ORIGINAL ARTICLE
Indacaterol/Glycopyrronium Versus Salmeterol/Fluticasone for COPD Exacerbations
Jadwiga A. Wedzicha, M.D.1, Donald Banerji, M.D.2, Kenneth R. Chapman, M.D.3, Jørgen Vestbo, M.D., DMSc.4, Nicolas Roche, M.D.5, R. Timothy Ayers, MSc.2, Chau Thach, PhD.2, Robert Fogel, M.D.2, Francesco Patalano, M.D.6 and Claus F. Vogelmeier, M.D.7, for the FLAME-COPD Investigators*
1National Heart and Lung Institute, Imperial College London, London, UK; 2Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA; 3Asthma and Airway Centre, University Health Network and University of Toronto, Toronto, Ontario, Canada; 4Centre for Respiratory Medicine and Allergy, The University of Manchester and University Hospital South Manchester NHS Foundation Trust, Manchester, UK; 5Service de Pneumologie AP-HP, University Paris Descartes (EA2511), Paris, France; 6Novartis Pharma AG, Basel, Switzerland; 7Department of Medicine, Pulmonary and Critical Care Medicine, University Medical Center Giessen and Marburg, Philipps-Universität Marburg, Marburg, Germany; Member of the German Center for Lung Research (DZL)
*Investigators in the FLAME study are listed in the Supplementary Appendix, available at NEJM.org.
Key words: COPD; dual bronchodilation; exacerbations; chronic bronchitis; emphysema; FLAME
Word count = 2680 [journal word-count limit = 2700]
Abstract word count = 250 [strict journal word-count limit = 250]
References = 33 [journal limit = 40]
Figures and tables = 5 [3 figures and 2 tables] [journal limit = 5]
Supplementary figures and tables = 14 [8 figures and 8 tables]
ABSTRACT
Background
Most guidelines recommend a long-acting β2-agonist/inhaled corticosteroid (LABA/ICS) combination or a long-acting muscarinic antagonist (LAMA) as first-choice treatment for patients with chronic obstructive pulmonary disease (COPD) with high exacerbation risk. The role of a LABA/LAMA combination for these patients is unclear.
Methods
In this 52-week, randomized, double-blind, double-dummy trial, we compared once-daily LABA/LAMA indacaterol/glycopyrronium (IND/GLY) 110/50 μg with twice-daily LABA/ICS salmeterol/fluticasone combination (SFC) 50/500 μg on exacerbations in COPD patients with a history of ≥1 exacerbation in the preceding year.
Results
Patients were randomized to IND/GLY (n=1680) or SFC (n=1682). Meeting non-inferiority versus SFC (primary endpoint), IND/GLY further showed superiority in reducing the annual rate of all exacerbations by 11% versus SFC (3.59 vs. 4.03; rate ratio [RR], 0.89; 95% confidence interval [CI], 0.83 to 0.96; P=0.003) and prolonged time to first exacerbation, with a 16% risk reduction (hazard ratio [HR], 0.84; 95% CI, 0.78 to 0.91; P<0.001). IND/GLY reduced the annual rate of moderate or severe exacerbations (RR, 0.83; 95% CI, 0.75 to 0.91; P<0.001) and prolonged the time to first moderate or severe exacerbation (HR, 0.78; 95% CI, 0.70 to 0.86; P<0.001) and first severe exacerbation (HR, 0.81; 95% CI, 0.66 to 1.00; P=0.046) versus SFC. The effect on exacerbations was independent of baseline blood eosinophil count. Incidence of adverse events and deaths was similar in both groups. Incidence of pneumonia was 3.2% with IND/GLY and 4.8% with SFC.
Conclusions
IND/GLY is more effective than SFC in preventing exacerbations in COPD patients with a history of exacerbations (Funded by Novartis; FLAME ClinicalTrials.gov number, NCT01782326).
[word count = 250]
INTRODUCTION
Exacerbations of chronic obstructive pulmonary disease (COPD) are associated with accelerated lung function decline,1-3 impaired quality of life,4 hospitalization,5 and increased mortality.6 COPD exacerbations are costly to healthcare systems,7 thus prevention is a key goal in COPD management.8
Inhaled long-acting bronchodilators, in addition to providing symptomatic control, prevent COPD exacerbations.9-12 Inhaled corticosteroids (ICS) are also known to reduce exacerbation frequency and have been studied in combination with inhaled long-acting β2-agonists (LABAs).11,13,14 A LABA/ICS fixed combination (salmeterol/fluticasone propionate [SFC]) and an inhaled long-acting muscarinic antagonist (LAMA; tiotropium) had similar effects on exacerbations in exacerbation-prone patients.15 Consequently, treatment guidelines have recommended that either a LABA/ICS or a LAMA can be used to prevent exacerbations in at-risk patients.8
Long-term ICS use is associated with a small but significant risk of pneumonia16,17 and other steroid-related adverse effects.18 An alternative strategy to the LABA/ICS combination for exacerbation-prone patients is a dual bronchodilator regimen of LABA/LAMA.19
The hypothesis for the present study was that, in patients with COPD at risk of exacerbations, a dual bronchodilator (indacaterol/glycopyrronium (IND/GLY) 110/50 μg, once daily) would be at least as effective in preventing exacerbations as a LABA/ICS (SFC 50/500 μg, twice daily). As recent studies have indicated that exacerbation prevention with ICS may relate to blood eosinophil level,20-22 the relationship of baseline blood eosinophils to exacerbation reduction by both interventions was examined prospectively.
METHODS
Study design and oversight
FLAME (NCT01782326) was a multicenter, randomized, double-blind, double-dummy, parallel-group, non-inferiority study from July 2013 to September 2015 at 356 centers in 43 countries (Fig. S1). In the 4-week run-in (preceded by a 1-week screening period) all patients received tiotropium 18 μg once daily. After run-in, patients discontinued tiotropium and were randomized 1:1 to IND/GLY 110/50 μg once daily or SFC 50/500 μg twice daily for 52 weeks, with a 30-day follow-up. Salbutamol 100 μg was provided as rescue medication. Additional details in Supplementary Appendix.
The sponsor (Novartis) developed the protocol, with guidance from the first author and advice from the other academic authors. The first draft was written by JAW and DB. All authors reviewed and edited the manuscript and decided to submit the manuscript for publication. All authors contributed to the interpretation of the data and had access to the full data (non-disclosure agreements were in place). The trial was approved by the ethics committee at each study center and all patients provided written informed consent. All authors vouch for the accuracy and completeness of the data and the fidelity of the study to the protocol (available at NEJM.org). Statistical analyses were performed by a statistician at DataMap (funded by Novartis, who provided oversight and verified key results).
Patients
COPD patients enrolled in the study were ≥40 years old, with a modified Medical Research Council grade ≥2 for dyspnea, a post-bronchodilator forced expiratory volume in one second (FEV1) ≥25 and <60% predicted and a post-bronchodilator FEV1/forced vital capacity (FVC) <0.70. Patients were required to have a documented history of ≥1 COPD exacerbation (that required treatment with systemic corticosteroids and/or antibiotics) in the previous 12 months. Additional details in Supplementary Appendix.
Outcome measures
The primary objective was to demonstrate that IND/GLY 110/50 µg was at least non-inferior to SFC 50/500 µg regarding the rate of all COPD exacerbations (mild, moderate or severe).
Our protocol lists 24 secondary objectives; we report data from 16 of these here (additional details in Supplementary Appendix). An important secondary objective was to demonstrate the superiority of IND/GLY versus SFC for the same endpoint, if non-inferiority was met. Other secondary endpoints included time to first COPD exacerbation (all); rate and time to first moderate or severe COPD exacerbation; and rate and time to first severe COPD exacerbation. Trough FEV1 and FVC, FEV1 area under the curve from 0–12 hours (AUC0–12h; in a subset of patients), health status (via the St George’s Respiratory Questionnaire [SGRQ]) and rescue medication use were also assessed.
Exacerbations, defined according to Anthonisen criteria23, were categorized as mild (worsening of symptoms for more than two consecutive days, not treated with systemic corticosteroids and/or antibiotics), moderate (treated with systemic corticosteroids and/or antibiotics), or severe (requiring hospital admission [or an emergency room visit of >24 hours] in addition to treatment with systemic corticosteroids and/or antibiotics). Patients recorded daily symptoms and rescue medication use in an electronic diary. When worsening of symptoms met pre-defined exacerbation criteria, alerts were triggered and patients were advised to contact the site.
Safety of IND/GLY and SFC were also assessed. An independent adjudication committee assessed blinded safety data. Radiographic imaging was required to confirm pneumonia. Additional details in Supplementary Appendix.
Statistical analysis
Approximately 3332 randomized patients were required to rule out a 15% increase in COPD exacerbation rate (all) with IND/GLY versus SFC, with greater than 95% power at the level of 0.025 (one-sided), assuming 30% drop out and/or major protocol deviations.
The non-inferiority bound of 1.15 was based on a previous study,11 where the rate ratio for SFC versus placebo on moderate or severe exacerbations was 0.75. If FLAME could rule out a 1.15-fold increase with IND/GLY versus SFC on the same endpoint, the rate ratio for IND/GLY versus placebo would be 0.8625, leading to a meaningful rate reduction of >13.75% versus placebo.
The full analysis set (also known as a modified intention-to-treat set) included all randomized patients who received at least one dose of double-blind drug and did not have major violations of Good Clinical Practice identified prior to unblinding. The per-protocol set included all patients in the full analysis set without major protocol deviations, defined prior to unblinding. Main analysis of the primary endpoint was on the per-protocol set; the full analysis set was supportive. Conclusions made on the primary endpoint were strengthened by the concordance of results obtained in the per-protocol set and full analysis set. Analyses of all other efficacy endpoints were performed on the full analysis set. All efficacy analyses, unless stated otherwise, were based on on-treatment data.
The number of exacerbations during the treatment period was analyzed using a negative binomial model, including terms for treatment, baseline smoking status, prior ICS use, airflow limitation, and region as fixed effects, along with baseline total symptom score and COPD exacerbation history as covariates. The overall Type I error from the two hypothesis tests (i.e. non-inferiority test followed by superiority test) was maintained at level 0.05 (two-sided) with the following statistical testing hierarchy:
- Non-inferiority of IND/GLY versus SFC could be claimed if the upper limit of the confidence interval (CI) of the exacerbation rate ratio (IND/GLY versus SFC) was less than 1.15.
- If non-inferiority was demonstrated, superiority of IND/GLY versus SFC in reducing exacerbation rate could be claimed if the upper limit of the same CI was less than 1.
No multiplicity adjustments were performed on any other endpoints.
Exacerbations were also analyzed according to 18 prespecified subgroups, including baseline blood eosinophils, to assess the consistency of the treatment effect. Other exacerbation endpoints were analyzed using the same negative binomial model. The time-to-event endpoints were analyzed using a Cox regression model, which included the same terms as the negative binomial model. Additional details in Supplementary Appendix.
RESULTS
Patients
Fig. 1 and Fig. S3 display the reasons for discontinuation from screening, run in and treatment. 3.6% of subjects discontinued run in due to an exacerbation. See Supplementary Appendix for additional details. 3362 patients were randomized: 1680 to IND/GLY and 1682 to SFC. Four were excluded from all analyses for not taking any study medication. 3084 were included in the per-protocol set and 3354 in the full analysis set (Fig. 1). Treatment discontinuation rates were 16.6% versus 19.0% on IND/GLY and SFC, respectively (Fig. 1).
Demographics and disease history for both treatment groups were well balanced (Table 1). 19.3% of patients had a history of ≥2 moderate or severe exacerbations in the past year and 56.3% of patients were using ICS at screening. Treatment medication compliance was >99%. Additional details in Supplementary Appendix.
Primary endpoint
The rate ratio (RR) for all exacerbations with IND/GLY versus SFC in the per-protocol set was 0.89 (95% CI, 0.83 to 0.96; P=0.003; Fig. 2A and Fig. S5A). The upper bound of the two-sided 95% CI was less than the non-inferiority margin of 1.15, demonstrating non-inferiority and meeting the primary endpoint. Non-inferiority was also demonstrated in the full analysis set (RR, 0.88; 95% CI, 0.82 to 0.94; P<0.001; Fig. 2A). Similar results were observed with additional sensitivity analyses performed using both on-treatment and off-treatment data from those who discontinued treatment early (additional details in Supplementary Appendix).
Secondary endpoints
Analyses of all other efficacy endpoints were performed on the full analysis set. Superiority (a secondary endpoint adjusted for multiplicity) was demonstrated for IND/GLY versus SFC for all exacerbations, as the upper bound of the same two-sided 95% CI was less than 1 (Fig. 2A). IND/GLY prolonged the time to first exacerbation versus SFC, with a 16% reduction in risk (hazard ratio [HR], 0.84; 95% CI, 0.78 to 0.91; P<0.001; Fig. 2B). IND/GLY also reduced the rate of moderate or severe exacerbations (healthcare utilization) by 17% (RR, 0.83; 95% CI, 0.75 to 0.91; P<0.001; Fig. S5C) and prolonged the time to first moderate or severe exacerbation, with a 22% reduction in risk (HR, 0.78; 95% CI, 0.70 to 0.86; P<0.001; Fig. 2C) versus SFC. Additionally, IND/GLY significantly prolonged the time to first severe exacerbation versus SFC, with a 19% reduction in risk (HR, 0.81; 95% CI, 0.66 to 1.00; P=0.046; Fig. 2D). The rate ratio of severe exacerbations for IND/GLY versus SFC was 0.87 (95% CI, 0.69 to 1.09; P=0.231).
The rate of moderate or severe exacerbations was analyzed by baseline blood eosinophils at cut-offs of <2% and ≥2%. The exacerbation rate was significantly lower with IND/GLY than SFC in both the <2% (RR, 0.80; 95% CI, 0.68 to 0.93; P=0.004) and ≥2% subgroups (RR, 0.85; 95% CI, 0.75 to 0.96; P=0.010). The other three subgroup analyses based on different cut-offs of baseline blood eosinophils, not presented here, provided similar results. No meaningful interaction was seen between the rate of all or moderate or severe exacerbations and prior therapy, or other baseline characteristics (Fig. 3; Figs. S5A and B). Additional details in Supplementary Appendix.
Other secondary endpoints
Improvement in trough FEV1 was significantly greater with IND/GLY versus SFC, with a treatment difference of 62 mL at Week 52 (P<0.001; Fig. S7A and Table S5). FEV1 AUC0–12h subset results are presented in the Supplementary Appendix.
Improvement in SGRQ-C total score was greater with IND/GLY versus SFC, with treatment differences of –1.2 to –1.8 units between Weeks 12 and 52 (all P<0.01; Fig. S7C; Table S6). The percentage of patients achieving a clinically important improvement of ≥4 units in SGRQ-C total score was significantly higher with IND/GLY than with SFC at Week 52 (49.2% vs. 43.7%, respectively; odds ratio, 1.30; P<0.001). Rescue medication use decreased with IND/GLY versus SFC (Figure S7D).
Safety
The incidence of adverse events, including serious adverse events, was similar between the IND/GLY and SFC groups (Table 2). The number of deaths was 24 (1.4%) in both groups with respiratory or cardiovascular deaths being the most common (Tables S7 and S8). The incidence of pneumonia was 3.2% with IND/GLY and 4.8% with SFC. The median percentage change in the 24-hour urinary cortisol/creatinine ratio at Week 52 was 5.62% with IND/GLY and –10.39% with SFC in a subset of 535 patients (Fig. S8).
DISCUSSION
This is the first clinical trial powered to determine if a LABA/LAMA combination (IND/GLY) is as effective as a LABA/LAMA combination (SFC) in patients at risk of COPD exacerbations. Not only was the non-inferiority endpoint reached but, on subsequent superiority testing, the LABA/LAMA was found to be consistently superior for all exacerbation outcomes, lung function, and health status.
COPD clinical guidelines and strategy documents8,24 have recommended that in patients at risk of exacerbations, first-line therapy should either be a LABA/ICS combination or a LAMA. One previous study found no difference between a LABA/ICS and a LAMA monotherapy in exacerbation rates.15 However, a recent combination bronchodilator study using LABA/LAMA demonstrated greater exacerbation reduction efficacy than LAMA monotherapy19. ICS withdrawal in patients on triple therapy did not increase exacerbation rate significantly25, further supporting the hypothesis that ICS may not be essential for exacerbation prevention in patients on LABA/LAMA.
The LABA/LAMA combination was superior and consistent on exacerbation outcomes for all types of exacerbation, including exacerbations requiring healthcare utilization. COPD exacerbations were carefully monitored with daily symptom recordings and captured by electronic diaries4, allowing us to document all exacerbations, including those requiring healthcare utilization. Studies have shown under-reporting of exacerbation events (mild exacerbations), yet these unreported events impact on patients’ health status.26-28 Therefore ‘all exacerbations’ was studied as the primary endpoint to reflect the importance of preventing every exacerbation. Capturing all exacerbations is a major strength of this trial and we have shown a very consistent benefit of dual bronchodilation across exacerbation severities.
Post-hoc analyses from LABA/ICS trials in COPD have suggested that patients with elevated blood eosinophil counts, such as ≥2% blood eosinophil levels, obtain larger benefits from LABA/ICS medications on exacerbation reduction than patients with lower eosinophil counts.20-22 This may suggest that a higher eosinophil count may be associated with a greater response to ICS. Therefore, the FLAME trial prospectively examined the relationship between blood eosinophils and exacerbation outcomes. The rate of moderate or severe exacerbations and all exacerbations was significantly lower with IND/GLY than SFC in both the <2% and ≥2% eosinophil subgroups, suggesting that the LABA/LAMA combination is more effective at reducing exacerbations than SFC in both eosinophil subgroups.
The superiority of IND/GLY in terms of lung function was expected as two bronchodilators improve lung function to a greater degree than a LABA/ICS combination.29-31 Further evidence of symptomatic benefit was seen with the reduction in use of rescue medication and the improvement in health status (SGRQ-C score) in patients on IND/GLY compared with SFC.
A potential limitation of the study is that some of the patients initially treated with LABA/ICS before enrollment and then randomized into the IND/GLY treatment arm may have suffered withdrawal effects from long-term LABA/ICS use, with an increase of exacerbations. There was clearly no evidence that patients on ICS prior to the trial withdrew preferentially during run in, and exacerbation rates during run in were low. Additionally, analyses of exacerbation rates by prior therapy showed no meaningful interaction between treatment and the type of prior therapy taken by the patient.