Abstract
The identification, prevention, and prompt treatment of exacerbations are major objectives of asthma management. We looked at change in PEF, symptoms, and use of rescue β -agonists during the 425 severe exacerbations that occurred during a 12-mo parallel group study (FACET) in which low and high doses of budesonide with and without formoterol were compared in patients with asthma. Oral corticosteroids were prescribed for severe exacerbations, the main study end point, defined as the need for a course of oral corticosteroids (n = 311) or a reduction in morning PEF of > 30% on two consecutive days. PEF, symptoms, and bronchodilator use over the 14 d before and after the exacerbation were obtained from diary cards. Exacerbations were characterized by a gradual fall in PEF over several days, followed by more rapid changes over 2 to 3 d; an increase in symptoms and rescue β -agonist use occurred in parallel, and both the severity and time course of the changes were similar in all treatment groups. Exacerbations identified by the need for oral corticosteroids were associated with more symptoms and smaller changes in PEF than those identified on the basis of PEF criteria. Female sex was the main patient characteristic associated with an increased risk of having a severe exacerbation. Exacerbations may be characterized predominantly by change in symptoms or change in PEF, but the pattern was not affected by the dose of inhaled corticosteroid or by whether the patient was taking formoterol.
Asthma is a fluctuating disease and the unpredictability of exacerbations causes major problems for patients. Despite their clinical and economic importance exacerbations have been little studied and they have rarely been a primary end point in intervention studies. Furthermore, few studies have lasted long enough or have had sufficient power to determine whether an intervention affects exacerbation rates and in many studies exacerbations have not been clearly defined. A better understanding of the changes that precede a severe exacerbation might allow exacerbations to be predicted at an earlier stage so that preemptive treatment can be initiated.
A year-long study of more than 800 patients (the Formoterol and Corticosteroids Establishing Therapy [FACET] study) was designed specifically to determine the effect of four treatment regimens on severe and mild exacerbations of asthma (1). Following a 4-wk run-in period in which a high dose of inhaled budesonide (800 μg, twice daily) was given, patients were randomized to receive budesonide (100 or 400 μg, twice daily) with the long-acting inhaled β2-agonist formoterol fumarate (12 μg, twice daily) or placebo. Severe exacerbations were defined a priori and treatment standardized to a 10-d course of oral corticosteroids. Patients completed a daily diary throughout the year, recording twice daily peak expiratory flow (PEF) and symptoms. The main finding of the study was that both the addition of formoterol and the higher dose of inhaled corticosteroid reduced the rate of exacerbations.
We have now analyzed the changes in PEF and symptoms that preceded the 425 severe exacerbations that occurred during the study and the time course of the recovery. We also examined the change in symptoms and PEF 2 d before the height of the exacerbation to determine the predictive value of such changes for a severe exacerbation and we attempted to determine which patient characteristics might predict an increased likelihood of having an exacerbation.
The FACET study was designed to compare the effect of low- and high-dose budesonide with and without twice daily treatment with formoterol. Details of the study design and main findings have been published (1) and are summarized here.
The FACET study had a double-blind, randomized, parallel-group design and was carried out at 71 centers in nine countries after obtaining regulatory and ethics committee approval. Patients took 1,600 μg of budesonide per day (800 μg twice daily) by Turbuhaler during a 4-wk run-in period and if their asthma fulfilled criteria for stability and compliance with treatment during the last 10 d of the run-in period they were randomized to one of four treatment regimens twice daily for the next 12 mo.
1. Budesonide (Pulmicort) 100 μg + placebo: BUD200 group
2. Budesonide 100 μg + formoterol (Oxis) 12 μg: BUD200 + F group
3. Budesonide 400 μg + placebo: BUD800 group
4. Budesonide 400 μg + formoterol 12 μg: BUD800 + F group
Terbutaline sulfate (Bricanyl), 250 μg per inhalation, was used as rescue medication throughout the study. All inhaled medications (Astra, Södertälje, Sweden) were inhaled from a dry powder inhaler (Turbuhaler) and all doses refer to the metered dose from the Turbuhaler.
The patients who took part were 18 to 70 yr of age, had had asthma for at least 6 mo, had taken an inhaled corticosteroid for at least 3 mo, and gave verbal or written informed consent. Baseline FEV1 had to be 50% of predicted or more and had to increase by at least 15% from baseline after inhalation of 1 mg of terbutaline. Patients were excluded if they required more than 2,000 μg per day of beclomethasone diproprionate by metered dose inhaler or equivalent doses of inhaled budesonide or fluticasone and if they had had three or more courses of oral corticosteroids or been in the hospital because of asthma in the last 6 mo. Patients completed a daily diary card throughout the study, recording the best of three measurements of PEF with a Vitalograph peak flow meter (Vitalograph, Buckingham, UK) in the morning and evening before medication, symptoms of asthma during the day and night on a four-point scale (0 = no symptoms, 1 = mild [easily tolerated], 2 = moderate [interfering with daily life], 3 = severe [incapacitating]), awakening at night owing to asthma, use of terbutaline inhaler rescue therapy, and use of oral corticosteroids.
The primary end point was the number of severe and mild asthma exacerbations over the year of the study. A severe asthma exacerbation was defined as an exacerbation that required oral corticosteroids as judged by the clinical investigator or an episode in which morning PEF fell by more than 30% from mean morning PEF during the last 10 d of the run-in period (baseline) on two consecutive days. All severe exacerbations were to be treated with a 10-d course of oral steroids, either 30 mg of prednisolone or prednisone or 24 mg of methylprednisolone per day. Patients who had three severe exacerbations within 3 mo or five in total were withdrawn from the study.
Change in peak flow, symptoms, and rescue β-agonist use during a severe exacerbation are summarized over time in relation to the day the exacerbation was diagnosed (Day 0). Day 0 was defined as the day on which oral corticosteroids were started in patients deemed to require oral corticosteroids and, for exacerbations defined on the basis of PEF criteria, as the second day on which PEF was more than 30% below baseline. The changes before Day 0 are described as Day −1, −2, etc., and the days after it as Day +1, +2, etc. To compare the rate of change in PEF, symptoms, and bronchodilator use before an exacerbation the data were standardized by expressing the measurement on Day −14 as 0% and Day 0 as 100%. We have compared the pattern of change in PEF, symptoms, and use of rescue β-agonists according to whether an exacerbation was defined by a fall in PEF of more than 30% or the need for oral corticosteroids, and we also looked at the pattern of exacerbation in those who failed to take a course of oral corticosteroids despite a fall in PEF of more than 30%.
In an attempt to determine whether severe exacerbations can be predicted within subjects we determined how often the changes seen 2 d before an exacerbation occurred at other times during the study without developing into an exacerbation. To do this we calculated the median change in PEF, symptoms, and rescue β-agonist use 2 d before the peak of the exacerbation. We then looked at the number of 10-d periods in which the figures were breached over the year without developing into a severe exacerbation, choosing a 10-d period because exacerbations were treated for 10 d. We also attempted to determine whether patients at risk of having an exacerbation could be identified, using a stepwise logistic regression to determine which of the following factors, after including treatment, could predict a severe exacerbation: age, sex, previous dose of inhaled corticosteroid, smoking, and peak expiratory flow variability, β2-agonist use, and daytime and nocturnal symptoms during the run-in period. Baseline PEF variability was measured as the mean daily PEF variability over the last 10 d of the run-in period, with daily variability calculated as the difference between morning and evening PEF as a percentage of the mean of the two values.
Of the 1,114 patients who entered the run-in period, 852 patients (436 women and 416 men) were randomized to one of the four treatment groups and 694 (81%) completed the 12-mo study. There were 425 severe exacerbations altogether, ranging from 57 in the group receiving high-dose budesonide and formoterol to 153 in the group receiving low-dose budesonide and placebo (Table 1). The 425 severe exacerbations occurred in 242 patients, of whom 145, 48, and 31 had one, two, and three exacerbations, respectively, 15 had four or five, and 3 had more than five (six, six, and eight). Of 21 patients withdrawn owing to frequent severe exacerbations, 10 were in the BUD200 group, 7 were in the BUD200 + F group, 4 were in the BUD800 group, and none were in the BUD800 + F group.
| BUD200 | BUD200 + F | BUD800 | BUD800 + F | |||||
|---|---|---|---|---|---|---|---|---|
| No. of exacerbations | 153 | 125 | 90 | 57 | ||||
| No. of exacerbations diagnosed by more than a 30% fall in PEF on two consecutive days | 28 | 40 | 24 | 22 | ||||
| No. of patients withdrawn owing to exacerbations | 10 | 7 | 4 | 0 | ||||
| No. of patients having an exacerbation | 81 | 62 | 60 | 41 | ||||
| Mean PEF during run-in, L/min* | 373 | 388 | 361 | 350 | ||||
| Mean PEF on Day −14, L/min | 341 | 388 | 341 | 342 | ||||
| Maximum fall in PEF from Day −14, L/min (%) | 53 (16) | 78 (20) | 62 (18) | 63 (18) |
The mean morning and evening PEF in the four treatment groups for the 14 d before and 14 d after the peak of the exacerbation (Day 0) are shown in Figure 1. Fourteen days before the exacerbation the mean PEF in the four groups ranged from 341 to 388 L/min (Table 1), with BUD200 + F patients having higher values before the exacerbation. The mean maximum reduction in morning PEF from Day −14 ranged from 53 to 78 L/min (16 to 20%), with no significant differences between treatment groups on analysis of variance.
Fig. 1. Change in morning and evening PEF (percent fall from Day −14) and daytime and nocturnal symptoms over the 14 d before and 14 d after an exacerbation in relation to treatment. Budesonide, 200 μg daily (solid line); budesonide (200 μg) plus formoterol (12 μg), twice daily (dotted line); budesonide, 800 μg (dashed line); budesonide (800 μg) plus formoterol (12 μg) twice daily (dashed-dot line).
[More] [Minimize]The decline in mean morning and evening PEF occurred gradually between Day −10 and Day −3 and then more rapidly. The increase in symptom scores and rescue β-agonist use before the exacerbation was similar in pattern to the fall in PEF. The pattern of change in PEF, symptoms, and rescue medication (not shown) was similar in all four treatment groups (Figure 1). When the mean values of PEF, symptom scores, and rescue β-agonist use for all exacerbations before Day 1 were standardized the rate of change for the different end points in the 10 d before the exacerbation was similar (Figure 2).
Fig. 2. Comparison of change in morning and evening PEF, daytime and nighttime symptoms, and rescue β-agonist use during an exacerbation. Data have been standardized (Day −14 = 0%, maximum change = 100%) to allow comparison of changes with time between different end points.
[More] [Minimize]The proportion of exacerbations in which a 10, 20, and 30% fall in PEF from baseline occurred on at least 1 d on or in the 14 d before Day 0 is shown in Table 2. A fall of at least 20% was seen in 69% of all exacerbations and a fall in PEF of at least 30% in almost half (45%).
| Fall in Morning PEF (%) | Percentage of Exacerbations | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| BUD200 (n = 153) | BUD200 + F (n = 125) | BUD800 (n = 90) | BUD800 + F (n = 57) | Total (n = 425) | ||||||
| ⩽ 10 | 14 | 14 | 11 | 21 | 15 | |||||
| > 10 | 86 | 86 | 89 | 79 | 85 | |||||
| > 20 | 72 | 70 | 65 | 59 | 69 | |||||
| > 30 | 41 | 49 | 42 | 43 | 45 | |||||
Most of the severe exacerbations were defined by the need for a course of oral steroids (n = 311, 73%) rather than a fall in PEF of > 30% from the run-in period on two consecutive days (n = 114, 27%). Exacerbations defined by the need for a course of oral corticosteroids were characterized by a greater increase in symptoms and a smaller fall in PEF than were exacerbations identified from a fall in PEF of more than 30% (Figure 3).
Fig. 3. Changes in daytime asthma score and morning PEF in relation to whether the exacerbation was diagnosed by the need for oral corticosteroids (OCS; n = 311) or by a 30% fall in PEF (PEF; n = 114). The values for evening PEF and nocturnal symptoms (not shown) were similar.
[More] [Minimize]Of the 114 exacerbations diagnosed from a fall in PEF of more than 30% on two consecutive days, only 36 were treated by oral corticosteroids within 14 d despite instructions and 2 were treated after 14 d. Overall, therefore, 347 of the 425 exacerbations (82%) were treated with oral corticosteroids, 94 with prednisolone (30 mg/d), 230 with prednisone (30 mg/d), and 23 with methyl prednisolone (24 mg/d). The change in PEF and symptoms in the 38 patients diagnosed on the basis of PEF criteria and who were treated with oral corticosteroids, compared with the 76 who failed to take treatment, is shown in Figure 4. Patients with a fall in PEF of more than 30% and who did not take oral corticosteroids showed fewer symptoms and smaller and more short-lived falls in PEF than did the patients treated with oral corticosteroids. They also took fewer inhalations of rescue β-agonists than did the patients who took oral corticosteroids during both the day (mean maximum, 1.4 versus 2.5 inhalations) and night (0.5 versus 1.3 inhalations). The response after the 10-d course of oral corticosteroids was rapid over the first 2 d, followed by a more gradual improvement. Mean PEF was almost back to preexacerbation PEF levels by Day +14 irrespective of whether the exacerbation was defined by the need for oral corticosteroids (353 versus 356 L/min; p = 0.7) or not (340 versus 353 L/min; p = 0.08).
Fig. 4. Change in daytime symptoms and morning PEF in patients in whom the exacerbation was diagnosed by a fall in PEF of more than 30% according to whether they did (OCS; n = 38) or did not (no OCS; n = 76) take a course of oral corticosteroids.
[More] [Minimize]Two days before an exacerbation median values for change from baseline values during the run-in period for PEF, symptoms, and rescue β-agonist use were, respectively, as follows: fall in morning and evening PEF, 9.6 and 10.1%, increase in daytime and nocturnal symptom scores, 0.6 and 0.0; increase in nocturnal and daytime rescue β-agonist use (number of inhalations), 0.5 and 0.0. Table 3 shows the number of 10-d periods in which these median figures were exceeded, excluding the 14 d before and after an exacerbation. Patients who had an exacerbation during the study exceeded these cutoff values more often during the period outside their exacerbations than did the patients who never had an exacerbation. The percentage of 10-d intervals in which the Day −2 values were exceeded ranged from 30% for daytime symptoms to 6.4% for nocturnal symptoms in patients who had an exacerbation at some time during the study (the figures were 17 and 4.4% for patients with no exacerbation).
| Treatment Group | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| BUD200 | BUD200 + F | BUD800 | BUD800 + F | Total | ||||||
| PEF | ||||||||||
| Morning | 24 (32) | 16 (22) | 18 (29) | 11 (18) | 17 (26) | |||||
| Evening | 21 (29) | 14 (19) | 16 (24) | 9.8 (16) | 15 (23) | |||||
| Symptom | ||||||||||
| Night | 4.9 (8.8) | 4.0 (6.0) | 5.1 (6.2) | 3.7 (3.5) | 4.4 (6.4) | |||||
| Day | 25 (32) | 22 (30) | 22 (34) | 17 (21) | 22 (30) | |||||
| Rescue Treatment | ||||||||||
| Night | 21 (37) | 14 (26) | 21 (27) | 12 (13) | 17 (27) | |||||
| Day | 15 (28) | 11 (21) | 15 (21) | 11 (11) | 13 (21) | |||||
Of the patient characteristics examined, female sex was the main factor associated with an increased risk of having a severe exacerbation (odds ratio, 1.89, 95% confidence interval, 1.37–2.6). PEF variability, inhaled corticosteroid dose, and age were also positively associated with an increased risk of having a severe exacerbation, with odds ratios that were slightly but significantly more than 1. For a 1% increase in PEF variability, a 100-μg increase in prestudy inhaled corticosteroid dose, and a 1-yr increase in age the odds ratios (95% CI) for having an exacerbation were 1.023 (1.011, 1.035), 1.056 (1.016, 1.096), and 1.011 (1.001, 1.023), respectively. No other variables, including interaction terms for treatment and sex, were significant. The risks were similar irrespective of treatment groups.
The FACET study provides an unusual opportunity to look at the changes in PEF and symptoms that precede an exacerbation of asthma and to see how the two measures respond to standard treatment with a course of oral corticosteroids. Severe exacerbations were the main end point of the original study and were defined by a fall in PEF of more than 30% from baseline values during the run-in period on two consecutive days or as the need for a course of oral corticosteroids as judged by the patient or doctor. The criteria used to define an exacerbation, although somewhat arbitrary, are broadly similar to those used in previous large studies (2-5) and had, for safety reasons, to include the perceived need for oral corticosteroids by the patient or doctor.
The present study extends the original publication from the FACET study by looking in detail at the events that occur before and during the 425 severe exacerbations that occurred during the study. The exacerbations were characterized by a gradual decline in PEF over 5 to 7 d, followed by a more rapid fall over 2 to 3 d. The increase in symptoms and use of rescue β-agonists were similar in pattern to the fall in PEF. When the data were standardized the changes in PEF, symptoms, and rescue β-agonist use before the exacerbation showed a parallel change, suggesting that none of the measures provides an earlier warning of an exacerbation than the other measures. This contrasts with the findings in a small study by Chan-Yeung and coworkers, who found an increase in symptoms before a fall in PEF in exacerbations in 27 children and 14 adults (6).
Patients receiving high-dose budesonide plus formoterol compared with low-dose budesonide plus formoterol had lower peak flow values before the exacerbation (Day −14) and during the run-in period (Table 1). This is in contrast to the findings during treatment in the main FACET study, where mean PEF values were higher in both groups receiving formoterol. The difference is probably because the number of patients having an exacerbation in this group was small (19%) and the patients would not necessarily be representative of the group as a whole. The pattern of change in PEF and symptoms leading up to an exacerbation was similar in all four treatment groups. It has been suggested that treatment with a long-acting β-agonist might mask the symptoms of an impending exacerbation so that when they occur they would consequently be more severe (7). Although McIvor and coworkers (8) found some support for this in a small steroid reduction study of 13 subjects in whom they measured induced sputum, our study questions the clinical significance of their findings. The addition of formoterol in a relevant clinical setting reduced the exacerbation rate rather than the opposite and the exacerbations that occurred in patients taking formoterol did not differ in severity or in response to treatment compared with exacerbations in patients not taking formoterol.
The need for oral corticosteroids was the defining reason for almost three-quarters of the exacerbations in the FACET study and these exacerbations were associated with more symptoms and a smaller fall in PEF than those defined by a fall in PEF of more than 30% from the run-in period on two consecutive days. All exacerbations should have been treated with a standardized 10-d course of oral corticosteroids but this did not happen with many of the exacerbations diagnosed by a fall in PEF; 18% of all exacerbations were, therefore, not treated with an oral corticosteroid. The patients who failed to institute corticosteroid treatment had fewer symptoms and a more rapid and short-lived reduction in their PEF despite the lack of treatment. After treatment with oral corticosteroids there was a rapid improvement in PEF and symptoms and 14 d after starting treatment PEF values and symptom scores had almost recovered to those seen 14 d before the exacerbation. Most previous large prospective studies have shown a beneficial role for oral corticosteroids in treating acute exacerbations of asthma (9-11); the extent to which recovery was due to the corticosteroids in our patients cannot be determined, however, in a retrospective study. Many patients with exacerbations will recover without such treatment, as seen from the patients who failed to take oral corticosteroids in this study.
One of the goals in asthma therapy is to prevent exacerbations. Knowing the rate of change in PEF and symptoms as an exacerbation develops might help to determine whether exacerbations can be identified at an earlier stage. This would enable treatment to be started earlier, thereby reducing the severity of the exacerbation. The extent to which such a measure would be helpful would depend on the efficacy of the intervention and, if PEF or symptoms were used to predict an exacerbation, the number of false positives it produced. We decided to look at the change in PEF, symptoms, and rescue β-agonist use that had occurred 2 d before the peak of the exacerbation because a period of 2 d might give time for an intervention to be effective. These Day −2 values were exceeded fairly often during the study without an exacerbation developing, particularly in the subjects who had an exacerbation at some time during the study. The percentage of 10-d periods in which the values were exceeded outside the 28 d surrounding an exacerbation ranged from 30% for daytime symptoms to 6.4% for nighttime symptoms in the patients who had an exacerbation during the study. This suggests that the development of nocturnal asthma may have greater specificity for the prediction of exacerbations, although it lacks sensitivity because most patients who had an exacerbation had not developed nocturnal symptoms by Day −2.
We have not explored the predictive value of the Day −2 changes further because the baseline used to define an exacerbation in the FACET study was determined after a course of high-dose budesonide, which would not normally occur in clinical practice. Furthermore, half the patients were subsequently given a long-acting β-agonist that increased lung function while the other half received placebo and tended to show reduced lung function as the dose of inhaled steroid was reduced. Our data should be of use in designing studies in the future to explore how exacerbations can be predicted and prevented.
We also looked at various patient characteristics to see whether patients at risk of having a severe exacerbation could be identified. Increasing age, corticosteroid dose before the study onset, and peak flow variability during the run-in period showed a significant positive association with the risk of having an exacerbation, but the effects were small. Being female was the only large and significant risk factor. The reason why women are more prone to exacerbations is unclear, although differences in the pattern of asthma between men and women have been identified. Women are more likely to have asthma and are more likely to be admitted to the hospital with asthma (12, 13) even after allowing for differences in asthma prevalence and smoking (14). Women with asthma also report more symptoms and a poorer quality of life than do asthmatic men with similar lung function (15).
The long-term aim of asthma management is to be able to predict and prevent exacerbations without excessive doses of medication. This requires a better understanding of how severe exacerbations develop. By providing a detailed analysis of the changes before and after an exacerbation in relation to treatment with and without a long-acting β-agonist this study provides reassurance about the effect of formoterol on the severity and duration of exacerbations, and it provides data on which to base prospective studies designed to reduce or prevent exacerbations. More work needs to be done to enable the clinician and patient to be able to predict when small changes in PEF and symptoms or some other variable are likely to develop into a full exacerbation.
The authors thank the following physicians, listed by country, who contributed randomized patient data: Belgium: Prof. W. De Backer, Prof. M. Decramer, Dr. P.-M. Mengeot, Dr. L. Siemons, Dr. J. Verhaert, Prof. W. Vinken. Canada: Dr. M. Alexander, Dr. J. Bouchard, Dr. A. Day, Dr. A. Knight, Dr. J.-L. Malo, Dr. D. Marciniuk, Dr. J. G. Martin, Dr. S. Peters, Dr. B. Sanders, Dr. B. Sproule, Dr. D. Stubbing. Holland: Dr. A. Baas, Dr. T. A. Bantje, Dr. J. Creemers, Dr. H. Sinninghe Damsté, Dr. W. Evers, Dr. S. Gans, Dr. A. Greefhorst, Prof. H. Hassing, Dr. F. Maesen, Dr. M. J. Möllers, Dr. H. R. Pasma, Dr. Z. Pelikan, Dr. P. E. Postmus, Dr. J. Prins, Dr. B. M. Santana, Dr. M. Schrijver, Dr. A. P. Sips, Dr. R. Stallaert, Dr. L. van der Maas, Dr. A. J. van Harreveld. Israel: Dr. J. Greif, Dr. D. Heimer, Dr. N. Richiert, Prof. A.-H. Rubin, Dr. A. Wollner. Italy: Prof. F. Bariffi, Dr. F. Bonifazi, Prof. V. Brusasco, Prof. G. D'Amato, Prof. L. Fabbri, Dr. C. Franco, Dr. L. Gandola, Prof. C. Giuntini, Prof. E. Gramiccioni, Prof. V. Grassi, Prof. L. Marazzini, Dr. A. Rossi, Prof. A.M. Santolicandro, Dr. C. Sturani. Luxembourg: Dr. J.-P. Parini. Norway: Prof. L. Bjermer, Dr. N. Ringdal. Spain: Dr. J. L. Alvarez Sala, Dr. P. L. Cabrera Navarro, Dr. S. Romero, Dr. J. Sanchis, Dr. V. Sobradillo, Dr. H. Verea. UK: Dr. G. Basran, Dr. L. M. Campbell, Dr. D. Franklin, Prof. G. J. Gibson, Dr. R. C. Joshi, Dr. A. Knox, Dr. A. B. MacLean, Dr. R. Scott, Dr. R. Smith, Prof. A. Tattersfield, Dr. J. P. Vernon.
| 1. | Pauwels R. A., Löfdahl C.-G., Postma D. S., Tattersfield A. E., O'Byrne P., Barnes P. J., Ullman A.on behalf of the FACET International Study GroupAdditive effects of inhaled formoterol and budesonide in reducing asthma exacerbations: a one-year, controlled study. N. Engl. J. Med.337199714051411 |
| 2. | Sears M. R., Taylor D. R., Print C. G., Lake D. C., Li Q., Flannery E. M., Yates D. M., Lucas M. K., Herbison G. P.Regular inhaled β-agonist treatment in bronchial asthma. Lancet336199013911396 |
| 3. | Wilding P., Clark M., Thompson J., Coon, Lewis S., Rushton L., Bennett J., Oborne J., Cooper S., Tattersfield A. E.Effect of long term treatment with salmeterol on asthma control: a double blind, randomised crossover study. Br. Med. J.314199714411446 |
| 4. | Greening A. P., Ind P. W., Northfield M., Shaw G.Added salmeterol versus higher-dose corticosteroid in asthma patients with symptoms on existing inhaled corticosteroid. Lancet3441994219224 |
| 5. | Woolcock A., Lundback B., Ringdal N., Jacques L. A.Comparison of addition of salmeterol to inhaled steroids with doubling of the dose of inhaled steroids. Am. J. Respir. Crit. Care Med.153199614811488 |
| 6. | Chan-Yeung M., Chang J. H., Manfreda J., Ferguson A., Becker A.Changes in peak flow, symptom score, and the use of medications during acute exacerbations of asthma. Am. J. Respir. Crit. Care Med.1541996889893 |
| 7. | Arvidsson M., Larsson S., Löfdahl C.-G., Melander B., Svedmyr N., Wahlander L.Inhaled formoterol during one year in asthma: a comparison with salbutamol. Eur Respir. J.4199111681173 |
| 8. | McIvor R. A., Pizzichini E., Turner M. O., Hussack P., Hargreave F. E., Sears M. R.Potential masking effects of salmeterol on airway inflammation in asthma. Am. J. Respir. Crit. Care. Med.1581998924930 |
| 9. | Barnes N. C.Effects of corticosteroids in acute severe asthma. Thorax471992582583 |
| 10. | Fanta C. H., Rossing T. H., McFadden E. R.Glucocorticoids in acute asthma. Am. J. Med.741983845851 |
| 11. | Chapman K. R., Verbeek P. R., White J. G., Rebuch A. S.Effect of a short course of prednisone in the prevention of early relapse after the emergency room treatment of acute asthma. N. Engl. J. Med.3241991788794 |
| 12. | Skobeloff E. M., Spivey W. H., Clair S. S., Schoffstall J. M.The influence of age and sex on asthma admissions. J.A.M.A.268199234373440 |
| 13. | Hyndman S. J., Williams D. R., Marrill S. L., Lipscombe J. M., Palmer C. R.Rates of admission to hospital for asthma. Br. Med. J.308199415961600 |
| 14. | Prescott E., Lange P., Vestbo J.the Copenhagen City Heart Study GroupEffect of gender on hospital admissions for asthma and prevalence of self-reported asthma: a prospective study based on a sample of the general population. Thorax521997287289 |
| 15. | Osborne M. L., Vollmer W. M., Linton K. L. P., Buist A. S.Characteristics of patients with asthma within a large HMO: a comparison by age and gender. Am. J. Respir. Crit. Care Med.1571998123128 |
The following Astra employees were involved in the study: Dr. C.-A. Bauer (project leader), M. Best (data entry), Dr. C. Hultquist (medical advisor), F. Jackson (safety evaluation), Dr. A. Lennon (medical coordinator), Dr. S. Lindgren (safety evaluation), H. MacFarlane (computing), Dr. A. McLean (deputy project leader), M. Nevinson (medical coordinator), M.-Å. Persson (computing), K. Svensson (statistician).
National medical monitors: F. Bellemans, T. Ben-Or, S. Bordonaro, M. Chiesa, I. Garcia, J. Haddon, S. Holthe, Dr. M. Huybrechts, Dr. A. Ning, H. Rijskamp, F. Stafford, Dr. M. van den Dobbelsteen.
