Effects of Terbutaline and Budesonide on Sputum Cells and Bronchial Hyperresponsiveness In Asthma

Effects of Terbutaline and Budesonide on Sputum Cells and Bronchial Hyperresponsiveness In Asthma RUTH E. ALDRIDGE, ROBERT J. HANCOX, D. ROBIN TAYLOR, JAN O. COWAN, MARTI C. WINN, CHRIS M. FRAMPTON, and G. IAN TOWN Canterbury and Otago Respiratory Research Groups, Christchurch and Dunedin Schools of Medicine, University of Otago, Christchurch, New Zealand Previous studies have shown that the regular administration of short acting -agonists can be associated with adverse effects on airway caliber and bronchial hyperresponsiveness (BHR) and that this may occur through a proinflammatory mechanism. The aim was to explore possible adverse effects of high-dose -agonist therapy and to assess any adverse interaction with corticosteroids. We undertook a randomized, crossover study to investigate the effects of 6 wk of treatment with regular terbutaline (1 mg four times a day), regular budesonide (400 g twice a day), combined treatment, and placebo in subjects with mild to moderate asthma. Major endpoints were PD 15 saline, PD 20 methacholine, and induced sputum differential cell counts. Thirty-four subjects were randomized and 28 completed the study. PD 15 saline decreased on terbutaline alone compared with placebo treatment and on combined treatment compared with budesonide alone (mean fold decrease of 0.57 [95% CI 0.36, 0.90] and 0.65 [95% CI 0.43, 0.97], respectively). PD 20 methacholine was not affected by the use of terbutaline either alone or in combination with budesonide. The percentage of eosinophils in induced sputum increased during terbutaline treatment alone compared with placebo (median 8.3% versus 4.4%, p 0.049). The addition of terbutaline to budesonide did not affect the percentage of eosinophils compared with budesonide treatment alone. These findings support the hypothesis that short-acting -agonists have a permissive effect on airway inflammation and that when used in high dose there may be an unfavorable interaction with inhaled corticosteroids. Inhaled corticosteroids and short-acting -agonists are very widely used in the treatment of patients with asthma. The beneficial effects of corticosteroids, which include improvement in airway caliber, bronchial hyperresponsiveness (BHR), and airway inflammation, are well documented (1). Although shortacting -agonists are undoubtedly safe when used as required for the relief of acute symptoms, their regular use has been shown to have adverse effects (2). Continuous use of shortacting -agonist has been associated with deleterious effects on peak expiratory flow rates (PEFR), spirometric values, and BHR to methacholine, histamine, and allergen (3, 4). The clinical significance of these effects has been widely debated (3, 5). The adverse changes in BHR have been more readily demonstrated when bronchoconstriction involves indirect cellular mechanisms (adenosine monophosphate [AMP], allergen) rather than stimulation of smooth muscle directly (histamine, methacholine) (6 10). These findings have generated the hypothesis (Received in original form June 11, 1999 and in revised form October 5, 1999) The study was funded by an Otago Research Grant. Dr Aldridge and Dr. Hancox were supported by an educational grant from GlaxoWellcome NZ Ltd. Astra Draco supplied the study medication and matching placebos. Ipratropium bromide was supplied by Boehringer Ingelheim, Germany. Correspondence and requests for reprints should be addressed to Dr. G. I. Town, Christchurch School of Medicine, P.O. Box 4345, Christchurch, New Zealand. E-mail: ian.town@chmeds.ac.nz Am J Respir Crit Care Med Vol 161. pp 1459 1464, 2000 Internet address: www.atsjournals.org that the regular administration of short-acting -agonist may have a permissive effect on underlying airway inflammation (4). The effects of regular short-acting -agonist on airway inflammation have only been investigated in three studies to our knowledge. In one study, an increase in the total number of eosinophils and their activation status in bronchial biopsies was demonstrated (11). In another, no change in bronchoalveolar lavage cell count or eosinophil cationic protein level (12) was observed. In contrast, in a third study short-acting -agonist treatment resulted in a reduction in lymphocyte number with no change in eosinophil numbers in bronchial biopsies (13). The combination of short-acting -agonist with inhaled corticosteroid has not been extensively investigated; however, there is some evidence to suggest that -agonists may impair the anti-inflammatory effects of corticosteroids (8, 14, 15). Clinical interactions between short-acting -agonists and corticosteroids have recently been studied by our group (16). We found that morning PEFR, FEV 1, and BHR to methacholine were improved with corticosteroid, and that the addition of regular terbutaline improved clinical indices. This new study was designed to extend our observations on the effect of combining short-acting -agonists with corticosteroids, particularly in light of the proposed negative interaction between these two drugs at a molecular level (14). We wished to examine the effects of short-acting -agonists and corticosteroids on inflammatory indices in induced sputum, and to compare the effect of treatment on BHR to hypertonic saline and methacholine. We hypothesized that regular shortacting -agonist treatment would be associated with increased airway inflammation and that any such adverse effects would be more readily detected by assessing BHR to hypertonic saline (indirect cellular mechanism) than to methacholine (direct smooth muscle mechanism). METHODS Subjects Subjects age 16 to 64 yr with mild to moderate asthma for more than 1 yr were recruited. All subjects were atopic (positive skin test to Dermatophagoides pteronyssinus, mixed grass pollen, or cat pelt), and demonstrated BHR to methacholine (provocative dose causing a 20% fall in FEV 1 [PD 20 ] or provocative concentration causing a 20% fall in FEV 1 [PC 20 ] 8 mol or 12 mol if taking inhaled corticosteroid), but had an FEV 1 50% predicted. They also needed to be able to produce sputum after induction with 4.5% hypertonic saline. Subjects were excluded if they had any other serious concurrent medical illness, if they had a smoking history of more than 5 pack-years, or if they had smoked tobacco in the preceding year. Subjects taking more than 1,500 g of inhaled corticosteroid per day, or who had required oral corticosteroids during the previous 3 mo, were also excluded. All subjects gave written informed consent and the study was approved by the Canterbury and Otago Ethics Committees. Study Design This was a randomized, double-blind, placebo-controlled, doubledummy, crossover study. During a prerandomization run-in period, all asthma medications were withdrawn except short-acting bron-

1460 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 161 2000 chodilators, together with any intranasal corticosteroid treatment. For the following 4 wk subjects recorded their peak flow rates and symptoms twice daily in a diary. Provided they did not become clinically unstable, subjects were then randomized to receive the following treatments, each for 6 wk: (1) terbutaline Turbuhaler 500 g 2 puffs four times a day and placebo Turbuhaler 1 puff twice a day; (2) placebo Turbuhaler 2 puffs four times a day and budesonide Turbuhaler 400 g 1 puff twice a day; (3) terbutaline Turbuhaler 500 g 2 puffs four times a day and budesonide Turbuhaler 400 g 1 puff twice a day; (4) placebo Turbuhaler 2 puffs four times a day and placebo Turbuhaler 1 puff twice a day. The trial medication and matching placebos were supplied by Astra Draco, Lund, Sweden. During the treatment periods, subjects were instructed to use ipratropium bromide 40 g via metered-dose inhaler (Boehringer Ingelheim, Germany) for relief of symptoms but no other asthma medication was permitted except in the event of a severe asthma attack. At the end of each 6-wk period, subjects underwent a combined hypertonic saline challenge/sputum induction procedure. Three to 7 d later, they underwent a methacholine challenge test. At this visit, study medication for the next period was issued to the subjects. Whenever possible, the challenge tests were performed at the same time of day. Study medication and ipratropium were withheld either overnight (for subjects whose challenge tests were conducted in the morning), or for at least 6 h before each challenge test performed late in the day. Subjects were monitored closely throughout the study and had 24-h access to a medical investigator together with an emergency supply of prednisone tablets and a salbutamol inhaler. In the event of a subject experiencing an exacerbation sufficiently severe to warrant additional treatment, they were temporarily withdrawn from the study. Thereafter, if considered clinically appropriate, subjects were reentered into the next treatment period once their asthma had restabilized. Subjects had to have discontinued oral and inhaled corticosteroids for at least 4 wk before they could rejoin the study. Treatment periods interrupted by asthma exacerbations were not repeated. Measurements Airway caliber. PEFR were measured twice daily using a mini-wright peak flow meter (Clement Clarke Int. Ltd, Harlow, UK) as the highest of three values taken prior to morning and evening study medication. FEV 1 was measured using a rolling seal spirometer (Spirotech; Graseby, Smyrna, GA) and was taken as the highest of three reproducible maneuvers recorded at the end of each treatment period immediately before the saline challenge. Symptom diary. The following daytime symptoms were documented daily in a diary on a score of 0 to 3: wheeze, breathlessness on activity, cough, and sputum production. Similarly nighttime symptoms of nocturnal wakening, wheeze, and cough were recorded. The use of ipratropium was recorded twice daily. Hypertonic saline challenge/sputum induction procedure. Baseline FEV 1 was recorded. Nebulized 4.5% saline was generated from an ultrasonic nebulizer (Ultra-Neb 2000; DeVilbiss, Somerset, PA) using large bore tubing, a large two-way nonrebreathing valve (2700; Hans Rudolph Inc., Kansas City, MO) and rubber mouthpiece (mean nebulizer output to subjects 1.6 ml/min). The saline was inhaled for the following time periods: 0.5, 0.5, 1, 2, 4, 4, 4, 4 min (modified version of Iredale and coworkers [17]). FEV 1 was measured 1 min after each period of inhalation. Between each inhalation period, subjects rinsed their mouth with water and were encouraged to cough sputum into a plastic container. The challenge procedure was stopped if the FEV 1 fell more than 20% from baseline, and subjects were given salbutamol 200 g from a metered-dose inhaler via spacer. Spirometry was repeated after 5 min. At this point the inhalation regimen was restarted in order to collect more sputum, provided the subject s FEV 1 was greater than 90% of baseline value. The procedure was stopped after 20 min or after an adequate sputum sample had been obtained. The nebulizer output was calculated by weighing the nebulizer canister before and immediately after the challenge test. The cumulative dose of saline causing a 15% decrease in FEV 1 (PD 15 saline) was calculated by linear interpolation on the log dose response plot. Methacholine challenge test. BHR to methacholine was measured using a modified version of the rapid challenge procedure (18). After measurement of the baseline FEV 1, subjects inhaled nebulized 0.9% saline as a control, followed by a series of doubling doses of methacholine (0.044 to 44 mol taken in one or two breaths), from a Nebicheck dosimeter (P.K. Morgan Ltd, Gillingham, Kent, UK). FEV 1 was measured 1 min after each dose, and immediately after this the subjects inhaled the next dose of methacholine. The test was stopped when the FEV 1 had fallen by greater than 20% from the postsaline value or when the highest dose had been administered. The cumulative provocation dose of methacholine causing a 20% reduction in FEV 1 (PD 20 methacholine) was calculated by linear interpolation between the last two readings on the log dose response plot. Sputum processing and analysis. The entire expectorated sample was homogenized by the addition of 1% dithiothreitol (the volume of 1% dithiothreitol added equating to 3 times the weight of the sputum sample) (Sputolysin; Calbiochem, La Jolla, CA) (19). The mixture was then placed in a rocking waterbath at 37 C for 30 min to ensure complete dispersion before being filtered through a 48- m nylon mesh (B&SH Thompson, Mississauga, ON, Canada). The total cell count, percentage squamous cells, and percentage cell viability (trypan blue exclusion) were determined using a hemocytometer. An aliquot was diluted to obtain a concentration of approximately 1 10 5 cells/ml from which cytospins were prepared (Shandon 2; Shandon Southern Products Ltd, Runcorn, Cheshire, UK). The remaining sputum mixture was centrifuged at 1,350 g for 5 min, the supernatant decanted and frozen at 80 C for future analysis. Sputum cell counts. Cytospins were stained with May Grunwald Giemsa and a 400 differential cell count (excluding squamous cells) was determined in duplicate. Statistical Analysis Results were analyzed for all subjects who completed two or more treatment periods (n 28). Diary data from the first 2 wk of each treatment period were excluded to minimize the influence of any carry-over effect and to remove the need for a washout period. Each symptom was analyzed as the percentage of symptomatic days or nights reported during the treatment period. Similarly, the use of ipratropium was analyzed as the percentage of days and nights during which it was taken. Normally distributed data (FEV 1, PEFR, log PD 15 saline, and log PD 20 methacholine) were initially assessed for an overall treatment effect using a general linear model repeated-measures analysis. If a significant treatment effect was revealed, further analysis of each of the six pairwise comparisons using the paired t test was undertaken. Nonparametric data (percentage of days with each symptom and in which ipratropium was used and the percentages of each cell type in sputum) were initially assessed with Friedman s test, and if appropriate, analyzed in pairs using the Wilcoxon signed-rank test. A p value of 0.05 was considered statistically significant. Parametric data are plotted as the mean with the least significant difference, nonparametric data as the median with the interquartile range. All statistical tests were performed using SPSS version 8.0 software (SPSS Inc., Chicago, IL). RESULTS Fifty-two subjects were initially recruited to the study. Eighteen of these were withdrawn before randomization because of deteriorating asthma (11), the inability to produce sputum (4), social reasons (2), or because a PD 20 methacholine was not achieved (1). The clinical characteristics of the 34 randomized subjects are shown in Table 1. Of these, 28 (82%) completed the study, with four subjects withdrawn on account of poor asthma control and two for social reasons. As these subjects all withdrew during the first or second treatment periods, treatment comparisons were only performed on the 28 remaining subjects. During the study proper, six exacerbations occurred in five of the 28 subjects. Three of these occurred in the terbutaline alone treatment period, and three in the placebo period. One subject withdrew in the placebo period after completing the saline challenge test but before the methacholine challenge was performed.

Aldridge, Hancox, Taylor, et al.: -Agonist Therapy and Airway Inflammation 1461 PEFR TABLE 1 BASELINE CLINICAL CHARACTERISTICS (n 34) Age, mean, yr 39 (range 18 61) Sex, male:female 18:16 Inhaled corticosteroids* at entry None 12 (35.0%) 1 250 g/d 8 (23.5%) 251 500 g/d 8 (23.5%) 501 1,000 g/d 6 (18.0%) FEV 1, mean, L 3.05 (95% CI 2.76 3.35) FEV 1, % predicted, mean 89.5 (95% CI 85.2 93.9) * Beclomethasone and budesonide regarded as equivalent. Analysis of the PEFR data on a week-by-week basis revealed no time trends. Morning PEFR (Table 2 and Figure 1) was significantly higher on budesonide than on terbutaline or placebo treatment ( 19 L/min, p 0.043 and 17 L/min, p 0.001, respectively). Morning PEFR was further increased on combined treatment compared with budesonide alone ( 22 L/min, p 0.001). There was no significant difference in morning PEFR between terbutaline and placebo. Evening PEFR (Table 2) was significantly higher on budesonide compared with placebo treatment ( 15 L/min, p 0.005), and significantly higher on terbutaline versus placebo treatment ( 30 L/min, p 0.001). There was no significant difference in evening PEFR between budesonide and terbutaline. Combined treatment resulted in a higher evening PEFR than any of the other treatments (p 0.001). Symptoms and Rescue Medication Use Analysis of the diary data on a week-by-week basis revealed no time trends. No significant treatment effects were seen for nighttime wakening, cough, sputum production, or breathlessness on activity. The percentage of days and nights with wheeze decreased on budesonide treatment, when given alone or in combination with terbutaline (Table 2). The addition of terbutaline to placebo or budesonide did not result in any significant changes. The percentage of days on which ipratropium was used fell for all three treatments compared with placebo (p 0.04). FEV 1 FEV 1 data are shown in Table 2 and Figure 2. FEV 1 was higher on budesonide treatment compared with terbutaline or placebo treatment ( 0.31 L, p 0.001, and 0.26 L, p 0.002, respectively). Similarly, FEV 1 was higher on the combined treatment compared with terbutaline or placebo treatment alone ( 0.17 L, p 0.019, and 0.12 L, p 0.042, respectively). FEV 1 was lower on combined treatment compared with budesonide alone ( 0.14 L, p 0.030). There was a trend toward a lower FEV 1 on terbutaline compared with placebo but this did not reach statistical significance. Hypertonic Saline PD 15 A PD 15 saline was achieved in 88 (82%) of the 107 saline challenge tests conducted during the randomized periods. Linear extrapolation, equivalent to a further 4 min of saline administration, allowed an additional two cases to be included in the analysis. For the remaining challenges an arbitrary value of 30 ml ( maximal calculated PD 15 value) was assigned. PD 15 saline data are shown in Table 2 and Figure 3. PD 15 saline was higher on budesonide treatment compared with terbutaline or placebo treatment: 3.99-fold increase (95% confidence interval [CI] 2.32, 6.88) and 2.32 (95% CI 1.48, 3.63), respectively. PD 15 saline was lower on terbutaline compared with TABLE 2 EFFECT OF ALL FOUR TREATMENT GROUPS ON OUTCOME MEASURES (n 24 28)* Treatment Treatment Effects (p Value) P T B C P versus T P versus B P versus C T versus B T versus C B versus C Morning PEFR, L/min 452 450 469 491 0.484 0.001 0.001 0.043 0.001 0.001 412 491 411 490 429 509 451 531 Evening PEFR, L/min 452 482 467 502 0.001 0.005 0.001 0.168 0.001 0.001 412 492 443 521 427 507 460 543 Daytime wheeze, % days present 25 18 9 4 0.287 0.064 0.001 0.267 0.020 0.126 4 74 3 51 0 32 0 18 Nighttime wheeze, % nights present 16 12 4 3 0.879 0.033 0.001 0.099 0.003 0.131 1 48 0 63 0 36 0 12 Daytime ipratropium use, % days used 6 0 0 0 0.006 0.039 0.002 0.777 0.184 0.049 0 40 0 13 0 25 0 6 FEV 1, L 2.95 2.90 3.21 3.07 0.824 0.002 0.042 0.001 0.019 0.030 2.60 3.29 2.58 3.21 2.88 3.53 2.77 3.37 PD 15, saline, ml 5.34 3.18 11.44 8.00 0.019 0.001 0.102 0.001 0.001 0.039 3.39 8.42 2.06 4.92 8.17 16.0 5.36 11.9 PD 20 methacholine, mol 0.72 0.52 2.81 2.29 0.228 0.001 0.001 0.001 0.001 0.254 0.46 1.13 0.30 0.92 1.59 5.11 1.24 4.26 Sputum % eosinophils 4.4 8.3 1.7 2.1 0.049 0.021 0.056 0.001 0.001 0.409 1.7 9.8 3.1 14.7 1.0 3.6 0.7 3.9 Sputum % neutrophils 28.5 23.0 23.4 27.9 No significant treatment effect 20.0 36.7 15.8 44.4 13.6 42.1 13.5 46.5 Sputum % lymphocytes 0.5 0.5 0.5 0.2 No significant treatment effect 0.4 1.2 0.5 0.9 0.2 0.9 0.0 0.7 Sputum % macrophages 57.0 46.8 63.8 53.0 No significant treatment effect 42.5 69.0 26.7 69.3 40.8 75.7 40.6 68.4 Sputum % epithelial cells 4.6 4.4 6.1 5.6 No significant treatment effect 2.2 8.1 2.4 12.9 1.5 13.6 3.0 16.8 Definition of abbreviations: B budesonide; C combined treatment; P placebo; T terbutaline. * PEFR, FEV 1 data displayed as mean and 95% CI. PD 15 saline and PD 20 methacholine data displayed as geometric mean and 95% CI. Sputum % cells, wheeze, and ipratropium use data displayed as median and interquartile range.

1462 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 161 2000 Figure 1. Treatment effects for morning PEFR. Data are displayed as mean (square symbol) with error bars representing the least significant difference at the 5% level. Figure 3. Treatment effects for PD 15 saline. Data are displayed as geometric mean (square symbol) with error bars representing the least significant difference at the 5% level. placebo treatment and on combined treatment versus budesonide alone (0.57-fold decrease [95% CI 0.36, 0.90] and 0.65 [95% CI 0.43, 0.97], respectively). There was no significant difference in PD 15 saline on combined treatment compared with placebo. There was no relationship between changes in FEV 1 and changes in BHR to saline in any of the treatment groups (data not shown). Methacholine PD 20 A PD 20 methacholine was achieved in 103 (97%) of the 106 methacholine challenge tests conducted during the randomized periods. For the three remaining challenge tests an arbitrary value of 54 mol ( maximal calculated PD 20 value) was assigned. PD 20 methacholine data are shown in Table 2 and Figure 4. PD 20 methacholine was higher on budesonide treatment compared with terbutaline or placebo treatment (5.08- fold increase [95% CI 2.73, 9.48] and 3.86 [95% CI 2.39, 6.22], respectively). Similarly PD 20 methacholine was higher on the combined treatment compared with terbutaline or placebo treatment (4.23-fold increase [95% CI 2.49, 7.19] and 3.31 [95% CI 2.17, 5.05], respectively). There was no significant difference in PD 20 methacholine on combined treatment versus budesonide alone, or for terbutaline compared with placebo. Sputum Differential Cell Counts Sputum quality (total cell count, percent squamous cell, percent cell viability) did not significantly differ between treatment periods (Table 3). There was a significant treatment effect for the percentage of eosinophils in sputum (Table 2, Figure 5) but not for any of the other cell types. Eosinophils were lower on budesonide treatment compared with terbutaline or placebo treatment ( 6.6%, p 0.001 and 2.7%, p 0.012, respectively). The percentage of eosinophils was higher on terbutaline compared with placebo treatment ( 3.9%, p 0.049). There was no significant difference for eosinophils on combined treatment compared with budesonide alone. Figure 2. Treatment effects for FEV 1. Data are displayed as mean (square symbol) with error bars representing the least significant difference at the 5% level. Figure 4. Treatment effects for PD 20 methacholine. Data are displayed as geometric mean (square symbol) with error bars representing the least significant difference at the 5% level.

Aldridge, Hancox, Taylor, et al.: -Agonist Therapy and Airway Inflammation 1463 TABLE 3 SPUTUM QUALITY IN EACH OF THE FOUR TREATMENT PERIODS (n 24 28)* Treatment P T B C Total cell count, 10 9 /L 1.27 1.97 1.50 1.96 (0.87 2.83) (0.99 2.89) (1.14 2.70) (1.38 2.90) Squamous cells, % 5.3 5.3 5.6 4.0 (3.1 11.4) (2.9 8.9) (3.4 10.6) (1.6 8.5) Cell viability, % 80 77 74 77 (74 84) (72 85) (67 84) (74 84) Definition of abbreviations: B budesonide; C combined treatment; P placebo; T terbutaline. * Data are displayed as median, with interquartile range in parentheses. DISCUSSION The results of this investigation provide additional data concerning the proinflammatory as well as the anti-inflammatory effects of the two drug therapies most commonly used in the management of chronic asthma, namely inhaled -agonists and corticosteroids. Although it is no longer recommended that short-acting -agonists be taken regularly, many patients use short-acting agents frequently in addition to their anti-inflammatory medication. For this reason investigating possible interactions between the two drugs is clinically relevant, and the design of our study reflected this objective. The principal finding was that high-dose terbutaline alone resulted in a significant increase in sputum eosinophils and BHR to saline when compared with placebo. There were no significant changes in symptoms, BHR to methacholine, morning PEFR, or FEV 1. These data indicate that when given as monotherapy in our patients, -agonist alone had a proinflammatory effect although no major impact on lung function was observed. The effects of regular short-acting -agonists on measurements of lung function and BHR to methacholine, histamine, and allergen have been reviewed previously (3), although not all published studies have reported negative effects consistently. The adverse effects of -agonist on more specific measures of airway inflammation have not been studied as extensively. The increase in sputum eosinophils seen in the present study is in keeping with the results of a previous investigation in which an increase in both the total number of eosinophils and their activation status was observed with regular shortacting -agonist (11). Similarly, short-acting -agonist has been shown to augment the increase in sputum eosinophils and eosinophilic cationic protein observed after allergen challenge (20). Moreover, the administration of regular short-acting -agonist to subjects with an early asthmatic response to allergen has been shown to increase the likelihood of developing a dual asthmatic response, suggesting a proinflammatory action (21). In vitro, exposure of blood eosinophils to short-acting -agonist resulted in a dose-dependent increase in their respiratory burst, suggesting that short-acting -agonists may affect eosinophil activation status directly (22). Terbutaline treatment resulted in increased BHR to saline but not to methacholine. A similar difference between saline and methacholine has been observed previously in a study which examined the effects of regular inhaled salbutamol in nonasthmatic atopic subjects (23). Short-acting -agonists have been shown to have a greater effect on BHR when the mechanism of induced bronchoconstriction involves indirect pathways such as allergen rather than direct stimulation of bronchial smooth muscle (6 10). Moreover, the loss of the acute protective effect of short-acting -agonists during regular treatment is more obvious using indirect provocation stimuli such as AMP, than with directly acting nonspecific stimuli such as methacholine (6). Hypertonic saline is thought to cause bronchoconstriction via indirect cellular mechanisms, notably stimulation of mast cells and local sensory nerves (24). Our data suggest that saline challenge may be a more sensitive measurement of airway inflammation. Our study was also designed to test the hypothesis that the benefits of inhaled corticosteroid might be compromised when inhaled -agonist was given in combination. When given as monotherapy, budesonide treatment resulted in a reduction in sputum eosinophils, a reduction in BHR to hypertonic saline and methacholine, and improved lung function. These effects are well-documented in the literature (25 28). The combination of budesonide with terbutaline, however, provided contrasting outcomes. On the one hand, there was no significant difference in sputum eosinophils with combined treatment compared with budesonide alone. On the other, combined treatment appeared to negate the improvement in BHR to saline but not to methacholine obtained with budesonide: PD 15 saline on combined treatment was not significantly different from placebo. This pattern of differential effects on BHR has been noted previously. In a crossover study in 13 asthmatics, Cockcroft and coworkers demonstrated that compared with treatment with budesonide alone, BHR to allergen but not methacholine was increased when salbutamol was simultaneously administered (8). Similarly, in a parallel group Figure 5. Treatment effect on % sputum eosinophils. Data are displayed as median with interquartile range, 10th and 90th percentiles, and outliers. * P significantly different from placebo, * T significantly different from terbutaline, * B significantly different from budesonide, * C significantly different from combined treatment, p 0.05.

1464 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 161 2000 study of 41 patients, Wong and colleagues found that the protective effect of 2 to 4 wk treatment with budesonide against allergen challenge was reduced in patients receiving regular terbutaline (15). In our study, we also observed that FEV 1 was significantly lower on combined treatment than with budesonide alone. This pattern is different from that for mean morning PEFR, an observation which has been made previously (15, 29). The explanation for these differential effects on measurements of lung function is unclear. The clinical relevance of our findings remains uncertain. Our earlier crossover study did not reveal any adverse changes in asthma control with terbutaline treatment, either as monotherapy or in combination with budesonide (16). In fact, the combination of terbutaline and budesonide was ranked best of the four treatments using a hierarchy of relevant clinical endpoints. In this study, despite the proinflammatory effects of terbutaline, no clinically important negative effect has been identified. However, patients recruited into these two studies were necessarily required to tolerate the withdrawal of inhaled corticosteroid during the run-in period, and thus only patients with mild asthma were enrolled. It is possible that the proinflammatory actions of short-acting -agonists may counterbalance the benefits of inhaled corticosteroid treatment when subjects with more severe asthma are studied, or when different dose combinations of the two drugs are used, or when subjects are investigated over longer periods of time. This might explain the adverse effect on asthma control which has been observed in some studies despite the use of anti-inflammatory treatment (30, 31). Certainly, our observations reinforce current guidelines that short-acting agents should be used only as required, and ought to prompt clinicians to carefully review patients whose asthma is poorly controlled and whose use of -agonist as reliever is excessive. Acknowledgment: The authors thank the volunteers for participating in this study, Timothy Chan for performing the analyzis of induced sputum, and Erin Flannery and Christine McLachlan for helping with the challenge testing. References 1. Barnes, P. J. 1998. Efficacy of inhaled corticosteroids in asthma. J. Allergy Clin. Immunol. 102:531 538. 2. Sears, M. R., and D. R. Taylor. 1994. The beta 2-agonist controversy: observations, explanations and relationship to asthma epidemiology. Drug Safety 11:259 283. 3. Taylor, D. R., M. R. Sears, and D. W. Cockcroft. 1996. The -agonist controversy. Med. Clin. North Am. 80:719 748. 4. Cockcroft, D. W., and V. A. Swystun. 1996. Functional antagonism: tolerance produced by inhaled beta-2 agonists. Thorax 51:1051 1056. 5. Barrett, T. E., and B. L. Strom. 1995. Inhaled beta-adrenergic receptor agonists in asthma: more harm than good? Am. J. Respir. Crit. Care Med. 151:574 577. 6. O Connor, B. J., S. L. Aikman, and P. J. Barnes. 1992. Tolerance to the non-bronchodilator effects of inhaled beta 2-agonists in asthma. N. Engl. J. Med. 327:1204 1208. 7. Taylor, D. A., M. W. Jensen, S. L. Aikman, J. G. Harris, P. J. Barnes, and B. J. O Connor. 1997. Comparison of salmeterol and albuterol-induced bronchoprotection against adenosine monophosphate and histamine in mild asthma. Am. J. Respir. Crit. Care Med. 156:1731 1737. 8. Cockcroft, D. W., V. A. Swystun, and R. Bhagat. 1995. Interaction of inhaled beta 2 agonist and inhaled corticosteroid on airway responsiveness to allergen and methacholine. Am. J. Respir. Crit. Care Med. 152: 1485 1489. 9. Cockcroft, D. W., C. P. McParland, S. A. Britto, V. A. Swystun, and B. C. Rutherford. 1993. Regular inhaled salbutamol and airway responsiveness to allergen. Lancet 342:833 837. 10. Bhagat, R., V. A. Swystun, and D. W. Cockcroft. 1996. Salbutamolinduced increased airway responsiveness to allergen and reduced protection versus methacholine: dose response. J. Allergy Clin. Immunol. 97:47 52. 11. Manolitsas, N. D., J. Wang, J. L. Devalia, C. J. Trigg, A. E. McAulay, and R. J. Davies. 1995. Regular albuterol, nedocromil sodium, and bronchial inflammation in asthma. Am. J. Respir. Crit. Care Med. 151:1925 1930. 12. Adelroth, E., L. Rosenhall, S. A. Johansson, M. Linden, and P. Venge. 1990. Inflammatory cells and eosinophilic activity in asthmatics investigated by bronchoalveolar lavage: the effects of antiasthmatic treatment with budesonide or terbutaline. Am. Rev. Respir. Dis. 142:91 99. 13. Laitinen, L. A., A. Laitinen, and T. Haahtela. 1992. A comparative study of the effects of an inhaled corticosteroid, budesonide, and a beta-2 agonist, terbutaline, on airway inflammation in newly diagnosed asthma: a randomized, double-blind, parallel-group controlled trial. J. Allergy Clin. Immunol. 90:32 42. 14. Adcock, I. M., D. A. Stevens, P. J. Barnes. 1996. Interactions of glucocorticoids and beta 2 agonists. Eur. Respir. J. 9:160 168. 15. Wong, C. S., I. Wahedna, I. D. Pavord, and A. E. Tattersfield. 1994. Effect of regular terbutaline and budesonide on bronchial reactivity to allergen challenge. Am. J. Respir. Crit. Care Med. 150:1268 1273. 16. Hancox, R. J., J. O. Cowan, E. M. Flannery, G. P. Herbison, C. R. McLachlan, C. S. Wong, and D. R. Taylor. 1999. A randomized trial of inhaled agonist, inhaled corticosteroid and their combination in the treatment of asthma. Thorax 54:482 487. 17. Iredale, M. J., S. A. R. Wanklyn, I. P. Phillips, T. Krausz, and P. W. Ind. 1994. Non-invasive assessment of bronchial inflammation in asthma: no correlation between eosinophilia of induced sputum and bronchial responsiveness to inhaled hypertonic saline. Clin. Exp. Allergy 24:940 945. 18. Yan, K., C. Salome, and A. J. Woolcock. 1983. Rapid method for measurement of bronchial responsiveness. Thorax 38:760 765. 19. Fahy, J. V., J. Liu, H. Wong, and H. A. Boushey. 1993. Cellular and biochemical analysis of induced sputum from asthmatic and from healthy subjects. Am. Rev. Respir. Dis. 147:1126 1131. 20. Gauvreau, G. M., M. Jordana, R. M. Watson, D. W. Cockroft, P. M. O Byrne. 1997. Effect of regular inhaled albuterol on allergen-induced late responses and sputum eosinophils in asthmatic subjects. Am. J. Respir. Crit. Care Med. 156:1738 1745. 21. Lai, C. K., O. P. Twentyman, and S. T. Holgate. 1989. The effect of an increase in inhaled allergen dose after rimiterol hydrobromide on the occurrence and magnitude of the late asthmatic response and the associated change in nonspecific bronchial responsiveness. Am. Rev. Respir. Dis. 140:917 923. 22. Nielson, C. P., and N. E. Hadjokas. 1998. Beta-adrenoceptor agonists block corticosteroid inhibition in eosinophils. Am. J. Respir. Crit. Care Med. 57:184 191. 23. Evans, D. W., C. M. Salome, G. G. King, S. J. Rimmer, J. P. Seale, and A. J. Woolcock. 1997. Effect of regular inhaled salbutamol on airway responsiveness and airway inflammation in rhinitic non-asthmatic subjects. Thorax 52:136 142. 24. Smith, C. M., and S. D. Anderson. 1989. Inhalation provocation tests using nonisotonic aerosols. J. Allergy Clin. Immunol. 84:781 790. 25. Juniper, E. F., P. A. Kline, M. A. Vanzieleghem, E. H. Ramsdale, P. M. O Byrne, and F. E. Hargreave. 1990. Effect of long-term treatment with an inhaled corticosteroid (budesonide) on airway hyperresponsiveness and clinical asthma in nonsteroid-dependent asthmatics. Am. Rev. Respir. Dis. 142:832 836. 26. Gauvreau, G. M., J. Doctor, R. M. Watson, M. Jordana, and P. M. O Byrne. 1996. Effects of inhaled budesonide on allergen-induced airway responses and airway inflammation. Am. J. Respir. Crit. Care Med. 154:1267 1271. 27. Claman, D. M., H. A. Boushey, J. Liu, H. Wong, and J. V. Fahy. 1994. Analysis of induced sputum to examine the effects of prednisone on airway inflammation in asthmatic subjects. J. Allergy Clin. Immunol. 94:861 869. 28. Djukanovic, R., J. W. Wilson, K. M. Britten, S. J. Wilson, A. F. Walls, W. R. Roche, P. H. Howarth, and S. T. Holgate. 1992. Effect of an inhaled corticosteroid on airway inflammation and symptoms in asthma. Am. Rev. Respir. Dis. 145:669 674. 29. Wilding, P. J., M. M. Clark, J. Oborne, J. A. Bennett, and A. E. Tattersfield. 1996. Effect of regular terbutaline on the airway response to inhaled budesonide. Thorax 51:989 992. 30. Sears, M. R., D. R. Taylor, C. G. Print, D. C. Lake, E. M. Flannery, M. K. Lucas, D. M. Yates, G. P. Herbison, and Q. Li. 1990. Regular inhaled -agonist treatment in bronchial asthma. Lancet 336:1391 1396. 31. Taylor, D. R., M. R. Sears, G. P. Herbison, E. M. Flannery, C. G. Print, D. C. Lake, D. M. Yates, M. K. Lucas, and Q. Li. 1993. Regular inhaled -agonist in asthma: effects on lung function and exacerbations. Thorax 48:134 138.