The Effect of Salmeterol on Nocturnal Symptoms, Airway Function, and Inflammation in Asthma

The Effect of Salmeterol on Nocturnal Symptoms, Airway Function, and Inflammation in Asthma* Monica Kraft, MD; Sally E. Wenzel, MD; Charlotte M. Bettinger, BS; and Richard]. Martin, MD

Study objective: To determine the efficacy of salmeterol alone in a group of patients with moderate asthma with nocturnal worsening of symptoms. Design: Double-blind, randomized, placebo-controlled crossover study. Setting: Tertiary care hospital specializing in respiratory diseases. Participants: Ten patients with nocturnal asthma. Interventions: Subjects were randomized to salmeterol, 100 J.tg twice daily, or placebo for 6 weeks with a 1-week washout between treatment periods. Symptoms, nocturnal awakenings, and 13 2 -agonist use were recorded daily. Spirometry was performed at weeks 1 and 6 of each period at bedtime and at 4 AM, and methacholine challenge was performed at 4 AM followed by bronchoscopy with BAL. BAL fluid analysis included cell count and differential count, eosinophil cationic protein, Charcot-Leyden crystal protein, leukotriene B 4 , and thromboxane B 2 • Results: The percentage of nights with awakenings decreased significantly with salmeterol (69.8±8.7% vs 30.6±10.8% for placebo and salmeterol, respectively; p=0.02). The percentage of 24-h days with supplemental inhaled 13 2 -agonist use significantly decreased with salmeterol (85.9±9.4% vs 70.4±10.1% for placebo and salmeterol, respectively; p=0.04). There were no significant differences in bronchial reactivity, 4 AM FEV 1, overnight percentage change in FEV 1, or indexes of airway inflammation. Conclusions: Salmeterol alone improves the number of nocturnal awakenings and supplemental 24-h 13 2 -agonist use in nocturnal asthma without significantly altering lung function and airway (CHEST 1997; 111:1249-54) inflammation. Key words: inflammation; nocturnal asthma; salmeterol

Abbreviations: AMP=adenosine monophosphate; DMEM=Dulbecco's modified eagle medium; ECP=eosinophil cati onic protein; LTB 4 = leukotriene B4 ; NA=nocturnal asthma; PC 20 =provocative concentration of methacholine required to produce a 20% fall in FEV1 ; PEFR=peak expirat01y flow rate; TXB 2 =thromboxane B2

s

almeterol, an inhaled [3 2 -agonist with a prolonged duration of action, has been used extensively in Europe and now in the United States for the treatment of asthma. 1-4 Because of its long duration of action, it also has the potential to treat nocturnal worsening of asthma, referred to as nocturnal asthma (NA).5 .6 In moderately severe NA, it is unknown if salmeterol alone can improve this condition. Several studies have shown efficacy of salmeterol in NA, but mainly in *From the D epartment of Medicine, Kational Jewish Center for Immunology and Respiratory Medicine, and the Pulmonary and Ctitical Care Division, Unive rsity of Colorado H ealth Sciences Center, Denver. Supported by grants from the Glaxo-Wellcome Company and the Parker B. Francis Foundation. Manuscript received October 22, 1996; revision accepted December 27, 1996. I Reprint requests: Dr. Kraft, National J ewish Center f or mmunology and Respiratory Medicine, 1400 Jackson Street, Room J107, Denver, CO 80206

combination with inhaled corticosteroids. 3 ·5 -8 The ability of salmeterol to improve asthma symptoms is thought to be due to smooth muscle relaxation, and possibly by exerting anti-inflammatory effects, although the latter mechanism is somewhat controversial. Gardiner and colleagues9 evaluated BAL fluid from subjects with asthma and found no change in total BAL cell count and differential count after salmeterol use. However, Butchers and colleagues 10 showed that salmeterol reduced the release of inflammatory mediators such as histamine, leukotriene C4 , and leukotriene D4 from human lung tissue in vitro. Given the paucity of information on the solitmy efficacy of salmeterol in NA and the conflicting literature on its anti-inflammatory properties, the goal of this study was to determine if salmeterol independently improves symptoms and lung function while decreasing airway inflammation in patients ·with moderate to severe NA. CHEST / 111 /5/ MAY, 1997

1249

MATERIALS AND METHODS

3

Subjects Ten asthmatics were recruited from the general Denver com munity. They met diagnostic criteri a fo r asthma. 11 All 10 experienced NA, defined as a docum ented f all in ove rnight peak expiratory flow rate (PEFR) of >20% on at least four of seven nights of testing at home during the run-in period. Exclusion criteria included use of oral or inhaled steroids, methotrexate, cromolyn, or antibiotics within 6 weeks, astemizole within 12 weeks, any investigational drug within 30 days, upper respiratmy tract infection within 6 weeks, vmi ab le shift schedule, any histoty of significant nonasthmatic pulmona1y disease, significant medical illness as determined by the principal inves tigator, or pregnancy. In addition, subjects could have no history of tobacco use over the past 1 year and <5 total pack-years of use. Informed consent was obtained for this Institutional Review Board-approved study. Subject characte1istics are shown in Table l. P-rotocol The study was a randomized double-blind placebo-controlled crossover study. The study consisted of two treatmen t periods, each 6 weeks in length, \vith a 1-week washout period. Prior to randomization, subjects unde twent screening, including ahistory and physical examination, chest radiograph, ECG, and analysis of CBC count and che mis try. Subjects were also instructed on the use of a peak flowm eter and keeping their peak flows durin g the run-in pe1iod (7 to 14 days) to document a nocturnal decre men t in lung fun ction prior to randomi zati on. They were th en randomized to treatmen t \vith salmeterol, 100 f.Lg/d (two puffs at 7 AYI and 7 PM ) or placebo during first treatm ent period, fo llowed by the alternate therapy during th e second period after a 1-week washout phase. Subjects kept di a1y cards, in which they recorded their PEFRs ptior to bedtime and in the morning, and rated their symptoms of chest tightness, wheezing, shminess of breath, and cough from 0 to 4. The washout phase consisted of treatment with a short-acting inhaled 13 2 -agonist as needed only. Each treatment period included a 4 AM methacholine challenge followed by bronchoscopy with BAL at th e end of weeks 1 and 6. Methacholine Challenge Meth acholine challenge was perform ed as previously described.12 Methacholine was administered as an aerosol in increasing concentrations (0.0175 to 25.0 mgldL) at 5-min intervals via a nebulizer (De Vilbiss 646; Somerset, Pa) powered by pressurized air (20 psi) delivered through a dosimeter (Rosenthal-French; Baltimore) that was tri gge red by a solenoid

Table !- Subj ect Characteristics Characteristics

No.

Age, yr Gender, M/F Nocturnal awakenings Medications % pred FEV 1 4 PM % pred FEV 1 4 AM Overnight fall PEFR, %, measured bedtime to AM Years with asthma

29.9::':: 1.9 8/2 > 4/wk Inhaled short-acting !32 agonists only 76.3::'::4. 1% 57.7::'::3.9% 24.3::'::3.4%

1250

valve set to remain open 0.6 s. Subjects performed five inspiratory capacity inhalations at each concentration of methacholine followed b yspirometry (Moose; Cybermedic; Louisville, Colo) min later. The challenge was concluded after reaching the concentration of methach oline that provoked a 20% or greater reduction in FEV1 from prechallenge baseline (PC 20 ).

> 10

Bronchoscopy With BAL Bronchoscopy with BAL was performed within 30 min of methacholine challenge using identical procedures previously described. 12 Prior to th e procedure and after methacholine challenge, subjects received 0.18 mg albuterol from a metered dose inhale r, 60 mg codeine and 0.6 mg atropine IM. In addition, 4% lidocaine (Xylocaine) was used to anesthetize the upper atrway an d 1 %lidocaine (Xylocaine) was applied to the laryngeal area, trachea, and orifice of the right middle lobe or lingula via th e bronchoscope. Subjects were randomi zed to undergo bronchoscopy with BAL of the right middle lobe or lingula the first week \vith th e alternate lobe lavaged during the se cond bronchoscopy at week 6 of th e treatm en t p eriod. A imilar s randomization was perform ed for the bronchoscopy site during the second treatment period. BAL was performed using five 60-mL aliquots of sterile normal saline solution at 37°C. Lavage fluid was obtained by immediate gentle hand suction applied to each instilling syringe. Nasal oxygen at 3to 4 Umin and pulse oximetry were used throughout the procedure. Subjects were monitored postbronchoscopy via heart rate, BP, and pulse oximetry until hospital discharge 4 to 6 h l ater. BAL Flu·id Analysis Cell Count and Differential Count: The lavage fluid was immedi ately put on ice, and the aliquots were combined and centrifuged for 10 min at 1,200 rpm at 4°C to separate cells from fluid. Differential cell co unts were done from a known volume of lavage \vi th a Giemsa-type stain ( Diff~Quick ; Dade Diagnostics, Inc; Aguada, PR). Cell counts were done \vith fresh lavage fluid and at least 500 cells were counted to obtain the differential cell co unt. Results are expressed as cells per milliliter of BAL fluid. Eosinophil Studies-Charcot-Leyden Crystal Protein and Major Basic Protein: Charcot-Leyden crystal protein was measured as previously described.l3. 14 One hundred microliter BAL fluid aliquots were placed in a "double-sandwich" radioimmunoassay. Results are expressed in nanograms per milliliter. Major basic protein was also measured as previously described via radioimmun oassay. Results are expressed in nanograms per milliliter. l4 Macrophage Studies-Leukotriene B4 and Thromboxane B0 : The BAL flu id rnacrophages were resuspended at aconcentr~­ tion of 1 X 106 macrophages per milliliter in Dulbecco's modified eagle medium (DMEM ) with 10% fetal bovine serum and 100 f.Lg/mL peni cillin/100 f.Lg/mL streptomycin. These cells were then plated in a 24- or 48-well plastic culture plate (depending on yield) and allowed to adhere for 2 h at37°C/10% C0 2 . After this 2-h adherence, the cells were washed three times with cold phosphate-buffered saline solution and new media (DMEM/ 0.1% bovine serum albumin) containing cal cium ionophore (10 f.LJllOlJL A23187) or a vehicle control (0.001% DMSO ) was then added. The supernatants were then harvested and analyzed for the eicosanoids leukottiene B4 (LTB 4 ) and thromboxane B2 (TXB 2 ). T~e plated cells were washed again three times (phosphate-buffered saline solution Ca++ and Mg++ ), lysed (0.2 NaOH ), and protein content per well was determined. LTB 4 and TXB 2 were qum1tified in th e supernatant of these adherent cells by a modified enzyme immunoassay, and they were indexed to BAL protein levels. All eicosanoids were reported as picogram per microgram protein \Vith and without the addition of A23187. Clinical Investigations

Eicosanoid production was determined by enzyme immunoassay, LS and cellular lysate protein was d etermined by using a modified Lowry assay.t 6 LTB 4 was measured as a representative of the 5-lipoxygenase pathway and TXB 2 was measured as a representative of the cyclo-oxygenase pathway. The LTB 4 antibody was purchased (from Advanced M agnetics Inc; Cambridge, Mass), while the TXB 2 antibody was a generous gift (of Dr. Frank Fitzpatrick; UCHSC ). The sensitivity of the LTB 4 assay was routinely 30 pglmL and TXB 2 was 15 pg/mL. Protein values averaged 89.9 f1g/1 X 106 cells.

women with an average age of 29.9:±. 1.9 years. All patients were maintained on a regimen of 13 2 -agonists only. All patients had a history of asthma of> 10 years and all exhibited nocturnal symptoms of asthma more than four times a week. No subject had a history of cigarette use.

Statistical Analysis

The mean number of nocturnal awakenings decreased while the subjects took salmeterol and this decrease trended toward significance (0.9:±.0.1 awakenings per night with placebo vs 0.4:±.0.1 awakenings per night with salmeterol; p=0.08) (Fig 1). In addition, the percentage of nights with awakenings decreased with salmeterol (69.8±8.7%) vs 30.6±10.8% for placebo and salmeterol, respectively; p=0.02) (Fig 1). The percent of days (24-h period) with the use of supplemental inhaled 13 2 -agonists also significantly decreased with salmeterol (85.9:±.9.4% vs 70.4:±. 10.1% for placebo and salmeterol; respectively; p=0.04). Symptoms of wheezing, shortness of

Analysis of physiologic and lavage variables listed above were performed via a two-period crossover analysis of variance modelP Carryover effects were tested for significance at the 10% alpha level and period/treatment effects were tested for significance at the 5% alpha level. Results are expressed as mean±SEM.

RESULTS

Subjects

The subject characteristics are shown in Table 1. The 10 subjects consisted of eight men and 1:\vo

Symptoms

/o nights with

#Awakenings/night

0

100

1

p = 0.08

0.8

60

0.4

40

0.2

20

s

p

p = 0.02

80

0.6

0

awakening

0

p

supplemental albuterol use 100

p = 0.04

80

% days

s

60 40

0"'----

p

s

FIGURE 1. The number of awakenings per night due to symptoms of asthma (chest tightness, wheezing, and/or coughing), the percentage of nights with awakening, and the percentage of days (24-h periods ) of supplemental albuterol use while subjects used placebo (P ) and salmeterol (S). CHEST I 111 I 5 I MAY, 1997

1251

breath, and cough did not change significantly v.rith salmeterol. There were no adverse events directly related to salmeterol use.

Pulmonary Function The mean percent predicted 4 AM FEV1 improved, while patients received salmeterol as compared v.rith placebo, but only during week 1 was there a trend toward significance (57.7:±:3.9% with placebo vs 69.0:±:3.9% with salmeterol; p=0.06). After week 6, the percent predicted FEY1 at 4 AM was not significantly different between tl1e two groups, as the placebo group improved, and the improvement v.rith salmeterol was maintained (62.3:±:4.13% vs 68.7:±:4.72% for placebo and salmeterol, respectively; p=0.28). The FVC at 4 AM did not change significantly v.ritb salmeterol during weeks 1 or 6 (81.7:±:4.2% v.rith placebo vs 88.8:±:4.2% witl1 salmeterol after week 1, p=0.27; 86.3:±:3.6% with placebo vs 88.6:±:3.9% v.rith salmeterol after week 6, p=0.69). Furthermore, the percent overnight fall in FEY 1> measured from bedtime to 4 AM, did not significantly improve v.rith salmeterol dming weeks 1 or 6 (20.9:±:3.2% \vith placebo vs 16.8:±:5.9% v.ritl1 salmeterol after week 1, p=0.56; 23.8:±:5.0% v.rith placebo vs 18.0:±:5.9% \vitl1 salmeterol after week 6 ,p=0.28). The PC 20 for methacholine was higher after week 1 v.rith salmeterol (0.04:±:0.04 mglmL v.rith placebo vs 0.18:±:0.06 mglmL \vitl1 salmeterol; p=0.05). After week 6,the PC 20 results were not significantly different between salmeterol and placebo as the placebo group improved, and tl1e PC 20 \vith salmeterol did not change significantly (0.14:±:0.09 mglmL \vith placebo vs 0.13:±:0.07 mglmL v.ritl1 salmeterol; p=0.84).

BAL Fluid Analysis Three subjects were unable to undergo bronchoscopy at all four time points due to significantly reduced FEY1 . Of these three subjects, two subjects underwent two of four bronchoscopies and the third subject unde1went three of four bronchoscopies. Analysis of the BAL fluid cell count and differential count showed a trend toward a significant improvement in the number of eosinophils when subjects used salmeterol, but only after week 6 (3.3 ±0.1X104/mL v.rith placebo vs 13.2±0.6X104 /mL \vith salmeterol after week 1, p=O.ll; 2.6±0.1Xl04/mL v.rith placebo vs 1.3±0.1 X104/mL v.rith salmeterol after week 6, p=0.08 ). Within the salmeterol group, LTB 4 and TXB 2 production after stimulation of BAL macrophages v.rith A23187 decreased from week 1 to week 6, and approached statistical significance (LTB 4 : 2,092±878 pglf.lg protein after week 1 to 623±418 pglf.lg protein after week 6, p=0.07; TXB 2 : 44.4± 19.4 pg/f.lg protein after week 1 to 9.9:±:3.8 pg/f.lg protein after week 6, p=0.12). There were no significant differences between the absolute number and percentage of macrophages, lymphocytes, or neutrophils, eosinophil cationic protein (ECP), and Charcot-Leyden crystal protein after either week 1 or week 6. BAL fluid results are summarized in Table 2.

DISCUSSION

Our study reveals that salmeterol does improve symptoms of NA by decreasing the percentage of nights v.rith awakenings. However, the objective mea-

Table 2-BAL Variables* Week 1

Cell coun t and differential' Total WBC, X 104 / mL o/oeos %macs %lymphs %pmns Eosinophi l products 1 CLC, nglmL MBP, f.Lg/m L Macrophage products 1 LTB 4 , pg/f.Lg prote in , control LTB 4 , pg/f.Lg protein , sti m.l TXB 2 , pg/f.Lg protein , control TXB 2, pg/f.Lg protein , stim. 1

Week 6

Placebo

Salmeterol

Placebo

Salmeterol

10.1 :!:: 1.5 3.3:!::0.7 84.4:!::2.6 10.8:!::2.7 1.4:!::0.4

15.4:!::2.2 7.3:!::2.7 79.0:!::5.3 10.0:!::2.6 3.9:!:: 1.7

10.8:!::2.5 2.9:!::0.8 84.2:!::3.3 12.0:!::3.7 0.8:!::0.2

10.2:!::1.3 1.7 :!::0.6 88.4:!::2.5 7.8:!::2.6 2.1 :!::0.8

8.5:!::2.6 6.6:!::1.6

20.0:!::10.4 6.4:!::0.8

7.1 :!::2.0 5.3:!::0.4

7.7:!::3.9 6.3:!:: 1.4

46.0:!::28.3 1,341 :!:: 601 46.9:!::22.9 22.0:!::9.0

77.8:!::71.8 2,092:!::878 77.8:!::38.3 44.4:!:: 19.4

18.1 :!:: 11.3 1,149 :!::321 62.1:!::33.0 22.4:!::7.9

49.4 :!::15.7 623:!::418 85.7:!::46.0 9.9:!::3.8

*eos=eosinophils; macs= macrophages; lymphs = lymp hocytes ; pmns = neutrophils; CLC=Charcot-Leyden crystal protein ; MBP = major basic p rotein. 1 ~o statistically significant difference between placebo and salme te rol. 1Stim = Macrophages stimulated with A23 187.

1252

Clinical Investigations

surements of 4 AM FEV1 and overnight fall in lung function were not significantly improved. Except for a trend toward a d ecrease in BAL eosinophils and LTB 4 at week 6 ,there were no significant anti-inflammatory effects. However, salmeterol did not increase aitway inflammation or worsen lung function and bronchial reactivity over the 6-week trial. Several studies have shovvn efficacy of salmeterol in NA. Three studies b yGreening et al,3 Fitzpatrick et al,5 and Brambilla e t al6 showed improvement of early morning PEFRs, number of nocturnal awakenings, and overall circadian variation of lung function with salmeterol. Britton et al7 and Lundbeck et al8 also noted similar findings. In addition, Fitzpatrick and colleagues5 demonstrated that subjects experienced increased stage 4 sleep when t aking the 50-f.Lg salmeterol dose, inferring an improvement in sleep quality. Importantly, most or all of the subjects in these three studies were using oral and/or inhaled corticosteroids and many subjects were also u sing tl1eophylline preparations, ipratropium bromide, and/or oral [3 2 -agonists. This is an important difference from our study in that our subjects were not taking a ny anti-inflammatory or other types of m edications. This difference may also explain why wedid not find any objective measurement of improvement in overnight lung function vvith salmeterol. Therefore, salmeterol may be auseful adjunct to treatment of NA asthma if anti-inflammatory medications are already in use, particularly in patients with moderate to severe asthma. We believe that this is an impmtant point and, in fact, it is supported by Faurschou et al 18 and Greening e t al. 3 They evaluated the effect of salmeterol in patients with poorly controlled NA who were using other agents, including corticosteroids. In these studies, adding salmeterol significantly improved nocturnal lung function compared vvith placebo. This study was conducted as a pilot study to evaluate the effect of salmeterol on both lung function and BAL inflamm atory parameters, and hence the sample size used was small (10 subjects in a crossover design). As in many small studies, our power to detect statistically s ignificant differences in lung function for the salmeterol group compared v.rith the placebo group is also small. Power calculations were performed at the study's completion and showed that we would need to study 90 subjects to have 80% power to detect the observed difference between active and placebo treatment in mean 4 AM FEV 1 at 6 weeks of 6.4% predicted (SD = 18.ll ). However, more interesting to us than the limited power was the small effect size we observed. The difference in 4 AM FE\\ (6.4% predicted) between placebo and salmeterol at week 6 that we observed was smaller than what is usually considered a clinically relevant difference (10 to 12%) . This finding

suggests that the efficacy of salmeterol in a group of patients with moderately severe asthma may not be of a clinically significant magnitude when salmeterol is used alone, although 6 weeks may not have been a sufficiently long treatment period to demonstrate anti-inflammatory effects. A larger s ample size in this study would increase our power to detect a statistically significant difference in means between the salmeterol and placebo groups, but it is unlikely that an increased sample size will change the magnitude of the effect being observed. Perhaps the addition of an anti-inflammatory medication is needed to achieve significant improvement in lung function and inflammatory parameters in patients with moderately severe asthma. The anti-inflamm atory effect of long-acting inhaled [3 2 -agonists is controversial, with studies both supporting and refuting this mechanism of action. Pedersen and coworkers 19 illustrated that salmeterol inhibited th e early and late asthmatic responses to allergen, with a concomitant decrease in serum ECP as compared with placebo. This work in part supports the findings by Wong and colleagues,20 in which formoterol, another long-acting inhaled [3 2 agonist, also inhibited the rise in serum ECP 24 h after exposure to allergen, but neither budesonide nor form oterol inhibited sputum eosinophils or CD25-positive lymphocytes. Finally, Calhoun 21 demonstrated that s almeterol reduced BAL eosinophils by approximately 40% 48 h after segmental allergen challenge. Conversely, other investigators have not shovvn anti-inflammatmy effects by salmeterol. Soler and colleagues 22 showed that salmeterol provided bronchoprotection to histamine, but not adenosine 5' monophosphate (AMP ), a direct mast cell stimulator. Subjects undetwent bronchoprovocation t esting with histamine and AMP, each 14 h after salmeterol, salbutamol, or placebo in a randomized, doubleblind, crossover fashion. The PC 20 for histamine was significantly improved with salmeterol , but the PC 20 for AMP was no different among salmeterol, salbutamol, and placebo. These results suggest alack of a cell stabilizing effect of salmeterol, favoring a functional antagonist effect at the level of airway smooth muscle. W empe and coworkers 23 found no effect of bambuterol on the circadian increase in blood eosinophils and ECP levels in asthmatics v.rith nocturnal worsening of symptoms . Gardiner and colleagues9 evaluated BAL fluid from subjects v,rith asthma after 8 weeks of salmeterol. They found no change in th e BAL total cell count, differential count, lymphocyte subsets, and mast cell tryptase. Their study differs from ours in that all subjects were using 400 to 1,000 f.Lg beclomethasone dipropionate daily for at least 2 CHEST / 111 / 5 / MAY, 1997

1253

months prior to study, and nocturnal symptoms and inflammation were not evaluated. Although salmeterol initially improved the PC 20 to methacholine at the end of week l as compared with placebo, the PC 20 was not significantly different than the placebo group by week 6 due to a change in the placebo PC 20 . Our results suggest that salmeterol exerts no significant bronchoprotective effect to methacholine, nor does it increase bronchial hyperresponsiveness. These results support those of Booth and coworkers, 24 who showed no change in bronchial hyperresponsiveness during 8 weeks of treatment, and no rebound increase in responsiveness 2 weeks after treatment was discontinued. However, these results conflict with the report by Cheung and colleagues, 25 who showed that PC 20 to methacholine improved after the first dose of salmeterol, but decreased significantly by day 28 of treatment. Of note, the 19 of 26 patients in the Booth et al study were using inhaled corticosteroids, as compared with none of the 24 subjects in the Cheung et al report, again suggesting a possible synergistic relationship between salmeterol and inhaled corticosteroids. Our study of moderately severe NA evaluated the independent effect of salmeterol in this group of patients. There were fewer nocturnal awakenings with salmeterol, but this was not accompanied by improved overnight lung function. Additionally, over a 6-week treatment interval, salmeterol neither improved nor worsened multiple inflammatory parameters that were studied. Salmeterol decreased BAL eosinophils, but not significantly so. It appears that physiologic improvement in moderately severe NA using salmeterol occurs with concomitant steroid therapy.3,5,6,18 ACKNOWLEDGMENT: The authors would like to acknowledge Dr. Steven Ackerman for assistance with the eosinophil studies and John Trudeau for assistance with the macrophage studies. REFERENCES 1 D'Alonzo GE, Nathan RA, Henochowicz S, eta!. Salmeterol xinafoate as maintenance therapy compared with albuterol in patients with asthma. JAMA 1994; 271:1412-16 2 Castle W, Fuller R, Hall J, et a!. Serevent nationwide surveillance study: comparison of salmeterol with salbutamol in asthmatic patients who require regular bronchodilator treatment. BMJ 1993; 306:1034-37 3 Greening AP, Ind PW, Northfield M, eta!. Added salmeterol versus higher-dose corticosteroid in asthma patients with symptoms on existing inhaled corticosteroids. Lancet 1994; 344:219-24 4 Jones KP. Salmeterol xinafoate in the treatment of mild to moderate asthma in primary care. Thorax 1994; 49:971-75 5 Fitzpatrick MF, Mackay T, Driver H, et a!. Salmeterol in nocturnal asthma: a double blind, placebo controlled trial of a long acting inhaled ~ 2 agonist. BMJ 1990; 301:1365-68 6 Brambilla C, Chastang C, Georges D , et a!. Salmeterol compared with slow-release terbutaline in nocturnal asthma. Allergy 1994; 49:421-26 1254

7 Britton MG, Earnshaw JS, Palmer JBD. A 12-month comparison of salmeterol with salbutamol in asthmatic patients. Eur Respir J 1992; 5:1062-67 8 Lundbeck B, Rawlinson DW, Palmer JBD. Twelve month comparison of salmeterol and salbutamol as dry powder formulations in asthmatic patients. Thorax 1993; 48:148-53 9 Gardiner PV, Ward C, Booth H, et a!. Effect of 8 weeks of treatment with salmeterol on bronchoalveolar lavage inflammatory indices in asthmatics. Am J Respir Crit Care Med 1994; 150:1006-11 10 Butche rs PR, Vardey CJ, Johnson M. Salmeterol: a potent and long-acting inhibitor of inflammatory mediator release from human lung. Br J Pharmacol1991; 104:672-76 11 American Lung Association. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am Rev Respir Dis 1987; 136:225-44 12 Kraft M, Bettinger CM, Wenzel SE, et a!. Methacholine challenge does not affect bronchoalveolar fluid cell number and many indices of cell function in asthma. Eur Respir J 1995; 8:1966-71 13 Ackerman SJ, Leogering DA, Gleich GJ. The human eosinophil Charcot-Leydon crystal protein biochemical characteristics and measurement by radioimmunoassay. J Immunol 1980; 125:2113-16 14 Dor PJ, Ackerman SJ, Gleich GJ. Charcot-Leyden crystal protein and eosinophil granule major basic protein in sputum of patients with respiratory diseases. Am Rev Respir Dis 1984; 130:1072-1173 15 Pradelles P, Grass J, Maclouf J. Enzyme immunoassay of eiconsanoids using an acetylcholine esterase as label: an alternative to radioimmunoassay. Anal Chern 1985; 57:1170-73 16 Markwell M, Haas S, Bieber C, et a!. A modification of the Lowry procedure to simplifY protein determination in membrane and lipoprotein samples. Anal Biochem 1978; 87:206-10 17 Jones B, Kenward MG, eds. Design and analysis of crossover trials. New York: Chapman & Hall, 1989 18 Faurschou P, Engel A, Haanaes 0. Salmeterol in two different doses in the treatment of nocturnal bronchial asthma poorly controlled by other therapies. Allergy 1994; 49:827-32 19 Pedersen B, Dahl R, Larsen BB, et a!. The effect of salmeterol on the early- and late-phase reaction to bronchial allergen and postchallenge variation in bronchial reactivity, blood eosinophils, serum eosinophil cationic protein, and serium eosinophil protein X. Allergy 1993; 48:377-82 20 Wong BJ, Dolovich J, Ramsdale EH, et a!. Formoterol compared with beclomethasone and placebo on allergeninduced asthmatic responses. Am Rev Respir Dis 1992; 146:1156-60 21 Calhoun W. The role of salmeterol in segmental allergen challenge [abstract]. Am J Respir Crit Care Med 1996; 153:A804 22 Soler M, Joos L, Bolliger CT, et a!. Bronchoprotection by salmeterol: cell stabilization or functional antagonism?: comparative effects on histamine- and AMP-induced bronchoconstriction. E ur Respir J 1994; 7: 1973-77 23 Wempe JB, Tammeling EP, Koeter GH, et a!. Blood eosinophil numbers and activity during 24 hours: effects of treatment with budesonide and bambuterol. J Allergy Clin Immunol 1992; 1992:757-65 24 Booth H, Fishwick K, Harkawat R, e t a!. Changes in methacholine induced bronchoconstriction with the long acting ~ 2 agonist salmeterol in mild to moderate asthmatic patients. Thorax 1993; 48:1121-24 25 Cheung D, Timmers C, Zwinderman AH, et a!. Long-term effects of a long-acting ~ 2 -adrenoceptor agonist, salmeterol, on airway hyperresponsiveness in patients with mild asthma. N Eng! J Med 1992; 327:1198-1203 Clinical Investigations