Respiratory Medicine (2014) 108, 1498e1507

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier .com/locate /rmed

Pooled safety analysis of the fixed-dosecombination of indacaterol andglycopyrronium (QVA149), itsmonocomponents, and tiotropium versusplacebo in COPD patients

Jadwiga A. Wedzicha a,*, Ronald Dahl b, Roland Buhl c,Agnes Schubert-Tennigkeit d, Hungta Chen e, Peter D’Andrea e,Robert Fogel e, Donald Banerji e

a Airway Disease Section, National Heart and Lung Institute, Imperial College London, London, UKb Allergy Department, Odense University Hospital, Odense, Denmarkc Pulmonary Department, Mainz University Hospital, Mainz, Germanyd Novartis Pharma AG, Basel, Switzerlande Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA

Received 5 June 2014; accepted 28 July 2014Available online 4 August 2014

KEYWORDSGlycopyrronium;Indacaterol;Pooled analysis;QVA149;Safety;Tiotropium

* Corresponding author. National H(mobile); þ44 207 594 7947 (office).

E-mail address: j.wedzicha@imper

http://dx.doi.org/10.1016/j.rmed.200954-6111/ª 2014 The Authors. Pubcreativecommons.org/licenses/by-nc-

Summary

Background: To further assess the safety profile of the fixed-dose combination of indacateroland glycopyrronium (QVA149) and its monocomponents; we investigated the impact of individ-ual patient-level factors and time by integrating the patient-level safety data from the QVA149clinical programme with relevant information from the independent indacaterol and glycopyr-ronium safety databases.Methods: Data from 11,404 patients with chronic obstructive pulmonary disease (COPD) werepooled from 14 clinical studies of QVA149, indacaterol and glycopyrronium of �3 month’s dura-tion with at least two of the treatment groups: QVA149 110/50 mg, glycopyrronium 50 mg, in-dacaterol 150 mg, placebo or tiotropium 18 mg. Overall hazard ratio (HR) was assessed betweenthe active treatments and placebo and in various subgroups related to severity of airwaysobstruction, inhaled corticosteroid use, cardiovascular risk factors, sex, age and body mass in-dex for death, serious cases of cardio- and cerebrovascular (CCV) events, major adverse

eart and Lung Institute, Imperial College London, London SW3 6LR, UK. Tel.: þ44 07850 630745

ial.ac.uk (J.A.Wedzicha).

14.07.011lished by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://nd/3.0/).

Safety of QVA149 in a pooled analysis 1499

cardiovascular events (MACEs), pneumonia, COPD exacerbations requiring hospitalisation oratrial flutter/fibrillation (AF/F).Results: The HR for QVA149 versus placebo showed no significant increase in the overall risk fordeath (HR [95% confidence interval]: 0.93 [0.34e2.54]); CCV events (0.60 [0.29e1.24]); MACE(1.04 [0.45e2.42]); pneumonia (1.10 [0.54e2.25]); COPD exacerbations (0.60 [0.40e0.91]);and AF/F (1.03 [0.49e2.18]). Similar results were observed for indacaterol, glycopyrroniumand tiotropium versus placebo for overall risk and in analysed subgroups.Conclusions: There was no increase in the risk for the investigated safety endpoints for thefixed-dose combination QVA149, and it had a comparable safety profile as its monocomponentsand tiotropium versus placebo.ª 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Introduction

Long-acting bronchodilators (b2-agonists and/or muscarinicantagonists) are the foundation of the pharmacologicalmanagement strategy for patients with chronic obstructivepulmonary disease (COPD) at all severity levels of the disease(alternative choice forGOLDgroupA and first choice forGOLDB-D) [1]. Dual bronchodilation with a long-acting b2-agonist(LABA) and long-acting muscarinic antagonist (LAMA) is rec-ommended for patients with a high symptomatic burden and/or risk of exacerbations, as well as for patients who remainsymptomatic on mono-bronchodilator therapy [1,2]. It pro-vides additional clinical benefits in terms of improving effi-cacy and has the potential to decrease the rate of side effectscompared with increasing the dose of a mono-bronchodilator[1] and avoiding risks such as pneumonia associated with theuse of LABA/inhaled corticosteroid (ICS) combinations [3].

The safety profile of bronchodilators in COPD remains ofprime importance [4,5]. Studies such as Towards a Revo-lution in COPD Health (TORCH) and Understanding Long-term Impacts on Function with Tiotropium (UPLIFT) haveprovided the evidence that long-acting bronchodilators attherapeutic doses in stable COPD do not increase risk of thecardio- and cerebrovascular (CCV) mortality or morbidity inpatients with COPD [6,7].

The efficacy and safety of free-dose combinations oflong-acting bronchodilators such as salmeterol with tio-tropium, formoterol with tiotropium, indacaterol with tio-tropium, and indacaterol with glycopyrronium comparedwith either one or both monocomponents have been re-ported in several studies of 6e26 weeks of duration in pa-tients with COPD. In these studies, the free combinationshowed a similar safety profile compared with theirrespective monocomponents [8e11]. QVA149 is an inhaledonce-daily (o.d.) dual bronchodilator containing a fixed-dose combination of the LABA indacaterol maleate andLAMA glycopyrronium bromide which along with its mono-components, is approved in the European Union, Japan,Canada and other countries for the maintenance treatmentof patients with COPD. The efficacy and safety of QVA149[12e14], indacaterol [15e19] and glycopyrronium [20,21]versus placebo and active comparators have been demon-strated in their clinical development programmes. In thepivotal Phase III Indacaterol and GlycopyrroNium bromideclInical sTudiEs (IGNITE) clinical trial programme, QVA149o.d. demonstrated an acceptable safety profile, with

adverse events (AEs) and serious AEs (SAEs) being similar toplacebo, its monocomponents, tiotropium or salmeterol/fluticasone combination (SFC) [12e14,22,23].

Here, we report a pooled analysis that was conducted toanalyse specific safety events that are of importance inpatients with COPD, so as to provide a more comprehensiveevaluation of the safety profile of QVA149. The presentanalysis integrated the patient-level safety data across theQVA149 IGNITE programme with the available relevant in-formation from the indacaterol and glycopyrronium safetydatabases so that individual components and their fixeddose combination can be examined together along with astandard of care, tiotropium. This analysis investigated theimpact of individual patient risk factors and duration ofexposure to study medications and provided additional as-sessments on infrequently occurring safety endpoints.

Methods

Study design and selection

All randomised clinical trials from QVA149, indacaterol andglycopyrronium programmes were reviewed for the purposeof this pooled analysis. The focus of the pooled analysis wasto provide information on key safety endpoints for QVA149,and the monocomponents studies were included as a part ofits evaluation. The individual study designs of 14 includedstudies for this pooled analysis are summarised inSupplementary Appendix 1 Table A1.

This pooled analysis included data from all completedphase III randomised clinical trials of �3 month’s duration(study completion date as of September 30, 2012) with atleast two of the following treatment groups: QVA149 110/50 mg o.d.; glycopyrronium 50 mg o.d.; indacaterol 150 mgo.d.; placebo (all delivered via the Breezhaler� device) ortiotropium 18 mg o.d. (delivered via the HandiHaler� device).

For studies involving indacaterol, only patients thatreceived the 150 mg dose were included in the analysis asthis corresponds to the equivalent indacaterol dose used inQVA149. In order to match the fine particle dose of inda-caterol in both the combination and the 150 mg mono-therapy product, the dose of indacaterol in QVA149 is110 mg and the two doses can be considered clinicallyequivalent [24]. Further details on dose adjustment aregiven in Supplementary Appendix 2.

1500 J.A. Wedzicha et al.

The methods and the main results of the includedstudies have been previously published or presented indetail. The studies included from the indacaterol pro-gramme were: INdacaterol: vs tiotropium to Help AchieveNew COPD treatment Excellence (INHANCE) [19,25], INda-caterol: Toward Establishment of cliNical SuperiorITY (IN-TENSITY) [15], INdacaterol: Double blind One yeaR SafetyEvaluation (INDORSE) [16], INdacaterol: efficacy evaLuationusInG 150 mg doses witH COPD paTients �1 (INLIGHT1) [18],INLIGHT2 [17], B1302 [26], and B2333 [27]; glycopyrroniumprogramme: GLycopyrronium bromide in COPD airWaysclinical study 1(GLOW1) [20], GLOW2 [21], and GLOW4 [28];and QVA149 IGNITE programme: SHINE [12], SPARK [13],ENLIGHTEN [14], and ARISE [29]. In the INTENSITY [15]study, blinded tiotropium was used as an active compar-ator while INHANCE [19], GLOW2 [21], SHINE [12], SPARK[13] and ARISE [29] studies had open-label tiotropium as anactive comparator. All included studies were approved byinstitutional review boards and ethics committees atparticipating centres, and were conducted in accordancewith the Declaration of Helsinki and Good Clinical Practice.All the patients provided written informed consent.

Patients/study population

The patients’ inclusion and exclusion criteria in thispooled analysis were similar across all studies. The pro-tocols of the included studies allowed patients with stableco-morbid cardiovascular conditions to be screened andincluded in studies at the discretion and judgment of theinvestigator.

The studies enrolled men and women aged �40 yearswith moderate-to-severe COPD (Stage II or III according to

Table 1 Safety population by study and treatment (number of

Study TreatmentDuration

QVA149110/50 mg

Indac150 m

Indacaterol

INHANCE [19,25] 26 weeks 0 272INDORSE [16] 52 weeks 0 144INLIGHT1 [18] 12 weeks 0 211INLIGHT2 [17] 26 weeks 0 330INTENSITY [15] 12 weeks 0 794B2333 [27] 26 weeks 0 187B1302 [26] 12 weeks 0 114Glycopyrronium

GLOW1 [20] 26 weeks 0 0GLOW2 [21] 52 weeks 0 0GLOW4 [28] 52 weeks 0 0QVA149

SHINE [12] 26 weeks 474 476SPARK [13] 64-76 weeks 729 0ENLIGHTEN [14] 52 weeks 225 0ARISE [29] 52 weeks 119 0Total 1547 2528

INDORSE, INdacaterol: Double blind One yeaR Safety Evaluation; INtreatment Excellence; INLIGHT1, INdacaterol: efficacy evaLuation usIToward Establishment of cliNical SuperiorITY; GLOW, GLycopyrroniuma All tiotropium treatment arms were open-label, except for INTEN

the GOLD 2005 and 2008 criteria; post-bronchodilatorforced expiratory volume in 1 s [FEV1] of �30% and <80%of the predicted normal; post-bronchodilator FEV1 toforced vital capacity (FVC) ratio of <0.70 at screening),with the exception of the SPARK study that had patientswith severe-to-very severe COPD (Stage IIIeIV according tothe GOLD 2008 criteria; post-bronchodilator FEV1 �50%predicted) and a documented exacerbation history in thelast year [13]. Patients were current or ex-smokers with asmoking history of �10 pack-years (�20 pack-years in B1302[26], INHANCE [19,25], INDORSE [16], INLIGHT1 [18], andINLIGHT2 [17]).

Patients with conditions such as ischaemic heart dis-ease, left ventricular failure, history of myocardial infarc-tion (MI) or arrhythmia (excluding chronic stable atrialflutter/fibrillation [AF/F]) were not specifically excludedfrom the studies, unless these cardiovascular conditionswere regarded as to affect the patient’s safety, complianceor ability to complete the study as per the investigator’sdiscretion.

In most of the studies, the key exclusion criteria were:a history of the long QT syndrome or a prolonged QTc in-terval at screening (>450 ms for males and females), ahistory of asthma or a clinically abnormal electrocardio-gram (ECG) that could potentially place patients at risk ifenrolled into the study as per the investigator’sdiscretion.

Evaluations and outcome measures

The present analysis focused on infrequently occurringsafety endpoints that included the risk of death (all-causemortality), CCV events, major adverse cardiovascular

patients).

aterolg

Glycopyrronium50 mg

Tiotropium18 mga

Placebo

0 415 2940 0 1240 0 2050 0 3350 799 00 0 1860 0 117

550 0 267525 267 268123 40 0

473 480 232740 737 0

0 0 1130 39 0

2411 2777 2141

HANCE, INdacaterol: vs tiotropium to Help Achieve New COPDnG 150 mg doses witH COPD paTients �1; INTENSITY, INdacaterol:bromide in COPD airWays clinical study.

SITY study.

Safety of QVA149 in a pooled analysis 1501

events (MACEs), pneumonia, COPD exacerbations requiringhospitalisation and atrial flutter/fibrillation. MACEs werecategorised as non-fatal MI, unstable angina, non-fatalstroke, heart failure requiring hospitalisation and coro-nary revascularisation. The details of the terms areincluded in the Supplementary Appendix 2.

The overall hazard ratio (HR) was assessed betweenplacebo and active therapy arms that included o.d.QVA149, indacaterol, glycopyrronium or tiotropium byincorporating the relevant covariates that may have impacton the safety outcome.

In addition, it was assessed whether there was any evi-dence that the HR of QVA149 or any other drug versusplacebo changes over time in various subgroups based on

Table 2 Demographics and baseline patient characteristics acr

Parameter QVA149 110/50 mg(n Z 1547)

Indacaterol 15(n Z 2528)

Mean (SD) agein years

63.8 (8.51) 63.8 (8.90)

Age (years)<65 818 (52.9) 1298 (51.3)65 to <75 554 (35.8) 927 (36.7)�75 175 (11.3) 303 (12.0)

SexMale 1206 (78.0) 1799 (71.2)Female 341 (22.0) 729 (28.8)

RaceCaucasian 1093 (70.7) 1894 (74.9)Black 5 (0.3) 42 (1.7)Asian 395 (25.5) 522 (20.7)Other 54 (3.5) 70 (2.8)

BMI (kg/m2)�30.0 1289 (83.3) 1968 (77.9)>30.0 258 (16.7) 560 (22.2)

Mean (SD) BMI (kg/m2) 25.3 (5.32) 26.2 (5.52)Severity of COPD

Moderate 539 (34.8) 1511 (59.8)Severe or very severe 1007 (65.1) 1017 (40.2)

ICS use 948 (61.3) 1152 (45.6)Smoking history

Ex-smoker 943 (61.0) 1460 (57.8)Current smoker 604 (39.0) 1068 (42.3)

Number of CV risk factorsa

0e2 1072 (69.3) 1122 (44.4)�3 475 (30.7) 1405 (55.6)

History of diabetes mellitus 151 (9.8) 254 (10.1)CCV conditionb 157 (10.2) 302 (12.0)Hyperlipidaemia 376 (24.3) 602 (23.8)Hypertension 700 (45.2) 1027 (40.6)

Values are n (%) unless otherwise indicated. BMI, body mass index; Cmonary disease; CV, cardiovascular; ICS, inhaled corticosteroid; Meddeviation.Missing values not included.a CV risk factors include: 1. presence of CCV history/condition at b

hyperlipidaemia at baseline; 4. presence of history of diabetes mellitu�65 years; 7. current smoker.b CCV condition determined based on the following pre-defined s

infarction (20,000,047, narrow), other ischaemic heart disease (20,0and haemorrhagic cerebrovascular conditions (20,000,063 and 20,000,

the following baseline factors: severity (moderate [FEV1

>50% and <80% of the predicted normal] and severe or verysevere based on GOLD 2008); ICS use (yes or no); cardio-vascular (CV) risk factors (0e2 or �3); sex (male or female);age group (<65 or �65 years) and body mass index (BMI;<30 or �30 kg/m2).

CV risk factors were determined based on the followingbaseline conditions: presence of CCV history/condition atbaseline based on the pre-defined Standardised MedicalQueries (SMQs) including MI, ischaemic heart disease, car-diac failure, ischaemic and haemorrhagic cerebrovascularconditions, and cardiac arrhythmias; presence of hyper-tension at baseline; presence of hyperlipidaemia at base-line; presence of history of diabetes mellitus (type II);

oss all studies (safety population).

0 mg Glycopyrronium 50 mg(n Z 2411)

Tiotropium 18 mg(n Z 2777)

Placebo(n Z 2141)

63.9 (8.82) 63.8 (8.33) 64.0 (8.93)

1235 (51.2) 1467 (52.8) 1082 (50.5)900 (37.3) 1025 (36.9) 810 (37.8)276 (11.4) 285 (10.3) 249 (11.6)

1821 (75.5) 1963 (70.7) 1551 (72.4)590 (24.5) 814 (29.3) 590 (27.6)

1725 (71.6) 2277 (82.0) 1456 (68.0)33 (1.4) 42 (1.5) 33 (1.5)

573 (23.8) 360 (13.0) 601 (28.1)80 (3.3) 98 (3.5) 51 (2.4)

1903 (78.9) 2143 (77.2) 1699 (79.4)507 (21.0) 633 (22.8) 441 (20.6)25.9 (5.82) 26.3 (5.67) 25.8 (5.69)

1030 (42.7) 1250 (45.0) 1288 (60.2)1380 (57.2) 1526 (55.0) 852 (39.8)1458 (60.5) 1587 (57.2) 919 (42.9)

1481 (61.4) 1639 (59.0) 1259 (58.8)930 (38.6) 1138 (41.0) 882 (41.2)

1282 (53.2) 1449 (52.2) 961 (44.9)1129 (46.8) 1328 (47.8) 1180 (55.1)283 (11.7) 282 (10.2) 220 (10.3)273 (11.3) 314 (11.3) 253 (11.8)593 (24.6) 725 (26.1) 514 (24.0)

1079 (44.8) 1221 (44.0) 877 (41.0)

CV, cardio- and cerebrovascular; COPD, chronic obstructive pul-DRA, Medical Dictionary for Regulatory Activities; SD, standard

aseline; 2. presence of hypertension at baseline; 3. presence ofs; 5. presence of obesity at baseline (i.e., BMI >30 kg/m2); 6. age

tandardised MedDRA queries (SMQs) search criteria: myocardial00,168, narrow), cardiac failure (20,000,004, narrow), ischaemic064, narrow), and cardiac arrhythmia terms (20,000,050, broad).

Figure 1 Forest plot of HRs for QVA149 (n [ 1547) and tiotropium (n [ 2777) versus placebo (n [ 2141) (a) Death and

serious CCV events BMI, body mass index; CCV, cardio- and cerebrovascular; CI, confidence interval; CV, cardiovascular; HR,hazard ratio; ICS, inhaled corticosteroid (b) MACE and pneumonia BMI, body mass index; CI, confidence interval; CV, cardiovas-cular; HR, hazard ratio; ICS, inhaled corticosteroid; MACE, major adverse cardiovascular event; MACE events included non-

Safety of QVA149 in a pooled analysis 1503

presence of obesity (i.e., BMI >30 kg/m2); age �65 years;or current smoker.

MACEs included non-adjudicated events from the inda-caterol and glycopyrronium studies and adjudicated eventsfrom the QVA149 programme (performed by an indepen-dent adjudication committee). Adjudicated MACEs, whenavailable or treatment-emergent SAEs as reported by in-vestigators or identified using pre-defined search criteria,within 30 days of the last dose, were included in theanalysis.

Statistical analysis

All analyses were based on the safety population defined asall randomised patients who received at least one dose ofthe study medication. The exposure data were collected onelectronic case report forms. Serious events of interestwere included only if the start date was within 30 days ofthe last study dose. As studies were completed at varioustimes, AEs were coded with different MedDRA versions. Forthe pooled analysis database, lower-level term codes weremapped to the latest MedDRA version (version 15.0). Datawere analysed using a Cox proportional hazard model, withtreatment, study, COPD severity, ICS use and the number ofCV risk factors as fixed covariates.

The subgroup analyses of the time to first event werealso performed using a Cox regression model with factorsfor treatment, study, baseline COPD severity (GOLD 2008),baseline ICS use, number of CV risk factors and the sub-group by treatment interaction term. For the subgroups ofsex, age and BMI, the model also included the factor forsubgroup.

The estimated HRs (active treatment versus placebo)along with its 95% confidence interval (CI) was calculatedfrom the model and was used to interpret the results. ForAF/F, only adjudicated data from QVA149 studiesdSHINE[12], SPARK [13], ENLIGHTEN [14], and ARISE [29] and gly-copyrronium studiesdGLOW1 [20], GLOW2 [21], andGLOW4 [28] were available for use (adjudicated by an in-dependent committee). Non-adjudicated data from inda-caterol studies were used in the analysis, as adjudicateddata were not available.

The model was based on the assumptions that all pa-tients share an equivalent baseline hazard function overtime, and the model accounts for the treatment and theimpact of different studies, COPD severity, ICS use statusand CV risk factors on the hazard.

Results

Patient characteristics

The analysis included safety data from 11,404 patients,with 1547 receiving QVA149; 2528 receiving indacaterol;

adjudicated events from the indacaterol and glycopyrronium studieexacerbation requiring hospitalisation and Atrial flutter/fibrillati

chronic obstructive pulmonary disease; CV, cardiovascular; HR, hadjudicated.

2411 receiving glycopyrronium; 2777 receiving tiotropium;and 2141 receiving placebo (Table 1). Baseline patient de-mographics and other clinical characteristics were compa-rable across all the treatment groups, with the notableexception that more patients in the QVA149 group versusthe other treatment groups had severe or very severe COPD(65.1% versus 39.8%e57.2%), due to the inclusion of datafrom the SPARK study. There was also a higher percentageof patients who used ICS at baseline (61.3% versus 42.9%e60.5%) in the QVA149 treatment arm while fewer patientson QVA149 had �2 CV risk factors at study entry (30.7%versus 46.8%e55.6%). The tiotropium group had moreCaucasian patients (82.0% versus 68.0%e74.9%) and lessAsian patients (13.0% versus 20.7%e28.1%) than the othertreatment groups (Table 2).

Safety

The mean exposure was highest in the QVA149 group(333.2 days) compared with its monocomponents indaca-terol (138.4 days) and glycopyrronium (283.1 days), tio-tropium (228.8 days) and placebo (178.6 days), due to theinclusion of the SPARK study that had a treatment durationof 64e76 weeks. A higher proportion of patients in theQVA149 group (83.6%) were treated for more than 6months compared with the other treatment groups(30.6%e75.7%), with 51.7% of the patients in the QVA149group being treated for more than 12 months(Supplementary Appendix 3 Table A2).

The HRs for QVA149 versus placebo and tiotropiumversus placebo showed no significant increase in the overallrisk for death, serious CCV events, MACEs, pneumonia orAF/F (Fig. 1). The overall risk of COPD exacerbation wassignificantly lower for QVA149 versus placebo (HR: 0.60;95% CI: 0.40e0.91), tiotropium versus placebo (HR: 0.49;95% CI: 0.32e0.74; Fig. 1c) and glycopyrronium versusplacebo (HR: 0.61; 95% CI: 0.41e0.89) while it was non-significant for indacaterol versus placebo (HR: 0.82; 95%CI: 0.47e1.42; Supplementary Appendix 4 Fig. A1c).

There was no significant increase in the risk for QVA149or tiotropium versus placebo for any of the analysed CVsafety endpoints for all subgroups related to severity ofCOPD (GOLD 2008), baseline ICS use, CV risk factors, sex,age or BMI (Fig. 1aec). Overall, the HR values for QVA149versus placebo were comparable to those observed withtiotropium versus placebo subgroups in a cross-comparisonmanner.

The risk of COPD exacerbation was significantly lower forQVA149 versus placebo in the subgroups (moderate COPD,baseline ICS use, female, age <65 years, BMI <30 kg/m2).Similar results were seen with those observed with tio-tropium versus placebo and glycopyrronium versus placebo(Fig. 1c and Supplementary Appendix 4 Fig. A1c). However,as mentioned earlier, tiotropium was open label in 5 of 6studies.

s and adjudicated events from the QVA149 programme (c) COPDon (AF/F) BMI, body mass index; CI, confidence interval; COPD,azard ratio; ICS, inhaled corticosteroid; All AF/F events were

1504 J.A. Wedzicha et al.

The risk of serious CCV events was significantly lower inpatients with baseline ICS use for QVA149 versus placebo(HR: 0.35; 95% CI: 0.14e0.86). The risk of serious CCVevents was also shown significantly lower in patients with�3 CV risk factors for tiotropium (HR: 0.45; 95% CI:0.22e0.96, Fig. 1a) and glycopyrronium versus placebo (HR:0.47; 95% CI: 0.23e0.99, Supplementary Appendix 4Fig. A1a), with similar risk with QVA149 versus placebo(HR: 0.67; 95% CI: 0.29e1.53, Fig. 1a).

The HR values in some of the subgroups investigated forall treatments versus placebo had wide 95% CI around theoutcomes (Fig. 1 and Supplementary Appendix 4 Fig. A1),which is likely due to low number of events and/or lownumber of patients with in these subgroups or variation inthe data. Most importantly, none of the differences in theHRs suggested significantly increased risk.

There was also no significant increase in the overall riskand in the subgroups for death, serious CCV events, MACEs,pneumonia, COPD exacerbations requiring hospitalisationor AF/F for indacaterol versus placebo and glycopyrroniumversus placebo except for MACE in patients with BMI�30 kg/m2 for glycopyrronium versus placebo where HRwas in favour of placebo (HR: 3.03; 95%CI: 1.02e8.97)(Supplementary Appendix 4 Fig. A1).

Discussion

The current pooled analysis included data from 11,404patients from 14 randomised clinical trials across multiplesafety databases. It was observed that the fixed-dosecombination QVA149 indirectly compared with its mono-components (indacaterol and glycopyrronium), tiotropiumand versus treatment with placebo did not increase therisk of death (all-cause mortality), serious cases of car-dio- and cerebrovascular (CCV) events, major adversecardiovascular events (MACEs), pneumonia, COPD exac-erbations requiring hospitalisation or atrial flutter/fibril-lation (AF/F) events overall and in a diverse range ofsubgroups.

The technique of pooled analysis of randomisedcontrolled trials, in which data is pooled from trials havingat least one intervention in common with another, has beenused recently for analysing data, as it allows comparison ofmultiple interventions that have not been evaluateddirectly against each other [30]. The analysis of pooled datamay also help to identify safety events that might occur atlow frequencies in the individual data sets.

The results observed with QVA149 versus placebo in thecurrent analysis did not show any increase in risk for any ofthe investigated safety points. The results were consistentwith the QVA149 cardiac safety data from Phase III studies,and confirmed that no significant cardiovascular safety ormortality concerns were observed in these studies withQVA149 [12,13,22].

CCV diseases have been reported to occur at increasedrates in patients with COPD, perhaps because of shared riskfactors such as smoking [31]. Rates of incident cardiovas-cular events are also increased in these patients [32].

In the current analysis, there was no significant increasein the risk for MACEs and other cardiovascular safety endpoints (serious CCV and AF/F) with QVA149 versus placebo

despite the inclusion of more patients with severe and verysevere COPD in the QVA149 group (65.1%). QVA149 also didnot increase the risk of death versus placebo, and the re-sults were comparable to those observed with tiotropiumversus placebo. The results of the pooled analysis suggeststhat the risk of cardiovascular events for QVA149 versusplacebo was comparable between older patients (age �65years) and younger patients (age < 65 years); patients withmoderate COPD and those with severe COPD; patients withhigh CV risk (CV risk factor � 3) and low risk (CV risk factor0e2); and in obese patients (BMI � 30 kg/m2) or otherwise(BMI < 30 kg/m2). These data provide some level ofassurance of the cardiovascular safety of QVA149 in higherrisk patients.

The results observed for tiotropium versus placebo in thecurrent analysis were consistent with those observed pre-viously in the UPLIFT study [7,33] and other reported meta-analyses of tiotropium 18 mg administered via the Handi-Haler� device for the risk of all-cause mortality [34,35] andmajor cardiovascular events versus placebo [36].

Clinically important subgroups of patients based on at-tributes such as age, sex, co-morbid condition or somecombination of these attributes can influence the level ofresponsiveness to the treatment of COPD. Hence, it isimportant to analyse the effect of the treatment in thesubgroups [37].There was no significant increase in risk withQVA149, indacaterol or glycopyrronium versus placebo formost of the subgroups related to severity, baseline ICS use,CV risk factors, sex, age or BMI for any of the investigatedsafety end points. The results were comparable to the tio-tropium versus placebo data in a cross-comparison manner.

The results for indacaterol versus placebo were alsoconsistent with those of other reported meta-analysis andpooled studies of indacaterol 150 mg in terms of mortalityand exacerbations, and did not show any significant in-crease in the overall risk for any of the investigated sub-groups [38,39].

In the current pooled analysis, QVA149 was not associ-ated with an increase in the overall risk of pneumoniacompared with placebo (HR: 1.10; 95% CI: 0.54e2.25) andin the subgroup related to baseline ICS use (HR: 1.05; 95%CI: 0.43e2.56). This provides additional evidence that thefixed-dose LABA/LAMA combinations, such as QVA149,could be a useful treatment option for symptomatic pa-tients with COPD, as it demonstrated similar safety profilecompared with its monocomponents.

Potential limitations of this pooled analysis include thefact that the trials included in the pooled analysis were notspecifically designed and powered to evaluate CV eventsand patients with high CV risks were excluded from thestudies. Real-life long-term safety studies with more pa-tients and longer-term follow-up are required to analysethe safety of drugs in such patients. Events were recordedonly during the duration of the studies (until 30 days of lasttreatment) and vital status follow up was not performed inmost of the studies. Further, tiotropium was open-label inall studies included in the pooled analysis, except for IN-TENSITY, where blinded tiotropium was used. Also, thenumber of GOLD IV (very severe) patients was limited in theanalysis (SPARK study only).

However, the strength of the current pooled analysis liesin the fact that the safety data have been pooled from 12

Safety of QVA149 in a pooled analysis 1505

published and 2 reported randomised clinical trials, withmore than 11,000 patients with COPD being included in theanalysis. The trials included in the analysis had similar in-clusion and exclusion criteria, with the exception of theSPARK study, and there was no evidence of clinical or sta-tistical heterogeneity between the trials. Different sub-groups based on the severity, baseline ICS use, CV riskfactors, sex, age and BMI were explored, which reflect theoverall risk factors in a general population of COPD. Thestatistical analysis was conducted using the Cox propor-tional hazard model, which incorporated the importantcovariates such as study, baseline COPD severity, baselineICS use, number of CV risk factors. Further, the adverseevent reporting was complete for all the trials despite somebeing of shorter duration and extended to 30 days after thelast study medication. Adjudicated data were used for theanalysis, except for atrial fibrillation data for indacaterolwhere the non-adjudicated data were included in theanalysis.

Conclusions

In this pooled analysis across multiple safety databases thatincluded data from 11,404 patients, the incidence ofdeaths, serious CCV events, MACEs, pneumonia, exacerba-tions requiring hospitalisation and atrial flutter/fibrillationwas comparable between QVA149 and placebo, with noincrease in the overall risk being observed for any of theinvestigated safety endpoints for any of the drugs versusplacebo.

There was no increase in infrequent AEs versus placebofor the fixed-dose combination QVA149, and additionally itssafety profile was comparable to its monocomponents(indacaterol and glycopyrronium) and tiotropium. Thesedata provide further reassurance that the use of dualbronchodilation may be of clinical benefit to patients withCOPD, without a safety penalty.

Role of funding source

AST, HC, PD, RF and DB, as employees of the sponsor,contributed to the design and preparation, analysis andinterpretation of the pooled analysis. JW, RD and RB madesubstantial contributions to the conception and design ofthe individual studies from which the data were pooled forthis analysis. All the authors had full access to raw data,contributed to the writing of each draft of the manuscriptand were responsible for the decision to submit forpublication. No restrictions were placed on the authorsregarding the statements made in the manuscript.

Statement of authorship

All authors had access to the data, were involved in inter-pretation and/or presentation of the data for this report,reviewed and revised the initial draft and subsequent ver-sions of the manuscript, had final responsibility for thedecision to submit for publication and approved the versionsubmitted.

Conflict of interest statements

In past three years, JAW has received fees for speakingand/or advisory boards from GlaxoSmithKline, Novartis,Bayer, Pfizer, Takeda, Boehringer Ingelheim, and Vectura aswell as travel reimbursements from Boehringer Ingelheim.She has received research grants from GlaxoSmithKline,Johnson and Johnson, Chiesi, Takeda and Novartis. She hasno stock holdings in pharmaceutical companies and hasnever received grant support from the tobacco industry.

In past three years, RD has received compensation forconsulting with Boehringer Ingelheim, Novartis, Vectura,and Norpharma. He has also undertaken research funded byBoehringer Ingelheim, Chiesi, GlaxoSmithKline, Novartisand ALK-Abello and has participated in educational activ-ities sponsored by AstraZeneca, Boehringer Ingelheim,GlaxoSmithKline, ALK-Abello, Novartis, and Almirall.

In past three years, RB has served on advisory boards andis a member of speakers’ bureau for AstraZeneca, Boeh-ringer Ingelheim, Chiesi Farmaceutici, Grifols, Glax-oSmithKline, Novartis, Roche, and Takeda.

AST is an employee of Novartis Pharma AG, Switzerland.HC, PO, RF and DB are employees of Novartis Pharma-

ceuticals Corporation, USA.

Acknowledgements

The authors were assisted in the preparation of thismanuscript by Kevin Roche and Saurabh Aggarwal atNovartis. Statistical support for the study was provided bySanthosh Lyathakula and Xiaohong Zhang (Novartis).Thisstudy was sponsored by Novartis Pharma AG.

Appendix A. Supplementary data

Supplementary data related to this article can be found athttp://dx.doi.org/10.1016/j.rmed.2014.07.011.

References

[1] Global Initiative for Chronic Obstructive Lung Disease (GOLD).Global strategy for the diagnosis, management, and preven-tion of chronic obstructive pulmonary disease; 2014. At:http://www.goldcopd.org [accessed 08.05.14].

[2] National Clinical Guideline Centre (NICE). Chronic obstructivepulmonary disease: Management of chronic obstructive pul-monary disease in adults in primary and secondary care; 2010.At: http://guidance.nice.org.uk/CG101/Guidance [accessed22.02.13].

[3] Rodrigo GJ, Castro-Rodriguez JA, Plaza V. Safety and efficacyof combined long-acting beta-agonists and inhaled cortico-steroids vs long-acting beta-agonists monotherapy for stableCOPD: a systematic review. Chest 2009;136(4):1029e38.

[4] Singh S, Loke YK, Furberg CD. Inhaled anticholinergics and riskof major adverse cardiovascular events in patients withchronic obstructive pulmonary disease: a systematic reviewand meta-analysis. JAMA 2008;300(12):1439e50.

[5] Nelson HS, Weiss ST, Bleecker ER, Yancey SW, Dorinsky PM.The Salmeterol Multicenter Asthma Research Trial: a com-parison of usual pharmacotherapy for asthma or usual

1506 J.A. Wedzicha et al.

pharmacotherapy plus salmeterol. Chest 2006;129(1):15e26.

[6] Calverley PM, Anderson JA, Celli B, Ferguson GT, Jenkins C,Jones PW, et al. Cardiovascular events in patients with COPD:TORCH study results. Thorax 2010;65(8):719e25.

[7] Tashkin DP, Celli B, Senn S, Burkhart D, Kesten S, Menjoge S,et al. A 4-year trial of tiotropium in chronic obstructive pul-monary disease. N Engl J Med 2008;359(15):1543e54.

[8] Vincken W, Aumann J, Chen H, Henley M, McBryan D, Goyal P.Efficacy and safety of co-administration of once-daily inda-caterol and glycopyrronium versus indacaterol alone in COPDpatients: the GLOW6 study. Int J Chron Obstruct Pulmon Dis2014;9:215e28.

[9] Mahler DA, D’Urzo A, Bateman ED, Ozkan SA, White T,Peckitt C, et al. Concurrent use of indacaterol plus tiotropiumin patients with COPD provides superior bronchodilationcompared with tiotropium alone: a randomised, double-blindcomparison. Thorax 2012;67(9):781e8.

[10] van Noord JA, Aumann JL, Janssens E, Smeets JJ, Zaagsma J,Mueller A, et al. Combining tiotropium and salmeterol inCOPD: effects on airflow obstruction and symptoms. RespirMed 2010;104(7):995e1004.

[11] Tashkin DP, Pearle J, Iezzoni D, Varghese ST. Formoterol andtiotropium compared with tiotropium alone for treatment ofCOPD. COPD 2009;6(1):17e25.

[12] Bateman ED, Ferguson GT, Barnes N, Gallagher N, Green Y,Henley M, et al. Dual bronchodilation with QVA149 versussingle bronchodilator therapy: the SHINE study. Eur Respir J2013;42:1484e94.

[13] Wedzicha JA, Decramer M, Ficker JH, Niewoehner DE,Sandstrom T, Taylor AF, et al. Analysis of chronic obstructivepulmonary disease exacerbations with the dual bronchodilatorQVA149 compared with glycopyrronium and tiotropium(SPARK): a randomised, double-blind, parallel-group study.Lancet Respir Med 2013;1(3):199e209.

[14] Dahl R, Chapman KR, Rudolf M, Mehta R, Kho P, Alagappan VK,et al. Safety and efficacy of dual bronchodilation with QVA149in COPD patients: the ENLIGHTEN study. Respir Med 2013;107(10):1558e67.

[15] Buhl R, Dunn LJ, Disdier C, Lassen C, Amos C, Henley M, et al.Blinded 12-week comparison of once-daily indacaterol andtiotropium in COPD. Eur Respir J 2011;38(4):797e803.

[16] Chapman KR, Rennard SI, Dogra A, Owen R, Lassen C,Kramer B. Long-term safety and efficacy of indacaterol, along-acting beta(2)-agonist, in subjects with COPD: a ran-domized, placebo-controlled study. Chest 2011;140(1):68e75.

[17] Kornmann O, Dahl R, Centanni S, Dogra A, Owen R, Lassen C,et al. Once-daily indacaterol versus twice-daily salmeterol forCOPD: a placebo-controlled comparison. Eur Respir J 2011;37(2):273e9.

[18] Feldman G, Siler T, Prasad N, Jack D, Piggott S, Owen R, et al.Efficacy and safety of indacaterol 150 microg once-daily inCOPD: a double-blind, randomised, 12-week study. BMC PulmMed 2010;10:11.

[19] Barnes PJ, Pocock SJ, Magnussen H, Iqbal A, Kramer B,Higgins M, et al. Integrating indacaterol dose selection in aclinical study in COPD using an adaptive seamless design. PulmPharmacol Ther 2010;23(3):165e71.

[20] D’Urzo A, Ferguson GT, van Noord JA, Hirata K, Martin C,Horton R, et al. Efficacy and safety of once-daily NVA237 inpatients with moderate-to-severe COPD: the GLOW1 trial.Respir Res 2011;12:156.

[21] Kerwin E, Hebert J, Gallagher N, Martin C, Overend T,Alagappan VK, et al. Efficacy and safety of NVA237 versus

placebo and tiotropium in patients with COPD: the GLOW2study. Eur Respir J 2012;40(5):1106e14.

[22] Vogelmeier CF, Bateman ED, Pallante J, Alagappan VK,D’Andrea P, Chen H, et al. Efficacy and safety of once-dailyQVA149 compared with twice-daily salmeterol-fluticasone inpatients with chronic obstructive pulmonary disease (ILLUMI-NATE): a randomised, double-blind, parallel group study.Lancet Respir Med 2012;1(1):51e60.

[23] Dahl R, Jadayel D, Alagappan VK, Chen H, Banerji D. Efficacyand safety of QVA149 compared to the concurrent adminis-tration of its monocomponents indacaterol and glyco-pyrronium: the BEACON study. Int J Chron Obstruct Pulmon Dis2013;8:501e8.

[24] European Public Assessment Report (EPAR) Ultibro Breezhaler.At :http://www.ema.europa.eu/ema/index.jsp?curlZ/pages/medicines/human/medicines/002679/human_med_001691.jsp. [accessed on 26.02.14].

[25] Donohue JF, Fogarty C, Lotvall J, Mahler DA, Worth H,Yorgancioglu A, et al. Once-daily bronchodilators for chronicobstructive pulmonary disease: indacaterol versus tiotropium.Am J Respir Crit Care Med 2010;182(2):155e62.

[26] Kinoshita M, Lee SH, Hang LW, Ichinose M, Hosoe M, Okino N,et al. Efficacy and safety of indacaterol 150 and 300 mg inchronic obstructive pulmonary disease patients from six Asianareas including Japan: a 12-week, placebo-controlled study.Respirology 2012;17(2):379e89.

[27] Yao W, Wang C, Zhong N, et al. Effect of once-daily indaca-terol in a predominantly Chinese population with chronicobstructive pulmonary disease: a 26-week Asia-Pacific study.Respirology 2014. http://dx.doi.org/10.1111/resp.12211[Epub ahead of print].

[28] Sekiya M, Kawayama T, Fukuchi Y, et al. Safety and efficacy ofNVA237 once daily in Japanese patients: the GLOW4 trial.Presented at European Respiratory Society Congress; 2012.Abstract number 853970.

[29] Asai K, Minakata Y, Hirata K, et al. QVA149 once-daily is safeand well tolerated and improves lung function and healthstatus in Japanese patients with COPD: the ARISE study. Pre-sented at European Respiratory Society Congress; 2013. Posternumber P3392.

[30] Caldwell DM, Ades AE, Higgins JP. Simultaneous comparison ofmultiple treatments: combining direct and indirect evidence.BMJ 2005;331(7521):897e900.

[31] Mullerova H, Agusti A, Erqou S, Mapel DW. Cardiovascularcomorbidity in chronic obstructive pulmonary disease: sys-tematic literature review. Chest 2013;144(4):1163e78.

[32] Schneider C, Bothner U, Jick SS, Meier CR. Chronic obstructivepulmonary disease and the risk of cardiovascular diseases. EurJ Epidemiol 2010;25(4):253e60.

[33] Tashkin DP. Impact of tiotropium on the course of moderate-to-very severe chronic obstructive pulmonary disease: theUPLIFT trial. Expert Rev Respir Med 2010;4(3):279e89.

[34] Celli B, Decramer M, Leimer I, Vogel U, Kesten S, Tashkin DP.Cardiovascular safety of tiotropium in patients with COPD.Chest 2010;137(1):20e30.

[35] Mathioudakis AG, Kanavidis P, Chatzimavridou-Grigoriadou V,Gialmanidis IP, Amanetopoulou SG, Christopoulou E, et al.Tiotropium HandiHaler improves the survival of patients withCOPD: a systematic review and meta-analysis. J Aerosol MedPulm Drug Deliv 2014;27(1):43e50.

[36] Rodrigo GJ, Castro-Rodriguez JA, Nannini LJ, Plaza MV,Schiavi EA. Tiotropium and risk for fatal and nonfatal cardio-vascular events in patients with chronic obstructive pulmo-nary disease: systematic review with meta-analysis. RespirMed 2009;103(10):1421e9.

Safety of QVA149 in a pooled analysis 1507

[37] Cazzola M, Segreti A, Rogliani P. Comparative effectiveness ofdrugs for chronic obstructive pulmonary disease. Drugs Today2012;48(12):785e94.

[38] Worth H, Chung KF, Felser JM, Hu H, Rueegg P. Cardio- andcerebrovascular safety of indacaterol vs formoterol,

salmeterol, tiotropium and placebo in COPD. Respir Med 2011;105(4):571e9.

[39] Chung VC, Ma PH, Hui DS, Tam WW, Tang JL. Indacaterol forchronic obstructive pulmonary disease: systematic review andmeta-analysis. PLoS One 2013;8(8):e70784.