METHODS

Section:

ChooseTop of pageAbstractMETHODS <<RESULTSDISCUSSIONReferencesCITING ARTICLES

Potential participants were identified from our volunteer database and general practice lists in and around Nottingham by searching for people aged 16 or older with a current diagnosis of asthma. Letters of invitation and information sheets were sent to all potentially suitable participants. To be enrolled, participants had to give written informed consent, be taking a maintenance dose of inhaled corticosteroid between 200 and 1,000 μg beclometasone dipropionate (or equivalent) per day, have temporarily increased their usual dose of inhaled corticosteroid (common practice in the UK at that time), or have taken a course of oral corticosteroids in the preceding 12 months in response to worsening of asthma symptoms but not in the preceding 4 weeks. Potential participants were excluded if they were taking a maintenance dose of oral corticosteroid (there were no other restrictions based on asthma treatment), had other clinically significant medical conditions, were pregnant or lactating, or had smoked for more than 20 pack-years. Participants had to complete a 2-week run-in period recording morning PEF using a mini-Wright PEF meter as the best of three measurements. To be randomized the lowest morning peak flow in the run-in period had to be greater than 90% of the mean peak flow over the run-in period. FEV₁ and FVC were measured (Micro Plus Spirometer; Micromedical Ltd., Kent, UK), taking the highest of two readings within 100 ml.

Randomized participants received an individualized asthma management plan and either an active or placebo corticosteroid inhaler. The management plan instructed participants to record their morning PEF if they believed their asthma control was deteriorating or they developed symptoms of an upper respiratory tract infection and to start using the study inhaler if PEF fell by 15% or more on 2 consecutive days, or 30% on 1 day, from the mean morning PEF measured during the run-in period. The study inhaler was to be taken for 7 days in addition to participants' normal asthma treatment and a daily diary of morning PEF was to be completed. The study inhaler and PEF diary were continued for a further 7 days if morning PEF had not returned to the prestudy baseline value after the initial 7 days. Participants were also instructed to commence oral prednisolone 30 mg daily if their asthma deteriorated to a point where they would normally start systemic corticosteroids, if their general practitioner advised them to do so or, as a safety precaution, if their PEF fell by 40% or more from baseline. The original protocol specified that participants would remain in the study for a 6-month period. This was extended to 12 months after 81 participants had been enrolled because a large proportion of participants were completing the study without starting the study inhaler or needing oral corticosteroids. Participants continued in the study for the 12 months unless they withdrew consent or were withdrawn from the study for prespecified safety reasons, namely the need for more than two courses of oral corticosteroids or the need to temporarily increase their dose of inhaled corticosteroid more than five times within a 6-month period.

The study inhaler matched each participant's usual inhaler in terms of drug and device and, when started, provided a quadrupling or no change in dose of inhaled corticosteroid in the active and placebo groups, respectively. Participants were asked to contact the research team after using the study inhaler to arrange collection of the used inhaler and completed diary card and to issue replacements.

The primary outcome was exacerbations of asthma treated with oral corticosteroids (intention-to-treat analysis). Prespecified secondary outcomes were exacerbations of asthma requiring oral corticosteroids among the people who started the study inhaler (per-protocol analysis) and adverse effects.

An allocation sequence of random permuted blocks of 10 was generated using a random number table by an independent pharmacist and implemented by one of the study investigators once participants were enrolled into the trial. Active and placebo inhalers were obtained from AstraZeneca (Luton, UK), Chiesi (Cheadle, UK) and GlaxoSmithKline (Slough, UK) and were identical apart from the presence or absence of inhaled corticosteroid, to achieve allocation concealment and blinding of investigators and participants.

Based on previous studies we estimated that 35% of participants in the placebo group would have an exacerbation of asthma requiring oral corticosteroids while taking part in the study (8, 9). The minimum size of effect we considered important to detect was a reduction in the need for oral corticosteroids of one-third. The desired level of significance and power was 0.05 and 0.8, respectively. We therefore set a target sample size of 500, which allowed for a drop-out rate of about 10%.

Before unblinding the results we compared the proportion of participants who had required oral corticosteroids for an asthma exacerbation in the active and placebo groups among all randomized participants (intention-to-treat analysis) and in the subset of the active and placebo groups who had started their study inhaler (per-protocol analysis) using χ² test.

Nottingham Research Ethics Committee 1 and the relevant Research and Development departments in Nottinghamshire and Derbyshire approved the study.

RESULTS

Section:

ChooseTop of pageAbstractMETHODSRESULTS <<DISCUSSIONReferencesCITING ARTICLES

Of the 1,015 subjects who replied to our letter of invitation 428 were enrolled in the study, of whom 403 were randomly allocated to receive an asthma management plan and either active (n = 197) or placebo (n = 206) inhaler (see flow diagram in Figure 1). All participants received their allocated intervention, although three were lost to follow-up with no outcome data (all in the placebo group) leaving 197 and 203 in the groups receiving active and placebo inhalers, respectively, for the intention-to-treat analysis. Thirty-eight and 39 participants in the active and placebo groups withdrew from the study but contributed data for the intention-to-treat analysis up to the point at which they left the study.

Figure 1. Trial flow diagram.

Download Figure | Download Powerpoint

[More] [Minimize]

Participants were enrolled in the study between May 2004 and September 2007 and follow-up was completed in September 2008. Trial recruitment took longer than planned and was discontinued after 403 participants were randomized because it was increasingly difficult to find eligible participants in and around Nottingham and to obtain a continuing supply of some of the placebo inhalers. There were no interim analyses.

The two intervention groups were well matched in terms of baseline demographic and clinical characteristics but for a slight predominance of ever smokers in the placebo group (Table 1). The mean (SD) daily dose of inhaled corticosteroid was 517 (241) μg and 521 (219) μg and mean FEV₁ was 84 (19) and 83 (18) percent of predicted in the active and placebo groups, respectively.

Of the 403 participants, 94 (23%) started the study inhaler, 56 (28%) in the active and 38 (19%) in the placebo group. Of these, 22 (23%) participants started the study inhaler on more than one occasion, 11 and 7 twice, 2 and 1 three times, and 1 and 0 four times in the active and placebo groups, respectively. All but four participants fulfilled the peak flow criteria before starting the study inhaler.

The number of participants having an exacerbation of asthma requiring treatment with oral corticosteroids was 18 (9%) and 29 (14%) in the active and placebo groups, respectively, giving a risk ratio of 0.64 (95% confidence interval [CI], 0.37–1.11, P = 0.11) (Table 2). Eight participants experienced more than one exacerbation (maximum three), three in the active and five in the placebo group, so that the total number of exacerbations requiring treatment with oral corticosteroids was 22 in the active and 34 in the placebo group. The reasons for starting oral corticosteroids on these 56 occasions were a 40% fall in PEF (5 active, 6 placebo), doctor's advice (14 active, 23 placebo), deterioration in asthma control to a point at which the participant would normally take oral corticosteroids (3 active, 3 placebo), or a combination of these reasons (0 active, 2 placebo).

Of the participants who used the study inhaler, 12 of 56 (21%) in the active group and 19 of 38 (50%) in the placebo group required treatment with oral corticosteroids, giving a risk ratio of 0.43 (95% CI, 0.24–0.78, P = 0.004) (Table 2). Among this subgroup 6 participants experienced more than one exacerbation (maximum three), 2 in the active and 4 in the placebo group, so that the total number of exacerbations requiring treatment with oral corticosteroids was 15 in the active and 23 in the placebo group. Most (86) participants recorded their study inhaler use and PEF values in the diary provided. Mean (SD) peak flow had fallen by 22.3% (12.5) and 22.9% (12.0) on the morning the study inhaler was started in the active and placebo groups, respectively, and returned toward baseline over the next 14 days with no major difference in rate or pattern of recovery between the two groups despite the greater use of oral corticosteroids in the placebo group (Figure 2). Forty-four participants used the study inhaler for up to 7 days and 42 for 8 to 14 days with similar distributions of duration of use in the active and placebo groups.

Figure 2. Mean fall in PEF (95% confidence interval) expressed as a percentage fall from the mean baseline PEF value (dashed line, active; solid line, placebo). Includes participants who started oral corticosteroids.

Download Figure | Download Powerpoint

[More] [Minimize]

Adverse events were reported by 12 of the participants during or within 4 weeks of starting the study inhaler, 9 in the active (headache, glossitis, upper respiratory tract infection, laryngitis, gastroenteritis, indigestion, depression, and two cases of nausea) and 3 in the placebo group (hay fever, sinusitis, vomiting).

DISCUSSION

Section:

ChooseTop of pageAbstractMETHODSRESULTSDISCUSSION <<ReferencesCITING ARTICLES

This is the first trial to explore whether, in patients receiving an asthma self-management plan, quadrupling the dose of inhaled corticosteroid when asthma control starts to deteriorate is an effective strategy for preventing asthma exacerbations. Overall we found that participants randomized to the active group had fewer exacerbations requiring treatment with oral corticosteroids, although the difference was not significant (9 vs. 14%, relative risk reduction, 36%). Among the 94 participants who started the study inhaler, quadrupling the dose of inhaled corticosteroid led to a marked reduction in the need for oral corticosteroids (21 vs. 50%, relative risk reduction, 57%).

We used a pragmatic clinical trial design to explore the effectiveness of quadrupling the dose of inhaled corticosteroids to prevent asthma exacerbations in a real-life setting. This involved accepting factors such as variable compliance that impact on patients in real life and meant that additional testing and monitoring that might be included in an efficacy study but would not happen in normal clinical practice were not appropriate. Our selection criteria were kept as broad as possible and patients remained on their usual inhaled corticosteroid treatment to ensure generalizability. We excluded people taking more than 1,000 μg inhaled beclometasone dipropionate equivalent a day because, without any evidence for efficacy, we wanted to avoid the use of very high doses of inhaled corticosteroids. We used symptom-driven criteria for starting PEF measurements rather than continuous PEF recordings throughout the study, again to reflect real life and because symptom-driven self-management plans have been shown to be as effective as PEF-driven plans (10).

Although this is the largest study to date to explore whether an increase in dose of inhaled corticosteroid when asthma control is deteriorating can prevent asthma exacerbations we failed to reach our original recruitment target of 500. Recruitment was more difficult than expected mainly because we were unable to obtain a continuing supply of placebo inhalers for some of the commonly used inhaled corticosteroids. We also found that fewer participants than expected started the study inhaler despite extending the time in the study from 6 to 12 months, possibly because participants had milder asthma than in our previous trial and therefore more stable control. These two factors probably explain why we failed to find a significant effect in the intention-to-treat population despite a large effect in the subjects who started the study inhalers. One weakness of our study is that more subjects started the study inhaler in the active than the placebo group raising the possibility that the two groups behaved differently after randomization. Participants in the active group did not appear to have a lower threshold for starting the study inhaler because the mean fall in peak flow before starting the study inhaler was similar for the two groups (22 vs. 23%). We also do not believe that participants were less likely to start the placebo inhaler because they had identified the inhaler as a placebo, because the active and placebo inhalers were identical, similar numbers in both groups continued to use the study inhaler for up to 14 days, and the drop-out rate from the study was very low.

Asthma management guidelines no longer recommend doubling the dose of inhaled corticosteroid when asthma control deteriorates after two large randomized placebo-controlled clinical trials showed no benefit from temporarily doubling the dose of a patient's usual maintenance inhaled corticosteroid (5) or doubling the dose of inhaled budesonide (6). Some guidelines suggest a larger increase may be effective because a number of small studies have suggested that the addition of inhaled budesonide (200 μg (8) or 600 μg (11) four times a day) or fluticasone propionate (1 mg twice daily) (12) appears to be effective for treating exacerbations. Trials exploring the effects of two maintenance doses of inhaled corticosteroids have also found that a fourfold (9) but not twofold (13) increase in maintenance dose reduces exacerbations. Failure to find an effect from doubling the dose of inhaled corticosteroids may be because the relationship between inhaled corticosteroid dose and a reduction in exacerbations is nonlinear but may simply reflect the size of studies looking at doubling the dose.

In conclusion, our trial suggests that providing patients with asthma with a self-management plan incorporating a fourfold increase in dose of inhaled corticosteroid when asthma control starts to deteriorate may be effective at preventing exacerbations. A larger, widely inclusive, multicenter clinical trial is now needed to confirm these findings.