A 6-Month Comparison Between Formoterol and Salmeterol in Patients with Reversible Obstructive Airways Disease

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

RESPIRATORY MEDICINE (1998) 92, 836-842

A 6-month comparison between formoterol and salmeterol in patients with reversible obstructive airways disease

D. VERVLOET”, T. EKSTR~M+, R. PELA*, F. DUCE GRACIA~, C. Korrll, B. D. SILVERT~, E. QUEBE-FEHLING”“, G. DELLA CIOPPA** AND G. DI BENEDETTO”’

“Hopital Sainte Marguerite, Marseilles, France ‘University Hospital, Linkiiping, Sweden *Ospedale Civile, Osimo, Italy ‘Hospital Clinico ‘Lozano Bless’, Zaragoza, Spain “Muenchenstein, Switzerland “Stonehill Medical Centre, Bolton, U.K. ““Novautis Pharma AG, Basel, Switzerland

The aim of this randomized, open, parallel group study was to compare the clinical efficacy of formoterol dry powder capsule 12pg b.i.d. and salmeterol dry powder 5Opg b.i.d. in the treatment of patients with reversible obstructive airways disease. The 6-month treatment was preceded by a 2 week run-in period. Morning pre-dose peak expiratory flow (PEF) during the last 7 days of treatment was the primary variable. Throughout the study, patients recorded morning and evening pre-dose PEF, use of rescue medication, respiratory symptoms and adverse events. Clinic visits were scheduled at monthly intervals. Of the 482 patients randomized (equal numbers in the two treatment groups), 428 completed the study. Four hundred and twenty-five patients were included in the efficacy analysis for the primary variable. For mean morning pre-dose PEF during the last 7 days of treatment, the 95% confidence interval (CI) for the treatment contrast formoterol minus salmeterol was included entirely in the pre-defined range of equivalence (CI limits= - 8.69, +9.84 1 min - ‘). This was also the case for the morning PEF during the last week before each clinic visit. For mean evening pre-dose PEF, the estimated treatment contrasts showed a trend towards superiority of formoterol over salmeterol, which became statistically significant at 2, 3 and 4 months (PcO.05; estimated contrasts 7.27, 10.45 and 10.51 1 min- ‘, respectively). No treatment group differences were found in use of rescue medication and respiratory symptom scores. The incidence of adverse events was similar in the two groups. These findings demonstrate that formoterol 12pg b.i.d. and salmeterol 5Opg b.i.d., both formulated as dry powders, have similar long-term efficacy and safety profiles in patients with reversible obstructive airways disease.

RESPIR. MED. (1998) 92, 836-842

Introduction Formoterol and salmeterol are the first drugs of a new generation of long-acting ,&-agonists (2). Both compounds, &agonists continue to be a cornerstone in the treatment of when administered as an aerosol by a metered dose inhaler bronchial asthma. These agents are the most effective drugs (MDI) or when inhaled as a powder, induce symptomatic available for the relief of acute asthma symptoms (1). Until relief of wheezing and breathlessness, with a duration of recently, their therapeutic usefulness was limited by a action of at least 12 h. Whereas short-acting /&-agonists relatively short duration of action, necessitating frequent have been recommended for use only as a relief medication, administration and resulting in insufficient protection inhaled long-acting &adrenoceptor agonists are usually against, for example, nocturnal asthma. prescribed for regular use in patients with asthma who are still symptomatic while receiving inhaled cortico- steroid therapy. In these patients, clinical studies have Received 10 October 1997 and accepted in revised form 5 February 1998. indicated that chronic treatment with long-acting Correspondence should be addressed to: G. Di Benedetto, Novartis &-agonists improves symptom score, decreases nocturnal Pharmaceuticals U.K. Ltd, Frimley Business Park, Frimley, asthma, improves lung function, and decreases the use of Camberley, Surrey GU16 5SG, U.K. short-acting inhaled &-agonists (336).

0954.6111/98/060836+07 $12.0010 0 1998 W. B. SAUNDERS COMPANY LTD FORMOTEROLAND SALMETEROLINLONG-TERMUSE 837

In vitro comparative studies have shown evidence of or 5Opg salmeterol dry powder (SEREVENT@ differences between formoterol and salmeterol with respect DISKHALERB). A computer-generated randomization to onset and duration of action and dose-dependency of scheme was used to provide balanced blocks of patient effects (7). At the present time, four comparative single- numbers for the two treatment groups within each country. dose studies have been published as full papers, two in A one-to-one treatment allocation and a block size of eight patients with bronchial asthma (8,9) and two in patients were used. No formal checks of patient compliance with with chronic obstructive pulmonary disease (COPD) dosing regimens were made. (10,ll). A direct comparison of efficacy and safety during long-term treatment is still lacking. We conducted a multinational, randomized, open METHODS study using a population of patients with reversible At home, using a mini-Wright peak flow meter, patients obstructive airways disease who could benefit from the use measured their PEF at about 12 h intervals, in the morning of regular inhaled long-acting &agonists. We compared (6 to 9 a.m.) and evening (6 to 9 p.m.). PEF measurements the clinical efficacy and safety of &month treatment with (highest of three consecutive measurements) had to be either formoterol dry powder capsule 12pg twice daily or performed before intake of trial medication and the values salmeterol dry powder 50 pg twice daily. entered in diary cards. Patients were asked not to take rescue medication 6 h prior to the PEF measurements. If rescue medication was used, this was noted in the diary Methods card. In addition, respiratory symptoms and adverse events were recorded. Day- and night-time respiratory symptoms STUDY SUBJECTS were scored on a five grade scale (O=best to 4=worst). Out-patients suffering from moderate-to-severe reversible Patients visited the clinic before and after the run-in period obstructive airways disease were entered into the study. and every 4th week of the treatment period. At each clinic Patients were of both sexes, aged 18 years or older. Patients visit, medication were reviewed and diary cards were col- were required to have had a documented diagnosis of lected. The study was conducted in accordance with the reversible obstructive airways disease for 1 yr or more and Declaration of Helsinki and the principles of Good Clinical to be using regular inhaled corticosteroids at a constant Practice and was approved by the Ethics Committees of dose of at least 400 pug day ~ ’ (or 200 ,ug day- i fluticasone) the participating centres. Signed informed consent was for at least 1 month before inclusion. Patients were obtained from all patients before recruitment. excluded from participation if they had evidence of other clinically relevant diseases. Pregnant or lactating women were also excluded. Patients on P-blocker therapy or with ANALYSIS hypersensitivity to sympathomimetic amines or inhaled The primary variable was the mean morning pre-dose PEF lactose were excluded, as were those who were considered during the last 7 days of treatment and the minimum unable to comply with the study protocol. Inhaled cortico- sample size required for the study was determined in steroids were continued at a constant dose throughout the relation to its statistical power to detect treatment differ- study. ences in mean PEF. Assuming an upper limit for the true Although the vast majority of patients selected according standard deviation of mean morning PEF of 48 1 min - ’ to the above inclusion criteria would have been patients and the use of two-sided significance test at the 5% level, with bronchial asthma, no attempt was made to exclude (or then a total of 180 evaluable patients (90 per treatment) differentiate) patients with COPD, as long as bronchial would have given the study a power of 80% to detect mean reversibility had been demonstrated. This was done in order differences in PEF of approximately 20 1 mini i. The to obtain a sample which was representative of the patient sample size was then doubled in order to allow meaningful population receiving inhaled long-acting &agonists in comparisons in a concomitant assessment of quality of daily clinical practice. life (manuscript in preparation). Secondary end-points included mean morning and evening pre-dose PEF during STUDY DESIGN the last week before each clinic visit, overall mean morning and evening pre-dose PEF for the entire treatment period, The trial was a randomized, parallel group, open study. day- and night-time use of rescue medication, and day- and A total of 41 centres, in France, Italy, Spain, Sweden, night-time symptom scores. Switzerland and the U.K., participated. Following a 2 week The primary efficacy variable and the corresponding run-in period, when patients recorded baseline symptoms, variable derived from the evening pre-dose PEF were use of rescue medication (salbutamol MD1 100 pg puff - * analysed for the population of patient completing the entire or salbutamol dry powder 200 pug dose - ‘) and twice daily treatment period. All other efficacy variables were analysed peak expiratory flow (PEF) measurements, patients who, for those patients who received at least one dose of trial according to the investigator(s), could benefit from the medication and who had at least one measurement during regular use of a long-acting &-agonist were allocated to the treatment period. All randomized patients who received 6 months of twice daily treatment with 12pg formoterol at least one dose of trial medication were included in the dry powder (FORADIL@) administered via the Aerolizer@ safety analysis. Tests for baseline homogeneity of treatment 838 D. VERVLOET ET AL.

TABLE 1. Summary of demographic and baseline data

Formoterol Salmeterol Total (n=241) (n=241) (n=482)

Age (years) Mean (range) 48 (18-78) 47 (18-77) 48 (18-78) Sex (12 %) Males 108 (45) 113 (47) 221 (46) Females 133 (55) 128 (53) 261 (54) Smoking history (n %) Current smokers 36 (14.9) 38 (15.8) 74 (15.4) Ex-smokers 70 (29.0) 77 (32.0) 147 (30.5) Never smoked 135 (56.0) 126 (52.3) 261 (54.1) Duration of obstructive airways disease (years) Mean (range) 15.8 (1.1-54.0) 16.3 (0.7-64.0) 16.0 (0.7-63.0) Morning PEF (1 min - ‘) Mean (range) 377 (110-670) 371 (89-749) 374 (89-749) Evening PEF (1 min ~ ‘) Mean (range) 388 (97-744) 384 (1499800) 386 (97-800) Day-time intake of rescue medication (number of puffs)* Mean (range) 2.1 (O-17.6) 1.9 (O-15.1) 2.0 (O-17.6) Night-time intake of rescue medication (number of puffs)* Mean (range) 1.2 (O-10.9) 1.1 (O-10.9) 1.2 (O-10.9) Day-time symptom score* Mean (range) 0.9 (040) 0.8 (O-3.7) 0.9 (040) Night-time symptom score* Mean (range) 0.6 (04.0) 0.5 (O-3.0) 0.6 (O-4.0)

*Mean of the last seven days of the 2-week run-in phase.

groups were not performed. Instead, treatment group scores were analysed with the van Elteren test stratified differences at baseline were adjusted for by including for centre. appropriate variables in the statistical model. The primary efficacy variable was analysed by means of an analysis of covariance (ANCOVA) with fixed effects for Results treatment, country, centre within country and sex, and with the baseline value (last 7 days before the randomiz- PATIENTS ation visit) as a covariate. The treatment comparison was assessed by constructing, based on the ANCOVA, a This multicentre study enrolled 529 patients. Forty-seven two-sided 95% confidence interval (CI) for the contrast patients were not randomized; the most frequent reasons formoterol minus salmeterol. Equivalence was meant one- were failure to meet the inclusion criteria (17 patients), sided with a limit of 20 1 min - I. Treatment equivalence non-compliance with study procedures (eight patients) and would therefore have been demonstrated if the CI for the withdrawal of consent (eight patients). Thus, 482 patients contrast formoterol minus salmeterol laid entirely above were randomized to treatment. the -2Olmin’ value, i.e. if the mean morning pre-dose Two hundred and forty-one patients were randomized to PEF during the last 7 days of treatment in the formoterol receive formoterol 12pg b.i.d. and the same number to group was at most 20 1 min ~ ’ lower than in the salm- receive salmeterol 50,~g b.i.d. The groups were well- eterol group. To account for the possible effect of the matched for demographical details. Patients’ baseline char- intake of rescue medication, a secondary exploratory acteristics are described in Table 1. Approximately 80% of analysis was performed, where all measurements for patients used salbutamol MD1 as rescue medication. There which an intake of rescue medication within the previous were no significant differences in terms of use of concomi- 6 h had been recorded were excluded. All other efficacy tant medications, including inhaled corticosteroids, variables related to PEF were analysed in the same way as between the two treatment groups. the primary variable. Differences in use of rescue medica- Twenty-four patients from the formoterol group and 30 tion were analysed with the van Elteren test (12) stratified patients from the salmeterol group were withdrawn after for centre and, additionally, with the same ANCOVA randomization. Of these, 26 patients were withdrawn model as the primary efficacy variable. Mean symptom because of adverse events, five because of unsatisfactory FORMOTEROL AND SALMETEROL IN LONG-TERMUSE 839

430 25 1 (a) (a) 420 20 -

. 0-r I I I I I I -101 I I I I I + Baseline 1 2 3 4 5 6 1 2 3 4 5 6

430 25 I (b) (b) 20 1 420 c

370 -5 $ I I I I I I 0 I I I -10’ ’ Baseline 1 2 3 4 5 6 1 2 3 4 5 6 Time (months) Time (months) FIG. 1. Mean morning (a) and evening (b) pre-dose peak FIG. 2. Treatment contrasts (formoterol minus expiratory flow (PEF) over the run-in and treatment salmeterol) including 95% CI of mean morning (a) and periods (mean of last 7 days before each clinic visit). evening (b) pre-dose peak expiratory flow (PEF) over (+-), formoterol; (a-), salmeterol. time (mean of last 7 days before each clinic visit). therapeutic effect, one for not meeting the protocol criteria, For mean evening pre-dose PEF, the estimated treatment eight for non-compliance, six for withdrawal of consent, contrasts showed a trend towards superiority of formoterol two because of administrative problems and six were lost to over salmeterol, which became statistically significant at 2, follow-up. There were no differences between treatment 3 and 4 months (P~0.05, estimated contrasts 7.27, 10.45 groups in the pattern of reasons for premature discontinu- and 10.51 1 min- ‘, respectively) (Fig. 2). Also in this case, ation of trial medication. the analysis excluding measurements after an intake of The efficacy primary end-point could be analysed for 425 rescue medication led to similar results. of the 428 patients completing the study, since for three patients no PEF measurements were available in the diary USE OF RESCUE MEDICATION for the last 7 days of treatment. All 482 patients who used the study medication were included in the safety analysis. At all visits, mean intake of rescue medication was found to be less than half of the baseline value for both treatment groups (Fig. 3). This was seen both at day- and night-time. PEF MEASUREMENTS No statistically significant difference between the two treat- The mean values for the run-in and treatment periods in ment groups was seen at any visit. morning and evening pre-dose PEF are presented in Fig. 1. For the primary efficacy end-point, mean morning pre- dose PEF during the last 7 days of treatment, the 95% CI RESPIRATORY SYMPTOM SCORE for the treatment contrast formoterol minus salmeterol was Mean symptom scores improved in both treatment groups - 8.69, +9.84 1 min - i and was included entirely in the in a similar way (Fig. 4). The percentage of days with a pre-defined range of equivalence (Fig. 2). This was also the symptom score=0 increased in both treatment groups by case for mean morning pre-dose PEF during the last week about the same extent. The symptom score improvements before each clinic visit. The analysis excluding measurement were similar for both day- and night-time periods. There after intake of rescue medication essentially confirmed the was no statistically significant treatment group difference at primary analysis. any examination. 840 D. VERVLOETET AL.

31 SAFETY ( (a) Both treatments were safe and well tolerated. Although adverse events were reported by 190 (79%) patients in the formoterol group and 193 (80%) patients in the salmeterol group, this is not unexpected in a trial of 6 month duration. No treatment group differences were seen. Most frequent adverse events included viral infection, asthma exacerbation, headache, rhinitis and chest infection. Exacerbation of asthma was reported as an adverse event by 41 patients (17%) in the formoterol group and 54 patients (22%) in the 6 salmeterol group. 1When considering only adverse events which were assessed by the investigator as possibly/probably trial drug- related, events were reported in 32 (13%) patients who received formoterol and 21 (9%) patients who received salmeterol. The most frequent possibly/probably trial drug- related adverse event was headache (seven patients in the formoterol group and 11 in the salmeterol group). Other possibly/probably trial drug-related adverse events included tremor (seven patients, five with formoterol and two with salmeterol), asthma exacerbation (eight patients, four in I each group) and palpitations (four patients, all of them in 6 the formoterol group). Time (months) FIG. 3. Mean number of puffs of rescue medication inhaled during day- (a) and night-time (b) over time Discussion (mean of the last 7 days before each clinic visit). (u), formoterol; (m), salmeterol. To our knowledge, this is the first study in which the long-term effects of formoterol and salmeterol on lung function and symptom control have been directly compared. In this mixed population of patients with reversible airways obstruction, which is likely to be representative of l (a) patients receiving inhaled long-acting &agonists in clinical 0.8 practice, formoterol dry powder capsule 12pg b.i.d. and salmeterol dry powder 50 pug b.i.d. were equally effective in 8 * 0.6 terms of PEF values and symptom control. E This was an open-label trial. The use of a double-dummy design would have been necessary to achieve double- binding. This would have added to the already fairly lo.4 0.2 onerous long-term trial schedule. The use of an open design is unlikely, however, to have introduced significant bias. 2 The patient information document merely stated that the Baseline ill1 2 lli 3 4 5 6 two treatments would be compared, without indicating that formoterol was a newer medication. In addition, the main x (b) efficacy variable was derived from objective measurements 0.8 (PEF). m c In vitro studies have shown that formoterol is a more _._ potent &adrenoceptor agonist and has a faster onset of E 2 action than salmeterol, but that the duration of action is Ea 0.4 - longer for salmeterol (7). In addition, the in vitro duration k of action of formoterol, but not of salmeterol, was found to 0.2 - be dose-dependent. There have been a number of studies comparing 0 the effects of single doses of formoterol and salmeterol Baseline 1 2 3 4 5 6 on airway tone. Rabe and co-workers found that 12,ug Time (months) formoterol and 50 pug of salmeterol were equally effective in FIG. 4. Mean respiratory symptom score at day- (a) and mild asthmatics in protecting against methacholine-induced night-time (b) over time (mean of the last 7 days before bronchoconstriction for up to 24 h (8). In a study published each clinic visit). (o), formoterol; (I), salmeterol. as an abstract by Zellweger and colleagues, 50 pg salmeterol FORMOTEROLAND SALMETEROLIN LONG-TERM USE 841 and 24 pg formoterol produced comparable protection type of inhalation device (18) on patients’ compliance with against methacholine-induced bronchoconstriction during treatment. at least 16 h (13). Cazzola and co-workers compared the effects of 5Opg salmeterol and 24,ug formoterol in 16 patients with COPD and concluded that both compounds Acknowledgements were equivalent in terms of maximum bronchodilation and duration of action (10). However, in a second study in 12 The authors would like to thank the following physicians patients with COPD, the same authors reported that 5Opg for their co-operation. France: H. Bensaidane (Grenoble), salmeterol had a longer duration of action than 12 or 24 pug J. Birnbaum (Marseille), M. Grosclaude (Guilherand formoterol (ll), in contrast with their own previous find- Granges), C. Le Loet (Rouen), F. Macone (Nice), D. ings. In a preliminary report on six patients with mild Murciano (Clichy), J. P. Orlando (Aubagne), D. Piperno asthma, Dal Negro’s group has shown that formoterol (Lyon), A. Prud’homme (Tarbes), G. Thomas (Marseille); 24 ,ug and salmeterol 5Opg, but not formoterol 12pg, Italy: R. Corsica (Montescano), C. F. Donner (Veruno), G. protected against methacholine-induced bronchoconstric- Moscato (Pavia), M. Neri (Tradate), C. M. Sanguinetti tion for 12 h (14). In contrast, another preliminary report (Osimo), P. Zanon (Busto Arsizio); Spain: P. Casan Clara has claimed a longer duration of action of 12pg formoterol (Barcelona), F. Manresa Presas (Hospitalet De Llobregat), over 50 ,ug salmeterol in nine patients with partially revers- P. Martin Escribano (Madrid), C. Picado VallCs ible COPD (15). Recently, van Noord and co-workers have (Barcelona), H. Verea Hernando (La Coruna), J. L. Viejo compared single doses of formoterol 24,ug, salmeterol Banuelos (Burgos); Sweden: P. Arvidsson (Goteborg), T. 50 pug and salbutamol 2OOpg and have concluded that, in Wegener (Eskilstuna), 0. Zetterstom (Stockholm); patients with moderately severe asthma, formoterol and Switzerland: U. Aebi (Biel), W. Graf (Bern), M. Hacki salmeterol had an equal bronchodilatory capacity, which (Zuerich), A. Paky (St. Gallen), M. Tschan (Laufen); U.K.: lasted for at least 12 h and that formoterol had a more B. Davies (South Glamorgan), R. J. Davies (London), rapid onset of action than salmeterol, equal to that of D. 0. Edwards (Merseyside), M. Rudolf (London), S. C. salbutamol (9). However, the clinical significance of the Stenton (Newcastle-upon-Tyne), J. R. Stradling (Oxford), study was hampered by the lack of a formoterol 12 pug arm, J. A. Turner (Bournemouth). which is the dose of formoterol recommended in the vast This study was supported by Novartis Pharma AG. majority of patients (16). During recent years, several guidelines for asthma and COPD treatments have been published. Although the role References of inhaled long-acting &agonists in the management of COPD patients remains to be determined (17) asthma 1. National Heart, Lung, and Blood Institute. Guidelines guidelines agreed that regular bronchodilator drug use for the diagnosis and management of asthma. NIH should be recommended when asthma symptoms and func- Publication Number 974051A, May 1997. tional abnormalities are not completely controlled by regu- 2. Lijfdahl CG, Chung KF. Long-acting a-adrenoceptor lar use of inhaled corticosteroids (1). One of the issues to agonists: a new perspective in the treatment of asthma. address is, therefore, the choice between formoterol and Eur Respir J 1994; 4: 218-226. salmeterol. 3. Britton MG, Earnshaw JS, Palmer, JB. A 12-month Our study in a large number of patients with reversible comparison of salmeterol with salbutamol in asthmatic airways obstruction demonstrates that, during 6 months of patients. Euv Respir J 1992; 5: 1062-1067. treatment, formoterol 12pg b.i.d. and salmeterol 5Opg 4. Kesten S, Chapman KR, Broder I. A 3-month com- b.i.d. have similar clinical efficacy in terms of morning PEF, parison of twice daily inhaled formoterol versus four use of rescue medication and symptom score. This sug- times daily inhaled albuterol in the management of gests that the in vitro longer duration of action of salmeterol stable asthma. Am Rev Respir Dis 1991; 144: 622-625. does not have clinical relevance. A surprizing result of 5. Kesten S, Chapman KR, Broder I et al. Sustained the present study was the trend towards superiority improvement in asthma with long-term use of formot- of formoterol over salmeterol in terms of evening PEF, erol fumarate. Ann Allergy 1992; 69: 415420. which became statistically significant at 2, 3 and 4 months. 6. Pearlman DS, Chervinsky P, LaForce C et al. A Further studies, including assessment of patients’ com- comparison of salmeterol with albuterol in the treat- pliance, are needed to investigate the reasons for this ment of mild-to-moderate asthma. N Engl J Med 1992; finding. 327: 1420-1425. In conclusion, our study demonstrates that formoterol 7. Naline E, Zhang Y, Qian Y et al. Relaxant effects and dry powder capsule 12pg b.i.d. and salmeterol powder durations of action of formoterol and salmeterol on 50 fig b.i.d. are equally effective in controlling lung function the isolated human bronchus. Eur Respir J 1994; 7: and symptoms in patients with reversible airways disease 914-920. who take regular inhaled steroids. In addition, they have 8. Rabe KF, Jijrres R, Nowak D, Behr N, Magnussen H. similar safety profiles. Features other than the ilz vitro Comparison of the effects of salmeterol and formoterol longer duration of action of salmeterol should, therefore, on airway tone and responsiveness over 24 hours guide the therapeutic choice between these two drugs, in bronchial asthma. Am Rev Respir Dis 1993; 147: including the impact of speed of onset of action (16) and 143661441. 842 D.~ERVLOETETAL.

9. van Noord JA, Smeets JJ, Raaijmakers JAM, Bommer single-dose trial. Eur Respir J 1994; 7 (Suppl. 18): AM, Maesen FPV. Salmeterol versus formoterol in 422s. patients with moderately severe asthma: onset and 14. Dal Negro R, Micheletto C, Pomari C, Turco P. duration of action. Eur Respir J 1996; 9: 16841688. Inhaled salmeterol 50 mcg, formoterol 24 mcg, but not 10. Cazzola M, Santangelo G, Piccolo A et al. Effect of 12 mcg, produce a sustained protection against the salmeterol and formoterol in patients with chronic methacholine (MCH)-induced bronchial hyper- obstructive pulmonary disease. Pulm Pharmacol 1994; responsiveness. Eur Respir J 1996; 9 (Suppl. 23): 271s. 7: 103-107. 15. Celik G, Kalaycioglu 0, Beder S, Durmaz G. 11. Cazzola M, Matera MG, Santangelo G, Vinciguerra A, Formoterol versus salmeterol in partially reversible Rossi F, D’Amato G. Salmeterol and formoterol in chronic obstructive pulmonary disease. Eur Respir J partially reversible severe chronic obstructive pulmon- 1996; 9 (Suppl. 23): 272s. ary disease: a dose-response study. Respir Med 199.5; 16. Steffensen I, Faurschou P, Riska H, Rostrup J, 89: 357-362. Wegener T. Inhaled formoterol dry powder in the 12. van Elteren PH. On the combination of independent treatment of patients with reversible obstructive airway two-sample tests of Wilcoxon. Bulletin de l’lnstitute disease. Allergy 1995; 50: 657-663. International de Statistique 1960; 37: 351-361. 17. American Thoracic Society. Standards for the diag-- 13. Zellweger JP, Pfenninger M, Ruff P. 24-hour protec- nosis and care of patients with chronic obstructive tive effect of salmeterol xinafoate (50 mcg) versus pulmonary disease. Am J Respir Crit Care Med 1995; formoterol (24 mcg) against methacholine-induced 152 (Suppl). bronchoconstriction in mild to moderate asthmatic 18. Siafakas NM, Bouros D. Choice of inhalation therapy patients. A randomised, double-blind, cross-over, in adults. Eur Respir Rev 1994; 4: 78-81.