INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bronchoscopic studies in patients with mild asthma provide evidence that eosinophils, mast cells, T lymphocytes, and epithelial cells are involved in the pathogenesis of airway inflammation in mild to moderate severe asthma (1, 2). However, relatively little is known about airway inflammation in severe asthma, in part because bronchoscopy is more difficult in symptomatic asthmatics with severe airway obstruction. Recently, sputum induction using nebulized hypertonic saline has been used as an alternative method to obtain lower airway lining fluid, with evidence of good repeatability, reproducibility, and safety (3, 4). Inflammatory profiles assessed by analysis of induced sputum are in agreement with the findings in airway biopsy and BAL fluid (5). Sputum induction is now widely used as a relatively noninvasive method to assess airway inflammation in subjects with various chronic airway diseases.

There is increasing evidence that neutrophils may play a role in acute severe asthma. Prominent neutrophilic inflammation has been demonstrated in fatal asthma of sudden onset (8). Neutrophil numbers and activation are also increased in the airways of subjects with status asthmaticus (9) and during exacerbations of asthma (10). The pathogenesis of airway inflammation in severe persistent asthma has not been well characterized, but recent evidence from bronchial biopsies suggests that neutrophils may be involved (11). This might explain why some patients with severe asthma do not respond well to conventional asthma therapy, as neutrophilic airway inflammation may be resistant to corticosteroid treatment, whereas eosinophilic inflammation is sensitive to steroids (12).

Exhaled NO has been proposed as a noninvasive measure for monitoring asthma control (13). The levels are elevated in steroid-naïve asthmatic subjects, but they are decreased with corticosteroid treatment (14). The levels are not changed after the treatment with a bronchodilator alone (15). However, it is still not clear whether exhaled NO can reflect airway inflammation in subjects who remain symptomatic despite corticosteroid treatment.

We speculated that the mechanisms of airway inflammation in asthmatic patients who respond very well to corticosteroids could be different from those who remain symptomatic despite corticosteroid treatment. The aim of our study was to use induced sputum as a novel method of evaluating inflammatory cells and mediators in severe persistent asthma. Also, we wanted to determine whether exhaled NO could reflect airway inflammation in subjects with uncontrolled asthma despite continuous corticosteroid therapy.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Asthma was diagnosed by a history of recurrent wheezing and chest tightness and a previous physician diagnosis. This was subsequently confirmed by methacholine airway hyperresponsiveness (PC₂₀ < 4 mg/ ml) when FEV₁ was  70% or had evidence of bronchial reversibility after inhaled albuterol 200 µg (> 15% improvement in baseline FEV₁ or > 10% predicted FEV₁ when FEV₁ was < 70% predicted). Twelve normal nonatopic subjects were used as the control group. Normal subjects had no chest symptoms with FEV₁ > 90% predicted and PC₂₀ > 16 mg/ml. All subjects were lifelong nonsmokers, and asthmatics were stable with no changes in asthma symptoms and medication for at least 1 mo, except for the use of short-acting inhaled ₂- agonists. No subjects had a history of upper respiratory tract infection within the previous 6 wk. All subjects gave written informed consent, and the study was approved by the Ethics Committee of the Royal Brompton Hospital.

Asthma severity was documented by (1) asthma questionnaire and previous medical records (if available) involving asthma history and asthma medication required to maintain asthma control, (2) symptom diary card, and (3) spirometry. Asthma severity was classified based on a current guideline (16) into three groups. Group 1: mild persistent, if symptoms persisted > twice a week but < one time a day, night-time symptoms > twice a month, with normal lung function. Group 2: moderate persistent, if patients had daily asthma symptoms, required daily use of inhaled short-acting ₂-agonist, and FEV₁ or PEF > 60% but  80% predicted before treatment. Group 3: severe persistent, if patients had continuous symptoms, frequent exacerbations, limited physical activity, frequent night-time symptoms, with FEV₁ or PEF  60% predicted.

Subjects in Group 1 (n = 23) were recruited from advertisement. Subjects in Groups 2 and 3 (n = 16 in each group) were recruited from patients attending outpatient clinics of the Royal Brompton Hospital. Subjects in Group 1 received only a short-acting inhaled ₂-agonist as required. Subjects in Group 2 were considered stable for the preceding 3 mo using medium-dose inhaled corticosteroids (beclomethasone dipropionate, 400 to 1,000 µg daily, via a metered-dose inhaler [MDI], or equivalent). However, they continued to have daily asthma symptoms with ongoing need for short-acting ₂-agonist as recorded in diary cards, despite normal lung function. The inhaled corticosteroids used were beclomethasone dipropionate 400 (n = 3), 600 (n = 1), 800 (n = 4), 1,000 (n = 3) µg/d via MDI, and budesonide Turbohaler 400 (n = 1), 800 (n = 4) µg/d. Subjects in Group 3 were all treated with oral prednisolone (mean dose, 25 mg daily; range, 5 to 60 mg) and high dose inhaled steroids, consisting of fluticasone Diskhaler 1,000 µg/d (n = 7), fluticasone MDI 2,000 to 4,000 µg/d (n = 2), budesonide Turbohaler 800 to 1,600 µg/d (n = 6), and beclomethasone MDI 2,000 µg/d (n = 1). Other concomitant medications were oral theophylline (n = 3), inhaled salmeterol (n = 8), regular salbutamol, and/or ipratopium bromide nebulization twice daily (n = 8).

Study Design

We used a cross-sectional study design and induced sputum in asthmatic subjects with varying asthma severity. We also measured exhaled nitric oxide (NO) as another marker of airway inflammation. Subjects attended the laboratory on two occasions. The first was for screening. Each subject completed a standard questionnaire, followed, respectively, by skin prick test, exhaled NO measurement, spirometry, methacholine challenge test, or reversibility test. Atopic status was defined as having positive skin prick test to at least one of four common aeroallergens (grass pollen, cat dander, Dermatophagoides pteronyssinus, Aspergillus fumigatus). Diary cards were issued. Subjects prospectively recorded the following asthma symptoms: daytime, night time, and early morning chest tightness, ranging from 0 to 3 for each symptom (none, mild, moderate, severe). Morning and evening peak expiratory flow (PEF) and use of rescue short-acting ₂-agonist (puffs/day) were also recorded. Subjects returned after 2 wk with their diary cards and for sputum induction. Subjects with severe persistent asthma were observed for 24 h after sputum induction as a safety precaution in case an exacerbation of asthma occurred.

Compliance to asthma medications was reinforced, and inhaler technique was checked at all clinic visits by asthma nurses. Plasma prednisolone levels were detected in all nine of nine blood samples randomly taken from subjects with severe asthma studied.

Exhaled NO Measurement

End-exhaled NO was measured by a chemiluminescence analyzer (Model LR2000; Logan Research, Rochester, UK) using a previously described method (17). In brief, subjects exhaled slowly with exhalation flow 5 to 6 L/min from TLC over 20 to 25 s through a mouthpiece. NO was sampled from a side arm attached to the mouthpiece. The mean value was taken from the point corresponding to the plateau of end-exhaled CO₂ reading (5 to 6% CO₂) and representing the 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.

Airway Function and Airway Responsiveness

FEV₁ and VC were measured with a dry spirometer (Vitalograph Ltd., Buckingham, UK). The best value of the three maneuvers was expressed as a percentage of predicted value.

If FEV₁  70%, bronchial responsiveness was determined by inhalation methacholine challenge test, using a dosimeter (Mefar, Bovezzo, Italy). Doubling concentrations of methacholine (0.06 to 32 mg/ml) were inhaled at tidal breathing while patients wore noseclips. A total of five inhalations of each concentration were administered (inhalation time, 1 s; breathholding time, 6 s). FEV₁ was measured 2 min after the last inhalation, until there was a fall in FEV₁ of  20% compared with the control inhalation (0.9% saline solution) or until the maximal concentration was inhaled. The PC₂₀ was calculated by interpolation of the logarithmic dose-response curve; a value of 8 mg/ml or less indicated airway hyperresponsiveness (18). If FEV₁ was < 70%, 400 µg of albuterol was given via a large-volume spacer in order to demonstrate a reversible airway obstruction.

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

Sputum Induction and Processing

Sputum was induced using the method described by Keatings and colleagues (12). Inhaled albuterol 200 µg was given via a metered-dose inhaler 15 min before sputum induction. After spirometry was recorded, subjects were instructed to wash their mouths thoroughly with water. They then inhaled 3.5% saline at room temperature nebulized by an ultrasonic nebulizer (DeVilbiss Co., Heston, UK) at the maximal 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 > 15% drop in FEV₁, the subject would be required to stay for observation until it had returned to baseline.

For sputum processing, 1 ml Hank's 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, further diluted with HBSS to 5 ml, vortexed briefly, and centrifuged at 400 g for 10 min at 4° C. The final concentration of DTT in all specimens was 0.2%.

Sputum supernatants were kept at 70° C for subsequent cytokine assays. The cell pellets were resuspended. Total cell counts were performed on a hemacytometer using Kimura stain. Slides were prepared by using cytospin (Shandon, Runcorn, UK) and stained with May-Grunwald-Giemsa for differential cell counts, which was performed by an observer blind to the clinical characteristics of the subjects. At least 500 inflammatory cells were counted in each sample. An adequate sample was defined as having less than 50% of squamous epithelial cells on cytospin.

The reproducibility of differential cell count 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.

Mediator Assays

Eosinophil cationic protein (ECP) concentrations were measured by radioimmunoassay (Pharmacia & Upjohns Diagnostics, Uppsala, Sweden). The detection limit of the assay was < 2 mg/L. Reproducibility of the assay assessed from 18 paired samples collected from asthmatics within a 2-wk interval was acceptable (19), with an intraclass coefficient of 0.8

Interleukin-5 (IL-5) concentrations were measured using an amplified sandwich enzyme-linked immunosorbent assay (ELISA). Ninety-six-well microtiter plates (Greiner Labortecnik Ltd., Dursley, UK) were coated with 50 µl of rat monoclonal antihuman IL-5 antibody (Pharmingen, Cambridge, UK) at 1:250 dilution and left overnight at 4° C. Plates were then washed with PBS containing 0.05% vol/vol Tween and immediately blocked with PBS/Tween containing fetal calf serum 10% vol/vol for 2 h at 37° C. After further washing, IL-5 standards and samples were added to the plates and incubated overnight at 4° C. The plates were washed and incubated for 45 min at room temperature with 100 µl of biotinylated rat antihuman IL-5 monoclonal antibody (Pharmingen), washed again, and incubated with avidin-peroxidase (Sigma Chemicals) for 30 min. The plates were washed and developed with 100 µl of ABTS substrate solution (0.547 mM 2,2'-azino-bis(3-ethybenzthiazoline-6 sulfonoc acid) and 0.1 M citric acid at pH 4.35 + 0.1% vol/vol (30%) H₂O₂. The optical density of the wells was read using a plate photometer at 405 nm. The detection limit of the assay was 32 pg/ml. Reproducibility of the assay assessed from 12 paired samples collected from steroid-free asthmatics within a 2-wk interval was acceptable (19) with an intraclass coefficient of 0.52.

Interleukin-8 (IL-8) concentrations were measured using sandwich ELISA. Ninety-six-well microtiter plates (Greiner Labortecnik Ltd.) were coated with 100 µl of mouse monoclonal antihuman IL-8 antibody (Genzyme, Cambridge, UK) at 1:200 dilution and left overnight at 4° C. Plates were then washed with PBS containing 0.05% vol/ vol Tween and immediately treated with bovine serum albumin 1% wt/vol for 2 h at 37° C. After decanted blocking buffer was blotted dry, IL-8 standards and samples were added to the plates and incubated for 1 h at 37° C. The plates were washed and incubated for 1 h at 37° C with 100 µl of rabbit antihuman IL-8 biotinylated antibody, washed again, and incubated with streptavidin-horseradish peroxidase (diluted 1:2,000). The plates were washed and developed with 100 µl of tetramethylbenzidine (TMB) and hydrogen peroxide substrate (Sigma Chemicals) at room temperature for 20 min. Stop solution (2N H₂SO₄) 100 µl was added into each well. The optical density of the wells was read using a plate photometer at 450 nm. The detection limit of the assay was 32 pg/ml. Reproducibility of the assay assessed from 12 paired samples collected from asthmatics within a 2-wk interval was acceptable (19), with an intraclass coefficient of 0.77.

Neutrophil myeloperoxidase (MPO) concentrations were measured by sandwich ELISA (Oxis International Inc., Portland, OR) according to the manufacturer's instructions. The detection limit of the assay was < 1.6 ng/ml.

Statistical Analysis

Data are expressed as median (25-75 percentile). The values of morning PEF, PEF variability [(highest PEF - lowest PEF) × 100/highest PEF], total symptom scores, and reliever inhaler use (puffs/day) were averaged from the last 7 d before sputum induction. PC₂₀ values were log-transformed prior to analysis. The differences between normal, mild, moderate, and severe persistent asthma were determined using Kruskall Wallis with Dunn's multiple comparison test for nonparametric data or one-way analysis of variance (ANOVA) with bonferroni's correction for parametric data. Analysis of correlation was achieved using Spearman's rank correlation test. Two-tailed tests were performed, and a p value of less than 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Parameters

At the time of sputum induction, subjects with moderate asthma (Group 2) had stable asthma based on FEV₁ and PEF variability. However, all reported either persistent asthma symptoms daily or daily use of rescue inhaled short-acting ₂-agonist (Table 1). All subjects tolerated sputum induction well. Only one subject with severe persistent asthma developed significant bronchospasm, but this was reversed quickly by albuterol nebulization. There were no asthma exacerbations after sputum induction. The subjects with severe persistent asthma had significantly lower FEV₁, higher PEF variability, greater symptom scores, and used more short-acting ₂-agonist to control symptoms (p < 0.001) (Table 1) compared with subjects with mild or moderate asthma. However, there were no differences in duration of asthma between groups. Although the subjects with severe asthma were older, this was not different significantly from those with mild or moderate asthma, or normal control subjects.

Sputum Inflammatory Cells and Fluid Phase Measurements

There were no significant differences between the groups in terms of sputum volume and squamous epithelial cell and lymphocyte numbers. Compared with normal subjects (Table 2), subjects with severe asthma had increases in total inflammatory cell count (p < 0.01 (Figure 1A), total eosinophil number (p < 0.01) (Figure 1B), and total neutrophil number (p < 0.001) (Figure 1C). Total eosinophil numbers were also greater in subjects with mild asthma (p < 0.01) than in normal control subjects. Among asthmatic subjects, those with severe disease had more increases in total inflammatory cell count (p < 0.05) and total neutrophil number (p < 0.01) than did those with mild asthma. Total numbers of macrophages (Table 2) were not different between asthmatic and normal subjects. The proportion of macrophages (Table 2), however, was significantly lower in induced sputum of subjects with severe asthma than in either normal subjects (p < .001) or subjects with mild asthma (p < 0.05).

View larger version (16K): [in this window] [in a new window]   Figure 1.   Total numbers of inflammatory cells ( panel A), eosinophils ( panel B), and neutrophils ( panel C ) in induced sputum of normal and asthmatic subjects. Symbols: normal (open circles), mild asthma receiving occasional inhaled 2-agonist treatment (open triangles), moderate asthma requiring medium dose of inhaled corticosteroids to maintain control (closed diamonds), severe asthma despite high dose corticosteroid treatment (closed circles). Horizontal lines represent median values. *p < 0.05; **p < 0.01, ***p < 0.001.

Eosinophils, ECP, and IL-5. Compared with normal subjects (Table 2 and Figure 2A), the proportions of eosinophils in sputum were elevated in subjects with mild (p < 0.001) or severe (p < 0.01) asthma. The proportions, however, were not different between normal subjects and subjects with moderate asthma. There was also a significant difference in the proportion of sputum eosinophils between asthmatic groups, with a higher proportion in those with mild asthma than in those with moderate asthma (p < 0.05), but this was not different from those with severe asthma.

View larger version (14K): [in this window] [in a new window]   Figure 2.   The proportion of eosinophils (panel A), eosinophil cationic protein (ECP) (panel B), and interleukin-5 (IL-5) (panel C ) in induced sputum of normal and asthmatic subjects. Symbols same as those in Figure 1.

ECP levels (Table 2 and Figure 2B) in induced sputum were significantly higher in those with mild (p < 0.01) or severe (p < 0.001) asthma compared with normal subjects, but there was no difference between normal and moderate asthma. Between asthmatic patients, there were significantly higher ECP levels in those with severe asthma than in those with moderate asthma (p < 0.01), but the levels were not different between those with severe and those with mild asthma.

IL-5 was detected ( 32 pg/ml) in 15 of 19 samples available for mild asthma. The numbers of samples positive for IL-5 were six of 16 for moderate asthma, and seven of 16 for severe asthma even though the patients were receiving oral prednisolone plus high dose inhaled steroids. There were no significant differences in IL-5 levels between the asthma groups (Figure 2C).

Neutrophils, IL-8, and MPO. The proportions of neutrophils (Table 2 and Figure 3A) were increased in subjects with severe asthma compared with normal subjects (p < 0.05) and subjects with mild asthma (p < 0.05). The levels of IL-8 (Table 2 and Figure 3B) were higher in both moderate (p < 0.01) and severe (p < 0.001) asthmatic groups compared with normal subjects. MPO levels (Table 2 and Figure 3C) were also greater in the groups with moderate (p < 0.05) and severe (p < 0.001) asthma than in the normal group. Among asthmatic subjects, the levels of IL-8 and MPO were also higher in those with severe asthma than in those with mild asthma (p < 0.01 and p < 0.05, respectively).

View larger version (14K): [in this window] [in a new window]   Figure 3.   The proportion of neutrophils ( panel A), interleukin-8 (IL-8) ( panel B), and myeloperoxidase (MPO) ( panel C ) in induced sputum of normal and asthmatic subjects. Key to symbols same as those in Figure 1.

Exhaled NO

Exhaled NO concentrations (Table 1 and Figure 4) were significantly increased in the groups with mild and severe asthma than in the normal group (p < 0.001). The levels, however, were not different between the patients with moderate and those with severe persistent asthma.

View larger version (13K): [in this window] [in a new window]   Figure 4.   Exhaled NO in mild, moderate, and severe asthma. Symbols used same as those in Figure 1.

Correlations

Only the correlations with r value  0.4 are presented. Data from all asthmatic subjects demonstrated correlations between eosinophils (%) and ECP (r = 0.4, p < 0.001), total eosinophils and ECP (r = 0.52, p < 0.001), ECP and IL-8 (r = 0.44, p = 0.001), IL-8 and MPO (r = 0.53, p < 0.001) (Figure 5A), IL-8 and total neutrophils (r = 0.58, p < 0.001) (Figure 5B), MPO and total neutrophils (r = 0.47, p < 0.001) (Figure 5C), PEF variability and IL-8 (r = 0.41, p < 0.01), PEF variability and MPO (r = 0.51, p < 0.001).

View larger version (11K): [in this window] [in a new window]   Figure 5.   The correlations between MPO and IL-8 ( panel A), IL-8 and total neutrophil number ( panel B), and MPO and total neutrophil number  ( panel C ) in asthmatic subjects.

In severe asthma, there were correlations between eosinophils and ECP (r = 0.53, p < 0.05), ECP and total eosinophils (r = 0.57, p < 0.05), IL-8 and total neutrophils (r = 0.54, p < 0.01). However, there were no significant correlations between FEV₁, PEF variability, and symptom scores with fluid phase measurements. In moderate asthma, there were correlations between eosinophils and ECP (r = 0.63, p < 0.01), total eosinophils and ECP (r = 0.7, p < 0.01), IL-8 and eosinophils (r = 0.56, p < 0.05), IL-8 and total eosinophils (r = 0.54, p < 0.05), IL-8 and PC₂₀ (r = 0.63, p < 0.05), MPO and ECP (r = 0.51, p < 0.05), and MPO and PEF variability (r = 0.51, p < 0.05). In mild asthma, there were correlations between total neutrophils and PC₂₀ (r = 0.45, p < 0.05), MPO and PC₂₀ (r = 0.48, p < 0.05), IL-8 and PC₂₀ (r = 0.49, p = 0.01), IL-8 and total neutrophils (r = 0.46, p < 0.05), IL-8 and MPO (r = 0.78, p < 0.001), and total neutrophils and MPO (r = 0.60, p < 0.01).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have demonstrated that sputum induction is a safe method for investigating airway inflammation in the patients with severe persistent asthma, and hospitalization for sputum induction is unnecessary. The inflammatory profiles in sputum suggest that both eosinophils and neutrophils may contribute to airway inflammation in severe asthma, and this may be in part regulated by IL-8. Although the subjects with severe persistent asthma were rather older, this was not significantly different from the other groups.

Attempts have been made to categorize severity of asthma based on symptoms, impairment of activity, lung function, degree of bronchial hyperreactivity, number of emergency visits, number of hospitalizations, and medication use. Although there is no standard agreement on classification of asthma severity, a combination of asthma symptoms and lung function before treatment has been used as a guide to asthma severity (16). After treatment, assessment of asthma severity can be more difficult, as current asthma medication may confound this. Our patients requiring a medium dose of inhaled steroids continued to have daily symptoms with an ongoing need for short-acting ₂-agonists to maintain control. Therefore, they had moderate asthma based on symptom criteria but mild asthma based on lung function. Guidelines, however, recommend that a patient should be assigned to the most severe grade in which any feature occurs. We have therefore grouped them as moderate asthma. Another set of guidelines indicates that overall severity of asthma after treatment is reflected by the minimum medication required to maintain asthma control (20).

The highest levels of ECP in severe persistent asthma reflect persistent eosinophil activation despite high dose corticosteroid therapy. Persistent eosinophil infiltrate in airway biopsies has been described previously in two of seven subjects despite high doses of oral steroids (11). It remains unclear why eosinophilia persists, as corticosteroids are effective inhibitors of eosinophil recruitment and activation, and they reduce eosinophil survival by increasing apoptosis (21). This could be due to inadequate doses of corticosteroids because of the lack of objective indicators for monitoring the effectiveness of anti-inflammatory treatment. Some asthmatics may require larger doses of anti-inflammatory drugs to control airway inflammation because of reduced corticosteroid responsiveness (22). IL-8 may also potentiate eosinophil recruitment and activation. There is evidence that IL-8 can function as a chemotactic factor for cytokine-primed eosinophils (23), and a dose- dependent migration of eosinophils to IL-8 has been demonstrated (24). In addition, elevated IL-8 concentrations have been correlated with elevated ECP levels in blood samples of subjects with severe asthma (25) and in sputum of patients with cystic fibrosis (26). The correlation between IL-8 and ECP was also demonstrated in our study. Corticosteroids may decrease eosinophil recruitment and activation by inhibiting IL-5 gene transcription (21). IL-5 is a cytokine that is specific for eosinophilic inflammation. A reduction in IL-5 concentration in sputum in association with a reduction in eosinophil numbers and activation has been demonstrated after treatment with corticosteroids (27). In patients with corticosteroid-resistant asthma, however, there may be resistance to IL-5 gene repression by corticosteroids (28). This may contribute to persistent elevation of IL-5 in sputum in some patients with severe asthma.

Increased neutrophil numbers in airway biopsies and BAL fluid has been described in chronic severe asthma (11). However, the cause of neutrophilic inflammation has not been determined. In a previous study it has been shown that the local IL-8 production may exceed the blocking capacity of IL-8 autoantibodies in severe asthma (25). We have demonstrated that the concentrations of IL-8 in sputum were higher in severe asthma and there were correlations between sputum neutrophil numbers, IL-8, and MPO. This suggests that IL-8 may be associated with neutrophil recruitment and activation. A dose-dependent migration of neutrophils in response to IL-8 has been previously demonstrated (29). For reasons that are unclear, sputum IL-8 concentrations remain significantly higher in severe persistent asthma despite high-dose steroid treatment, as corticosteroids can inhibit airway epithelial cell IL-8 secretion (30). It is possible that neutrophils themselves may be an important source of IL-8 (31). IL-8 secretion from neutrophils may be enhanced through local autocrine regulation by IL-1 (32). Evidence from patients with COPD suggests that neutrophilic inflammation may be insensitive to corticosteroid therapy (12). This may account for the sustained elevation of IL-8 and neutrophilic airway inflammation in severe persistent asthma. The complex proinflammatory milieu of the airway lining fluid may enhance neutrophil recruitment and reduce responsiveness to corticosteroid treatment. Elevated IL-8 is unlikely to result from infection, as all the patients had stable asthma, with no changes in asthma symptoms at the time of sputum induction. Although viral infection as a cause of neutrophilic inflammation and elevated IL-8 levels may not be excluded, this was unlikely as the subjects with a history of respiratory tract infection within the preceding 6 wk were excluded from the study.

It is possible that neutrophilic inflammation in severe asthma may be the consequence of high dose corticosteroid treatment. Corticosteroids increase neutrophil survival by reducing apoptosis (33, 34), whereas they increase apotosis of eosinophils (34). However, short-term oral prednisolone treatment does not appear to cause increased airway neutrophil numbers (35). Indeed, there is increasing evidence that neutrophilic airway inflammation may reflect asthma severity (9, 36). This is supported by our findings that there were correlations between neutrophils, IL-8, and MPO with PEF variability and FEV₁. However, we could not demonstrate a clear correlation between clinical markers of asthma severity such as FEV₁, PEF variability, and symptom scores with sputum inflammatory markers in severe persistent asthma. If they are present, corticosteroid treatment may confound this.

The consequence of persistent activation of neutrophils in asthma remains unclear. Neutrophils can be an important source of proinflammatory cytokines and proteolytic enzymes (31, 37). Sustained release of these inflammatory products in excess of the capacity of their inhibitors implicates the role of neutrophils in airway injury and remodeling in chronic persistent asthma. Reactive oxygen species (38) and neutrophil elastase (36) have been shown to increase with asthma severity.

Exhaled NO has been proposed as a noninvasive measure for monitoring asthma control (13). We have found that NO levels were highest in mild asthma during short-acting ₂-agonist treatment occasionally. The levels, however, were not different between symptomatic asthmatic subjects with persistent airway obstruction and those with moderate asthma whose lung function was relatively normal. This suggests that exhaled NO could be more sensitive to inhibition by corticosteroid therapy than other indices of inflammation. This may limit the clinical utility of exhaled NO as an accurate inflammatory markers for monitoring the control of airway inflammation once corticosteroid treatment has been started. In this study, we could not demonstrate significant correlations between exhaled NO and other markers of airway inflammation in each asthma group. This was in contrast to our previous study in mild asthma in which the correlations between NO with sputum eosinophils and methacholine airway responsiveness were demonstrated (39). The difference could be due to the heterogeneity of mild asthma that was involved in our previous study, as the number of patients studied was larger. The correlations between exhaled NO and other inflammatory markers in the patients with moderate or severe asthma may be confounded by corticosteroid treatment, as sputum eosinophils could be less sensitive to corticosteroid treatment than exhaled NO. The validity of exhaled NO for monitoring asthma control remains to be established in long-term studies.

In summary, evidence of neutrophil activation in more severe asthma suggests that neutrophils may play a role, particularly in severe persistent asthma. This could result from proinflammatory cytokines and enzymes released directly from neutrophils upon activation. Indirectly, neutrophils may enhance eosinophil recruitment and degranulation through secretion of IL-8 secretion. This neutrophilic inflammation appears to be relatively resistant to corticosteroids and may account for the high doses of steroids needed to control asthma in these patients.