INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Current guidelines advocate the use of inhaled corticosteroids (ICS) as first line treatment for patients with moderate to severe persistent asthma (1, 2). In patients insufficiently controlled with a low dose of ICS, several therapeutic options can be considered. These include increasing the dose of ICS or, alternatively, combining long-acting (LA) inhaled ₂-agonists with a low dose of ICS. Several studies have indicated that the latter combination offers better symptom control and lung function improvement than does increasing the dose of ICS (3). Moreover, the combination with LA ₂-agonists has the added advantage of avoiding the potential risk of steroid-associated systemic side effects. In addition, in vitro and in vivo animal data suggest that LA ₂-agonists might have some antiinflammatory effect (6). From the data currently available, it would, however, appear that, despite offering significant symptomatic relief, monotherapy with LA ₂-agonists has limited or no antiinflammatory effect in asthma (7, 8). Therefore, there is concern that adding LA ₂-agonists to a low dose of ICS might be insufficient to control the underlying airway inflammation, or might even mask its progression. To date this issue has not been directly addressed.

The aim of the present study was to compare in patients with asthma the effect of a 1-yr treatment with budesonide (100 µg, twice daily) plus the LA ₂-agonist formoterol (12 µg, twice daily) versus budesonide (400 µg, twice daily) on markers of airway inflammation in induced sputum.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Patients with an established diagnosis of asthma for at least 6 mo were enrolled. They were between 18 and 70 yr of age, and had all been treated with ICS for at least 3 mo. The dose of ICS had to be constant for at least 1 mo before enrollment in the study. Patients were excluded from the study if they were treated daily with more than 2,000 µg of beclomethasone, more than 1,600 µg of budesonide via pressurized metered dose inhaler, more than 800 µg of budesonide via Turbuhaler (Astra, Lund, Sweden), or more than 800 µg of fluticasone. Patients who had needed at least three courses of oral steroids or who had been hospitalized owing to asthma during the 6 mo preceding the study were also excluded. Baseline forced expiratory volume in 1 s (FEV₁) had to be at least 50% of the predicted value. The increase in FEV₁ in response to an inhalation of 1 mg of terbutaline (Bricanyl; Astra) was at least 15% from baseline or 9% of the predicted value.

Study Design

The study consisted of a 1-mo run-in period, followed by a 1-yr double-blind, randomized, parallel group study with two treatment arms. The study was conducted in three centers: the Department of Respirology, McMaster University, Hamilton, Ontario, Canada; the Clinical Research Unit, Brompton Hospital, London, UK; and the Department of Respiratory Diseases, University Hospital Ghent, Ghent, Belgium. Ethics committee approval was obtained from the three centers, and all patients gave written informed consent.

During the 1-mo run-in period, patients were treated with inhaled budesonide (Pulmicort, 800 µg twice daily; Astra) in addition to terbutaline (0.25 mg as needed). Compliance with the study medication was assessed by means of a hidden counter built into the inhaler and accessible only by the investigator. Patients were randomized only if they had taken between 75 and 125% of the recommended number of doses of budesonide and if their asthma had been stable for the last 10 d of the run-in period. Asthma was not considered stable if diurnal variation in peak flow exceeded 20% on two consecutive days, or if ₂-agonist use exceeded four inhalations per day, or if awakenings due to asthma occurred on two consecutive nights, or if the patient had needed oral corticosteroids.

Eligible patients were randomized to one of two possible treatment regimens, namely 100 µg of budesonide plus 12 µg of formoterol (Oxis; Astra) twice daily (BUD200+F) or 400 µg of budesonide plus placebo twice daily (BUD800). All drugs were delivered from a multidose Turbuhaler. Doses are expressed as metered dose.

During the study, 10 visits were scheduled: at start of run-in; 14 d into run-in; at start of treatment; after 2 wk of treatment; and after 1, 2, 3, 6, 9, and 12 mo of treatment. In addition, there were scheduled telephone contacts with the subjects after 2-5 d of treatment and after 4, 5, 7, 8, 10, and 11 mo of treatment.

Outcome Measures

Induced sputum. On each of the study visits, sputum was induced and processed as previously described (9). The protocol was strictly standardized for identical use in the three participating centers. In brief, patients were pretreated with inhaled terbutaline, 0.5 mg. Increasing concentrations of hypertonic saline (3, 4, and 5%) were inhaled during 7-min periods (Ultra-Neb 2000; De Vilbiss, Somerset, PA). After each concentration step, patients were encouraged to expectorate sputum, after rinsing the mouth and blowing the nose. Spirometry was monitored throughout the induction procedure. Induction was stopped once a sample was produced, FEV₁ fell by > 20%, or the highest concentration was reached. From the expectorated sample, the more viscid portions were selected for further processing. A volume of dithiothreitol (DTT, 0.1%) equal to four times the weight of the sample was added and mixed on a bench rocker for 15 min. An equal volume of phosphate-buffered saline (PBS) was added and rocked for five more minutes. The sample was then filtered (70-µm pore size nylon mesh) and centrifuged. The supernatant was discarded and kept at 70° C. The cell pellet was resuspended in PBS, supplemented with 1% human serum albumin (HSA). Cytospins were prepared, air dried, fixed, and stained with May-Grünwald Giemsa or toluidine blue. Immunocytochemical staining with a monoclonal antibody directed against cleaved eosinophil cationic protein (ECP) (EG2; Kabi Pharmacia, Uppsala, Sweden) was performed on cytospins fixed with periodate- lysine-paraformaldehyde. The concentration of ECP in the supernatant was measured by a commercially available radioimmunoassay (Kabi Pharmacia). All samples were analyzed in duplicate.

The primary efficacy variable was the proportion of eosinophils. Secondary efficacy variables included the proportion of EG2⁺ cells, the differential cell count, and ECP levels in sputum supernatant. Cell profiles were determined on the basis of a count of 400 nonsquamous nucleated cells, except for metachromatic cells, which required a count of 1,500 cells. Counting of EG2⁺ cells and measurement of ECP levels were performed at one center for all samples. The other outcome measures were assessed at each center separately.

Other outcome measures. In addition, clinical and functional outcome measures were monitored.

Asthma exacerbations. Exacerbations were defined as being severe if oral glucocorticosteroids were required either as judged by the investigator, or after a decrease in morning or evening peak flow by more than 30% below baseline on two consecutive days. Baseline peak expiratory flow (PEF) was defined as the mean morning PEF during the last 10 d of the run-in period.

Mild exacerbations were defined according to the following criteria: (1) morning or evening PEF > 20% below baseline, (2) rescue terbutaline use of more than four inhalations per 24 h above baseline, or (3) awakenings due to asthma. Single, isolated days of mild exacerbations were not counted.

Diary record data. A daily diary was completed throughout the run-in and treatment periods, recording premedication PEF (peak flow meter; Vitalograph, Buckingham, UK) in the morning and evening, asthma symptoms (night and day, according to a 0-3 scale, where 0 = no symptoms and 3 = incapacitating symptoms), awakenings due to asthma, rescue inhalation of terbutaline, and use of oral glucocorticosteroids.

Episode-free days. A day with no rescue terbutaline use, asthma symptoms = 0, morning PEF  80% of baseline, and no adverse events was defined as an episode-free day.

Spirometry. On each of the clinic visits, prebronchodilator FEV₁ was measured. The best of three consecutive attempts was noted.

Statistical Analysis

For laboratory variables, area under the curve (AUC) values during the randomized period were calculated. For diary variables, mean values for the last 10 d before each visit were used. For comparison of changes within groups, the Wilcoxon signed rank test was used. For comparisons between groups, a two-way analysis of variance (ANOVA) model based on ranks was used. The analysis is based on a "per protocol" approach. Data from patients violating the protocol were included up to the violation time.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Seventy patients were enrolled in the study, 60 of whom were randomized to either of the 2 treatment regimens (Table 1). Mean age and mean dose of steroids were comparable in both groups; baseline FEV₁ at the start of the run-in period was slightly (but not significantly) lower in the BUD800 group.

Proportion of Eosinophils in Induced Sputum

At the start of the run-in period, the median proportion of eosinophils was 6.45% in the patients subsequently randomized to the BUD800 group versus 1.88% in the BUD200+F group) (Table 2). This difference was not significant. Treatment with 1,600 µg of budesonide during the 1-mo run-in period significantly reduced the proportion of sputum eosinophils from 4.5 to 0.68% (p < 0.0005). At randomization, sputum eosinophils were similar in both groups (0.88% in the BUD800 group versus 0.60% in the BUD200+F group) (Figure 1). During the randomized treatment period, a slight, nonsignificant increase in the proportion of eosinophils was observed in the BUD200+F group (3.41 versus 1.74% in the BUD800 group). The change over the entire study period in either of the two treatment groups was not significant. The proportion of patients with a median sputum eosinophil count < 2.5% was not significantly different between both groups (41.7% in the BUD200+F group versus 65.4% in the BUD800 group).

View larger version (22K): [in this window] [in a new window]   Figure 1.   Proportion of eosinophils in induced sputum expressed as a percentage of total nucleated nonsquamous cells, during the run-in (shaded area) and treatment periods. Values are expressed as medians ± interquartile range. During the run-in period, all patients were treated with budesonide (800 µg, twice daily). Patients were then randomized to treatment with budesonide (100 µg) plus formoterol (12 µg) or budesonide (400 µg), twice daily. A significant decrease was observed over the 1-mo run-in period. During the randomized treatment period, no significant changes occurred either within or between both treatment groups.

Proportion of EG2⁺ Cells in Induced Sputum

The proportion of EG2⁺ cells showed a similar pattern: the percentage of EG2⁺ cells was different, albeit nonsignificantly, at enrollment (median, 9.7% in the BUD800 group versus 3.5% in the BUD200+F group), and decreased significantly to equal levels during the run-in period (p = 0.0003). At randomization, the median percentage of EG2⁺ cells in sputum was 0.75% in the BUD800 group and 1.5% in the BUD200+F group. No significant change was observed throughout the treatment period in either group (Table 2).

Differential Cell Count on Induced Sputum

The differential cell counts on induced sputum were not significantly different between treatment groups, nor were any changes observed either within or between groups over the whole treatment period (Table 3).

Eosinophil Cationic Protein Levels in Sputum Supernatant

ECP levels were not significantly different at the start of the run-in period and did not change throughout the study in either treatment group (Table 2).

Lung Function Data

Throughout the randomized treatment period, the prebronchodilator FEV₁ was slightly, but not significantly, lower in the BUD800 than in the BUD200+F group (Figure 2). Evolution of the prebronchodilator FEV₁ over the 1-yr treatment period was not significantly different in the subgroups of patients with median sputum eosinophil counts below or above 2.5%. Peak flow recordings were higher in the BUD200+F than in the BUD800 group. The difference in the evening recordings reached significance (p < 0.05).

View larger version (22K): [in this window] [in a new window]   Figure 2.   Prebronchodilator forced expiratory volume in 1 s (FEV1) as a percentage of the predicted value during the run-in (shaded area) and treatment periods. Data are expressed as means ± SEM. During the run-in period, all subjects were treated with budesonide (800 µg, twice daily). Subjects were then randomized to treatment with budesonide (100 µg) plus formoterol (12 µg) or budesonide (400 µg), twice daily. No significant differences were observed throughout the randomized treatment period, either within or between groups.

Clinical Outcome Measures

Both morning and evening symptom scores were consistently lower in the BUD200+F group than in the BUD800 group. The difference was not significant and no significant changes occurred during the treatment period. The use of rescue inhalers, nocturnal awakenings, the number of severe or mild exacerbations, and the number of episode-free days did not differ between both treatment groups (Table 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, control of airway inflammation in patients with moderate asthma was not found to be significantly different during a 1-yr treatment period with either the long-acting inhaled ₂-agonist formoterol, combined with a low dose of the inhaled glucocorticosteroid budesonide, or a four-fold higher dose of budesonide.

Adding long-acting inhaled ₂-agonist to a low dose of steroids, instead of increasing the dose of steroids administered to patients with insufficiently controlled asthma, seems a particularly attractive therapeutic option. The functional and symptomatic superiority of this "add-on" approach has been clearly illustrated both in mild to moderate and moderate to severe asthma (3). However, long-acting inhaled ₂-agonists are less effective than inhaled corticosteroids at modifying airway inflammation in asthma. In one study, the effect of a 2-mo period of treatment with formoterol was compared with budesonide at a dose of 800 µg daily. Although formoterol reduced the number of eosinophils in bronchial biopsies from a subgroup of patients, the effect of budesonide was more pronounced (8). In a previous study, it had been shown that an 8-wk treatment with salmeterol does not influence the cellular composition of broncoalveolar lavage (BAL) fluid in asthma (7). Concern therefore exists that in uncontrolled asthma, adding long-acting inhaled ₂-agonists might mask progression of the inflammatory process, allowing the disease to worsen before symptoms develop (10). Control of airway inflammation cannot be deduced from clinical or lung function criteria. A number of studies have shown a correlation between eosinophil numbers and biopsies or sputum and functional characteristics such as baseline FEV₁ or PC₂₀ (methacholine provocative dose of methacholine causing a 20% fall in FEV₁ (11) but this is not invariably true (14), and this correlation is insufficiently strong to use any of the functional parameters as a surrogate marker of the underlying inflammation.

In this study airway inflammation was assessed by examining the proportion of eosinophils in induced sputum. Analysis of induced sputum has been increasingly validated as a representative measure of airway inflammation in asthma (15). This validation process has focused predominantly on the eosinophil, as this cell has been attributed a key role in the pathophysiology of the disease (18). Sputum from patients with asthma contains increased proportions of eosinophils, in comparison with healthy volunteers or other chronic airway disorders such as chronic obstructive pulmonary disease (COPD) (12, 19). In addition, the proportion of eosinophils in sputum has been shown to correlate with eosinophil numbers in BAL and bronchial washing (22). Finally, sputum eosinophils also respond to therapeutic interventions, including treatment with oral or inhaled steroids, thus further validating this outcome measure as a marker of airway inflammation (26, 27). The present study extends previous observations, showing that the sputum eosinophil count as outcome measure is sufficiently sensitive to reflect modulation of the dose of steroids (28). All subjects were treated at the time of enrollment with inhaled steroids, at a mean daily dose of approximately 700 µg. Increasing the dose of inhaled steroids to 1,600 µg of budesonide per day, delivered by Turbuhaler for 1 mo, markedly reduced the proportion of eosinophils.

Over the ensuing 1-yr treatment period, clinical and functional outcome measures were better, some of them significantly, in the group treated with low-dose budesonide plus formoterol, compared with the higher dose of budesonide. This is in line with previous observations (5). In addition, this is the first study to establish that this clinical superiority is not accompanied by a loss of control of the underlying inflammation, as judged by sputum eosinophils and EG2⁺ cells as well as ECP levels in sputum supernatant. The value of the EG2 monoclonal antibody in distinguishing between resting and activated eosinophils is somewhat debated (29). However, provided tissue specimens are properly fixed, most studies would indicate that activated eosinophils show a higher degree of EG2⁺ staining (30, 31). Moreover, it has been reported that macrophages, although they can ingest apoptotic eosinophils, do not show demonstrable EG2⁺ staining (32). It has been reported that neutrophils can also ingest ECP. The total amount per cell is, however, far less than in eosinophils and the staining with ECP antibodies is far more faint, if at all demonstrable (33). Therefore, we believe that EG2⁺ cells in sputum in the present study do represent eosinophils, the majority of which are activated.

Induced sputum has been validated as a marker of airway inflammation in a number of short-term studies. The present study adds useful information concerning the use of this technique as an outcome measure in long-term clinical trials. Several studies have illustrated the wide range of eosinophil numbers in sputum from patients with asthma, even in samples obtained from patients with mild asthma that appears well controlled (9, 19, 20). This has an impact on study design, as this clinically undetectable variability in sputum eosinophil counts could result in significant differences at randomization. The present study illustrates this point. Although clinical characteristics were similar in both groups at the start of the run-in period, median sputum eosinophil numbers were different, although not significantly so. Moreover, this difference was no longer apparent after treatment with a high dose of budesonide during the nonrandomized 1-mo run-in period. As all measurements made during the randomized part of the study were compared with the post run-in measurements, the nonsignificant difference in baseline sputum eosinophilia between the two groups did not influence these comparisons.

The variability in sputum eosinophilia can also affect the power of the study. Although over the 1-yr treatment period the median proportion of eosinophils in induced sputum was not significantly different between both groups, and the proportion of EG2⁺ cells even less so, a slight increase was noted in the group treated with 200 µg of budesonide plus formoterol. The lack of significance could in part be attributed to the variability of the measurement and emphasizes the need for even larger numbers of patients in long-term clinical studies such as the present one. However, apart from the statistical significance, a more important question relates to the pathophysiological significance of a given eosinophil count in sputum, and what constitutes a sufficient reduction in sputum eosinophil numbers. It has been reported, using the same sputum-processing techniques as in the present study, that the upper limit for eosinophils in healthy subjects is 2.5% (34). Whether this represents a clinically important cutoff point remains to be further established. In the present study, no differences were found in the proportion of patients in each study group with a median sputum eosinophil count above or below 2.5%. Of note, the change in FEV₁ over the 1-yr treatment period was not different in patients with sputum eosinophil counts above or below this treshold value.

Finally, it must be remembered that the patients randomized in the present study had been stabilized during a 1-mo treatment period with a high dose of inhaled steroids, during which sputum eosinophil counts decreased. This observation seems to support the concept of the "hit hard" approach (2) when initiating asthma therapy. It cannot be ascertained that a similar evolution of sputum eosinophil numbers in both treatment groups would have been observed without this run-in phase.

In conclusion, the present study illustrates that treatment of moderate asthma for 1 mo with a high dose of budesonide significantly reduces activated eosinophil numbers in induced sputum and that no significant differences in sputum markers of airway inflammation are observed during the ensuing 1-yr treatment period with a low dose of budesonide plus formoterol compared with a fourfold higher dose of budesonide. In addition, the data emphasize the potential of combining a low dose of inhaled steroids with LA ₂-agonists as opposed to increasing the dose of inhaled steroids as the preferred long-term treatment strategy in chronic asthma.