Introduction
The Global Initiative for Asthma (GINA) guideline emphasises the combination of inhaled
corticosteroids (ICSs) and long-acting β2-agonists (LABAs) as the first-choice treatment in patients with asthma who are not
controlled with ICS alone (GINA treatment Step 3 and higher) [1]. These treatments can be administered via two separate inhalers or as a fixed-dose
combination via a single inhaler; the latter approach may have benefits with regard
to patient adherence, if both components are clinically indicated [2]. This is particularly important in the era of uncertainty and continuous discussion
about rare, potentially fatal side effects of LABA monotherapy, given that fixed-combination
dosing with a single inhaler ensures concomitant delivery of ICS and LABA [3].
The fixed-dose combination of budesonide/formoterol (Symbicort® Turbuhaler®, AstraZeneca) can be used as both maintenance and reliever therapy in adults [4]. This treatment strategy is known as SMART™ (Symbicort® maintenance and reliever therapy), and is recommended in doses of 160 /4.5 μg per
inhalation, administered as either one or two inhalations twice daily and additional
as-needed reliever doses. A series of randomised, controlled trials (RCTs) has demonstrated
the clinical superiority of the SMART™ regimen over the conventional free combination
of ICS and LABA with additional short-acting β2-agonist (SABA) as required, via separate inhaler [5]
[6]
[7]. However, the external validity of data gathered from carefully controlled RCTs
in the real world is questionable [8]. Indeed, the effectiveness of the SMART™ treatment regimen in comparison with traditional
treatment (ICS, LABA and reliever from three discrete inhalers) under real-life conditions
is not clear.
This non-interventional study, performed in a real-life clinic setting, therefore
aimed to evaluate the use of rescue medication and health-related quality of life
(HRQL) in adult patients with asthma who received budesonide/formoterol as maintenance
therapy plus additional inhalations as needed to control symptoms (SMART™, single-inhaler
therapy). These patients were compared with patients who received a free combination
of an ICS and an inhaled LABA plus as-needed use of SABA, all via separate inhalers.
The prescribed doses used in this study were at the discretion of the treating physician
according to the recommendations of the package inserts of each medication.
Methods
Study design and patients
Patients were recruited from the German centres involved in the EuroSMART study [9]. By collecting data from a real-life population, very few restrictions by inclusion
and exclusion criteria were applied. However, to ascertain that only asthma patients
with an indication for GINA Step 3 treatment were included, (historical) proof of
reversible airway flow limitation and asthma exacerbation history were required. Therefore,
this multicentre, non-interventional, real-life study (AstraZeneca study code: NIS-RDE-SYM-2009/1;
ClinicalTrials.gov identifier: NCT00884689) included male and female outpatients aged
≥ 18 years who had a diagnosis of asthma with reversible obstruction confirmed within
the previous 5 years. Patients were receiving either budesonide/formoterol SMART™
therapy (single inhaler) or a free combination of ICS plus LABA with SABA rescue medication
(via three separate inhalers) for ≥ 3 months, with no change in treatment within 1
month of study entry. Patients were also required to have had a severe asthma exacerbation
within the past 24 months, but not within 1 month of study entry. Severe asthma exacerbations
were defined as occurrence of at least one of the following: use of systemic glucocorticosteroids
due to asthma for ≥ 3 days; unscheduled healthcare visit due to asthma; and hospitalisation
and/or emergency room visit due to asthma requiring treatment with systemic glucocorticosteroids.
Patients who had used oral, rectal or parenteral glucocorticosteroids within 30 days,
used β-blockers or who had a respiratory infection within 30 days were excluded from
participation in the study.
Treatments
Asthma medication and doses used were not assigned randomly and were administered
at the discretion of the prescribing physician, according to standard medical practice,
in a naturalistic study design. Treatments were obtained as commercially available
medication; the sponsor did not provide any medication for this study. Investigators
were provided with the package insert for budesonide/formoterol including dosage and
SMART™ treatment principles. No other restrictions were applied. No concomitant therapies
were disallowed (with the exception of systemic corticosteroids and β-blockers), but
investigators were asked to take into account relevant information from the budesonide/formoterol
summary of product characteristics [4]. Likewise, the only direction given to investigators regarding the comparator group
was that these patients had to be treated with ICS plus LABA and as-needed SABA via
separate inhalers and should be treated according to the relevant information in the
product package inserts. The physician had to document relevant data during a routine
visit at inclusion, approximately three (visit 2) and six months (visit 3) after start
of treatment.
Outcomes
The primary endpoint of this trial was the use of rescue medication, defined as the
mean daily number of rescue puffs of budesonide/formoterol SMART™ (regular morning
and evening maintenance treatment with budesonide/formoterol was not included) vs
use of SABA in the free-combination arm. This was recorded using daily patient diaries
during the 6-month follow-up period.
Secondary endpoints included: HRQL (differences from baseline and between group differences,
not accounting for baseline differences); the number of, and time to, severe asthma
exacerbations (defined as deterioration of asthma leading to at least one of the following:
oral/systemic glucocorticosteroid treatment for ≥ 3 days; unscheduled healthcare visit;
or hospitalisation or emergency room visit requiring treatment with oral/systemic
glucocorticosteroids); change in forced expiratory volume in 1 s (FEV1) from Visit 1 to Visit 3; the number of, and time to, hospitalisation and/or emergency
room treatments; level of patient satisfaction with treatment; and tolerability.
HRQL was assessed using the standardised version of the validated, asthma-specific
Asthma Quality of Life Questionnaire [AQLQ(S)] [10]. This 32-item questionnaire is scored on a 7-point scale where 1 = severe impairment
and 7 = no impairment; therefore, higher scores on the AQLQ(S) indicate more favourable
HRQL. The average daily use of ICS over the entire study period was calculated in
terms of beclomethasone propionate powder equivalents. The calculation also included
the corticosteroids inhaled during rescue puffs in the SMART™ group; patients in the
free-combination group did not use ICS for reliever inhalation. Patients were asked
to rate their overall satisfaction with their current treatment by choosing one of
the following options: very satisfied, fairly satisfied, satisfied, dissatisfied or
very dissatisfied.
Tolerability was investigated by reporting of adverse events (AEs), serious AEs and
treatment-related AEs. No clinical laboratory parameters were assessed.
Patients were followed up at 3 and 6 months. The study was managed by ClinResearch
GmbH, Cologne, Germany, an independent clinical research organisation, and was funded
by AstraZeneca.
Statistical methods
The mean daily number of as-needed puffs of rescue medication (primary endpoint) was
analysed using an analysis of covariance (ANCOVA) model with treatment as factor.
Potential confounders in the analysis were identified as body mass index, gender,
smoking status and season of the year at the individual study start. Therefore, a
logistic regression model with treatment assignment as the dependent variable and
the confounders found in the first step as independent factors was applied. Only ‘season
at study start’ had a p-value below the predefined limit of 0.1. Therefore, the mean
number of puffs of rescue medication per day was calculated using ANCOVA with treatment
and propensity score as the factor, where the propensity score was calculated as the
probability of having SMART™ assigned as treatment given the season at study start.
The difference between the two treatment groups in the change from baseline in AQLQ(S)
scores was evaluated using a t-test, which did not take baseline differences into
account. To assess the impact of baseline differences, the mean change from baseline
in the overall AQLQ(S) score was analysed using ANCOVA with treatment as factor and
the baseline value as covariate.
It was planned in the study protocol that the time to first severe asthma exacerbation
and the time to first hospitalisation/emergency room treatment would be evaluated.
However, due to the small number of patients who experienced these endpoints, the
planned analyses could not be performed and no robust statistical comparisons could
be made.
Sample size considerations
Approximately 500 patients with asthma were planned to be included in this non-interventional
study across approximately 60 sites in Germany. From a maximum of 20 patients, each
site was expected to enrol two-thirds of patients who were treated according to the
budesonide/formoterol SMART™ principle and one-third of patients who had been prescribed
a free-combination treatment.
While all analyses were carried out for exploratory purposes, some statistical considerations
concerning the sample size were provided by the protocol. A sample size of 333 patients
in the SMART™ group and 167 patients in the free-combination group was predicted to
have 80 % power to detect a difference in the mean number of rescue medication puffs
per day of 0.506, using a two-group t-test with a two-sided significance level of
p < 0.05. A mean of 2.2 puffs per day and a common standard deviation (SD) of 1.9
were assumed based on data from a subgroup analysis of GINA treatment Step 3 and 4
patients from a previous study (DESOLO) [11]. It was estimated that even if the true enrolment was 350 /151 patients, there would
be 80 % power to detect a difference in daily use of rescue medication between the
two treatment groups of ≥ 0.52 actuations.
Analysis sets
The safety analysis set included all patients enrolled into the study; the full analysis
set (FAS) was defined as all patients who had at least one efficacy assessment; and
the sensitivity analysis set (SAS) consisted of all patients with none of the following
protocol deviations: at least one inclusion criteria not fulfilled; date of written
informed consent after date of Visit 1; patient prescribed a treatment other than
Symbicort SMART™ or a free combination of an ICS plus LABA plus as-needed SABA; no
severe asthma exacerbations documented in the last 24 months before Visit 1; at least
one exclusion criteria fulfilled; and the administered therapy was not the same as
the prescribed therapy.
Ethical considerations
This non-interventional study was performed in accordance with the ethical principles
outlined in the Declaration of Helsinki and is consistent with ICH/Good Clinical Practice,
applicable regulatory requirements in Germany and AstraZeneca policy. All patients
provided written, informed consent prior to study inclusion. The Ethics Committee
of the Hessen Medical Association (Landesärztekammer Hessen) also agreed on the protocol.
Results
Patients
The flow of patients through this study is shown in [Fig. 1]. The first patient was enrolled into the study on 21 April 2009 and the last patient’s
final visit was 02 June 2010. A total of 498 patients were enrolled at 49 sites in
Germany; of these, 482 patients were included in the FAS. Patients’ demographic and
clinical characteristics were similar between the SMART™ and free-combination treatment
groups ([Table 1]). However, patients in the SMART™ group were more likely to be very satisfied with
their current treatment (39.0 % vs 23.3 %) than free-combination treatment recipients.
Fig. 1 Flow of patients who received SMART™ or a free combination of ICS plus LABA and as-needed
SABA through the study. The safety analysis set consisted of all patients enrolled
in the study; the FAS was defined as all enrolled patients who had at least one efficacy
assessment. The SAS included patients with no protocol deviations. Abbreviations: FAS, full analysis set; ICS, inhaled corticosteroid; LABA, long-acting β2-agonist; SABA, short-acting β2-agonist; SAS, sensitivity analysis set; SMART™, Symbicort® [budesonide/formoterol] maintenance and reliever therapy.
Table 1
Baseline demographic and clinical characteristics of patients who received SMART™
or a free combination of ICS plus LABA and as-needed SABA (full analysis set).
|
Characteristics
|
SMART™
(n = 310)
|
Free combination
(n = 172)
|
|
Age, years
|
49.1 (15.2)
|
51.4 (15.4)
|
|
Male, n (%)
|
119 (38.4)
|
59 (34.3)
|
|
Smoking status, n (%):
|
|
|
|
Current smokers
|
31 (10.0)
|
18 (10.5)
|
|
Ex-smokers
|
63 (20.3)
|
37 (21.5)
|
|
Never-smokers
|
216 (69.7)
|
117 (68.0)
|
|
Pulmonary function
|
|
|
|
FEV1, l
|
2.64 (0.87)
|
2.45 (0.73)
|
|
PEF, l/s
|
6.16 (2.16)
|
5.65 (2.08)
|
|
VC, l
|
3.45 (1.02)
|
3.33 (0.91)
|
|
Severe asthma exacerbations within 24 months
|
1.4 (0.7)
|
1.7 (1.5)
|
|
Asthma treatment Step, n (%):[1]
|
|
|
|
GINA Step 3
|
238 (76.8)
|
138 (80.2)
|
|
GINA Step 4
|
72 (23.2)
|
34 (19.8)
|
|
Comorbid allergy, n (%)[2]
|
27 (8.7)
|
13 (7.6)
|
|
AQLQ(S) global score
|
5.37 (1.02)
|
4.89 (1.19)
|
|
Therapy satisfaction, n (%):
|
|
|
|
Very satisfied
|
121 (39.0)
|
40 (23.3)
|
|
Fairly satisfied
|
94 (30.3)
|
55 (32.0)
|
|
Satisfied
|
89 (28.7)
|
58 (33.7)
|
|
Unsatisfied
|
6 (1.9)
|
19 (11.0)
|
|
Prior medication, n (%)[3]
|
35 (11.3)
|
7 (4.1)
|
|
Concomitant medication, n (%)[3]
|
|
|
|
Systemic corticosteroids
|
13 (4.2)
|
5 (2.9)
|
|
Systemic antibacterials
|
5 (1.6)
|
0 (0.0)
|
|
Cough and cold preparations
|
5 (1.6)
|
0 (0.0)
|
|
Systemic antihistamines
|
0 (0.0)
|
1 (0.6)
|
Values presented are mean (SD), unless otherwise stated.
Abbreviations: AQLQ(S), standardised version of the Asthma Quality of Life Questionnaire; FEV1, forced expiratory volume in 1 s; GINA, Global Initiative for Asthma; ICS, inhaled
corticosteroid; LABA, long-acting β2-agonist; PEF, peak expiratory flow; SABA, short-acting β2-agonist; SD, standard deviation; SMART™, Symbicort® [budesonide/formoterol] maintenance and reliever therapy; VC, vital capacity.
1 Although all included patients had to be on ICS + LABA, investigators were asked
to assess asthma severity retrospectively (i. e. before treatment was introduced).
2 Allergy included: seasonal allergy, multiple allergies, hypersensitivity and house
dust allergy.
3 Prior medication was defined as medication that was discontinued on the calendar
day prior to the first study visit, or earlier; only drugs for asthma treatment were
recorded.
273 (88.1 %) of the patients in the SMART™ group had been receiving Symbicort® for more than 6 months before study entry; 152 (88.4 %) of the patients in the free-combination
group had been taking the ICS product they received during the study for 6 months
before entering the study.
Overall, patients had high mean ± SD AQLQ(S) scores (5.37 ± 1.02 and 4.89 ± 1.19 points
in the SMART™ and free-combination group, respectively) and high levels of treatment
satisfaction: 98.1 % of the SMART™ and 89.0 % of the free-combination group were very
satisfied, fairly satisfied or satisfied with treatment.
Use of rescue medication (primary endpoint)
In the FAS, the least squares mean daily number of rescue medication puffs (primary
endpoint) was 0.557 in SMART™-treated patients and 0.822 in free-combination recipients.
This represents a difference of – 0.266 (95 % confidence interval [CI] – 0.474, – 0.057)
puffs per day in favour of SMART™ (p = 0.013). The difference between the treatment
groups in the SAS was – 0.209 (95 % CI – 0.414, – 0.004), again favouring SMART™ (p = 0.046).
The unadjusted mean daily number of rescue medication puffs was 0.548 for SMART™ and
0.839 for the free-combination cohort.
AQLQ(S)
The mean ± SD AQLQ(S) global score was higher in the SMART™ group than in the free-combination
treatment group at study end (5.60 ± 0.99 vs 5.29 ± 1.15 points). However, the mean
AQLQ(S) global score change from baseline to study end (in-group difference) was greater
in the free-combination group than in SMART™ recipients ([Table 2]).
Table 2
Change from baseline to 6 months (study end) in AQLQ(S) scores in patients who received
SMART™ or a free combination of ICS plus LABA and as-needed SABA (full analysis set;
last observation carried forward)[1].
|
SMART™
(n = 310)
|
Free combination
(n = 172)
|
|
AQLQ(S) score, mean (SD)
|
Baseline
|
Change from baseline
|
Baseline
|
Change from baseline
|
|
Global score
|
5.37 (1.02)
|
0.25 (0.82)
|
4.89 (1.19)
|
0.42 (0.89)
|
|
Symptoms subscore
|
5.34 (1.12)
|
0.23 (0.97)
|
4.84 (1.26)
|
0.41 (1.01)
|
|
Activity limitation subscore
|
5.38 (1.04)
|
0.24 (0.82)
|
4.91 (1.21)
|
0.41 (0.94)
|
|
Emotional function subscore
|
5.56 (1.22)
|
0.29 (1.00)
|
5.00 (1.41)
|
0.38 (1.00)
|
|
Environmental stimuli subscore
|
5.20 (1.31)
|
0.24 (1.01)
|
4.84 (1.43)
|
0.46 (1.17)
|
Abbreviations: AQLQ(S), standardised version of the Asthma Quality of Life Questionnaire; ICS, inhaled
corticosteroid; LABA, long-acting β2-agonist; SABA, short-acting β2-agonist; SD, standard deviation; SMART™, Symbicort® [budesonide/formoterol] maintenance and reliever therapy.
1 The differences between the two treatment groups in terms of the changes from baseline
were evaluated by a t-test. While some of these differences attained statistical significance,
the test does not take into account differences in baseline scores; the results of
this testing are, therefore, not shown.
Improvements in the AQLQ(S) domain scores for symptoms, activity limitation, emotional
function and environmental stimuli were seen within both treatment groups (in-group
differences in terms of change from baseline; [Table 2]). While some statistically significant differences between SMART™-treated patients
and the free-combination group were seen with regard to the mean AQLQ(S) global and
domain scores (data not shown), the test used to generate the p-values did not take
into account differences between the treatment groups in terms of the AQLQ(S) scores
at baseline. As the SMART™ group had higher AQLQ(S) scores at baseline, the change
in the AQLQ(S) global score during the study was tested by ANCOVA with treatment as
factor and baseline as covariate, to investigate the impact of these baseline differences.
This test resulted in a p-value of 0.9545, indicating no significant difference between
the treatment groups in terms of the improvement in the global score of the AQLQ(S)
when taking into consideration baseline AQLQ(S) scores.
Health care utilization and asthma exacerbations
Overall, the incidence of severe asthma exacerbations was similar in the SMART™ and
free-combination treatment groups, both up to 3 months and between 3 – 6 months’ follow-up
([Table 3]). This corresponds to an estimated mean annualised severe exacerbation rate of 0.20
(95 % CI 0.14, 0.29) among SMART™ recipients and 0.17 (95 % CI 0.10, 0.29) among those
who received the free combination. Throughout the study duration, two SMART™ recipients
(0.6 %) required emergency room treatment and one patient in the free-combination
group (0.6 %) was hospitalised due to asthma.
Table 3
Severe asthma exacerbations in patients who received SMART™ or a free combination
of ICS plus LABA and as-needed SABA (full analysis set).
|
SMART™
(n = 310)
|
Free combination
(n = 172)
|
|
Exacerbations at 0 – 3 months:
|
|
|
|
Patients, n (%)
|
10 (3.2)
|
5 (2.9)
|
|
Episodes per patient, range
|
1 – 3
|
1 – 2
|
|
Exacerbations at 3 – 6 months:
|
|
|
|
Patients, n (%)
|
13 (4.2)
|
8 (4.6)
|
|
Episodes per patient
|
1
|
1
|
|
Extrapolated mean annual exacerbation rate, n episodes (95 % CI)
|
0.20 (0.14, 0.29)[1]
|
0.17 (0.10, 0.29)[1]
|
The annual exacerbation rate was estimated by translating the reported numbers of
episodes into estimated mean numbers per year using a Poisson regression model with
treatment as factor and total time in study (in years) as offset variable.
Abbreviations: CI, confidence interval; ICS, inhaled corticosteroid; LABA, long-acting β2-agonist; SABA, short-acting β2-agonist; SMART™, Symbicort® [budesonide/formoterol] maintenance and reliever therapy.
1 Between-treatment p = 0.66.
The time to the first severe asthma exacerbation and the time to first hospitalisation/emergency
room treatment were not evaluated as no robust statistical comparisons could be made
due to the small number of patients who experienced these endpoints.
Lung function
The baseline FEV1 values were similar between both groups (SMART™, 2.64 ± 0.87 l; free combination
2.45 ± 0.73 l [mean ± SD]). FEV1 improved in both groups during the study. The mean ± SD improvement was slightly
higher with SMART™ than in the free-combination group at both 3 (0.06 ± 0.33 vs 0.03 ± 0.33 l)
and 6 (0.13 ± 0.48 vs 0.07 ± 0.43 l) months (in-group differences were assessed, as
per protocol, due to lack of randomisation).
Daily inhaled steroid use
The prescribed mean ± SD daily dose of ICS (including rescue ICS) throughout the study
duration was approximately 10 % lower in the SMART™ group than in the free-combination
group (615 ± 318 μg vs 678 ± 380 μg, respectively).
Use of systemic corticosteroids
During the study, 13 (4.2 %) SMART™ and five (2.9 %) free-combination recipients used
systemic corticosteroids.
Treatment satisfaction
High levels of treatment satisfaction were reported for both groups at baseline, although
the proportion of patients who reported being very satisfied with their treatment
was higher for SMART™ than for free combination treatment (39.0 % vs 23.3 %, respectively;
[Fig. 2]). In both treatment arms, the number of patients who were either very or fairly
satisfied with their asthma treatment increased slightly over time. At all timepoints,
the proportion of patients who were very satisfied with treatment was higher (unadjusted
for baseline) in the SMART™ group than among patients who received free-combination
therapy. The proportion of patients who were very satisfied was 50.0 % in the SMART™
group and 34.3 % in the free-combination group at Visit 2, and 54.2 % and 39.0 %,
respectively, at Visit 3 ([Fig. 2]). By Visit 3, 258 (83.2 %) patients in the SMART™ group and 120 (69.8 %) patients
in the free-combination group confirmed a treatment satisfaction level of very satisfied
or fairly satisfied.
Fig. 2 Patients’ level of satisfaction with SMART™ or a free combination of ICS plus LABA
and as-needed SABA at baseline, 3 months and 6 months (full analysis set). Abbreviations: ICS, inhaled corticosteroid; LABA, long-acting β2-agonist; SABA, short-acting β2-agonist; SMART, Symbicort® [budesonide/formoterol] maintenance and reliever therapy.
Healthcare resource utilisation and sick leave
During the observation period, the proportion of patients who consulted a physician
for any reason was similar with SMART™ and free-combination therapy (105 [33.9 %]
vs 62 [36.0 %] of patients, respectively). Additional diagnostic tests (e. g., pulmonary
function tests) were performed in 18 (5.8 %) patients in the SMART™ group and 16 (9.3 %)
patients in the free-combination group.
Confirmed sick leave was reported for 27 (8.7 %) patients in the SMART™ group and
11 (6.4 %) patients in the free-combination group.
Safety
Overall, seven (2.2 %) patients in the SMART™ group and seven (3.9 %) patients in
the free-combination group reported one or more AEs. Three AEs were judged to be related
to asthma treatment (one incidence of hoarseness and one of oral candidiasis in the
free-combination group, and one case of hoarseness in the SMART™ group). None of the
serious AEs reported during the trial were considered to be related to either SMART™
or free-combination therapy.
Discussion
There is strong evidence of the benefits of SMART™ in the treatment of patients with
asthma from the largest ever performed international asthma RCT program [5]
[6]
[7]
[9]
[12]
[13]
[14]
[15]
[16], and these findings have been supported by real-life studies [9]
[17]
[18]. However, all three of these real-life studies have assessed SMART™ in randomised
studies where the treatment arm was predefined. Data confirming the effectiveness
of this treatment regimen in a real-life situation where all the patients are treated
at the complete discretion of the physician are still sparse. This naturalistic non-interventional
study, performed in a real-life clinical setting, therefore aimed to evaluate the
use of rescue medication and HRQL in adult patients with a confirmed asthma diagnosis.
In this study, routine assessments were performed under real-life conditions, with
no intervention by the sponsor with regard to patient selection, diagnostic procedures
or therapeutic decisions. Thus, real-life patient management with no dose restrictions
could be studied for both study arms (SMART™ and free combination) [19]. Real-life observations from non-interventional studies can be strikingly different
from data gathered in RCTs. For example, recently published data show that leukotriene
antagonists can be as effective as ICS in the routine asthma clinical setting, irrespective
of a myriad of clinical trial data showing superiority of ICS [20]. Archibald Cochrane stated that “[b]etween measurements based on RCTs and benefit
… in the community there is a gulf which has been much under-estimated” [21].
Evaluation of asthma treatment success in patients in a real-life clinical situation
is much more challenging than assessment of outcomes within the context of a RCT.
The usual endpoints such as assessment of lung function at standardised control visits
or physician assessment of exacerbations are not feasible in a non-interventional
study, and daily measurements of peak expiratory flow would be too demanding for real-life
conditions. Moreover, self-assessment of exacerbations is less reliable than physician
assessment in RCTs. Thus, exacerbations could not be used as a primary outcome. Therefore,
the number of rescue puffs per day was chosen as the most reliable, responsive and
ascertainable single primary variable for assessment of asthma control. The choice
of this parameter as the primary outcome is supported by a recent publication, which
showed that airway obstruction lability may allow for a more complete assessment of
disease activity [22]. Furthermore, this endpoint also allows evaluation of dropouts without introducing
a ‘healthy survivor’ bias. Evaluations of HRQL and other conventional asthma-related
efficacy parameters were chosen as secondary variables in this study.
In our setting, a sample size with 2 : 1 allocation (SMART™ and free combination)
was calculated assuming a between-treatment difference of 0.506 puffs of rescue medication
per day, based on data attained from a previous RCT (DESOLO), which had a very similar
design [11]. In the 498 patients enrolled in the present study, the least squares mean for the
difference between treatment groups was only one-half of this estimate (– 0.266; p = 0.013).
The smaller between-treatment difference may have been caused by the smaller mean
daily number of rescue medication puffs than expected based on data from DESOLO, and
further highlights the difference between pragmatic non-interventional studies and
RCTs. Nevertheless, a significant benefit of SMART™ can be shown for the primary outcome.
Shortly after completion of the EuroSMART RCT [9], which had similar inclusion criteria to those for the present non-interventional
study, the same investigators were recruited. Consequently, the EuroSMART cohort and
the patient population included in the present study are comparable. Prior to inclusion,
> 88 % of patients in both treatment arms had been receiving the same asthma treatment
for > 6 months. This resulted in a population of patients who had good HRQL and high
levels of satisfaction with treatment in both groups. We therefore may have observed
a ceiling effect for further clinically meaningful changes, consecutively seeing a
small effect, even in cases where the recorded changes achieved statistical significance.
It is also important to note that the AQLQ(S) endpoint comparison is biased due to
baseline differences since the study was not randomised). While the mean AQLQ(S) global
score was slightly higher in the SMART™ group than in the free-combination group at
study end, this value at baseline was considerably higher in the SMART™ vs the free-combination
treatment group (with the resulting difference being greater in the free-combination
group). When baseline AQLQ(S) score was included in the statistical analysis there
was no difference between the treatment groups.
Similarly, the mean annualised number of exacerbations in the present study was low
(0.20 episodes for SMART™ vs 0.17 episodes for free-combination treatment). Again,
this incidence is less than one-half of the expected, given that patients who had
at least one severe exacerbation in the previous 24 months were recruited. However,
this has previously been observed in large RCTs and could be partially related to
the precision of retrospectively collected data. Regardless of the cause of this discrepancy,
asthma exacerbations requiring systemic corticosteroids were rare in this population
of intensively pre-treated patients who were satisfied with their treatment and had
good HRQL. Therefore, robust statistical analysis of the difference between treatment
groups could not be performed. The reasons outlined above, together with the difference
in exacerbation reporting between RCTs and this real-life study, would also explain
the lack of difference between treatments in the reduction in number of exacerbations
observed here compared with the greater reduction in number of exacerbations observed
with SMART™ vs comparator arms in RCTs [23].
As expected, clinic assessments of FEV1, with no standardised washout periods as is the case in real life, are not suitable
for evaluation of between-group differences. There was, however, a marked mean improvement
from baseline, especially in the SMART™ group, despite the fact that this population
had better FEV1 values at baseline. This could also be related to the better adherence observed with
treatment with just one inhaler.
If the number of rescue medication puffs per day, HRQL and the number of exacerbations
are a surrogate marker for asthma control, it is notable that the same, or better,
control can be achieved with a lower corticosteroid burden, despite the fact that
ICS was also used (in addition to LABA) in the SMART™ group as a reliever medication,
as well as part of maintenance treatment.
The strengths of this non-interventional study include close quality assurance and
on-site monitoring, experienced investigators who participated in the previous EuroSMART
RCT [9] and verified asthma diagnosis, resulting in a relatively homogenous population with
GINA Step 3 treatment and presumed severity grade of 3 or 4. This is particularly
important given prior findings that one-third of patients treated with asthma drugs
in the community do not have asthma [24]. The real-life setting of this study may be considered a bias; however, we consider
this to be a strength, as it means that we were able to clarify the effectiveness
of SMART™ outside the stringent RCT environment.
Limitations inherent to the nature of the non-interventional study design should be
considered. Real-life patient adherence is not comparable with surveillance in RCTs
for a number of reasons including drug holidays, running out of drugs and drug pref-erences,
among many. Indeed, register studies of 1-year treatment duration have shown that
patients take far less than one-half of their prescribed doses [25]
[26]
[27]. The most important limitation of this non-interventional study is the fact that
patients who previously took part in a RCT were included. As already discussed, these
patients could have been better controlled and more adequately pre-treated in comparison
with treatment-naïve or non-study patients. A confounding factor to this is that the
prescribed doses used in the free-combination arm of the study were at the discretion
of the treating physician according to the recommendations of the package inserts
of each medication; this could result in possible bias and create a wide variety of
ICS/LABA treatment regimens (and possibly outcomes) for the free-combination arm.
Despite this, benefits in the daily number of rescue medication puffs used (primary
endpoint) were observed.
Most secondary variables showed either significant or numerical benefit in favour
of SMART™ or the two treatment regimens were, for other secondary variables, at least
equally effective. The differences between the treatment groups that were observed
in this non-interventional study are not as marked as those previously seen in the
SMART RCTs, particularly with regard to HRQL [6]
[18]. A number of factors may underlie these blunted effect sizes. Firstly, RCTs select
the ‘best candidates’ for treatment; i. e., pre-treated, but still symptomatic, patients
who have room for improvement [5]
[6]
[7]
[9]
[12]
[13]
[14]
[15]
[16]
[18]. Non-interventional studies offer treatments for patients who are less likely to
show improvements, such as the pre-treated population enrolled in the present study.
Secondly, real-life patient assessment is inferior to RCT patient assessment, as RCTs
can enforce a standardised treatment duration and washout periods, as well as being
able to assess trough and peak effects. In this study, we aimed to assess real-life
management of patients with asthma and the treating physicians were given limited
guidance regarding therapeutic decisions. Thirdly, in real-life naturalistic trials,
patients with significant co-morbidities and concomitant treatments with potential
interactions are included. While the physicians were asked to consider the prescribing
information for SMART™ and the medications used in the free combination plus as-needed
SABA therapy arm, we cannot guarantee that patients did not receive contraindicated
concomitant medications.
Treatment adherence with the simple SMART™ strategy could play an important role compared
with conventional three inhaler treatment; however, the non-interventional study design
is not suitable for assessment of adherence.
In this real-life setting, patients treated with SMART™ needed less rescue medication
despite a lower inhaled corticosteroid burden when compared with conventional asthma
management with three separate inhaler devices.
