Twelve Month Comparison of Salmeterol and Salbutamol As Dry

148 Thorax 1993;48:148-153 Twelve month comparison of salmeterol and as

salbutamol dry powder formulations in Thorax: first published as 10.1136/thx.48.2.148 on 1 February 1993. Downloaded from asthmatic patients

B Lundback, D W Rawlinson, J B D Palmer on behalf of a European study group

Abstract mol (400,ufg twice daily) over a further Background Salmeterol is a potent nine months. Neither salmeterol nor selective IA agonist that has been shown salbutamol was associated with any to have a duration of action in excess of worsening of control of asthma. 12 hours. In this study salmeterol and salbutamol were compared over a three (Thorax 1993;48:148-153) month period with a further extension of nine months. Methods Three hundred and eighty f2 Adrenoceptor agonists are established eight patients with mild to moderate treatment for asthma for patients of all ages. reversible airways obstruction (forced They are effective bronchodilators and pro- expiratory volume in one second (FEV,) tect against various constrictor challenges >50% predicted) were randomised to such as exercise, cold air, methacholine, and receive salmeterol (50 fig) twice daily or histamine.lA The inhaled route is preferred as salbutamol (400 ,ug) four times daily, this delivers the drug directly to the lung and both by dry powder, in a double blind thus minimises potential systemic side parallel group study. During the first effects.5 The major drawback of currently three months detailed assessment of effi- available inhaled drugs is their short duration cacy was made with recording of morn- of action-namely, less than six hours.6 ing and evening peak expiratory flow Salmeterol is a potent selective fl2 agonist that rates (PEF), asthma symptoms, and has a duration of action in excess of 12 hours bronchodilator use when necessary for in vitro and in vivo.8 In a four week parallel http://thorax.bmj.com/ the relief of symptoms. Patients contin- group study in 692 patients with mild to ued in the study for a further nine moderate asthma salmeterol increased morn- months with the salbutamol dose ing and evening peak flow, reduced the varia- reduced to 400 fug twice daily. Lung function in diurnal peak flow, reduced the tion was measured at the clinic and requirement for additional salbutamol, and safety data were collected during this lessened symptoms by comparison with period. placebo.9

Results Salmeterol produced a signifi- This study was designed to investigate the on September 28, 2021 by guest. Protected copyright. candy higher mean morning PEF (mean safety and efficacy of salmeterol (50 ,ug) twice difference compared with salbutamol 21 daily and salbutamol (400 ,ug) four times (95% CI 12-31) Vmin), and a significant daily, both as dry powders, in patients with reduction in mean diurnal variation in mild to moderate reversible airways obstruc- PEF (from 30 1min at baseline to 11 tion over a three month period. The double 1min during salmeterol compared with blind study continued for a further nine 34 1min at baseline to 32 1min during months with the salbutamol treatment group salbutamol treatment). Salmeterol also on a reduced dose of 400 ,ug twice daily. reduced day and night symptoms and use During this period lung function was of rescue bronchodilator. FEV, increased measured at the clinic and safety data were with both salmeterol and salbutamol collected. treatment over the 12 month treatment period. For both treatments the number Medical Division, National Institute of of patients reporting exacerbations of Methods Occupational Health, asthma and the frequency of these ex- PATIENTS Umea, Sweden acerbations remained constant during The inclusion criteria required patients (aged B Lundback the study. Thirty six patients in the 18 or over) to have a forced expiratory vol- Department of Respiratory Medicine, salmeterol and 49 in the salbutamol ume in one second (FEVI) or mean daily Glaxo Group Research group withdrew during the 12 months of peak expiratory flow rate (PEF; mean of the Ltd, Greenford, the study. highest morning and evening measurements) Middlesex UB6 OHE In D W Rawlinson Conclusions this study salmeterol greater than 50% predicted, an increase in J B D Palmer (50 4ug twice daily) was more effective FEV, greater than 15% after taking 200 ,ug Correspondence to: than salbutamol (400 ,ug four times daily) inhaled salbutamol by a pressurised inhaler or Mr D W Rawlinson in the control of asthma over three 400 ,ug inhaled salbutamol by a Diskhaler, a Accepted 6 July 1992 months, and more effective than salbuta- symptom score of 2 or more (for scale see Twelve month comparison ofsalmeterol and salbutamol as dry powderformulations in asthmatic patients 149

Table 1 Asthma symptom scores (200,g) twice daily under double blind, double dummy conditions. Night time asthma symptom score a O = No symptoms during the night I = Symptoms causing you to wake once or to wake early CONCOMITANT MEDICATION Thorax: first published as 10.1136/thx.48.2.148 on 1 February 1993. Downloaded from 2 = Symptoms causing you to awake twice or more (including waking early) At the start of the run in period all oral, par- 3 = Symptoms causing you to be awake for most of the night enteral, rectal, or inhaled ,B receptor agonists 4 = Symptoms so severe that you did not sleep at all (other than the salbutamol inhaler provided), Daytime asthma symptom score methylxanthines, and anticholinergic drugs 0 = No symptoms during the day were withdrawn. The patients continued I = Symptoms for one short period during the day inhaled corticosteroids (table 2), oral pred- 2 = Symptoms for two or more short periods during the day 3 = Symptoms for most of the day which did not affect your normal daily activities nisolone (<20 mg per day), sodium cromo- 4 = Symptoms for most of the day which did affect your normal daily activities glycate, nedocromil, or ketotifen at 5 = Symptoms so severe that you could not go to work or perform normal daily activities unchanged doses throughout the study. Short courses of increased inhaled or oral corticosteroids were allowed during exacerbations. table 1), or a diurnal variation in PEF of at least 15% on four of the last seven days of the PROTOCOL run in period. Patients who required >20 mg Patients attended the clinic on seven occa- prednisolone were excluded. A total of 460 sions during the first three months of the patients with a clinical history of mild to study with three further visits during the next moderate reversible airways obstruction and nine months. At the first visit demographic no other serious medical conditions entered details were recorded and blood samples were the run in period and of these 388 (table 2) taken for biochemical and haematological were randomised to receive treatment (190 analysis. Patients were given a daily record received salmeterol, 198 received salbuta- card to record their morning and evening mol). The study was performed in 47 centres PEF with a mini-Wright peak flow meter in 11 countries (see acknowledgements). All (before study or relief medication). They patients gave informed consent. Regulatory were asked to record their daytime and night permission was obtained from all countries time symptoms according to a scoring system where it was required and approval from the (see table 1), and to note their use of rescue appropriate ethics committee was obtained salbutamol. If patients met the inclusion crit- for all centres. eria they entered the treatment phase. They were seen initially at intervals of two weeks STUDY MEDICATION for four weeks (visits 3, 4, and 5) and then The study was of a randomised, double blind, every four weeks (visits 6 and 7) to three double dummy, parallel group design. There months. was a two week run in period in which The patients then continued a nine month http://thorax.bmj.com/ patients took either salbutamol by pressurised safety phase with visits at three-monthly inhaler or a salbutamol Diskhaler when nec- intervals (visits 8, 9, and 10). At each visit the essary for the relief of symptoms. This same physician recorded changes in medication; rescue medication was also available during intercurrent illnesses, adverse events, and the treatment period. After the run in period, withdrawal and measured blood pressure, patients were randomised for three months to heart rate, and FEV1 and forced vital capacity salmeterol (50 pg) twice daily and placebo (FVC). Blood and urine samples were col-

four times daily or salbutamol (400 ,ug) four lected at all visits (except visit 4) and 12 lead on September 28, 2021 by guest. Protected copyright. times daily and placebo twice daily. During electrocardiograms were recorded at visit 3 the final nine months patients received either (before study treatment), visit seven (after salmeterol (50 ,ug) twice daily or salbutamol three months), and visit 10 (after 12 months). Table 2 Patients' characteristics POWER CALCULATION Salmeterol Salbutamol From results of our previous experience with (50 pg twice daily) (400 pgfour times a day) salbutamol it was estimated that the residual standard deviation of mean morning or Number allocated to treatment 190 198 Men 89 (47) 100 (51) evening PEF was unlikely to exceed 45 /min. Women 101 (53) 98 (49) From this estimate of variability it was calcu- Median age (y) 42 46 lated that with 388 evaluable patients the Range 19-79 18-78 study had a power of 0 91 to detect a mean Median duration of asthma (y) 10-0 11-5 difference in mean PEF between the two Mean PEF (1/min) (am/pm) 367/397 349/384 groups of 15 1/min, assuming a two sided t Mean FEV, (1) 2-46 2-36 test and a 5% level of significance. (% predicted) 76 73 Smoking history: ANALYSIS Yes 23 (12) 27 (14) To be included in the analysis of a variable, No 108 (57) 110 (56) patients must have provided at least four days Ex-smoker 59 (31) 61 (31) of data for that variable during the run in Concurrent steroids period and data from at least half of the days Inhaled only 108 (57) 112 (57) in any period of assessment. The mean PEFs Oral only 7 (4) 11 (6) Inhaled and oral 13 13 in the morning and evening were calculated (7) (7) for each patient together with the mean dif- % in parentheses ference between salmeterol treatment and 150 Lundback, Rawlinson, Palmer

Figure 1 Changes in 50 Evening mean morning peak flow (PEF; left hand panel) and mean evening PEF (right

hand panel) during 12 Thorax: first published as 10.1136/thx.48.2.148 on 1 February 1993. Downloaded from weeks of treatment with sal- O Salmeterol meterol (50 pg twice daily; E c open circle) and salbutamol LL * Salbutamol (400 pgfour times daily; LL. U- closed circle). Changes in wc w morning PEF were signifi- CLlU cantly higher in patients 0~ as taking salmeterol CU (p < 0 001). E E C 0, 0,cm C CUco to 0

-1 o 1 2 3 4 5 6 7 8 9 11 Weeks of treatment Weeks of treatment

salbutamol treatment for these responses. ma (seven salmeterol, eight salbutamol). The The mean values obtained during treat- two treatment groups were well matched for ment were subjected to analysis of covariance demographic details, smoking habits, and with the mean values recorded in the second duration and severity of their reversible air- run in period as covariates. Changes in the ways disease (table 2). During the treatment proportion of nights with asthma score 0, period 57% of patients receiving salmeterol proportion of days with asthma score less and 57% of patients receiving salbutamol also than 2, proportion of nights with no addition- received inhaled corticosteroids. al salbutamol, and proportion of days with no Figure 1 shows the changes in mean morn- additional salbutamol were subjected to ing and evening PEFs over the 12 weeks of analyses of variance after transformation to treatment. Increase in mean morning PEF ranks. was significantly higher in patients taking salmeterol than in patients taking salbutamol

(p < 0 001); the mean difference was 21 http://thorax.bmj.com/ Results I/min (95% confidence interval (95% CI) Of the 388 patients who were randomised to 12-31). Salmeterol also caused a significant treatment (190 salmeterol, 198 salbutamol) reduction in the diurnal variation in PEF 303 patients (154 salmeterol, 149 salbuta- compared with salbutamol (p < 0-001); the mol) completed the 12 months of treatment. mean variation was from 30 1/min at baseline Forty five patients were withdrawn during the to 11 1/min during salmeterol treatment com- three month treatment period (20 salmeterol, pared with 34 1/min at baseline to 32 1/min 25 salbutamol). The primary reasons for during salbuterol treatment. on September 28, 2021 by guest. Protected copyright. withdrawal were poor compliance with proto- Figure 2 shows the weekly median percent- col (seven salmeterol, 10 salbutamol) or asth- age of nights with no symptoms of asthma.

Figure 2 Median % of °°0r- 100 symptom free nights during Salmeterol 12 weeks' treatment with salmeterol (50 g twice daily; left hand panel) and 80 salbutamol (400 ugfour 80 times daily; right hand panel). The median value was significantly greater for patients taking salmeterol 60 60 (p < 0 001). 0, .0

0 40H 40

20 20

1-. . O I i** i -----.r. . t-.14. 1-t...- .-.-r.-.-.- a4: 0 -2 0 2 4 6 8 10 12 Weeks of treatment Weeks of treatment Twelve month comparison ofsalmeterol and salbutamol as dry powderformulations in asthmatic patients 151

Figure 3 Median number 4 of rescue doses of/2 agonist Salmeterol taken duting 12 weeks' treatment with salmeterol CO) c 3 (50,ug twice daily; 0 Thorax: first published as 10.1136/thx.48.2.148 on 1 February 1993. Downloaded from left hand panel) and Co salbutamol (400 pgfour 4- times daily; right hand 0 2 panel). The number of o doses was significantly less .0 for patients taking n salmeterol (p = 0-013). E z

Weeks of treatment Weeks of treatment The number of nights with no symptoms was the overall incidence and nature of adverse higher during salmeterol treatment than with events in the two groups were similar. The salbutamol treatment (p < 0001). Median most commonly reported adverse event was values were 29% during the run in period for exacerbation of asthma defined as episodes both groups. During treatment with salmet- of acute asthma requiring additional or erol the median value increased to 100%, and increased treatment other than relief /32 agon- with salbutamol it increased to 71%. ists. Tables 3 and 4 show the number of Treatment with salmeterol was associated patients with exacerbations of asthma and the with significantly less use of rescue salbuta- frequency of exacerbations during each of the mol during the day (fig 3). The median per- three monthly intervals for salmeterol and centage of days when no additional salbutamol respectively. The frequency of salbutamol was used was zero for both groups exacerbations was highest during the first during the run in period. During treatment three months of the study but this almost cer- with salmeterol the median value increased to tainly reflects the greater frequency of study 75% and during salbutamol treatment it visits during this period. For both treatments increased to 67% (p = 0-013). The percent- the number of patients reporting exacerba- age of nights when no additional salbutamol tions and the frequency of exacerbations did was used was also significantly lower in the not increase during the last nine months of salmeterol group than in the salbutamol the study. This was also the case when the group (p < 0001). The percentage of nights study population was subdivided into those when no salbutamol was used during the run not taking concomitant steroids, those taking

in period was 43% for both groups. During inhaled steroids, and those taking oral http://thorax.bmj.com/ salmeterol treatment the weekly median value steroids. increased to 86% after the first week of treatment and for the next 11 weeks of treatment Table 3 Salmeterol: number ofpatients with asthma it was 100%. During treatment with salbuta- exacerbations during particular time intervals mol the median value increased to 86%. Time interval (months) FEV, increased with both salmeterol and salbutamol treatment over the 12 month 0-3 3-6 6-9 9-12

treatment period (mean change for salmeterol No of exacerbations: on September 28, 2021 by guest. Protected copyright. varied from 0-23 to 0 28 1 and for salbutamol 0 149 136 138 138 from 0 1 to 0-15 1). These changes were sig- 1 29 19 15 18 2 8 6 4 2 nificantly greater for salmeterol at some but 3 2 1 3 0 not all visits (see fig 4). 4 0 1 1 0 Both study drugs were well tolerated and 5 1 1 0 0 10 1 0 0 0 No of patients 190 164 161 158 No of patients with at 41 28 23 20 least one exacerbation (22%) (17%) (14%) (13%)

Table 4 Salbutamol: number ofpatients with asthma exacerbations during particular time intervals wU LL. Time interval (months) as 0-3 3-6 6-9 9-12 CD -c No of exacerbations: 0 0 144 124 133 126 1 36 26 22 20 2 9 9 2 3 3 6 1 1 2 4 2 0 0 0 5 1 1 0 0 6 0 1 0 0 0 weeks 2 weeks 4 weeks 8 weeks 12 weeks 6 months 9 months 12 months 10 0 0 1 0 No of patients 198 162 159 151 Figure 4 FEV, during 12 months' treatment with salmeterol (50 pg twice daily; closed No of patients with at 54 38 26 25 circle) and salbutamol (400 pgfour times daily) for 12 weeks and 400 pg twice dailyfor least one exacerbation (27%) (23%) (16%) (17%) the last nine months (open circle). *p < 0-05. 152 Lundback, Rawlinson, Palmer

Table 5 summarises the incidence of the in lung function as has been reported recently pharmacologically predictable and most com- in patients having regular fl2 agonist mon (>5%) adverse events for the first three treatment.'° 11 Although this is reassuring

months. There were no significant differences there was no placebo group having treatment Thorax: first published as 10.1136/thx.48.2.148 on 1 February 1993. Downloaded from between the two groups. Twenty seven for relief of symptoms only for comparison patients (12 receiving salmeterol, 15 receiving and therefore further studies are required to salbutamol) experienced serious adverse answer the question over the long term safety events requiring admission to hospital. of regular fl2 agonist treatment. This sus- Seventeen of these events involved exacerba- tained control of asthma was achieved despite tions of asthma, of which seven occurred in the fact that 40% of the patients were not patients receiving salmeterol and 10 in receiving concurrent inhaled glucocortico- patients receiving salbutamol. One patient in steroids. At the outset of this study it was the salbutamol group died as a result of a considered appropriate to relate salmeterol to myocardial infarction, which was considered an active comparator given at its recommen- to be unlikely to be related to the study drug. ded regular dosage. Direct comparisons in There were also no significant changes in adults in challenge studies between salme- haematological and biochemical indices in terol and salbutamol, both administered as either group and no significant changes were dry powders, have not been reported. noted in heart rate, systolic and diastolic Salmeterol (50 pg) given as a dry powder or blood pressure, or 12 lead electrocardio- metered dose aerosol, however, gave equal grams. protection against histamine challenge'2 and salbutamol (200 pg) by metered dose aerosol gave acute protection equivalent to that of Discussion salmeterol (50 pg by metered dose aerosol) in The data presented from this large multi- most but not all'2 challenge studies.'3-16 centre clinical trial show that compared Salbutamol given as 200 ,ug by metered dose with salbutamol (400 pg four times daily) aerosol has been shown to be equivalent to salmeterol (50 pg twice daily) produced signi- salbutamol given as 400 pg of dry powder.'7 ficantly greater improvements in lung func- We therefore believe that this dose of salbuta- tion, including a reduction in the diurnal mol would give an acute effect roughly equal variation of peak flow to a normal level. Also to that of salmeterol (50 pg) although the both the requirement for bronchodilator effect would not be maintained for the same treatment to relieve symptoms and the symp- length of time. The change in the dose of toms of asthma during the day and night were salbutamol from four times to twice daily reduced. These improvements occurred dur- after three months was to aid continued com- ing the first week of treatment and were sus- pliance. During the final nine months lung

tained throughout the 12 weeks of detailed function was measured at the clinic at regular http://thorax.bmj.com/ study with no evidence of deterioration in intervals and this was usually in the morning control of asthma. Both treatments were well so the maintained effect of salbutamol is to be tolerated with a similar low incidence of expected. adverse events in the two groups. Thus salmeterol given either as a dry pow- The study was continued for a further nine der or by metered dose aerosol'8 is an effec- months with the patients receiving salbutamol tive treatment for asthma over at least 12 at a lower frequency (twice daily). This facili- months compared with salbutamol.

tated the continued use of a double dummy on September 28, 2021 by guest. Protected copyright. technique to maintain blinding. During this We express thanks for assistance with the study to Dr N extension the control of asthma as assessed by Vetter and Dr H Zwick (Austria), Dr R Cordier and Dr Deman (Belgium), Dr P Faurschou, Dr 0 Davidsen, Dr UG exacerbation rate and lung function at the Svendsen, and Dr B Weeke (Denmark), Dr J -Kotaniemi and clinic was maintained. The incidence of Dr M I Havu (Finland), Dr Balgaries, Dr Lebas, and Dr de Muizon (France), Dr D-H Mahlo, Dr M Kaeppel, Dr J exacerbations of asthma in both groups was Ebeling, Dr T Koumba, Dr N Scheffler, Dr H P Brackmann, higher during the first three months (almost and Dr H L Hahn (Germany), Professor E A Pastorello and Professor Bruno (Italy), Dr P Mendes and Dr F Ribeiro certainly as a result of the greater frequency (Portugal), Dr P 0 Rydstrom, Dr D Huberman, Dr K of clinic visits), and then declined during the Osterman, Dr J Boe, Dr B-A Hermansson, Dr G Wikman, Dr Y Hornblad, Dr J Ankerst, Dr L Grettve, Dr 0 Robertsson, next nine months. Thus there was no evi- Dr J Kiviloog, Dr B Lundback, Dr Granstrom, Dr Persson, dence of waning control of asthma or decline and Dr Nemcek (Sweden), Dr J M Tschopp, Dr 0 Brandli, and Dr Rochat (Switzerland), Dr R Clark, Dr R Dent, Dr M G Britton, Dr J Siddorn and Dr J A McM Turner (United Kingdom). Table 5 Most common ( > 5%) and pharmacologically predictable adverse events 1 Clark TJH. Choices of drug treatment in asthma. Salmeterol Salbutamol Pharmacol Ther 1982;17:221-8. (50 g twice daily) (400 pgfour times a day) 2 McFadden ER Jr. Aerosolized bronchodilators and steroids in the treatment of airways obstruction in No of patients 190 198 adults. Am Rev Respir Dis 1980;122:89-96. 3 Svedmyr N, Lofdahl CG. Physiology and Pharmaco- Asthma and related events 29 (15-3) 42 (21-2) dynamics of beta-adrenergic agonists. In: Jenne JW, Acute nasopharyngitis 12 (6 3) 20 (10-1) Murphy S, eds. Drug therapy for asthma. New York: Respiratory and chest infections 6 (3 2) 12 (6-1) Dekker, 1987:177-212. Tremor 2 (1 1) 5 (2 5) 4 Tattersfield AE. Bronchodilator drugs. Pharmacol Ther Tachycardia (noted by physician) 0 3 (1-5) 1982;17:299-313. Palpitations 1 (0 5) 3 (1-5) 5 Larsson S, Svedmyr N. Bronchodilating effect and side Headache 11 (5 8) 6 (3 0) effects of f,-adrenoceptor stimulants by different modes Muscle cramp 2 (1-1) 3 (1-5) of administration (tablets, metered aerosol and combi- nations thereof). A study with salbutamol in asthmatics. % in parentheses. Am Rev Respir Dis 1977;116:861-9. Twelve month comparison ofsalmeterol and salbutamol as dry powderformulations in asthmatic patients 153

6 Svedmyr N. Anti-asthma xanthines and adenosine. In: 13 Taylor IK, O'Shaughnessy KM, Choudry NB, Adachi M, Andersson KE, Persson CGA, eds. Current clinical Palmer JBD, Fuller RW. A comparative study in atopic practice. Series No 19. Amsterdam; Excerpta Medica, subjects with asthma of the effects of salmeterol and 1985:135-45. salbutamol on allergen-induced bronchoconstriction, 7 Svedmyr N. Fenoterol: A beta2-adrenergic agonist for use increase in airway reactivity, and increase in urinary in asthma. Pharmacol Ther 1985;5:109-26. leukotriene E4 excretion. Allergy Clin Immunol 1992; Thorax: first published as 10.1136/thx.48.2.148 on 1 February 1993. Downloaded from 8 Ball DI, Brittain RT, Coleman RA, Denyer LH, Jack D, 82:575-83. Johnson M, et al. Salmeterol, a novel, long-acting l2- 14 Anderson SD, Rodwell LT, Du Toit J, Young IH. adrenoceptor agonist: characterization of pharmacologi- Duration of protection by inhaled salmeterol in exer- cal activity in vitro and in vivo. Br Pharmacol 1991; cise-induced asthma. Chest 1991 ;100: 1254-60. 104:665-71. 15 Malo JL, Ghezzo H, Trudeau C, L'Archeveque J, Cartier 9 Dahl R, Earmshaw JS, Palmer JBD. Salmeterol: a four A. Salmeterol, a new inhaled beta2-adrenergic agonist, week study of a long-acting beta-adrenoceptor agonist has a longer blocking effect than albuterol on hyperven- for the treatment of reversible airways disease. Eur tilation-induced bronchoconstriction. Allergy Clin RespirJ 1991;4:1178-84. Immunol 1992;89:567-74. 10 Sears MR, Taylor RD, Print CG, Lake DC, Li Q, 16 Pauwels R, Derom E. Salmeterol's early clinical develop- Flannery EM, et al. Regular Inhaled beta-agonist treatment and challenge studies. European Respiratory Review ment in bronchial asthma. Lancet 1990;336: 1391-6. 1991;1:261-4. 11 van Schayck CP, Dompeling E, van Herwaarden CL. 17 Mathieu M, Goldman M, Lellouche N, Sartene R. Bronchodilator treatment in moderate asthma or Kinetics of action of salbutamol inhaled from a metered chronic bronchitis: continuous or on demand? BMJ dose inhaler (MDI) and a 'diskhaler'. Eur Clin 1991;303:1426-31. Pharmacol 199242:435-8. 12 Campos Gongora H, Antoni F, Wisniewski Z, Tattersfield 18 Britton MG, Earmshaw JS, Palmer JBD. A twelve-month A. A single-dose comparison of inhaled albuterol and comparison of salmeterol with salbutamol in asthmatic two formulations of salmeterol on airwty reactivity in patients. Eur RespirJt 1992;5:1062-7. asthmatic subjects. Am Rev Respir Dis 1991;144:626-9. http://thorax.bmj.com/ on September 28, 2021 by guest. Protected copyright.