INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The cysteinyl-leukotrienes (Cys-LTs) LTC₄, LTD₄, and LTE₄ have been implicated in the pathogenesis of asthma. They can be recovered in increased concentrations from bronchoalveolar lavage fluid (BALF) and urine of patients with asthma. Administration of exogenous Cys-LTs causes airflow obstruction (1), increases mucus production (2), and leads to infiltration of inflammatory cells into asthmatic airways (3). Further evidence for the role of Cys-LTs in the pathogenesis of asthma has been provided by studies in which antileukotriene drugs have blunted the magnitude of the obstructive response following exercise (4), allergen challenge (7), cold-air hyperventilation (8), and administration of nonsteroidal antiinflammatory drugs to sensitized subjects (9). In addition, drugs directed against synthesis of the Cys-LTs, and selective receptor antagonists of the Cys-LTs, improve baseline pulmonary function both acutely (10, 11) and with long-term treatment (12, 13). Furthermore, treatment directed against Cys-LTs can reduce airway inflammation (14) and bronchial hyperresponsiveness (BHR) (15), as well as improving asthma control (12).

Long-term clinical trials with antileukotriene drugs have reported improved pulmonary function, decreased symptoms and need for rescue ₂-agonists, and a reduction in asthma exacerbations, predominantly in patients with mild to moderately severe asthma who had previously received only ₂-agonists (9, 12, 13).

Current guidelines recommend inhaled corticosteroids as first-line control therapy in asthma, and emphasize the combination of symptom-controlling drugs such as long-acting ₂-agonists with higher doses of corticosteroids in patients with more severe asthma (16). However, not all asthma can be adequately controlled, despite the use of high-dose inhaled corticosteroids; moreover antiinflammatory treatment options in these patients are limited to increasing the dose of inhaled corticosteroid or introducing systemic corticosteroids, with their increased risk of systemic side effects. Although deteriorations in asthma control following reduction of inhaled corticosteroid therapy have been prevented with a leukotriene antagonist (17), there is currently no evidence that antileukotriene therapy can improve asthma control and reduce exacerbations when added to high doses of inhaled corticosteroids. To test the hypothesis that antileukotriene therapy can improve asthma control in patients who still have symptoms despite the regular use of high-dose inhaled corticosteroids ( 1,200 µg/d), we investigated the effects on pulmonary function, symptoms, and asthma exacerbations of adding a leukotriene antagonist in a high dose (zafirlukast, 80 mg twice daily; Zeneca Pharmaceuticals, Macclesfield, UK) to existing asthma therapy for a period of 6 wk.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

The study was designed as a multicenter, double-blind, randomized, parallel-group trial. The study protocol was approved by the respective ethics committee for each clinical center. The study was conducted in accordance with the Declaration of Helsinki and with guidelines for good clinical practice issued by the European Commission.

Patients

Patients eligible to enter the trial were adults 17 to 71 yr of age who had not smoked during the preceding 6 mo and had a diagnosis of asthma according to published criteria (16, 18). Other eligibility requirements included an FEV₁% predicted of 50% to 75%, a reversibility in clinic measurements of peak expiratory flow rate (PEFR) or FEV₁ of  15% after inhalation of  400 µg albuterol, and current therapy with inhaled corticosteroids (beclomethasone  1,200 µg/d or equivalent). To be eligible for randomization, a patient had to have a daytime symptom score of  10 wk at the end of the screening period (using a 0 to 3/d scale on which 0 = no symptoms, 1 = mild symptoms not interfering with activities, 2 = moderate symptoms interfering with some activities, and 3 = severe symptoms interfering with most activities. None of the patients had a history of other significant respiratory disease or respiratory tract infections preceding entry into the study, or had been hospitalized with asthma within 3 mo before the study. Written informed consent was obtained from all patients.

Study Protocol

Following a 2-wk prerandomization (baseline) period, during which patients were maintained on their existing inhaled corticosteroid therapy and supplemental short-acting ₂-agonist bronchodilators, eligible patients were randomized to 6 wk of treatment with either zafirlukast 80 mg or a matching placebo taken twice daily in addition to their existing therapy, which remained unchanged from that taken at study entry. Assessment of inhaled corticosteroid use was based on the assumption that 1,000 µg of beclomethasone is equipotent to 1,000 µg of budesonide or 500 µg of fluticasone. If asthma was insufficiently controlled during the study period, requiring a change in a patient's medication, the responsible physician was allowed to withdraw the patient from the study or to provide any additional asthma therapy considered necessary.

Recording of Outcomes

At the first visit, patients' demographic details were recorded and their inhaler techniques were assessed. During both the baseline and treatment periods, patients were instructed to record the best of three consecutive measurements of daily morning and evening PEFR at home, using a peak flow meter (mini-Wright; Clement Clark International, Harlow, UK). Morning PEFR was the primary outcome measure of the study. Patients were also asked to record in their daily asthma diary their daytime asthma symptom score, the number of awakenings with asthma each night, mornings with asthma on arising, and number of puffs of supplemental short-acting ₂-agonist used each day. Patients' FEV₁ and PEFR were also measured in the clinic, at least 6 h after the last dose of ₂-agonist, at the beginning and end of the baseline period, and after 3 and 6 wk of treatment. At each visit, changes in medication, adverse events (AEs), and asthma exacerbations were recorded.

Compliance with tablet medication was assessed by the return of blister packs. The total number of empty blisters was expressed as a percentage of the total number of prescribed tablets to be taken. Patients continued to use their own inhaled corticosteroid medication during each period of the study. Because the corticosteroid inhalers were not available for collection, compliance with inhaled corticosteroids was not assessed.

Asthma Exacerbations

Exacerbations of asthma that required emergency room treatment or hospitalization, or that resulted in death, were classified as severe. Exacerbations that required additional controller therapy, such as oral steroids, an increase in the daily dose of inhaled corticosteroids, or addition of a long acting bronchodilator (salmeterol or theophylline) were classified as moderate, as were asthma exacerbations that the investigator judged to be severe or moderately severe AEs. An asthma-related AE classified as being of mild severity, a deterioration in asthma control defined as a decrease in clinically measured FEV₁ of  20% from the prerandomization visit FEV₁, or a deterioration in asthma control recorded by the clinician at the clinic visit, but not documented as an AE, qualified as a mild exacerbation.

Safety

AEs were recorded at each clinic visit. Furthermore, hematocrits; platelet, eosinophil, and leukocyte counts; and hemoglobin, total bilirubin, alkaline phosphatase, aspartate aminotransferase, alanine aminotransferase, -glutamyl transferase, creatinine, and urea levels were measured. Electrocardiograms and physical examinations were done at the start and end of the 6-wk randomized treatment period.

Statistical Analysis

Changes from baseline are expressed as mean ± SE. All asthma diary card variables and clinical pulmonary function assessments were compared through a two-way analysis of covariance (ANCOVA), with fitting of terms for treatment and country, and using the last 7 d of the baseline value as a covariate. Treatment differences are presented as the least-squares mean with associated 95% confidence interval (CI) and p value. If scatter plots of residuals suggested deviations from normality, a nonparametric analysis (Wilcoxon's rank sum test) was used to confirm the ANCOVA. The endpoint analysis included the last recorded values during treatment with the study drug, regardless of withdrawal from the study or completion of the study protocol. Data collected after changes in asthma treatment, other than increases in supplemental short-acting ₂-agonists, were excluded in the analysis of asthma diary variables and clinically measured pulmonary function. A two-way ANCOVA also was used to assess the number of symptom free days (days with a daytime asthma score of zero and no mornings with asthma) and ₂-agonist-free days (the number of days with no supplemental ₂-agonist use) in the final week of the treatment period as compared with the last 7 d of the baseline period. All values were expressed as days per calendar month (standardized to a 30-d period).

All data, irrespective of changes in asthma therapy, were included in the analysis of asthma exacerbations and the requirement for increased asthma therapy (intention-to-treat principle). Asthma exacerbations were characterized by severity (adapted from Greening and colleagues [19]) and were summarized with descriptive statistics. The risk of a patient having at least one exacerbation of any severity during randomized treatment was analyzed with a logistic regression model, which included treatment group as the explanatory variable. Country was not included as an explanatory variable because of the difficulty in fitting such a model with many centers and relatively infrequent events. However, summary statistics for each country were inspected to ensure consistency among them. This method was also used to assess the risk of patients requiring an increase in their existing inhaled corticosteroid therapy or needing additional asthma therapy. The rate (i.e., the frequency) of exacerbations during treatment, including data from patients with multiple exacerbations, is presented as the total number of exacerbations divided by the total number of patients at risk. The total number of exacerbations per patient was converted into an average number of asthma exacerbations per 30-d period, which was then analyzed with a country-adjusted Cochran-Mantel- Haenszel chi-square test in which the response was treated as the ordinal.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Characteristics

Of 496 patients screened, a total of 368 patients (from 82 centers) had impaired pulmonary function and sufficient symptoms during the baseline period to confirm the need for additional treatment. Eligible patients were randomized to receive either zafirlukast at 80 mg twice daily (n = 180) or placebo twice daily (n = 188) in addition to their own, usual inhaled asthma therapy. The demographic details of the two groups, including clinically evaluated pulmonary function, corticosteroid doses at entry, and patient-assessed pulmonary function and symptoms before randomization (Tables 1 and 2) were broadly similar. The male-to-female ratio was slightly higher in the zafirlukast group than in the placebo group. This difference was reflected in higher absolute values for pulmonary function in the zafirlukast group at baseline. A total of 32 patients (14 in the zafirlukast and 18 in the placebo group) failed to complete the study: 25 patients were withdrawn with adverse events (11 in the zafirlukast and 14 in the placebo group), five patients failed to comply with the study protocol or withdrew consent (two in the zafirlukast and three in the placebo group) and one patient in each group was lost to follow up.

Compliance

The mean compliances with tablet treatment, as assessed through returned empty blister packs of the total prescribed doses taken, were 95% and 94% for zafirlukast at 80 mg twice daily and placebo tablets twice daily, respectively. There were no differences in compliance between the study groups throughout the study.

Pulmonary Function Data

Morning PEFR increased from baseline by 18.7 ± 3.6 L/min (mean ± SE) and 1.5 ± 3.8 L/min after 6 wk of treatment with zafirlukast and placebo, respectively. There was a statistically significant least-squares mean treatment difference of 17.5 ± 5.1 L/min (p < 0.001 with a 95% CI for the mean treatment difference of 1.4 to 27.5 L/min) at endpoint (Figure 1, lower panel ). There was also a similar improvement in evening PEFR by 16.5 ± 3.8 L/min, after 6 wk of zafirlukast treatment, but only a 1.4 ± 3.8 L/min increase for patients given placebo. The least-squares mean treatment difference in evening PEFR, of 16.2 ± 5.3 L/min, was also statistically significant (p = 0.002; 95% CI: 5.8 to 26.6 L/min). As compared with that in the placebo group, evening PEFR improved similarly on a week-by-week basis in the zafirlukast-treated group (Figure 1, upper panel ). FEV₁ measured at clinic visits improved by 0.19 ± 0.04 L in patients treated with zafirlukast, versus 0.09 ± 0.03 L in the placebo group (Figure 2). The least-squares mean treatment difference (0.12 ± 0.05 L) for this parameter at endpoint was also statistically significant (p = 0.014; 95% CI: 0.02 to 0.21 L). The improvement in clinically measured PEFR was slightly greater with zafirlukast (24.9 ± 5.4 L/min) than with placebo (18.7 ± 5.8 L/min), but this difference did not reach statistical significance. Absolute values for the overall effects of zafirlukast and placebo on PEFR recorded in the patients' asthma diaries at endpoint are summarized in Table 2.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Change in PEFR in patients treated with high-dose inhaled corticosteroids plus zafirlukast at 80 mg twice daily or with high-dose inhaled corticosteroids plus placebo. Mean ± SEM change from baseline for morning PEFR (lower panel ) and evening PEFR (upper panel ) measured daily by patients are presented as the weekly least-squares mean values. The two-sided p values, obtained by analysis of covariance, are presented for the treatment comparison at endpoint (i.e., Week 6 or the last value carried forward before withdrawal or increases in asthma controller therapy).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Change in clinically measured FEV1 in patients treated with high-dose inhaled corticosteroids plus zafirlukast at 80 mg twice daily or with high-dose inhaled corticosteroids plus placebo. Mean ± SE values for FEV1 measured at clinic visits are presented as least-squares mean changes from baseline. The two-sided p values, obtained by analysis of covariance, are for the treatment comparison at endpoint (i.e., Week 6, the last value carried forward before withdrawal, or increases in asthma controller therapy).

Asthma Exacerbations

During the 6-wk double-blind study period, 36 of 179 patients treated with zafirlukast (20%) and 54 of 184 patients given placebo (29%) for whom data were available had one or more exacerbations of asthma as defined earlier. As compared with placebo, the difference with zafirlukast corresponds to a statistically significant (p = 0.046) reduction in the risk of a patient having a first exacerbation of any severity (odds ratio [OR]: 0.61; 95% CI: 0.38 to 0.99). The number of patients who experienced multiple exacerbations while taking zafirlukast was six (3.4%), versus 24 (13.0%) in the placebo group over the 6-wk treatment period (Table 3). The total number of asthma exacerbations, expressed as the rate per treatment group, and including exacerbations in patients with multiple exacerbations, showed that the frequency for patients in the placebo group was 0.45 (82 of 184) exacerbations per patient, versus 0.25 (44 of 179) exacerbations per patient treated with zafirlukast. This 44% reduction approached statistical significance (p = 0.053). A descriptive analysis of asthma exacerbations by severity suggested that the rate of moderate and severe exacerbations was 0.18 per patient given placebo, versus 0.07 per patient treated with zafirlukast, representing a 59% reduction (Figure 3). Mild asthma exacerbations, as defined earlier, were 0.27 per patient given placebo versus 0.17 per patient treated with zafirlukast, corresponding to a reduction of 34%.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Number of asthma exacerbations reported in patients treated with high-dose inhaled corticosteroids plus zafirlukast at 80 mg twice daily or with high-dose inhaled corticosteroids plus placebo. Exacerbations during the 6 wk trial were categorized by reported severity and order of occurrence (first or repeat attack in those patients with multiple exacerbations). Fractions presented in parentheses indicate the proportion of patients contributing to the overall totals for first exacerbations or repeat exacerbations. A severe asthma exacerbation was defined as a hospital admission, an emergency room visit, or a death from asthma. A moderate asthma exacerbation was defined as an exacerbation requiring additional controller medication or a moderate or serious asthma-related AE as judged by the clinician. A mild asthma exacerbation was defined as a decrease in clinically measured FEV1 after 3 wk or 6 wk of treatment of  20% from baseline, an exacerbation defined as a mild AE by the clinician, or a reported worsening in asthma control not classed as an AE.

Of the exacerbations leading to treatment changes, 32 of 36 (89%) resulted in an increased need for corticosteroids. Twenty five of 36 (70%) of the patients in whom these changes were made required the addition of oral corticosteroids, and seven of 36 (19%) required an increase in the dose of inhaled corticosteroids. The remaining four of 36 (11%) patients required the addition of theophylline or salmeterol or both. This need for an increase in asthma therapy was more than halved with zafirlukast treatment; it occurred on 26 occasions in 24 (13%) of the patients given placebo, as compared with 10 (5.6%) of the zafirlukast-treated patients (OR: 0.4; 95% CI: 0.2 to 0.8; p = 0.017).

Other Variables

Asthma symptom scores as recorded on patients' asthma diary cards decreased from the beginning to the end of the study in both study groups. In patients treated with zafirlukast, daytime symptom scores fell from a mean baseline value of 1.79 ± 0.03 to 1.22 ± 0.07 (a 32% decrease), as compared with a fall from 1.77 ± 0.03 to 1.48 ± 0.06 (a 16% decrease) in patients given placebo. The least-squares mean treatment difference for daytime asthma symptom scores at endpoint, of 0.28 (95% CI: 0.42 to 0.13), was statistically significant (p < 0.001). In addition, patient-reported use of rescue ₂-agonist decreased from a mean baseline value of 6.11 ± 0.38 to 4.78 ± 0.40 puffs/d (a 22% decrease) in the zafirlukast-treated group, and from 5.27 ± 0.36 to 5.06 ± 0.40 puffs/d (a 4% decrease) in patients receiving placebo. The least-squares mean between-treatment difference of 0.93 (95% CI: 1.60 to 0.26) puffs/d was statistically significant (p = 0.007). Mornings with asthma fell from a mean baseline value of 4.28 ± 0.22 to 3.14 ± 0.25 per week (27% reduction) in the zafirlukast-treated group, as compared with a decrease from 4.37 ± 0.23 to 3.66 ± 0.26 per week (a 16% reduction) in the placebo group, but the least-squares mean treatment difference was not statistically significant (p = 0.09). Because the scatter plots of residuals for mornings with asthma deviated from normality, a nonparametric test was performed, as specified in the study protocol. This test confirmed a significant treatment difference for the median number of mornings with asthma (p = 0.015; Wilcoxon's rank sum test). The overall effects of zafirlukast and placebo on patient-reported asthma symptoms and ₂-agonist use at endpoint are summarized in Table 2 as absolute means and in Figure 4 as the mean absolute change at endpoint expressed as a percentage of the mean baseline value.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Changes from baseline in pulmonary function, symptoms, and recue 2-agonist use in patients treated with high-dose inhaled corticosteroids plus zafirlukast at 80 mg twice daily or with high-dose inhaled corticosteroids plus placebo. All data are presented as the mean change at endpoint as a percent of the starting mean baseline value in each treatment group. All p values indicate significance for the between-treatment comparisons at endpoint, through analysis of covariance. Note that a treatment difference in favor of zafirlukast for mornings with asthma was significant according to an analysis of median change at endpoint (p = 0.015; Wilcoxon's rank sum test)

The mean number of symptom-free days and ₂-agonist- free days per month (standardized to a 30-d period) increased from baseline in both treatment groups. In the zafirlukast-treated group, symptom-free days increased from 0.67 ± 0.15 d (2.2% of days) to 6.86 ± 0.94 d per month (22.9% of days), compared with an increase from 0.20 ± 0.07 d (0.6%) to 3.68 ± 0.71 d (12.3%) in the placebo group. The least-squares mean between-treatment difference of 1.88 d (95% CI: 0.22 to 3.99) approached statistical significance (p = 0.079). There was a significant difference for ₂-agonist-free days in favor of zafirlukast, with an increase from 3.07 ± 0.59 d (10.2%) to 7.28 ± 0.99 d per month (24.3%), compared with an increase from 3.40 ± 0.64 d (11.3%) to 5.05 ± 0.79 d per month (16.8%) in the placebo group (between-treatment difference of 2.38 d; 95% CI: 0.37 to 4.40; p = 0.021).

Safety Analysis

Placebo and zafirlukast at 80 mg twice daily were equally well tolerated. A total of 82 (46%) patients treated with zafirlukast and 85 patients (45%) given placebo reported AEs. Data for patients with an incidence of AEs of  3% in either treatment group are summarized in Table 4. The only notable difference was a reduced incidence of asthma-related events in the zafirlukast-treated group. Twenty-five patients withdrew from the trial because of AEs, including worsening asthma. In the zafirlukast-treated group, seven patients were withdrawn because of a nonasthma-related event and four because of worsening asthma. In the placebo group, 10 patients were withdrawn because of nonasthma-related events and four because of worsening asthma. One patient in the placebo group suffered a fatal asthma attack. Serious AEs were reported for two patients in the zafirlukast group (detached retina; exacerbation of asthma) and four patients in the placebo group (gastroenteritis, sciatica, chest pain, abdominal pain). Two patients, both in the zafirlukast group, had transient asymptomatic increases in liver transaminases during treatment (which returned to normal in one patient while therapy was continued, and which resolved in the second patient after withdrawal of treatment).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we found that in asthma patients whose symptoms persisted despite treatment with high-dose inhaled corticosteroids (mean dose: 1,624 µg/d), the addition of a high dose of a leukotriene-receptor antagonist (zafirlukast, at 80 mg twice daily) resulted in a statistically significant improvement in pulmonary function, alleviation of asthma symptoms (daytime asthma symptom score and mornings with asthma), and decrease in the use of ₂-agonists. Patients treated with zafirlukast also had significantly fewer asthma exacerbations and required significantly fewer increases in asthma therapy to treat these exacerbations. These observations strongly suggest that leukotrienes contribute to the pathogenesis of moderate to severe asthma even in the presence of high-dose inhaled corticosteroids, and that the addition of a leukotriene-receptor antagonist can improve asthma control in patients with more severe asthma.

Morning and evening PEFR values of the patients treated with high-dose inhaled corticosteroids plus zafirlukast increased on a week-by-week basis as compared with those in the placebo group over the period of the study. There was no suggestion that the increase in morning PEFR group with zafirlukast treatment had reached a plateau by Week 6 of the treatment period. This observation contrasts with the more rapid onset of effects reported with zafirlukast in patients with more mild to moderate asthma treated with inhaled ₂-agonists alone (13, 20), and also contrasts with the effects of long-acting ₂-agonists added to a treatment regimen with moderate doses of inhaled corticosteroids (19, 21, 22). In our patients with moderate to severe asthma, the profile of zafirlukast's effect on pulmonary function appeared to share some similarities with that observed with inhaled corticosteroids. Gradual improvements in pulmonary function have been reported in response to inhaled corticosteroids both in mild to moderate asthma (23, 24) and in moderate to severe asthma (25, 26).

The week-by-week improvements with zafirlukast treatment in the present study contrasted with the lack of improvement observed in patients given placebo, in whom little or no change in home morning and evening PEFR or clinical FEV₁ was observed. The difference in favor of zafirlukast was statistically significant for these three measures of pulmonary function. The lack of standardization of clinical pulmonary function equipment across centers in the study may account for the observed lack of a significant difference between treatments in clinical PEFR, in contrast to PEFR measured at home (with standardized equipment). In contrast to pulmonary function, asthma symptoms, as assessed from patient-reported nighttime awakenings, as well as mornings with asthma, daytime asthma symptom score, and daily rescue ₂-agonist use decreased in both study groups. The effects in patients receiving zafirlukast were always consistently greater than in patients receiving placebo, and differences were statistically significant for all measures except nighttime awakenings. Discrepancies between objective measures of pulmonary function and subjective patient-reported symptoms are often observed in asthma studies. Improvements in subjective assessments might be more indicative of patients' expectations of the efficacy of a new treatment, and may therefore be more prone to placebo-related bias in patients despite a lack of change in objective measures of asthma control.

This is the first study to show improved asthma control with antileukotriene therapy in moderate to severe asthma already treated with high-dose inhaled corticosteroids. Although the study was performed with a higher than licensed dose of zafirlukast (80 mg twice daily, which is four times the standard dose) to provide clear evidence of the involvement of the Cys-LTs in the pathophysiology of chronic asthma in patients taking high doses of inhaled corticosteroids, a recent placebo-controlled study reported similar beneficial effects on pulmonary function and asthma exacerbations with lower doses of zafirlukast (20 mg and 40 mg twice daily) in combination with a range of background asthma therapies including inhaled and oral corticosteroids (27). It is reasonable to speculate that a similar degree of benefit as seen with zafirlukast in the present study is also possible in moderate to severe asthma when standard doses of zafirlukast are used in combination with high doses of inhaled corticosteroid. A further study, however, will be required to confirm this.

The improvement in pulmonary function in the zafirlukast-treated group in the present study is consistent with the observation that oral antileukotriene therapy can provide bronchodilation (10) as well as bronchoprotection against a number of specific and nonspecific stimuli (4). The size of this increase (19 L/min over baseline) in morning PEFR upon adding zafirlukast to a mean of 1,624 µg/d of inhaled corticosteroid in the present study compared favorably with that observed after substitution of 1,600 µg/d of budesonide for the more potent inhaled corticosteroid fluticasone dipropionate (26), and with the addition of theophylline to lower doses of inhaled corticosteroid (28). Neither of these two interventions was reported to have any beneficial effects on asthma exacerbations.

Inhaled corticosteroids are the only drugs that have been shown unequivocally to reduce asthma exacerbations. A recently published study, however, reported a significant reduction in mild and severe asthma exacerbations with the addition of a long-acting ₂-agonist (formoterol) to budesonide at either 200 or 800 µg/d, suggesting that treatment with a long-acting ₂-agonist in addition to inhaled corticosteroids can improve asthma control in patients with mild to moderate asthma. However, it appeared that increasing the maintenance dose of inhaled corticosteroids was a more appropriate first step in controlling severe exacerbations (22). Two other trials in which the addition of long acting ₂-agonists to inhaled corticosteroids were compared with double-dose corticosteroid therapy failed to show improvement in asthma exacerbations, despite good improvements in pulmonary function and symptom control with the long acting ₂-agonists (19, 21). Differences in study design and concomitant doses of corticosteroids do not allow further comparison between these studies and our own. Nevertheless, our study does provide additional evidence for the concept that in the treatment of asthma, the addition of alternative nonsteroidal antiasthma therapy to inhaled corticosteroids may offer an advantage over a further increase in the dose of inhaled corticosteroids.

In our study, additional asthma medication was needed for 78% of the moderate and severe asthma exacerbations, and this need was significantly reduced by zafirlukast. The characteristics of these treated exacerbations in our study resembled the definition of severe exacerbations in a recently published 1-yr trial of formoterol added to low or moderate doses of inhaled budesonide (200 µg/d or 800 µg/d) for treating such exacerbations (22). In that trial, the incidence of severe exacerbations was higher (range: 19 to 39% of patients per group) than in our study (22). Of these severe types of exacerbation, 73% were defined by clinician-led decisions to use oral corticosteroids and 27% were defined by a decrease in pulmonary function. In our study the incidence of treated exacerbations was 6 to 13% of patients per group. Of these patients, 89% received increased corticosteroid treatment. The lower incidence of treated exacerbations in our study most certainly reflects the shorter treatment period (6 wk versus 1 yr), but may also in part reflect the use of higher doses of concomitant corticosteroids in our study population. The mean dose of 1,624 µg/d of inhaled corticosteroid in our study was at least double that used in the trial by Pauwels and colleagues (22), in which most patients had their corticosteroid dose reduced during treatment compared with the dose taken at entry to the trial. This contrasts with our study, in which the significant effect on exacerbations of adding zafirlukast was obtained while maintaining existing antiinflammatory treatment.

Although the improvements in asthma control observed in our study suggest that the action of zafirlukast extends beyond inhibition of leukotriene-induced bronchoconstriction, the study does not provide any conclusive evidence about whether zafirlukast also has antiinflammatory or disease-modifying properties. Yet, the gradual improvement in morning and evening PEFR observed in patients treated with zafirlukast plus high-dose inhaled corticosteroids is not easily explained entirely by the antibronchoconstrictor effects of zafirlukast. The mechanisms by which these effects are achieved remain to be established. However, several challenge models have shown that Cys-LTs provoke eosinophil infiltration into the airways (3, 29), which can be blunted by pretreatment with zafirlukast (14), and leukotriene-receptor antagonists have recently been reported to reduce both induced sputum and peripheral blood eosinophilia (30). Because excretion of LTE₄ in the urine of asthmatic patients remains increased despite treatment with inhaled or oral corticosteroids (31), leukotriene drive does not appear to be significantly influenced by corticosteroids. Furthermore, increased urinary LTE₄ excretion has been associated with severe acute exacerbations of asthma requiring emergency treatment (34). Accordingly, antileukotriene drugs could exert both antibronchoconstrictor and antiinflammatory effects in addition to those of corticosteroids. Several short-term studies in which leukotriene receptor antagonists have been found to reverse chronic airflow obstruction when given in addition to inhaled corticosteroids (10, 35, 36) support an early onset of the antibronchoconstrictor effects of these drugs. Our study suggests that in more severe asthma, combined treatment with inhaled corticosteroids and a leukotriene-receptor antagonist might also extend to complementary antiinflammatory effects that increase with time.

Our results show that the addition of a leukotriene-receptor antagonist was beneficial in patients whose asthma was not adequately controlled by high-dose inhaled corticosteroids. In symptomatic patients receiving high-dose inhaled corticosteroids, the addition of zafirlukast in a high dose (80 mg twice daily) increased pulmonary function and reduced asthma exacerbations. This suggests that Cys-LTs contribute to asthma symptoms and exacerbations in patients whose asthma is not adequately controlled by high-dose inhaled corticosteroids.