糖心vlog

Key Points

Question听 In patients with poorly controlled asthma, is switching to single inhaler combination inhaled corticosteroid (ICS)鈥揻ormoterol maintenance and reliever therapy (SMART) associated with longer time to first severe asthma exacerbation compared with step up or continuation of Global Initiative for Asthma (GINA) treatment step with ICS鈥搇ong-acting 尾₂-agonist (ICS-LABA) plus short-acting 尾₂-agonist (SABA)?

Findings听 In this systematic review and meta-analysis of 5 randomized clinical trials that included 4863 patients with poorly controlled asthma, compared with maintenance ICS-LABA plus SABA reliever, use of SMART with budesonide-formoterol was associated with a longer time to first severe asthma exacerbation.

Meaning听 The findings suggest that SMART is associated with a longer time to first severe asthma exacerbation compared with step up or continuation of GINA treatment step with ICS-LABA plus SABA in patients with uncontrolled asthma.

Importance听 The Global Initiative for Asthma (GINA) recommends 2 alternative treatments for patients receiving treatment at steps 3 to 5: single inhaler combination inhaled corticosteroid鈥揻ormoterol as both maintenance and reliever (SMART) or inhaled corticosteroid鈥搇ong-acting 尾₂-agonist as maintenance plus short-acting 尾₂-agonist as reliever.

Objective听 To assess whether switching to SMART is associated with longer time to first severe asthma exacerbation compared with a step up or continuation of GINA treatment step with maintenance inhaled corticosteroid鈥搇ong-acting 尾₂-agonist plus short-acting 尾₂-agonist reliever among patients with poorly controlled asthma.

Data Sources听 For this systematic review and meta-analysis, the literature, internal study databases at AstraZeneca and the Medical Research Institute of New Zealand, and references from a previous systematic review and meta-analysis on SMART were searched to identify randomized clinical trials published from January 1990 to February 2018, that compared budesonide-formoterol by SMART with maintenance inhaled corticosteroid鈥搇ong-acting 尾₂-agonist plus short-acting 尾₂-agonist reliever.

Study Selection听 Trials of at least 24 weeks鈥� duration were included if they reported baseline data on GINA treatment step, asthma control status, and efficacy measures of severe exacerbations. Included patients were adults and adolescents with asthma and baseline Asthma Control Questionnaire 5-item version scores of 1.5 or higher.

Data Extraction and Synthesis听 Patient-level data were identified by independent extraction, and analyses were performed using a fixed-effect model. Data analysis was performed from August 2018 to November 2021.

Main Outcomes and Measures听 The primary outcome was time to first severe asthma exacerbation associated with each treatment, analyzed by Cox proportional hazards regression.

Results听 Overall, 4863 patients were included (3034 [62.4%] female; mean [SD] age, 39.8 [16.3] years). Switching patients with uncontrolled asthma at GINA step 3 (n鈥�=鈥�1950) to SMART at either step 3 or 4 was associated with a prolonged time to first severe asthma exacerbation, with a 29% reduced risk compared with stepping up to step 4 inhaled corticosteroid鈥搇ong-acting 尾₂-agonist maintenance plus short-acting 尾₂-agonist reliever (hazard ratio, 0.71; 95% CI, 0.52-0.97). For patients with uncontrolled asthma at step 3 and step 4 (n鈥�=鈥�2913), switching to SMART was associated with a prolonged time to first severe asthma exacerbation and a 30% reduced risk compared with remaining at the same treatment step (hazard ratio, 0.70; 95% CI, 0.58-0.85).

Conclusions and Relevance听 In this systematic review and meta-analysis, for patients with poorly controlled asthma, SMART was associated with longer time to first severe asthma exacerbation compared with a step up or continuation of GINA step with maintenance inhaled corticosteroid鈥搇ong-acting 尾₂-agonist plus short-acting 尾₂-agonist reliever. These findings suggest that if an adult or adolescent receiving treatment at GINA step 3 or 4 has poorly controlled asthma, it is preferable to switch to the SMART regimen rather than to step up or continue the GINA treatment step with maintenance inhaled corticosteroid鈥搇ong-acting 尾₂-agonist plus short-acting 尾₂-agonist reliever therapy.

Inhaled corticosteroid (ICS)鈥搇ong acting 尾₂-agonist (LABA) therapy is the mainstay pharmacological treatment of moderate and severe asthma in adolescents and adults.^1,2 The Global Initiative for Asthma (GINA)³ recommends 2 ICS-LABA treatment regimens for patients at treatment steps 3, 4, and 5: single inhaler combination ICS鈥揻ormoterol as both maintenance and reliever therapy (SMART) or ICS-LABA as maintenance therapy plus short-acting 尾₂-agonist (SABA) as reliever (ICS-LABA maintenance plus SABA). As-needed ICS-formoterol is the preferred reliever at all GINA steps except when the ICS-LABA maintenance combination does not include formoterol as the LABA component.³

Meta-analyses of randomized clinical trials (RCTs)^4,5 have provided level I evidence that in adolescents and adults with asthma, the SMART regimen is associated with a lower risk of severe asthma exacerbations compared with ICS-LABA maintenance plus SABA. In a meta-analysis,⁴ SMART was associated with a 32% reduced exacerbation risk compared with the same dose of ICS-LABA maintenance therapy but with SABA as reliever and with a 23% reduced exacerbation risk compared with a higher-dose ICS-LABA maintenance plus SABA. Although there was no significant heterogeneity in treatment effects across the studies included in these meta-analyses, there were differences in terms of whether the randomized comparator treatments represented an escalation, continuation, or reduction in the level of treatment received compared with the patients鈥� prescribed treatment before study start. There were also differences in patients鈥� asthma control levels between and within studies, and thus, whether a change of treatment was clinically indicated is not clear. Furthermore, the inclusion of step down in maintenance treatment for patients with poor asthma control, which is not in line with asthma guidelines (eg, GINA³), may have biased the overall results. An important clinical question is whether, in patients with poorly controlled asthma, a change in treatment to SMART is more effective than stepping up to a higher ICS dose with ICS-LABA maintenance plus SABA reliever. The aim of this study was to assess whether, in patients with poorly controlled asthma at study start, the SMART regimen was associated with a longer time to first severe exacerbation compared with a step up or continuation in GINA step maintenance treatment with a SABA reliever.

This systematic review and meta-analysis used pooled individual data from studies evaluating budesonide-formoterol SMART vs equal or higher-dose conventional ICS-LABA maintenance therapy plus as-needed SABA or formoterol reliever (). Two changes were made to the original protocol (eMethods in the Supplement). First, to provide a more robust definition of uncontrolled asthma, only patients who had an asthma control questionnaire (ACQ) score of 1.5 or higher at baseline were included rather than additionally including patients who had used SABA on more than 5 occasions within 5 or more of the previous 7 days of a 2-week run-in. This meant that the study by O麓Byrne et al⁶ did not meet the prespecified criteria for inclusion (eTable 1 in the Supplement). Second, time to first severe exacerbation was chosen as the primary outcome variable for severe exacerbation risk because it was the primary end point in 4 of the 5 studies that were included in the meta-analysis.

References from a systematic review and meta-analysis of the SMART regimen were reviewed,⁴ and the literature and internal study databases at AstraZeneca and the Medical Research Institute of New Zealand were searched for RCTs published from January 1990 to February 2018 that compared budesonide-formoterol by SMART vs ICS-LABA plus SABA reliever. Studies that met the following criteria were included in the analysis: (1) included adults and/or adolescents with asthma; (2) reported baseline data on the GINA step of asthma treatment²; (3) reported baseline data on asthma control (ACQ 5-item version [ACQ-5]); (4) reported data on measures of efficacy for severe exacerbations as defined by the American Thoracic Society/European Respiratory Society⁷; (5) had at least a 24-week treatment period; (6) had access to individual patient data; (7) received institutional ethics review; and (8) had written informed consent provided by all participants. For each patient, the treatment before study entry was classified according to GINA 2018,² the current GINA version at the time of the relevant study search.

Patient-level data were obtained from each identified study by independent extraction. Patients were classified as being in the step-up or same-step arm according to the level of the ICS-LABA as maintenance therapy plus SABA reliever treatment to which patients in the control arm were randomly assigned. In the step-up control arm, patients stepped up in the GINA treatment hierarchy from GINA step 3 to GINA step 4 (medium-dose^8,9 or high-dose¹⁰ ICS-LABA plus SABA reliever). In the same-step control arm, patients remained at the same step in the GINA treatment hierarchy compared with their treatment level at enrollment (ie, GINA step 3 to GINA step 3 [low-dose ICS-LABA plus SABA reliever^11,12] and GINA step 4 to GINA step 4 [medium-dose^8,9 or high-dose¹⁰ ICS-LABA plus SABA reliever]) (Figure 1 and Table). This same-step arm could include patients who continued to receive GINA step 4 treatment even if they were randomly assigned to receive a higher ICS dose (ie, a step up from medium-dose ICS-LABA at baseline to high-dose ICS-LABA during the study) (Table).

Patients randomly assigned to the SMART regimen in the step-up comparison were receiving GINA step 3 treatment at enrollment and could either stay at the same GINA step (GINA step 3⁸) or step up as in the control arm (GINA step 4^9,10). Patients randomly assigned to the SMART regimen in the same-step comparison were either receiving GINA step 3^11,12 or step 4^8-10 treatment at enrollment and received the same GINA step after randomization (GINA step 3^8,11,12 or 4^9,10) or a lower GINA step treatment (GINA step 4 down to 3⁸).

Patients were included in this analysis if their ACQ-5 score was 1.5 or higher at baseline, representing poorly controlled asthma.^13,14 To reflect clinical practice, patients who were randomly assigned to a lower maintenance treatment than treatment at enrollment in the comparator ICS-LABA maintenance therapy plus SABA arm were excluded from the analyses.

Time to first severe asthma exacerbation was the primary outcome variable. A severe asthma exacerbation was defined as deterioration in asthma leading to treatment with an oral glucocorticoid for at least 3 days and/or hospitalization or an emergency department visit (or equivalent) in accordance with the American Thoracic Society/European Respiratory Society criteria.⁷ Secondary end points included number of severe exacerbations, ACQ-5 score, and forced expiratory volume in 1 second (FEV₁) measured at baseline and clinical visits at months 1 and 6. Because of trial design differences, for the COMPASS trial,⁸ the 2-month follow-up was used as 1-month data, and for the SMILE trial,¹¹ the 4-month follow-up was used as 6-month results. The number of symptom-free days (defined as a day with no morning and no evening symptoms and no nighttime awakening) was analyzed in the studies in which a daily symptom diary was included.^8,10-12

Data were analyzed from August 2018 to November 2021. A Cox proportional hazards regression model with treatment as a factor was used to analyze the time to first severe exacerbation and was performed separately for each study. For the pooled estimate, an analysis using all data stratified by study was performed. A fixed-effect model was used for analysis. For the number of severe exacerbations, a negative binomial model was used separately for each study, with treatment as a fixed factor and the logarithm of the time in study as offset. For the pooled estimate, a fixed factor term for study was included. Because of convergence problems, the study factor as a random factor could not be used. Longitudinal data, ACQ-5 score, and FEV₁ measured at clinical visits as well as symptom-free days (reduced to weekly mean values) were analyzed in a mixed model for repeated measurements with patient as a random effect below time point and time and treatment and treatment鈥壝椻�塼ime interaction as factors. A Toeplitz covariance matrix was used to model the correlation between measurements. This approach assumes that the correlation depends on the distance between the time units. In the COMPASS trial,⁸ the fluticasone-salmeterol and budesonide-formoterol maintenance plus SABA arms were pooled because there was no significant difference in time to first severe exacerbation between them (eTable 2 in the Supplement). Estimates and 95% CIs derived from the analysis models are shown in forest plots. All statistical analyses were performed with SAS, version 9.4 (SAS Institute). P values were 2-sided, and P鈥�<鈥�.05 was considered statistically significant.

A sensitivity analysis was undertaken without including the study by Patel et al,⁹ which was not double-blinded and used budesonide-formoterol via a pressurized metered-dose inhaler, thereby representing off-label use. Budesonide-formoterol administered via a Turbuhaler (AstraZeneca), not via a pressurized metered-dose inhaler, is registered worldwide for the SMART regimen.

Overall, 4863 patients (3034 [62.4%] female; mean [SD] age, 39.8 [16.3] years) were included in the analyses (Figure 1). The included patients had moderate and severe asthma treated at GINA step 3 or 4 at baseline, with a mean (SD) ACQ-5 of 2.46 (0.69) and mean (SD) FEV₁ of 72% (14%) of predicted normal values, as shown in eTable 3 in the Supplement. The study populations in the 5 studies included in the analysis^8-12 were broadly comparable, as shown in eTable 4 in the Supplement, and all studies included the criterion that patients had experienced at least 1 asthma exacerbation in the 12 months before enrollment. A summary of the included studies and the direction of the treatment steps is shown in the Table and Figure 1. A total of 98 patients receiving GINA step 2 treatment had erroneously been enrolled in the studies and were excluded.

In total, 1950 patients with uncontrolled asthma receiving GINA step 3 treatment were included in the step-up comparison of the SMART regimen vs step up to GINA step 4 ICS-LABA maintenance plus SABA (Table). In the COMPASS trial,⁸ the level of treatment in the SMART arm remained at GINA step 3 and was increased to step 4 in the ICS-LABA maintenance plus SABA arm. In the AHEAD trial¹⁰ and the trial by Patel et al,⁹ the level of treatment for both treatment groups was increased from step 3 to step 4. The maintenance ICS metered equivalence dose in the SMART arm was half that of the comparator ICS-LABA maintenance plus SABA arm in the AHEAD¹⁰ and COMPASS⁸ trials, whereas in the trial by Patel et al,⁹ the metered ICS dose was the same in both treatment arms.

Patients with uncontrolled asthma who were receiving GINA step 3 (n鈥�=鈥�1130) and GINA step 4 (n鈥�=鈥�1783) treatment were included in the same-step comparison of SMART vs remaining at the same step of ICS-LABA maintenance plus SABA as the step at baseline. In the COMPASS trial,⁸ the level of treatment was reduced from GINA step 4 to step 3 in the SMART arm, and in the AHEAD trial,¹⁰ there was an increase in the ICS dose in the ICS-LABA maintenance plus SABA arm (moderate to high ICS dose in GINA step 4). This resulted in the maintenance ICS metered-dose equivalence in the SMART regimen being half that in the ICS-LABA maintenance plus SABA reliever arm. In the SMILE¹¹ and SAKURA¹² trials and the trial by Patel et al,⁹ the metered ICS dose was the same in both treatment arms.

In the step-up comparison, switching patients with uncontrolled asthma receiving GINA step 3 treatment to the SMART regimen at either step 3 or 4 was associated with a prolonged time to first severe exacerbation, with a 29% reduced risk of exacerbation (hazard ratio [HR], 0.71; 95% CI, 0.52-0.97) compared with ICS-LABA maintenance plus SABA at step 4 (Figure 2A and eFigures 1 and 2 in the Supplement). Similarly, SMART was associated with a rate reduction in the number of severe exacerbations of 22% (risk ratio, 0.78; 95% CI, 0.56-1.10) (eFigures 1 and 3 in the Supplement).

The ACQ-5 scores, FEV₁, and proportion of symptom-free days per week improved over the study period in all trials analyzed in both treatment arms. The differences in ACQ-5 scores and FEV₁ in patients who received SMART compared with those stepping up to ICS-LABA maintenance therapy plus SABA at step 4 were 鈭�0.10 (95% CI, 鈭�0.19 to 鈭�0.01) and 35 mL (95% CI, 鈭�10 to 81 mL), respectively (Figure 3A and C and eFigure 1 in the Supplement). No differences between treatment arms were observed in terms of symptom-free days (Figure 3E).

In patients with uncontrolled asthma at GINA step 3 or 4, the SMART regimen at the same or a lower treatment step was associated with a prolonged time to first severe exacerbation compared with remaining at the same GINA treatment step with ICS-LABA maintenance plus SABA, with a 30% reduced risk (HR, 0.70; 95% CI, 0.58-0.85) (Figure 2B and eFigures 2 and 4 in the Supplement). The SMART regimen was also associated with a 40% severe exacerbation rate reduction (risk ratio, 0.60; 95% CI, 0.48-0.74) (eFigures 3 and 4 in the Supplement).

Patients randomly assigned to the SMART regimen experienced greater improvements in ACQ-5 scores (鈭�0.12; 95% CI, 鈭�0.20 to 鈭�0.05) and FEV₁ (47 mL; 95% CI, 15-78 mL) compared with patients remaining at the same GINA step with ICS-LABA maintenance plus SABA (Figure 3B and D and eFigure 4 in the Supplement). No differences between the 2 treatment strategies were observed in asthma symptom鈥揻ree days during each week of the study (Figure 3F).

This systematic review and meta-analysis showed that in adults and adolescents with poorly controlled asthma receiving GINA step 3 treatment, budesonide-formoterol taken according to the SMART regimen at step 3 or 4 was associated with a reduced risk of severe exacerbation compared with ICS-LABA maintenance plus SABA reliever at step 4 (medium- or high-dose ICS-LABA). The magnitude of the difference between the regimens was substantive, with a 29% reduction in risk of severe exacerbations. These findings suggest that for patients with poorly controlled asthma, switching to budesonide-formoterol administered by the SMART regimen at the same or higher level of treatment may be preferable to increasing to a higher level of treatment with maintenance ICS-LABA and SABA as required. This evidence supports the 2021 GINA Report,³ which recommends ICS-formoterol as reliever therapy at all steps of the treatment algorithm across the spectrum of asthma severity with the exception of use of an ICS-LABA that does not include formoterol.

These findings complement previous systematic reviews that demonstrated that ICS-formoterol outperformed SABA reliever therapy when taken alone as a reliever or in combination with maintenance ICS-LABA therapy.^4,5,15 Relevant to the comparisons made in this analysis, Sobieraj et al⁴ reported in a systematic review and meta-analysis that for patients receiving maintenance ICS-LABA therapy, use of an ICS-formoterol reliever was associated with a reduced risk of exacerbation of 32% compared with SABA reliever therapy at the same step and of 23% when compared with a higher dose of maintenance ICS-LABA plus SABA reliever therapy. More recently, in a systematic review and network analysis that compared efficacy among all possible treatment regimens included in relevant RCTs, Rogliani et al⁵ reported that ICS-formoterol SMART at step 3 was associated with a reduced risk of exacerbation compared with all other ICS-LABA maintenance plus SABA reliever treatment options at steps 3 and 4. The risk of a severe exacerbation associated with the low-dose ICS-formoterol SMART regimen was reduced by 45% compared with switching to low-dose ICS-LABA plus SABA, by 29% compared with stepping up to medium-dose ICS-LABA plus SABA, and by 22% compared with high-dose ICS-LABA plus SABA. The analyses by Sobieraj et al⁴ and Rogliani et al⁵ included RCTs in which all participants were included regardless of their level of symptom control, unlike our analysis, which was restricted to patients who had poorly controlled asthma at baseline with an ACQ-5 score of 1.5 or greater. This criterion was used in our study to address the potential criticism that the RCTs were not generalizable to clinical practice in which a switch or step up in treatment may not be considered if the patient鈥檚 asthma is well controlled and to address the clinical question regarding the preferred therapeutic approach for a patient with moderate or severe asthma that is not adequately controlled.

The issue of generalizability can be further assessed in this clinical scenario using the baseline characteristics of the patients included in our analysis. The mean (SD) baseline ACQ-5 score was 2.46 (0.69), with all patients in the trials requiring an ACQ-5 score of 1.5 or higher; the mean estimated FEV₁ was 72% of predicted normal reference values. All patients had experienced at least 1 asthma exacerbation in the 12 months before enrollment. Thus, the population can be considered a group at high risk of severe exacerbation for whom a change in treatment would be clinically indicated, with an increase in GINA step treatment being the main consideration.

Our findings suggest that the timing of the ICS-LABA dose may be more important than the total daily dose in determining the outcomes associated with ICS-LABA therapy. This interpretation is consistent with the limited available data, which suggest that budesonide-formoterol has both greater potency and efficacy when taken as needed compared with its use as regular maintenance therapy.^8,16,17 An RCT of budesonide-formoterol reliever monotherapy vs budesonide-formoterol fixed-dose maintenance plus SABA reliever therapy showed a 4.6-fold (95% CI, 2.9-7.3鈥揻old) greater potency in reducing the risk of severe exacerbations,^16,17 although the incidence of severe exacerbations was low overall (0.143 per patient per year), suggesting that both regimens were effective in controlling exacerbations.¹⁷ In an RCT that compared budesonide-formoterol SMART with higher fixed-dose maintenance budesonide-formoterol plus SABA reliever therapy,⁸ there was an additional 26% (95% CI, 4%-42%) reduction in severe exacerbation risk despite a 25% lower total budesonide-formoterol dose, suggesting a greater maximum effect.

This study has strengths. The external validity of our study was enhanced by the inclusion of the trial by Patel et al,⁹ a real-world, open-label study of high-risk asthma that used an off-label strength of pressurized metered-dose inhaler (200/6 渭g metered dose) to facilitate electronic monitoring. To allow regulatory authorities and other entities to be able to use the data from approved doses, a sensitivity analysis was undertaken without this study, leading to similar results.

Time to first severe asthma exacerbation was the primary outcome variable in this analysis, to be consistent with the primary end point in 4 of the 5 studies included in the meta-analysis, and the total number of severe exacerbations was also reported. However, data were also available for other key measures, such as ACQ-5 and FEV₁, which had greater improvement in patients who received the SMART regimen, but the differences were below the recognized minimal clinically important differences.¹⁸ This finding suggests that the main benefit associated with the SMART regimen is a reduction in risk of severe exacerbations, whereas symptom control and lung function may be maintained or modestly improved. All analyzed studies were of at least 24 weeks鈥� duration, but longer-term data spanning many years are not available for SMART or for RCTs of other ICS-LABA treatments.

This study also has limitations. First, because of the requirement to have access to individual patient data, the analysis was limited to RCTs of the budesonide-formoterol SMART regimen vs fixed-dose ICS-LABA plus SABA; this enhanced the internal validity but meant that the RCT of beclomethasone dipropionate鈥揻ormoterol SMART regimen could not be included.¹⁹ Although that study did not fulfill our prespecified treatment of interest (budesonide-formoterol) per the published PROSPERO protocol, the results from the subpopulation of patients with uncontrolled asthma would have been informative. In the overall study population of that study,¹⁹ the beclomethasone-formoterol SMART regimen significantly prolonged time to first exacerbation compared with beclomethasone-formoterol plus salbutamol as needed, with a risk reduction similar to that observed in the overall populations of the studies included in this meta-analysis. This suggests that inclusion of the study may not have substantially changed the results. In addition, 1 study evaluating budesonide-formoterol SMART⁶ was not included because ACQ-5 data as an indicator of asthma control were not available for comparison. However, the results were similar to those of the studies included, suggesting inclusion may not have substantially influenced the results.

In this systematic review and meta-analysis, in patients with poorly controlled asthma, switching to the SMART regimen was associated with longer time to first severe exacerbation compared with a step up or a continuation in level of ICS-LABA maintenance plus SABA reliever therapy.

Accepted for Publication: December 21, 2021.

Published: March 1, 2022. doi:10.1001/jamanetworkopen.2022.0615

Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License. 漏 2022 Beasley R et al. 糖心vlog Open.

Correction: This article was corrected on May 18, 2022, to fix an error in the Figure 2B label.

Corresponding Author: Richard Beasley, DSc, Medical Research Institute of New Zealand, Level 7 CSB Building, Wellington Hospital, Riddiford Street, Newtown, Wellington 6021, New Zealand (Richard.Beasley@mrinz.ac.nz).

Author Contributions: Dr Peterson and Mr Bengtsson had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Beasley, Harrison, Gustafson, Hamblin, Fager氓s.

Acquisition, analysis, or interpretation of data: Beasley, Peterson, Gustafson, Hamblin, Bengtsson, Fager氓s.

Drafting of the manuscript: Beasley, Gustafson, Hamblin, Fager氓s.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Peterson, Gustafson, Bengtsson, Fager氓s.

Obtained funding: Gustafson, Fager氓s.

Administrative, technical, or material support: Beasley, Harrison, Gustafson, Hamblin, Fager氓s.

Supervision: Beasley, Harrison, Gustafson, Fager氓s.

Conflict of Interest Disclosures: Dr Beasley reported receiving grants from AstraZeneca, Cure Kids, the Health Research Council of New Zealand, Genentech and GlaxoSmithKline; receiving personal fees from AstraZeneca, Avillion, Cipla, and Theravance; serving on advisory boards for AstraZeneca, Avillion, and Theravance; and serving as the chair of the adolescent and adult asthma guidelines group for the Asthma and Respiratory Foundation NZ outside the submitted work. Dr Harrison reported receiving personal fees and grants from AstraZeneca and being an employee and shareholder of AstraZeneca during the conduct of the study and receiving grants from AstraZeneca outside the submitted work. Dr Peterson reported receiving payment for statistical analysis from AstraZeneca and being an employee of StatMind during the conduct of the study. Dr Gustafson reported being an employee and shareholder of AstraZeneca during the conduct of the study and outside the submitted work. Mr Hamblin reported being an employee and shareholder of AstraZeneca during the conduct of the study and outside the submitted work. Mr Bengtsson reported receiving payment for statistical analysis from AstraZeneca and being an employee of StatMind during the conduct of the study. Dr Fager氓s reported being a full-time employee and shareholder of AstraZeneca during the conduct of the study.

Funding/Support: This study was supported by AstraZeneca. StatMind received funding from AstraZeneca to complete the statistical analysis.

Role of the Funder/Sponsor: AstraZeneca contributed to the design and conduct of the study; was responsible for the collection of data from 4 of the 5 studies included in the analysis; was responsible for management and analysis of the data; contributed to the interpretation of the data; contributed to the preparation, review, and approval of the manuscript; and supported the decision to submit the manuscript for publication.

Additional Contributions: David Candlish, BSc, and Stefan Courtney, MSc (inScience Communications, Springer Healthcare Ltd, Chester, UK), provided additional medical writing support and received compensation through funding from AstraZeneca.