. Data from both patient groups were pooled for analysis of correlation (Table 4). There were significant correlations between the change in sputum eosinophil count with the changes in morning PEF, PF variability, FEV1, exhaled NO, and amount of rescue inhaler use; a significant correlation was also found between the change in exhaled NO with the changes in FEV1 and the amount of rescue inhaler use. Multiple regression analysis (forward) indicated that the change in sputum eosinophil count was more useful than the change in exhaled NO for predicting both the changes in morning PEF and FEV1 (Table 5). We also selected change in FEV1 as another outcome variable for loss of asthma control, which is characterized by increasing airway obstruction. FEV1 is more sensitive to airway obstruction and its measurement is more reproducible than measurement of PEF (14).


OutcomeExacerbation of Asthma
Morning PEF change−2.8−8.9
 (% change)(−7.4, 1.7)(−13.8, −4.0)
PFvar change3.514.9
 (% change)(−1.2, 8.2) (9.8, 20.0)
FEV1 −1.1−27.4
 (% change)(−5.9, 3.8)(−36.8, −18.0)
Exhaled NO change* 0.818.6
 (ppb)(0.2, 3.9)(10.8, 32.3)
Sputum eosinophil change* 0.614.2
 (%)(0.2, 2.1) (6.4, 31.1)

Definition of abbreviations: PEF = peak expiratory flow; PFvar change = change in peak flow variability. Data are shown as mean (95% confidence interval).

*Geometric value.


MornPEF* PFvarFEV1 * Eos NOSymptom Score
FEV1 * 0.53−0.72
Eos −0.600.51−0.71
β2 −0.630.57−0.660.580.560.87

Definition of abbreviations: β2 = β2-agonist use; Eos = eosinophils; MornPEF = morning peak expiratory flow; PFvar = peak flow variability; Spt = sputum; symp = symptom score.

*Data shown as Pearson correlation coefficient (p value) between the change values from baseline.

Percent changes from baseline and log-transformed values were used for analysis, respectively.


ModelRR2 R2 ChangeF ChangePredictor VariablesCoefficient
10.6020.3620.3627.374* Constant  −6.479* 2.269
Eosinophil change−6.026* 2.219
Excluded variable
 ModelBeta inSignificantPartial correlation
 1.NO change−0.390.9040.036
10.707 0.4990.49912.969 Constant −8.477 3.452
Eosinophil change−12.157 3.376
Excluded variable
 ModelBeta inSignificantPartial correlation
 1.NO change−0.1840.570−0.166

Dependent variable = morning PEF (%change); dependent variable = FEV1 (%change).

*  p < 0.05.

p < 0.05.

p < 0.01.

Exacerbations of asthma are characterized by increasing airway obstruction. We showed that sputum eosinophils and exhaled NO are increased in parallel in association with decreases in morning PEF and FEV1, an increase in PF variability, and mild exacerbations of asthma. This suggests the potential use of changes in sputum eosinophil count and exhaled NO as objective markers for monitoring loss of asthma control during corticosteroid stepdown therapy. Monitoring sputum eosinophils may, however, be more accurate than monitoring exhaled NO in reflecting a loss of asthma control. Also, decreasing the dose of inhaled corticosteroids when airway inflammation is insufficiently controlled may lead to an early exacerbation of asthma.

An important goal of asthma treatment is to prevent exacerbations (2). Inflammation has been assumed to be an important factor underlying exacerbations of asthma and in determining asthma severity. However, few studies have shown increased inflammatory cell numbers during spontaneous exacerbations of asthma (15, 16). Sputum eosinophilia has been demonstrated in association with mild (10) and severe (7) exacerbations of asthma. However, there is no clear relationship between increased airway inflammation and exacerbations of asthma. Withdrawal from inhaled corticosteroid therapy has been safely used to examine mild exacerbations of asthma (10). We therefore induced exacerbations of asthma by reducing a medium or high dose of inhaled steroids required for asthma control to a lower dose, and monitored the resulting changes in airway inflammation prospectively. We chose not to reduce the dose gradually, since exacerbations can occur at varying doses of inhaled corticosteroids after variable periods. This may confound any analysis, since each inflammatory parameter may have a different dose-responsiveness and kinetic response to corticosteroid withdrawal. Also, we did not induce exacerbations by withdrawing subjects completely from inhaled steroids, since worsening of asthma in clinical practice usually occurs despite continuous corticosteroid therapy. By using induced sputum, we demonstrated an associated increase in sputum eosinophil numbers with increasing airflow limitation and exacerbations of asthma. Although our study was not done in a double-blind manner, all measurements and sputum differential cell counts were performed by a blinded investigator.

Current evidence indicates that corticosteroids suppress airway inflammation but do not cure its primary cause (17). Withdrawal from corticosteroid therapy is therefore expected to be associated with a relapse of inflammation and worsening of asthma. However, there is little information on the likelihood of asthma relapse after corticosteroid withdrawal. A study involving patients with moderate to severe asthma indicated that in terms of age, disease duration, and the degree of airflow obstruction, patients who developed a worsening of asthma did not differ from those whose asthma remained stable. Among patients with exacerbations, those who were treated with lower doses of inhaled steroids may develop early worsening (18). Our study indicates that insufficient control of airway inflammation at the time of reduction in the dose of inhaled corticosteroid may lead to an early loss of asthma control and eventually to an exacerbation of asthma. Asthmatic subjects who developed subsequent exacerbations had significantly higher sputum eosinophil counts at baseline, and this was the only marker that distinguished them from those who did not develop exacerbations. Lung function, methacholine PC20, exhaled NO, and asthma symptom scores at baseline were not different in these two groups. Moreover, sputum eosinophilia in patients who developed exacerbations persisted throughout subsequent sputum analysis, with a further increase at the time of exacerbations. This suggests that sputum eosinophil numbers may be a reliable marker of the control of airway inflammation. Sequential monitoring of sputum eosinophils could therefore be useful for predicting loss of asthma control after corticosteroid withdrawal.

The role of sputum eosinophils as an accurate marker of asthma severity has been demonstrated previously shown (19). Monitoring sputum eosinophils has also been shown to be useful in evaluating the antiinflammatory effects of inhaled corticosteroids (4-7). Although the method of sputum induction and processing used for this purpose has not yet been standardized, the reliability and validity of the resulting data do not seem to be influenced by any technical differences in methods used (20). Either sputum selected from the whole expectorate or the whole expectorate itself can be used for analysis. Increased eosinophil numbers in sputum have been previously demonstrated in asthmatic subjects with exacerbations (7, 21). The increased eosinophil numbers could be mediated by interleukin (IL)-5, a cytokine specific for eosinophil recruitment and activation. IL-5 has been found to be increased in sputum of asthmatic subjects with exacerbations, in association with sputum eosinophilia, and both IL-5 and eosinophil numbers were decreased after corticosteroid treatment, in association with an improvement in airway function (7). Increased sputum eosinophil counts in our patients could have been due to an insufficient dose of budesonide used to inhibit the release of IL-5.

Despite their greater baseline number of sputum eosinophils, subjects who developed subsequent exacerbations of asthma showed no increase in baseline exhaled NO levels. This may indicate a limitation of NO as an accurate marker of the control of inflammation in subjects using inhaled corticosteroids. It may also suggest that NO is more sensitive to the inhibitory effect of corticosteroids than are sputum eosinophil numbers, so that exhaled NO is suppressed while sputum eosinophil numbers remain increased (22). The rapid increase in NO levels prior to exacerbations suggests that exhaled NO could be more useful in monitoring loss of asthma control. In a previous study involving asthmatic subjects taking high-dose inhaled steroids, exhaled NO was increased by 6 wk after the initial doses were halved, in association with a loss of asthma control (23). The increase in exhaled NO preceding exacerbations is likely to result from airway inflammation, since exaled NO also increases in normal subjects with upper respiratory tract infections (24), and in our study was paralleled by an increase in sputum eosinophils. An increase in exhaled NO therefore requires careful interpretation, as it could be nonspecific. We also found that monitoring sputum eosinophil numbers may be more useful than measuring exhaled NO in reflecting loss of airway function. The role of NO in monitoring airway inflammation in asthma requires further long-term studies.

At present, there are no standardized asthma symptom scores. Our symptom scores included nocturnal, early morning, and daytime asthma symptoms. Our study suggests that monitoring both asthma symptoms and the amount of rescue medication required to control symptoms is useful in detecting worsening of airway inflammation, since both measures were increased in parallel with the changes in other objective parameters of asthma instability in our subjects. The sensitivity of asthma symptom scores in reflecting exacerbations of asthma has also been suggested in another study (10). However, some asthmatic patients may have a reduced perception of their symptoms (25). Because symptom-score grading is subjective, it should be used as an adjunct to objective markers for monitoring the control of inflammation. Our study also demonstrates the lack of accuracy of symptom scores and rescue medication use in reflecting control of airway inflammation determined by sputum eosinophil numbers. Both were scored as low in some patients despite high sputum eosinophil numbers. This may reflect a blunting of symptom perception.

It has previously been shown that the clinical course of asthma may change without any concomitant change in airway responsiveness (26). We were unable to demonstrate the usefulness of methacholine PC20 in reflecting loss of asthma control prior to exacerbations. It has been shown that the dose and duration of inhaled corticosteroid treatment may influence methacholine PC20 (27, 28). Exacerbations of asthma may also be associated with sputum neutrophilia (29, 30). Two of our subjects (Patients 2 and 3) had exacerbations with sputum neutrophilia, but there were no signs or symptoms to suggest infection. Also, there was no change in these subjects' total cell counts or in the macroscopic appearance of their sputum.

The change in morning PEF from an individual patient's personal best reading has been used as the standard method for monitoring exacerbations of asthma. Some asthmatic patients, however, may not reliably monitor their PEF (31). Measurement of airway function may be confounded by bronchodilator treatment, and particularly by the use of long-acting inhaled β2-agonists (5). In these situations, monitoring airway function may be less accurate than monitoring sputum eosinophil numbers, which reflects the control of airway inflammation more directly. Measurement of sputum eosinophil numbers may also be used for establishing adequate control of airway inflammation before an attempt is made to reduce the dose of inhaled corticosteroids, since spirometry may not indicate adequate control of airway inflammation. We previously have shown that in patients with mild asthma, airway function can be normal despite persistent inflammation in the airways (32), and in the present study we showed that increased sputum eosinophil numbers are found in some patients despite their having normal airway function. Although sputum induction is more complicated than measurement of airway function, it is not difficult. It remains to be established whether asthma therapy with additional guidance from sputum eosinophil counts may reduce the risk of exacerbations and/or hospital admissions for asthma or improve long-term asthma treatment outcome in comparison with conventional monitoring of airway function.

In summary, our study provides prospective evidence that after inhaled steroid dose reduction, exhaled NO and sputum eosinophil numbers are increased in parallel with loss of airway function. Sputum eosinophil numbers may, however, be more useful than exhaled NO in predicting loss of airway function and exacerbations of asthma.

1. National Heart, Lung, and Blood Institute. 1997. Guidelines for the diagnosis and management of asthma. National Institutes of Health, Bethesda, MD. NIH Publication No. 97-4051A.
2. British Thoracic SocietyThe British guidelines on asthma management. Thorax521997S1S21
3. Sont J. K., Han J., van Krieken J. M., Evertse C. E., Hooijer R., Willems L. N., Sterk P. J.Relationship between the inflammatory infiltrate in bronchial biopsy specimens and clinical severity of asthma in patients treated with inhaled steroids. Thorax511996496502
4. Claman D. M., Boushey H. A., Liu J., Wong H., Fahy J. V.Analysis of induced sputum to examine the effects of prednisone on airway inflammation in asthmatic subjects. J. Allergy Clin. Immunol.941994861869
5. Pizzichini M. M., Kidney J. C., Wong B. J., Morris M. M., Efthimiadis A., Dolovich J., Hargreave F. E.Effect of salmeterol compared with beclomethasone on allergen-induced asthmatic and inflammatory responses. Eur. Respir. J.91996449455
6. Keatings V. M., Jatakanon A., Worsdell Y. M., Barnes P. J.Effects of inhaled and oral glucocorticoids on inflammatory indices in asthma and COPD. Am. J. Respir. Crit. Care Med.1551997542548
7. Pizzichini M. M., Pizzichini E., Clelland L., Efthimiadis A., Mahony J., Dolovich J., Hargreave F. E.Sputum in severe exacerbations of asthma: kinetics of inflammatory indices after prednisone treatment. Am. J. Respir. Crit. Care Med.155199715011508
8. Kharitonov S. A., Yates D. H., Barnes P. J.Inhaled glucocorticoids decrease nitric oxide in exhaled air of asthmatic patients. Am. J. Respir. Crit. Care Med.1531996454457
9. Kharitonov S. A., Yates D. H., Chung K. F., Barnes P. J.Changes in the dose of inhaled steroid affect exhaled nitric oxide levels in asthmatic patients. Eur. Respir. J.91996196201
10. Gibson P. G., Wong B. J., Hepperle M. J., Kline P. A., Girgis A., Gabardo, Guyatt G., Dolovich J., Denburg J. A., Ramsdale E. H., Hargreave F. E.A research method to induce and examine a mild exacerbation of asthma by withdrawal of inhaled corticosteroid. Clin. Exp. Allergy221992525532
11. Kharitonov S. A., Chung K. F., Evans D., O'Connor B. J., Barnes P. J.Increased exhaled nitric oxide in asthma is mainly derived from the lower respiratory tract. Am. J. Respir. Crit. Care Med.153199617731780
12. Kharitonov S. A., Alving K., Barnes P. J.Exhaled and nasal nitric oxide measurements: recommendations. Eur. Respir. J.10199716831693
13. Chai H., Farr R. S., Froehlich L. A., Mathison D. A., McLean J. A., Rosenthal R. R., Sheffer A. L., Spector S. L., Townley R.Standardization of bronchial inhalation challenge procedures. J. Allergy Clin. Immunol.561975323327
14. Enright P. L., Lebowitz M. D., Cockroft D. W.Physiologic measures: pulmonary function tests: asthma outcome. Am. J. Respir. Crit. Care Med.1491994S9S18
15. Boulet L. P., Turcotte H., Boutet M., Montminy L., Laviolette M.Influence of natural antigenic exposure on expiratory flows, methacholine responsiveness, and airway inflammation in mild allergic asthma. J. Allergy Clin. Immunol.911993883893
16. Djukanovic R., Feather I., Gratziou C., Walls A., Peroni D., Bradding P., Judd M., Howarth P. H., Holgate S. T.Effect of natural allergen exposure during the grass pollen season on airways inflammatory cells and asthma symptoms. Thorax511996575581
17. Laitinen L. A., Laitinen A., Haahtela T.A comparative study of the effects of an inhaled corticosteroid, budesonide, and a beta2- agonist, terbutaline, on airway inflammation in newly diagnosed asthma: a randomized, double-blind, parallel-group controlled trial. J. Allergy Clin. Immunol.9019923242
18. Spence D. P., Johnston S. L., Calverley P. M., Dhillon P., Higgins C., Ramhamadany E., Turner S., Winning A., Winter J., Holgate S. T.The effect of the orally active platelet-activating factor antagonist WEB 2086 in the treatment of asthma. Am. J. Respir. Crit. Care Med.149199411421148
19. Pizzichini E., Pizzichini M. M., Efthimiadis A., Dolovich J., Hargreave F. E.Measuring airway inflammation in asthma: eosinophils and eosinophilic cationic protein in induced sputum compared with peripheral blood. J. Allergy Clin. Immunol.991997539544
20. Kips J. C., Peleman R. A., Pauwels R. A.Methods of examining induced sputum: do differences matter? Eur. Respir. J.111998529533
21. Baigelman W., Chodosh S., Pizzuto D., Cupples L. A.Sputum and blood eosinophils during corticosteroid treatment of acute exacerbations of asthma. Am. J. Med.751983929936
22. Jatakanon A., Kharitonov S. A., Lim S., Barnes P. J.Effect of differing doses of inhaled budesonide on markers of airway inflammation in patients with mild asthma. Thorax541999108114
23. Tamaoki J., Kondo M., Sakai N., Nakata J., Takemura H., Nagai A., Takizawa T., Konno K.Leukotriene antagonist prevents exacerbation of asthma during reduction of high-dose inhaled corticosteroid: The Tokyo Joshi-Idai Asthma Research Group. Am. J. Respir. Crit. Care Med.155199712351240
24. Kharitonov S. A., Yates D., Barnes P. J.Increased nitric oxide in exhaled air of normal human subjects with upper respiratory tract infections. Eur. Respir. J.81995295297
25. Kikuchi Y., Okabe S., Tamura G., Hida W., Homma M., Shirato K., Takishima T.Chemosensitivity and perception of dyspnea in patients with a history of near-fatal asthma. N. Engl. J. Med.330199413291334
26. Smith L. S., McFadden E. R.Bronchial hyperreactivity revisited. Ann. Allergy741995454469
27. Haahtela T., Jarvinen M., Kava T., Kiviranta K., Koskinen S., Lehtonen K., Nikander K., Persson T., Selroos O., Sovijarvi A., Stenius-Aarniala B., Svahn T., Tammivaara R., Laitinen L. A.Effects of reducing or discontinuing inhaled budesonide in patients with mild asthma. N. Engl. J. Med.3311994700705
28. Vathenen A. S., Knox A. J., Wisniewski A., Tattersfield A. E.Time course of change in bronchial reactivity with an inhaled corticosteroid in asthma. Am. Rev. Respir. Dis.143199113171321
29. Fahy J. V., Kim K. W., Liu J., Boushey H. A.Prominent neutrophilic inflammation in sputum from subjects with asthma exacerbation. J. Allergy Clin. Immunol.951995843852
30. Turner M. O., Hussack P., Sears M. R., Dolovich J., Hargreave F. E.Exacerbations of asthma without sputum eosinophilia. Thorax50199510571061
31. Chowienczyk P. J., Parkin D. H., Lawson C. P., Cochrane G. M.Do asthmatic patients correctly record home spirometry measurements? B.M.J.30919941618
32. Jatakanon A., Lim S., Chung K. F., Barnes P. J.An inhaled steroid improves markers of airway inflammation in patients with mild asthma. Eur. Respir. J.12199810841088
Correspondence and requests for reprints should be addressed to Prof. P. J. Barnes, Department of Thoracic Medicine, National Heart and Lung Institute, Dovehouse St., London SW3 6LY, UK. E-mail:

Dr. Jatakanon was the recipient of a research fellowship from the Royal Thai Government.


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record