ORIGINAL RESEARCH

Efficacy of Budesonide/Glycopyrronium/FormoterolFumarate Metered Dose Inhaler (BGF MDI) VersusOther Inhaled Corticosteroid/Long-Acting MuscarinicAntagonist/Long-Acting b2-Agonist (ICS/LAMA/LABA)Triple Combinations in COPD: A Systematic LiteratureReview and Network Meta-analysis

Gary T. Ferguson . Patrick Darken . Shaila Ballal . Mohd Kashif Siddiqui .

Barinder Singh . Sumeet Attri . Ulf Holmgren . Enrico de Nigris

Received: February 10, 2020 / Published online: April 25, 2020� The Author(s) 2020

ABSTRACT

Introduction: Triple inhaled corticosteroid/long-actingmuscarinic antagonist/long-acting b2-agonist (ICS/LAMA/LABA) combination therapy isrecommended for patients with chronic obstruc-tive pulmonary disease (COPD) who experiencefurther exacerbations/symptoms on dual LAMA/LABAor ICS/LABA therapy. The relative efficacy ofbudesonide/glycopyrronium/formoterol fuma-rate metered dose inhaler 320/18/9.6 lg (BGF

MDI) in COPD was compared with other ICS/LAMA/LABA fixed-dose and open combinationtherapies in a network meta-analysis (NMA).Methods: A systematic literature review wasconducted to identify randomized controlledtrials of at least 10-week duration, including atleast one fixed-dose or open combination tripletherapy arm, in patients with moderate to verysevere COPD. Studies were assessed formethodological quality and risk of bias. A three-level hierarchical Bayesian NMA model wasused to determine the exacerbation rate perpatient per year as well as the following out-comes at week 24: changes from baseline in pre-dose trough forced expiratory volume in 1 s(FEV1), post-dose peak FEV1, and St. George’sRespiratory Questionnaire (SGRQ) total score;proportion of SGRQ responders; and TransitionDyspnea Index focal score. Change from base-line in rescue medication use over weeks 12–24was also analyzed. Meta-regression and sensi-tivity analyses were used to assess heterogeneityacross studies.Results: Eighteen studies (n = 29,232 patients)contributed to the NMA. ICS/LABA dual com-binations were combined as a single treatmentgroup to create a connected network. Across alloutcomes, there were no statistically significantdifferences between BGF MDI and other tripleICS/LAMA/LABA fixed-dose (fluticasone furo-ate/umeclidinium/vilanterol and beclometha-sone dipropionate/glycopyrronium/formoterol

Enhanced Digital Features To view enhanced digitalfeatures for this article go to https://doi.org/10.6084/m9.figshare.12005604.

Electronic supplementary material The onlineversion of this article (https://doi.org/10.1007/s12325-020-01311-3) contains supplementary material, which isavailable to authorized users.

G. T. Ferguson (&)Pulmonary Research Institute of SoutheastMichigan, Farmington Hills, MI, USAe-mail: garytferguson@msn.com

P. Darken � S. BallalAstraZeneca, Morristown, NJ, USA

M. K. Siddiqui � B. Singh � S. AttriParexel International, Punjab, India

U. HolmgrenAstraZeneca, Gothenburg, Sweden

E. de NigrisAstraZeneca, Cambridge, UK

Adv Ther (2020) 37:2956–2975

https://doi.org/10.1007/s12325-020-01311-3

fumarate) and open combinations with dataavailable within the network. Results fromsensitivity analyses and meta-regression wereconsistent with the base-case scenario.Conclusion: This NMA suggested that BGF MDIhas comparable efficacy to other ICS/LAMA/LABA fixed-dose and open triple combinationtherapies in reducing exacerbations andimproving lung function and symptoms inpatients with moderate to very severe COPD.Further research is warranted as additional evi-dence regarding triple therapies, especiallyfixed-dose combinations, becomes available.

Keywords: Chronic obstructive pulmonarydisease; Exacerbations; Inhaled corticosteroid;Long-acting muscarinic antagonist; Long-actingb2-agonist; Lung function; Network meta-analysis; Patient-reported outcomes; Tripletherapy

Key Summary Points

Why carry out this study?

Budesonide/glycopyrronium/formoterolfumarate metered dose inhaler (BGF MDI)is a triple fixed-dose combination therapyfor chronic obstructive pulmonary disease(COPD).

Given the relatively recent introduction offixed-dose triple therapies for COPD, thereare no head-to-head randomizedcontrolled trials of their relative efficacy.

We performed a network meta-analysis tocompare the relative efficacy of BGF MDIversus other triple therapies (in fixed-doseor open combination) in patients withmoderate to very severe COPD.

What was learned from the study?

On the basis of evidence from 18 studies,BGF MDI was found to have similarefficacy to other fixed-dose and opentriple combination therapies in reducingexacerbations and improving lungfunction and symptoms in patients withmoderate to very severe COPD.

The results of this network meta-analysisprovide important context for healthcareproviders and payers in evaluating thecurrent evidence regarding triple therapiesin COPD.

INTRODUCTION

Chronic obstructive pulmonary disease (COPD)is a progressive disease that causes symptomsincluding dyspnea, sputum production, andchronic cough, and can be associated with sig-nificant comorbidities [1]. COPD is associatedwith significant morbidity and mortality: it wasreported to be the third leading cause of deathin 2016, causing an estimated 3.0 milliondeaths globally [2].

A range of pharmacological treatmentoptions exists for COPD, with the key treatmentgoals being to reduce symptoms, decrease therisk of exacerbations, and minimize the impactof exacerbations if they occur [1]. Patients witha high symptom burden and a history of exac-erbations may be treated with dual bron-chodilator therapy [long-acting muscarinicantagonist (LAMA)/long-acting b2-agonist(LABA)], or for patients who also have elevatedeosinophil levels, inhaled corticosteroid (ICS)/LABA therapy is a recommended initial treat-ment option [1]. Escalation to ICS/LAMA/LABAtriple therapy is recommended for patients whocontinue to experience symptoms such asbreathlessness or difficulty with physical activ-ity while on ICS/LABA treatment, or for patientson LAMA/LABA therapy who continue to haveexacerbations and who have eosinophil countsof at least 100 cells/lL [1]. Notably, the GlobalInitiative for Chronic Obstructive Lung Disease(GOLD) 2019 report includes, for the first time,a specific treatment pathway for the manage-ment of exacerbations, indicating the clinicalimportance of preventing their occurrence [3].

Combination therapies may be delivered viaseparate inhalers (in ‘‘open’’ combination) [4–6]or within a single inhaler (in fixed-dose com-bination) [7–10]. Given the relatively recent

Adv Ther (2020) 37:2956–2975 2957

introduction of fixed-dose triple therapies, thereare no head-to-head randomized controlledtrials (RCTs) of their relative efficacy in COPD.In the absence of head-to-head data, networkmeta-analysis (NMA) can be used to comparemultiple interventions by combining direct andindirect evidence, adjusting with the use ofcommon comparators [11]. NMA techniqueshave been applied to compare the efficacy oftriple therapy as a class with LAMA/LABA dualtherapy or bronchodilator monotherapy [12],and two pairwise meta-analyses have providedcomparisons of triple therapies with ICS/LABA[13, 14] or LAMA/LABA [14].

Budesonide/glycopyrronium/formoterolfumarate metered dose inhaler (BGF MDI), for-mulated with co-suspension delivery technol-ogy (AEROSPHERE�, AstraZeneca), is a tripleICS/LAMA/LABA fixed-dose combination thathas recently been approved in Japan and Chinafor the treatment of COPD [15, 16]. In thephase III KRONOS study, BGF MDI showedbenefits in improving lung function andsymptoms and reducing COPD exacerbationsversus dual LAMA/LABA and ICS/LABA thera-pies [10]. We performed a systematic literaturereview (SLR) and NMA to compare the relativeclinical efficacy of BGF MDI versus other tripleICS/LAMA/LABA therapies (in fixed-dose oropen combination) in patients with moderateto very severe COPD. To our knowledge, this isthe first NMA to assess the relative efficacy oftriple therapies in COPD. This article is based onpreviously conducted studies and does notcontain any studies with human participants oranimals performed by any of the authors.

METHODS

Systematic Literature Review

An SLR was conducted to identify evidence onthe efficacy of triple ICS/LAMA/LABA fixed-doseor open combination therapies in patients withmoderate to very severe COPD, in accordancewith the Preferred Reporting Items for System-atic Reviews and Meta-Analyses (PRISMA)guidelines. Searches of Embase�, MEDLINE�,MEDLINE� In-Process, and the Cochrane

Central Register of Controlled Trials (CENTRAL)were run from database inception to October2019. Only articles published in English wereincluded. The search strategies are shown inTable S1. Abstracts from the American ThoracicSociety, European Respiratory Society, andAmerican College of Chest Physicians confer-ence proceedings were hand-searched fromSeptember 2015 to August 2019 to retrievestudies that had not yet been published in full-text articles or abstracts reporting supplemen-tary results of previously published studies.Additionally, the following trial registries weresearched to capture unpublished clinical trials:ClinicalTrials.gov of the US National Institute ofHealth and the World Health OrganizationInternational Clinical Trials Registry Platform.

To be included in the SLR, studies had tomeet pre-defined eligibility criteria: the primarycriteria were RCTs with a duration of at least10 weeks, assessing patients of at least 40 yearsof age with moderate to very severe COPD,published in English, and including at least onetreatment arm with fixed-dose or open combi-nation triple therapy (Table 1). Titles andabstracts of publications identified in the searchwere screened, full-text copies of articles judgedto be potentially relevant reviewed, and data foreligible studies extracted using a pre-definedextraction grid, which included details on trialdesign, inclusion criteria, study populationcharacteristics, interventions, outcome mea-sures, and length of follow-up. Screening,review, and data extraction were conducted bytwo independent reviewers, with resultschecked and reconciled by a third independentreviewer. Where a single study was described bymore than one publication, the data werecompiled into a single entry in the dataextraction sheet to avoid duplication, with allpublications referenced.

The methodological quality of includedstudies was assessed using the concise criticalappraisal checklists provided by the NationalInstitute for Health and Care Excellence (NICE)in the Single Technology Appraisal user guide[17]. The risk of bias was assessed with respect tothe method of randomization and allocationconcealment, comparability of baseline charac-teristics, blinding, the balance of withdrawals

2958 Adv Ther (2020) 37:2956–2975

between groups, outcomes reporting, and sta-tistical analysis.

Network Meta-analysis

The NMA followed the recommended bestpractice of the NICE Decision Support Unit forevidence synthesis [18, 19]. Relevant resultswere combined using a three-level hierarchicalBayesian NMA model, which assumesexchangeability between treatments within thesame class, i.e., that all studies measure the

same underlying relative treatment effects[20, 21]. This model accounted for theexchangeability between interventions of thesame class (i.e., LAMA monotherapy, ICS/LABAdual combinations, ICS/LAMA/LABA triplecombinations) by assuming that underlyingtreatment effects within each class followed anormal distribution with class-specific meanand variance. Thus, estimates of treatmenteffects and their uncertainty are affected byboth the evidence propagated through thenetwork, as well as the borrowed strengthbetween treatments in the same class. The syn-thesis was conducted from a Bayesian perspec-tive, using WinBUGS (a Markov chain MonteCarlo simulation-based software for Bayesianinference) version 1.4.3. The NMA WinBUGScode, developed initially by Dias et al. [22], wasadapted to incorporate a three-level hierarchicalclass-effect model [20, 21, 23]. Results weregenerated using both random- and fixed-effectsmodels and compared for goodness-of-fit to thedata, calculated as the overall mean residualdeviance. The goodness of fit was assessed usingthe Deviance Information Criterion (DIC); themodel with the lowest DIC was considered themodel with the best fit to the data. If DIC andresidual deviance were comparable betweenmodels, a random-effects model was preferredas it takes into account additional heterogeneityin the network, but a fixed-effects model wasused when the number of contributing studieswas five or fewer. Results are presented as pos-terior median effect estimates with 95% crediblelimits: rate ratios for counts outcomes modeledwith a Poisson model (exacerbations), meandifferences for continuous outcomes [forcedexpiratory volume in 1 s (FEV1), St. George’sRespiratory Questionnaire (SGRQ), TransitionDyspnea Index (TDI), and rescue medicationuse], and odds ratios for binary outcomes (SGRQresponders). No type I error control was per-formed, as is common with other studies of thisnature using a Bayesian framework [24]. Forcontinuous outcomes, results were considerednon-significant if the credible interval (CrI)contained the null value; for Poisson/binomialoutcomes, results were considered non-signifi-cant if the CrI contained 1. Inconsistenciesbetween direct and indirect estimates were

Table 1 Population, interventions, comparators, out-comes, and study design criteria for inclusion in the net-work meta-analysis

Population Adult patients (C 40 years of age) of any

gender or race with moderate to very

severe COPD (predicted FEV1 B 80%)

Interventions Triple therapies (ICS ? LAMA ? LABA,

both fixed-dose and open combinations)

Comparators Any included intervention

Dual therapies (ICS ? LABA or

LAMA ? LABA both fixed-dose and

open combinations)

Monotherapies (ICS/LAMA/LABA)

Placebo

Outcomes Efficacy outcomes

Exacerbations (severe only, moderate to

severe)

Lung function (peak FEV1, trough FEV1)

SGRQ total score and SGRQ responders

TDI focal score

Use of rescue medication

Study designs Randomized controlled trials of

C 10 weeks duration

COPD chronic obstructive pulmonary disease,FEV1 forced expiratory volume in 1 s, ICS inhaledcorticosteroid, LABA long-acting b2-agonist, LAMA long-acting muscarinic antagonist, SGRQ St. George’s Respira-tory Questionnaire, TDI Transition Dyspnea Index

Adv Ther (2020) 37:2956–2975 2959

checked, where appropriate, for outcomeswhere both direct and indirect data were avail-able for the comparison of interest (furtherdetails are provided in the Supplementarymethods).

A separate NMA was performed for each ofthe outcomes of interest. Exacerbations wereassessed as the mean rate of exacerbations perpatient per year (moderate/severe and severeexacerbations, analyzed as relative ratios forBGF MDI vs comparators). Lung function end-points were the change from baseline in troughFEV1 and the change from baseline in peakFEV1, both at week 24. The following patient-reported outcomes were assessed: change frombaseline in SGRQ total score at 24 weeks, pro-portion of SGRQ responders (patients whoexperienced an improvement that met orexceeded the minimum clinically importantdifference of at least 4 units [25]) at 24 weeks,TDI focal score at 24 weeks, and change frombaseline in rescue medication use over 12–24weeks. Analysis of TDI responders could not beperformed because of variation between studiesin the reporting of this outcome.

Studies reporting data between 22 and26 weeks were considered for inclusion in the24-week analyses. For the exacerbation out-comes, there was no limit applied on the basisof maximum trial duration, as the treatmentswere compared using rates (events per patient-year). Networks were presented graphically witha ‘‘node’’ representing each intervention and an‘‘edge’’ representing the comparison betweenthem. Each node was weighted according to thenumber of patients receiving that intervention,and each edge was weighted according to thenumber of studies included for the comparison.

Several sensitivity analyses were undertaken.The base-case scenario included only double-blind studies; therefore, a sensitivity analysiswas performed, including both double-blindand open-label studies in one network. A sen-sitivity analysis was also performed, includingonly studies in which the majority of thepatient population was symptomatic (defined asa COPD assessment test (CAT) score of at least10 or a modified Medical Research Council(mMRC) dyspnea scale score of at least 2).

Additionally, a meta-regression analysis wasconducted for the efficacy outcomes with atleast 10 studies to account for differences inselected baseline patient characteristics thatcould be acting as the key effect modifiers.

RESULTS

Study Selection

The SLR process is shown in Fig. 1. Initial data-base searches identified 15,542 publications,with 2742 removed owing to duplication acrossdatabases searched. Initial screening of the12,800 remaining records (based on titles andabstracts) reduced eligible publications to 1589,which were subject to full-text review. A further32 citations were identified from conferenceproceedings and bibliographies of identifiedpublications, including a clinical study reportfor an RCT with BGF MDI. Following full-textreview, a total of 23 studies from 165 publica-tions met the inclusion criteria of the SLR. Fivestudies did not report any efficacy outcome ofinterest at specified time points, thus leaving18 studies [4–10, 26–34] that contributed to theNMA (Table 2; Table S2).

Study Characteristics

All 18 of the studies included were multicenter,the majority were phase III, two were open-la-bel, and the remainder were double-blind(Table 2). All studies included in the NMA wereadjudged to pose a low risk of bias with respectto randomization and allocation concealment,baseline characteristics, the balance of with-drawals between groups, and statistical analysis(Table S3). High risk of bias in blinding wasassociated with two open-label studies [4, 5],and one study was considered to have a highrisk of bias with respect to outcome selectionand reporting [31].

Analysis Assumptions

When all treatments reported in the includedstudies were considered, there was no

2960 Adv Ther (2020) 37:2956–2975

interlinked network allowing comparison ofBGF MDI with other triple therapies in thebase-case analysis, which included only double-blind studies (n = 16; Fig. 2a). Therefore, tocreate an interlinked network, all ICS/LABAdual combinations were considered as a singletreatment group (Fig. 2b). The networks ofevidence for each outcome varied according tothe endpoints available from each study(Figs. S1–S3).

Patient Baseline Clinical Characteristics

A total of 29,232 evaluable patients contributedto the NMA. Patient characteristics were gener-ally similar in terms of age, gender, body mass

index, and smoking status, but differences werenoted in disease duration, race, symptom bur-den, COPD severity, and exacerbation history(Table 3). These potential differences in keyeffect modifiers could not be adjusted in a meta-regression model because of the limited numberof studies contributing to the NMA, except formoderate/severe exacerbations and severeexacerbations. For these outcomes, meta-re-gression was feasible for prior exacerbation his-tory, smoking status, BMI, and disease severity,as classified by GOLD (III or IV). Thirteen of the18 studies included enrolled only symptomaticpatients, while the remainder did not describeany inclusion criteria regarding symptomburden.

Fig. 1 PRISMA flowchart. CSR clinical study report, NMA network meta-analysis

Adv Ther (2020) 37:2956–2975 2961

Table2

Studycharacteristics

Stud

yTreatment

Patient

number

Stud

y

duration

(weeks)

Primaryendp

oint

Pub

lication

type

Stud

y

phase

Blin

ding

Stud

ysetting

Aaron

2007

[26]

FP/SAL(FP500lg

?SA

L50

lg)

BID

?TIO

18lg

OD

145

52Proportion

ofpatientswho

experiencedaCOPD

exacerbation

that

required

treatm

entwithsystem

icsteroids

orantibioticswithin52

weeks

of

rand

omization

Journal

article

Unclear

Double-

blind

Multicenter

TIO

18lgOD

?SA

L50

lgBID

148

TIO

18lgOD

156

Bremner2018

a

[27]

FF/U

MEC/V

I(FF100lg

?UMEC

62.5lg

?VI25

lg)OD

527

24Changefrom

baselin

ein

trough

FEV1at

week24

Journal

article

III

Double-

blind

Multicenter

FF/V

I(FF100lg?

VI25

lg)

OD

?UMEC

62.5lg

OD

528

FULFILa[28]

FF/U

MEC/V

I(FF100lg

?UMEC

62.5lg

?VI25

lg)OD

911

24Changefrom

baselin

ein

trough

FEV1at

week24,changefrom

baselin

ein

SGRQ

totalscoreat

week24

Journal

article/

CSR

III

Double-

blind

Multicenter

international

BUD/FOR(BUD

320lg?

FOR9lg)

BID

899

Hanania2012

a

[29]

FP/SAL(FP250lg

?SA

L50

lg)

BID

?TIO

18lg

OD

173

24Changefrom

baselin

ein

morning

pre-dose

FEV1at

week24

Journal

article

IVDouble-

blind

Multicenter

TIO

18lgOD

169

IMPA

CTa[8]

FF/U

MEC/V

I(FF100lg

?UMEC

62.5lg

?VI25

lg)OD

4155

52Ann

ualrate

ofmoderateor

severe

COPD

exacerbationsover

52weeks

Journal

article/

CSR

III

Double-

blind

Multicenter

international

FF/V

I(FF100lg?

VI25

lg)

OD

4139

UMEC/V

I(U

MEC

62.5lg

?VI25

lg)

OD

2073

Jung

2012

[4]

FP/SAL(FP250lg

?SA

L50

lg)

BID

?TIO

18lg

OD

237

24Changefrom

baselin

ein

prebronchodilatorFE

V1at

week24

Journal

article

IVOpen-

label

Multicenter

TIO

18lgOD

242

2962 Adv Ther (2020) 37:2956–2975

Table2

continued

Stud

yTreatment

Patient

number

Stud

y

duration

(weeks)

Primaryendp

oint

Pub

lication

type

Stud

y

phase

Blin

ding

Stud

ysetting

KRONOSa

[10]

BUD/G

LY/FOR(BGFMDI;BUD

320lg?

GLY14.4lg

?FO

R10

lg)

BID

640

24Japan/China

Changefrom

baselin

ein

morning

pre-dose

trough

FEV1over

weeks

12–2

4

EU/Canada

Changefrom

baselin

ein

morning

pre-dose

trough

FEV1over

24weeks,F

EV1AUC0–

4over

24weeks

USA

FEV1AUC0–

4at

week24,change

from

baselin

ein

morning

pre-

dose

trough

FEV1at

week24

Journal

article/

CSR

III

Double-

blind

Multicenter

international

GLY/FOR(G

FFMDI;GLY

14.4lg

?FO

R10

lg)BID

627

BUD/FOR(BFF

MDI;BUD

320lg?

FOR10

lg)BID

316

BUD/FOR(BUD/FORDPI;BUD

320lg?

FOR9lg)BID

319

Lee

2016

[5]

BUD/FOR(BUD

320lg?

FOR9lg)

BID

?TIO

18lg

OD

287

12Changefrom

baselin

ein

pre-

bronchodilatorFE

V1atweeks

1,

6,and12

Journal

article

IVOpen-

label

Multicenter

international

TIO

18lgOD

291

Study

AC4116135a

[33]

FP/SAL(FP250lg

?SA

L50

lg)

BID

?UMEC

125lg

OD

205

12Changefrom

baselin

ein

trough

FEV1at

day85

(week12)

Journal

article

III

Double-

blind

Multicenter

international

FP/SAL(FP250lg

?SA

L50

lg)

BID

?UMEC

62.5lg

OD

204

FP/SAL(FP250lg

?SA

L50

lg)BID

205

Adv Ther (2020) 37:2956–2975 2963

Table2

continued

Stud

yTreatment

Patient

number

Stud

y

duration

(weeks)

Primaryendp

oint

Pub

lication

type

Stud

y

phase

Blin

ding

Stud

ysetting

Study

AC4116136a

[33]

FP/SAL(FP250lg

?SA

L50

lg)

BID

?UMEC

125lg

OD

202

12Changefrom

baselin

ein

trough

FEV1at

day85

(week12)

Journal

article

III

Double-

blind

Multicenter

international

FP/SAL(FP250lg

?SA

L50

lg)

BID

?UMEC

62.5lg

OD

203

FP/SAL(FP250lg

?SA

L50

lg)BID

201

Study200109

a

[6]

FF/V

I(FF100lg?

VI25

lg)

OD

?UMEC

125lg

OD

207

12Changefrom

baselin

ein

trough

FEV1at

day85

(week12)

Journal

article

III

Double-

blind

Multicenter

international

FF/V

I(FF100lg?

VI25

lg)

OD

?UMEC

62.5lg

OD

206

FF/V

I(FF100lg?

VI25

lg)

OD

206

Study200110

a

[6]

FF/V

I(FF100lg?

VI25

lg)

OD

?UMEC

125lg

OD

207

12Changefrom

baselin

ein

trough

FEV1at

day85

(week12)

Journal

article

III

Double-

blind

Multicenter

international

FF/V

I(FF100lg?

VI25

lg)

OD

?UMEC

62.5lg

OD

206

FF/V

I(FF100lg?

VI25

lg)

OD

206

SUNSE

Ta[30]

FP/SAL(FP500lg

?SA

L50

lg)

BID

?TIO

18lg

OD

526

26Changefrom

baselin

ein

post-dose

trough

FEV1at

week26

Journal

article

IVDouble-

blind

Multicenter

international

GLY/IND

(GLY43

lg?

IND

85lg)OD

527

TRIBUTEa[9]

BDP/GLY/FOR(BDP174lg

?GLY

18lg

?FO

R10

lg)BID

764

52Ann

ualrate

ofmoderate/severe

COPD

exacerbationsover

52weeks

Journal

article

III

Double-

blind

Multicenter

international

GLY/IND

(GLY43

lg?

IND

85lg)OD

768

2964 Adv Ther (2020) 37:2956–2975

Table2

continued

Stud

yTreatment

Patient

number

Stud

y

duration

(weeks)

Primaryendp

oint

Pub

lication

type

Stud

y

phase

Blin

ding

Stud

ysetting

TRILOGYa[7]

BDP/GLY/FOR(BDP200lg

?GLY

25lg

?FO

R12

lg)BID

687

52Changefrom

baselin

ein

pre-dose

(morning)FE

V1,change

from

baselin

ein

2-hpost-doseFE

V1,

andTDIfocalscoreat

week26

Journal

article

III

Double-

blind

Multicenter

international

BDP/FO

R(BDP200lg?

FOR12

lg)

BID

681

TRIN

ITYa[34]

BDP/GLY/FOR(BDP200lg

?GLY

25lg

?FO

R12

lg)BID

1078

52Moderate/severe

COPD

exacerbation

ratefor52

weeks

of

treatm

ent

Journal

article

III

Double-

blind

Multicenter

international

BDP/FO

R(BDP200lg?

FOR12

lg)

BID

?TIO

18lg

OD

538

TIO

18lgOD

1075

Welte

2009

[31]

BUD/FOR(BUD

320lg?

FOR9lg)

BID

?TIO

18lg

OD

329

12Changein

pre-dose

FEV1from

rand

omization(week0)

tothe

fulltreatm

entperiod

(mean

FEV1at

weeks

1,6,

and12

of

treatm

ent)

Journal

article

IVDouble-

blind

Multicenter

international

TIO

18lgOD

331

WISDOM

[32]

FP500lgBID

?SA

L50

lgBID

?TIO

18lg

OD

1244

52Tim

eto

thefirstmoderate/severe

COPD

exacerbation

Journal

article

IVDouble-

blind

Multicenter

international

FP500lgBID

(reducingb)?

SAL50

lg

BID

?TIO

18lg

OD

1244

Doses

representthetotalam

ount

peradministereddose,which

may

bethesum

oftwoactuations.Fixed-dose

combinationsarerepresentedwith‘‘/’’betweencomponents;open

combinationsarerepresentedwith‘‘?

’’betweencomponents

AUC0–

4area

underthecurvefrom

0to

4h,

BDPbeclom

ethasone

dipropionate,B

FFbudesonide/formoterol

fumarate,BGFbudesonide/glycopyrronium/formoterol

fumarate,BID

twicedaily,B

UDbudesonide,C

ATCOPD

Assessm

entTest,COPD

chronicobstructivepulmonarydisease,CSR

clinicalstudyreport,D

PIdrypowderinhaler,FE

V1forced

expiratory

volumein

1s,FF

fluticasone

furoate,FP

fluticasone

propionate,F

ORform

oterol,G

LYglycopyrronium

,IND

indacaterol,ITTintent-to-treat,MDImetered

dose

inhaler,mMRC

modified

MedicalResearchCouncildyspneascale,OD

once

daily,SALsalmeterol,SGRQSt.G

eorge’sRespiratory

Questionn

aire,T

DITransitionDyspn

eaIndex,TIO

tiotropium

,

UMECum

eclidinium,V

Ivilanterol

aAmajorityof

thepatientpopulation

was

classifiedas

symptom

aticat

baselin

e(based

onCAT

C10

ormMRC

C2)

bThe

BID

dose

ofFP

was

reducedevery6weeks

inastepwisewithdrawal,from

500to

250lg,thento

100lg,andfin

allyto

0lg

(placebo)[32]

Adv Ther (2020) 37:2956–2975 2965

Efficacy

For each outcome, findings are presented for allcomparisons with data available within thenetwork.

ExacerbationsModerate/severe and severe exacerbations werereported in 15 and 12 studies, respectively, withone study excluded from the severe exacerba-tions analysis as no events were reported

Fig. 2 Networks using treatments as reported (a), andusing all ICS/LABA treatments as a single treatment group(b). Fixed-dose combinations are represented with ‘‘/’’between components; open combinations are representedwith ‘‘?’’ between components. BDP beclomethasonedipropionate, BFF budesonide/formoterol fumarate,BGF budesonide/glycopyrronium/formoterol fumarate,BID twice daily, BUD budesonide, FF fluticasone furoate,

FOR formoterol, FP fluticasone propionate, GFF glycopy-rronium/formoterol fumarate, GLY glycopyrronium,ICS inhaled corticosteroid, IND indacaterol, LABA long-acting b2-agonist, MDI metered dose inhaler, OD oncedaily, red. reducing, SAL salmeterol, TIO tiotropium,UMEC umeclidinium, VI vilanterol

2966 Adv Ther (2020) 37:2956–2975

Table3

Patientbaselin

eclinicalcharacteristics

Stud

yMean

age

(years)

Male

(%)

Race

(%white)

Disease

duration

(years)

Current

smok

er(%

)

BMI

(kg/m

2)

COPD

severity

(%;GOLD

1/2/3/4)

Mod

erate/severe

exacerbation

history,

‡1exacerbation

s(%

)

Mean

CAT

score

Mean

mMRC

score

Aaron

2007

[26]

6856

98NR

2828

NR

100

NR

NR

Bremner2018

a[27]

6675

NR

NR

38NR

\1/35/49/15

100

19.9

NR

FULFILa

[28]

6474

85NR

4427

\1/33/54/13

6519.1

NR

Hanania2012

a[29]

6147

967

5827

NA/68/32/N

A29

bNR

2.5

IMPA

CTa[8]

6566

78NR

3527

\1/36/48/16

100

20.1

NR

Jung

2012

[4]

6798

NR

NR

NR

22NA/58/38/3

NR

NR

NR

KRONOSa

[10]

6571

507

4026

\1/49/43/8

2618.3

NR

Lee

2016

[5]

6796

NR

5NR

21NA/8/74/18

100

NR

NR

StudyAC4116135a

[33]

6366

88NR

5428

NA/46/44/11

21b

18.2

2.4

StudyAC4116136a

[33]

6563

82NR

3827

NA/40/48/12

31b

17.7

2.4

Study200109

a[6]

6466

98NR

4228

NA/40/46/14

15b

16.6

2.5

Study200110

a[6]

6363

86NR

5727

NA/48/41/11

14b

17.6

2.3

SUNSE

Ta[30]

6571

100

842

28NA/70/30/N

A34

NR

NR

TRIBUTEa[9]

6472

928

4526

NA/N

A/79/20

100

21.2c

NR

TRILOGYa[7]

6476

100

847

26NA/N

A/77/23

100

20.8

NR

TRIN

ITYa[34]

6376

998

4826

NA/N

A/79/21

100

21.6

NR

Welte

2009

[31]

6275

NR

6d44

26NA/25/64/11

100

NR

NR

WISDOM

[32]

6482

818

3325

\1/\

1/61/38

100

NR

1.8

Baselinecharacteristicswereobtained

from

publicallyavailableclinicalstudyreportswhennotavailablein

theprim

arypublication

BMIbody

massindex,

CAT

COPD

Assessm

entTest,

COPD

chronicobstructivepulmonarydisease,

mMRC

Modified

Medical

Research

Council,

NA

notapplicable,

NRnotreported

aAmajorityof

thepatientpopulation

was

classifiedas

symptom

aticat

baselin

e(based

onCATC

10or

mMRC

C2)

bModerateexacerbation

history(m

oderate/severe

notreported)

cReportedin

[47]

dMedian(m

eannotreported)

Adv Ther (2020) 37:2956–2975 2967

(networks shown in Fig. S1). The definitions ofmoderate and severe exacerbations werebroadly similar among the included studies.BGF MDI 320/18/9.6 lg twice daily (BID)showed comparable reductions in moder-ate/severe exacerbations to two other triple ICS/

LAMA/LABA fixed-dose combinations [be-clomethasone dipropionate/glycopyrro-nium/formoterol fumarate (BDP/GLY/FOR100/6/12.5 BID) and fluticasone furoate/ume-clidinium/vilanterol (FF/UMEC/VI 100/62.5/25 lgonce daily (OD))] and eight open triple

Fig. 3 Rate ratio of (a) moderate/severe exacerbations and(b) severe exacerbations. Fixed-dose combinations arerepresented with ‘‘/’’ between components; open combi-nations are represented with ‘‘?’’ between components.BDP beclomethasone dipropionate, BGF budesonide/glycopyrronium/formoterol fumarate, BUD budesonide,

CrI credible interval, FF fluticasone furoate,FOR formoterol, FP fluticasone propionate, GLY glycopy-rronium,MDI metered dose inhaler, red. reducing, RR rateratio, SAL salmeterol, TIO tiotropium, UMEC umecli-dinium, VI vilanterol

2968 Adv Ther (2020) 37:2956–2975

combinations (Fig. 3a). BGF MDI also showed acomparable reduction in severe exacerbations toFF/UMEC/VI, BDP/GLY/FOR, and six open triplecombinations (Fig. 3b).

Lung FunctionChanges frombaseline in trough andpeak FEV1 at24 weeks were reported in nine and four studies,respectively (networks shown in Fig. S2). At24 weeks, BGF MDI showed comparable effect ontrough FEV1 to BDP/GLY/FOR, FF/UMEC/VI, andfour open triple combinations [fluticasone propi-onate ? tiotropium ? salmeterol (FP/SAL250/50 lg BID ? TIO 18 lg OD and FP/SAL500/50 lgBID ? TIO18 lgOD), beclomethasonedipropionate ? tiotropium ? formoterol (BDP/FOR 200/12 lg BID ? TIO 18 lg OD), and fluti-casone furoate ? umeclidinium ? vilanterol (FF/VI 100/25 lg OD ? UMEC 62.5 lg OD)] (Fig. 4a).

At 24 weeks, BGF MDI showed comparableimprovement in peak FEV1 to BDP/GLY/FORand FF/UMEC/VI (Fig. 4b).

Quality of Life and SymptomsChanges from baseline in SGRQ total score at24 weeks were reported in eight studies (net-work shown in Fig. S3a). At 24 weeks, BGF MDIshowed comparable improvement in SGRQtotal score to BDP/GLY/FOR, FF/UMEC/VI, andthree open triple combinations (BDP/FOR200/12 lg BID ? TIO 18 lg OD, FP/SAL500/50 lg BID ? TIO 18 lg OD, and FF/VI100/25 lg OD ? UMEC 62.5 lg OD) (Fig. 5a).

Changes at week 24 in TDI focal score werereported in five studies (network shown inFig. S3b). At 24 weeks, BGF MDI showed compa-rable improvements in TDI focal score to BDP/GLY/FOR, FF/UMEC/VI, and an open triple com-bination (FF/VI 100/25 lg OD ? UMEC 62.5 lgOD) (Fig. 5b).

Rescue medication use over 12–24 weeks wasreported in six studies (network shown inFig. S3c). Over 12–24 weeks, BGF MDI showed acomparable reduction in mean puffs per day ofrescue medication versus BDP/GLY/FOR,FF/UMEC/VI, and an open triple combination(BDP/FOR 200/12 lg BID ? TIO 18 lg OD)(Fig. 5c).

Fig. 4 Lung function endpoints at 24 weeks. Change frombaseline in (a) trough FEV1 and (b) peak FEV1. Fixed-dosecombinations are represented with ‘‘/’’ between compo-nents; open combinations are represented with ‘‘?’’between components. BDP beclomethasone dipropionate,BGF budesonide/glycopyrronium/formoterol fumarate,

CrI credible interval, FEV1 forced expiratory volume in1 s, FF fluticasone furoate, FOR formoterol, FP flutica-sone propionate, GLY glycopyrronium, MD mean differ-ence, MDI metered dose inhaler, SAL salmeterol,TIO tiotropium, UMEC umeclidinium, VI vilanterol

Adv Ther (2020) 37:2956–2975 2969

Inconsistency Testing

Where feasible, assessments of inconsistencywere performed. For the majority of outcomes,no direct comparisons were available, so statis-tical inconsistency checks between direct andindirect comparisons were not possible. Formoderate/severe exacerbations, the results ofthe consistency assessment indicated noinconsistency between direct and indirect esti-mates for all comparisons, i.e., the 95%

confidence intervals for the inconsistency esti-mates contained 0.

Sensitivity Analyses and Meta-Regression

Meta-regression and sensitivity analyses wereconducted to examine the impact of hetero-geneity across the studies included in the NMA.Analyses of lung function, symptom, andexacerbation outcomes, including open-labelstudies produced results in line with the base-

Fig. 5 Health-related quality of life and symptom end-points. (a) Change from baseline in SGRQ total score at24 weeks, (b) TDI focal score at 24 weeks, and (c) changefrom baseline in daily rescue medication use over12–24 weeks. Fixed-dose combinations are representedwith ‘‘/’’ between components; open combinations arerepresented with ‘‘?’’ between components.BDP beclomethasone dipropionate, BGF budesonide/

glycopyrronium/formoterol fumarate, CrI credibleinterval, FF fluticasone furoate, FOR formoterol,FP fluticasone propionate, GLY glycopyrronium,MD mean difference, MDI metered dose inhaler, SAL sal-meterol, SGRQ St. George’s Respiratory Questionnaire,TDI Transition Dyspnea Index, TIO tiotropium,UMEC umeclidinium, VI vilanterol

2970 Adv Ther (2020) 37:2956–2975

case model, which only included double-blindstudies (Table S4). The studies included in thebase-case were conducted entirely in symp-tomatic patients, with the exception of onestudy that assessed the open combinationbudesonide (BUD)/FOR ? TIO [31] and twostudies that assessed open combinations of FP,SAL, and TIO [26, 32]. Sensitivity analysesexcluding these studies produced findings inline with the base-case model for all compar-isons (Table S4). There was no statistically sig-nificant association between the covariatesassessed in the meta-regression and treatmenteffects on severe exacerbations, which indicatedthat no linear relationship could be demon-strated between these covariates and treatmenteffect size (Table S5). For moderate/severeexacerbations, there were no significant associ-ations with BMI, disease severity, or priorexacerbation history, but there was a significantnegative association with current smokers atbaseline (Table S5). For both endpoints, resultsfrom the meta-regression were broadly similarto the base-case NMA. None of the covariate-adjusted models offered notable improvementin between-study variability compared tounadjusted models. The meta-regression resultsshould be interpreted with caution, as theanalyses were based on aggregate data, to allowfor accurate modeling of the effect of covariateson the treatment effect.

DISCUSSION

This SLR and NMA compared the efficacy ofBGF MDI with other triple ICS/LAMA/LABAfixed-dose or open combinations in the treat-ment of mostly symptomatic patients withmoderate to very severe COPD. NMAs areincreasingly recognized as an essential form ofevidence in developing healthcare guidelines,especially in areas of clinical practice wheredirect head-to-head trials are lacking [35, 36].To date, three triple fixed-dose combinationshave been developed and, as yet, no head-to-head trials of these therapies have been per-formed. Therefore, this NMA provides impor-tant context for healthcare providers and payersin evaluating the current evidence regarding

triple therapies in COPD. Our findings sug-gested that the efficacy of BGF MDI is compa-rable with all other fixed-dose (BDP/GLY/FORand FF/UMEC/VI) and open triple ICS/LAMA/LABA combination therapies with respect toreducing exacerbation rates and improving lungfunction, quality of life, and symptoms.

Exacerbations of COPD are associated withsignificant morbidity and mortality, andreducing the risk of future exacerbations is a keygoal of treatment [1, 37]. Reductions in mod-erate/severe exacerbations and severe exacerba-tions with BGF MDI were comparable to BDP/GLY/FOR, FF/UMEC/VI, and all open triplecombinations evaluated. The effect of BGF MDIin reducing exacerbations in a patient popula-tion with high exacerbation risk has beeninvestigated in the phase III ETHOS study,which reported topline results for more than8500 patients in August 2019 [38]. This studywill provide a more considerable body of evi-dence to evaluate the efficacy and safety profilesof two different doses of BGF MDI (160/18/9.6 lg and 320/18/9.6 lg) and will allow formore precise estimates of the comparisonbetween triple therapies with regard to exacer-bation rates [39].

When the relative efficacy of different treat-ments is comparable, other factors should beconsidered in choosing the right inhaler forindividual patients. COPD medications areavailable in a variety of device types, includingMDIs, dry powder inhalers, smooth mist inha-lers, and nebulizers. Each device has advantagesand disadvantages; the optimal inhaler for eachpatient depends on their preferences (e.g.,device familiarity, ease of use, size, cost), diseasecharacteristics (e.g., inspiratory flow), and abil-ities (e.g., hand–breath coordination, gripstrength, and dexterity) [1, 40–42]. While thisNMA suggests comparable efficacy with fixed-dose and open triple combinations in RCTsconducted in highly controlled settings, real-world studies have reported higher medicationadherence and persistence in patients withCOPD using fewer inhalers [43–45]. A 12-monthretrospective cohort study found higher adher-ence to triple therapy in patients with COPDwho were using two inhalers versus threeinhalers [43]. Similarly, a retrospective

Adv Ther (2020) 37:2956–2975 2971

observational study of patients with COPD whowere receiving LAMA, LABA, and/or ICS therapyin single or multiple inhalers showed that theproportion of adherent patients was higher forsingle-inhaler users versus multiple inhalerusers [44]. Persistence was also higher for singleversus multiple inhalers, with a significantlyhigher risk of discontinuation in users of mul-tiple inhalers [44]. The use of a dual combina-tion inhaler versus two individual inhalers hasalso been associated with a lower risk of respi-ratory-related hospitalization and reducedhealthcare costs [45]. Together, these studiessuggest that in real-world use, fixed-dose com-bination triple therapies may result in betterpatient outcomes than open triple therapy withmultiple inhalers despite comparable efficacy,owing to improved adherence and persistence.However, to our knowledge, specific compar-isons of adherence and persistence for fixed-dose triple therapies versus open triple therapiesin real-world use have not yet been reported.

Several limitations of the NMA methodologyshould be acknowledged. Different ICS/LABAswere grouped under a single treatment class toresolve the disconnected network. While thisapproach has been used in previous meta-anal-yses [12, 13], it means that intra-class differ-ences among ICS/LABAs would not have beencaptured within the analyses. However, intra-class differences in dual therapies were beyondthe objective of this NMA.

NMAs and traditional pairwise meta-analysesdepend upon an assumption of similaritybetween the included studies, including in theirpatient populations, study design, and outcomemeasures [11]. While the studies included inthis NMA were broadly similar, there were somedifferences in study design and patient popula-tions across studies, including disease severityand exacerbation history. While statisticalinconsistency checks were not possible for mostoutcomes because of the lack of direct compar-isons available within the networks, for the oneoutcome with direct comparisons available(moderate/severe exacerbations), no inconsis-tency was identified. Potential sources of clini-cal heterogeneity were explored in sensitivityanalyses and meta-regression where possible,and the results were consistent with the base-

case scenario. While some outcomes could notbe assessed with meta-regression because of asmaller number of studies reporting data, thestudy populations were broadly similar acrossstudies (moderate to very severe COPD), and themajority of studies included in the NMA wereconsidered to have a low risk of bias. Furtherresearch is warranted as additional evidenceregarding triple therapies, especially fixed-dosecombinations, becomes available.

In conclusion, this NMA of 18 studies sug-gests that BGF MDI has similar efficacy to otherICS/LAMA/LABA fixed-dose and open combi-nation therapies in reducing exacerbation ratesand improving lung function and symptoms inpatients with moderate to very severe COPD.

ACKNOWLEDGEMENTS

The authors thank Pragya Shukla (formerly ofParexel International) and Akanksha Sharma(Parexel International) for their assistance withthe analyses.

Funding. This systematic literature reviewand network meta-analysis was conducted byParexel International Ltd and funded by Astra-Zeneca. No authors received financial compen-sation in relation to authorship on thismanuscript. AstraZeneca funded the journal’sarticle Rapid Service Fee and Open Access fee.

Medical Writing Assistance. Medical writ-ing support, under the direction of the authors,was provided by Carole Evans, Ph.D., on behalfof CMC Connect, McCann Health MedicalCommunications, funded by AstraZeneca,Gaithersburg, USA in accordance with GoodPublication Practice (GPP3) guidelines [46].

Authorship. All named authors meet theInternational Committee of Medical JournalEditors (ICMJE) criteria for authorship for thisarticle, take responsibility for the integrity ofthe work as a whole, and have given theirapproval for this version to be published.

2972 Adv Ther (2020) 37:2956–2975

Authorship Contributions. Patrick Darken,Shaila Ballal, Ulf Holmgren, and Enrico deNigris made substantial contributions to theconception and design of the study. MohdKashif Siddiqui, Barinder Singh, and SumeetAttri performed the meta-analysis. All authorscontributed to the interpretation of the data,critically revised the manuscript, approved thefinal version to be submitted, and agree to beaccountable for all aspects of the work inensuring that questions related to the accuracyor integrity of any part of the work are appro-priately investigated and resolved.

Disclosures. Gary T Ferguson reports grants,personal fees, and non-financial support fromAstraZeneca during the conduct of the study;grants, personal fees and non-financial supportfrom AstraZeneca, Boehringer Ingelheim,Novartis, and Sunovion; grants and personalfees from Theravance; and personal fees fromCircassia, GlaxoSmithKline, Innoviva, Mylan,and Verona, outside of the submitted work.Patrick Darken, Ulf Holmgren, and Enrico deNigris are employees of AstraZeneca and holdstock in the company. Shaila Ballal is anemployee of AstraZeneca. Mohd Kashif Sid-diqui, Barinder Singh, and Sumeet Attri areemployees of Parexel International, the organi-zation that received funding from AstraZenecato perform the systematic literature review andnetwork meta-analysis.

Compliance with Ethics Guidelines. Thisarticle is based on previously conducted studiesand does not contain any studies with humanparticipants or animals performed by any of theauthors.

Data Availability. The datasets generatedduring and/or analyzed during the currentstudy are available from the correspondingauthor on reasonable request.

Open Access. This article is licensed under aCreative Commons Attribution-NonCommer-cial 4.0 International License, which permitsany non-commercial use, sharing, adaptation,distribution and reproduction in any mediumor format, as long as you give appropriate credit

to the original author(s) and the source, providea link to the Creative Commons licence, andindicate if changes were made. The images orother third party material in this article areincluded in the article’s Creative Commonslicence, unless indicated otherwise in a creditline to the material. If material is not includedin the article’s Creative Commons licence andyour intended use is not permitted by statutoryregulation or exceeds the permitted use, youwill need to obtain permission directly from thecopyright holder. To view a copy of this licence,visit http://creativecommons.org/licenses/by-nc/4.0/.

REFERENCES

1. Global Initiative for Chronic Obstructive LungDisease. 2020 Report: Global Strategy for the Diag-nosis, Management and Prevention of COPD; 2020.https://goldcopd.org. Accessed Nov 5, 2019.

2. World Health Organization. Global Health Esti-mates 2016: deaths by cause, age, sex, by countryand by region 2000–2016; 2018. https://www.who.int/healthinfo/global_burden_disease/estimates/en/. Accessed Apr 11, 2019.

3. Global Initiative for Chronic Obstructive LungDisease. 2019 Report: global strategy for the diag-nosis, management and prevention of COPD; 2019.https://goldcopd.org. Accessed Oct 28, 2019.

4. Jung KS, Park HY, Park SY, et al. Comparison oftiotropium plus fluticasone propionate/salmeterolwith tiotropium in COPD: a randomized controlledstudy. Respir Med. 2012;106:382–9.

5. Lee SD, Xie CM, Yunus F, et al. Efficacy and toler-ability of budesonide/formoterol added to tio-tropium compared with tiotropium alone inpatients with severe or very severe COPD: a ran-domized, multicentre study in East Asia. Respirol-ogy. 2016;21:119–27.

6. Siler TM, Kerwin E, Sousa AR, Donald A, Ali R,Church A. Efficacy and safety of umeclidiniumadded to fluticasone furoate/vilanterol in chronicobstructive pulmonary disease: results of two ran-domized studies. Respir Med. 2015;109:1155–63.

7. Singh D, Papi A, Corradi M, et al. Single inhalertriple therapy versus inhaled corticosteroid pluslong-acting b2-agonist therapy for chronic obstruc-tive pulmonary disease (TRILOGY): a double-blind,

Adv Ther (2020) 37:2956–2975 2973

parallel group, randomised controlled trial. Lancet.2016;388:963–73.

8. Lipson DA, Barnhart F, Brealey N, et al. Once-dailysingle-inhaler triple versus dual therapy in patientswith COPD. N Engl J Med. 2018;378:1671–80.

9. Papi A, Vestbo J, Fabbri L, et al. Extrafine inhaledtriple therapy versus dual bronchodilator therapy inchronic obstructive pulmonary disease (TRIBUTE):a double-blind, parallel group, randomised con-trolled trial. Lancet. 2018;391:1076–84.

10. Ferguson GT, Rabe KF, Martinez FJ, et al. Tripletherapy with budesonide/glycopyrrolate/for-moterol fumarate with co-suspension deliverytechnology versus dual therapies in chronicobstructive pulmonary disease (KRONOS): a dou-ble-blind, parallel-group, multicentre, phase 3 ran-domised controlled trial. Lancet Respir Med.2018;6:747–58.

11. Tonin FS, Rotta I, Mendes AM, Pontarolo R.Network meta-analysis: a technique to gather evi-dence from direct and indirect comparisons. PharmPract (Granada). 2017;15:943.

12. Cazzola M, Rogliani P, Calzetta L, Matera MG.Triple therapy versus single and dual long-actingbronchodilator therapy in COPD: a systematicreview and meta-analysis. Eur Respir J. 2018;52:1801586.

13. Calzetta L, Cazzola M, Matera MG, Rogliani P.Adding a LAMA to ICS/LABA therapy: a meta-analysis of triple combination therapy in COPD.Chest. 2019;155:758–70.

14. Zayed Y, Barbarawi M, Kheiri B, et al. Triple versusdual inhaler therapy in moderate-to-severe COPD: asystematic review and meta-analysis of randomizedcontrolled trials. Clin Respir J. 2019;13:413–28.

15. AstraZeneca. Breztri Aerosphere (PT010) approvedin Japan for patients with chronic obstructive pul-monary disease; 2019. https://www.astrazeneca.com/media-centre/press-releases/2019/breztri-aerosphere-pt010-approved-in-japan-for-patients-with-chronic-obstructive-pulmonary-disease-19062019.html. Accessed July 2, 2019.

16. AstraZeneca. AstraZeneca’s triple-combinationtherapy approved in China for patients with COPD;2019. https://www.astrazeneca.com/media-centre/press-releases/2019/astrazenecas-triple-combination-therapy-approved-in-china-for-patients-with-copd-23122019.html. Accessed Dec 23, 2019.

17. National Institute for Health and Care Excellence.Single technology appraisal: User guide for com-pany evidence submission template. last updated2015. https://www.nice.org.uk/process/pmg24/

chapter/5-clinical-effectiveness#quality-assessment-of-the-relevant-randomised-controlled-trials.Accessed June 4, 2019.

18. Dias S, Welton NJ, Sutton AJ, Ades AE. NICE DSUTechnical Support Document 2: a generalised linearmodelling framework for pairwise and networkmeta-analysis of randomised controlled trials. 2011,last updated September 2016. http://nicedsu.org.uk/wp-content/uploads/2017/05/TSD2-General-meta-analysis-corrected-2Sep2016v2.pdf. Acces-sed Jan 23, 2019.

19. Dias S, Sutton AJ, Welton NJ, Ades AE. NICE DSUTechnical Support Document 3: Heterogeneity:subgroups, meta-regression, bias and bias-adjust-ment; 2011, last updated April 2012. http://nicedsu.org.uk/wp-content/uploads/2016/03/TSD3-Heterogeneity.final-report.08.05.12.pdf. AccessedJan 23, 2019.

20. Owen RK, Tincello DG, Abrams KR. Network meta-analysis: development of a three-level hierarchicalmodeling approach incorporating dose-relatedconstraints. Value Health. 2015;18:116–26.

21. Phillippo DM, Dias S, Ades AE, Didelez V, WeltonNJ. Sensitivity of treatment recommendations tobias in network meta-analysis. J R Stat Soc Ser A StatSoc. 2018;181:843–67.

22. Dias S, Welton NJ, Caldwell DM, Ades AE. Checkingconsistency in mixed treatment comparison meta-analysis. Stat Med. 2010;29:932–44.

23. National Institute for Health and Care Excellence.Depression in adults: treatment and management.Appendix N1: network meta-analysis—detailedmethods and results; 2018. https://www.nice.org.uk/guidance/gid-cgwave0725/documents/addendum-appendix-19. Accessed 4 Apr 2020.

24. Schlueter M, Gonzalez-Rojas N, Baldwin M,Groenke L, Voss F, Reason T. Comparative efficacyof fixed-dose combinations of long-acting mus-carinic antagonists and long-acting b2-agonists: asystematic review and network meta-analysis. TherAdv Respir Dis. 2016;10:89–104.

25. Jones PW. St. George’s Respiratory Questionnaire:MCID. COPD. 2005;2:75–9.

26. Aaron SD, Vandemheen KL, Fergusson D, et al.Tiotropium in combination with placebo, salme-terol, or fluticasone-salmeterol for treatment ofchronic obstructive pulmonary disease: a random-ized trial. Ann Intern Med. 2007;146:545–55.

27. Bremner PR, Birk R, Brealey N, Ismaila AS, Zhu CQ,Lipson DA. Single-inhaler fluticasone furoate/umeclidinium/vilanterol versus fluticasone furoate/vilanterol plus umeclidinium using two inhalers for

2974 Adv Ther (2020) 37:2956–2975

chronic obstructive pulmonary disease: a random-ized non-inferiority study. Respir Res. 2018;19:19.

28. Lipson DA, Barnacle H, Birk R, et al. FULFIL trial:once-daily triple therapy for patients with chronicobstructive pulmonary disease. Am J Respir CritCare Med. 2017;196:438–46.

29. Hanania NA, Crater GD, Morris AN, Emmett AH,O’Dell DM, Niewoehner DE. Benefits of addingfluticasone propionate/salmeterol to tiotropium inmoderate to severe COPD. Respir Med. 2012;106:91–101.

30. Chapman KR, Hurst JR, Frent SM, et al. Long-termtriple therapy de-escalation to indacaterol/gly-copyrronium in patients with chronic obstructivepulmonary disease (SUNSET): a randomized, dou-ble-blind, triple-dummy clinical trial. Am J RespirCrit Care Med. 2018;198:329–39.

31. Welte T, Miravitlles M, Hernandez P, et al. Efficacyand tolerability of budesonide/formoterol added totiotropium in patients with chronic obstructivepulmonary disease. Am J Respir Crit Care Med.2009;180:741–50.

32. Magnussen H, Disse B, Rodriguez-Roisin R, et al.Withdrawal of inhaled glucocorticoids and exacer-bations of COPD. N Engl J Med. 2014;371:1285–94.

33. Siler TM, Kerwin E, Singletary K, Brooks J, ChurchA. Efficacy and safety of umeclidinium added tofluticasone propionate/salmeterol in patients withCOPD: results of two randomized, double-blindstudies. COPD. 2016;13:1–10.

34. Vestbo J, Papi A, Corradi M, et al. Single inhalerextrafine triple therapy versus long-acting mus-carinic antagonist therapy for chronic obstructivepulmonary disease (TRINITY): a double-blind, par-allel group, randomised controlled trial. Lancet.2017;389:1919–29.

35. Kanters S, Ford N, Druyts E, Thorlund K, Mills EJ,Bansback N. Use of network meta-analysis in clini-cal guidelines. Bull World Health Organ. 2016;94:782–4.

36. Faltinsen EG, Storebø OJ, Jakobsen JC, Boesen K,Lange T, Gluud C. Network meta-analysis: thehighest level of medical evidence? BMJ Evid BasedMed. 2018;23:56–9.

37. Epstein D, Barak-Corren Y, Isenberg Y, Berger G.Clinical decision support system: a pragmatic toolto improve acute exacerbation of COPD dischargerecommendations. COPD. 2019;16:18–24.

38. AstraZeneca. Breztri Aerosphere Phase III ETHOStrial met its primary endpoint in chronic obstruc-tive pulmonary disease; 2019. https://www.astrazeneca.com/media-centre/press-releases/2019/breztri-aerosphere-phase-iii-ethos-trial-met-its-primary-endpoint-in-chronic-obstructive-pulmonary-disease-28082019.html Accessed Jan 23, 2020.

39. Rabe KF, Martinez FJ, Ferguson GT, et al. A phase IIIstudy of triple therapy with budesonide/glycopy-rrolate/formoterol fumarate metered dose inhaler320/18/9.6 lg and 160/18/9.6 lg using co-suspen-sion delivery technology in moderate-to-very severeCOPD: the ETHOS study protocol. Respir Med.2019;158:59–66.

40. Scichilone N, Benfante A, Bocchino M, et al. Whichfactors affect the choice of the inhaler in chronicobstructive respiratory diseases? Pulm PharmacolTher. 2015;31:63–7.

41. Molimard M, Colthorpe P. Inhaler devices forchronic obstructive pulmonary disease: insightsfrom patients and healthcare practitioners.J Aerosol Med Pulmon Drug Deliv. 2015;28:219–28.

42. Rogliani P, Calzetta L, Coppola A, et al. Optimizingdrug delivery in COPD: the role of inhaler devices.Respir Med. 2017;124:6–14.

43. Bogart M, Stanford RH, Laliberte F, Germain G, WuJW, Duh MS. Medication adherence and persistencein chronic obstructive pulmonary disease patientsreceiving triple therapy in a USA commerciallyinsured population. Int J Chron Obstruct PulmonDis. 2019;14:343–52.

44. Yu AP, Guerin A, Ponce de Leon D, et al. Therapypersistence and adherence in patients with chronicobstructive pulmonary disease: multiple versussingle long-acting maintenance inhalers. J MedEcon. 2011;14:486–96.

45. Chrischilles E, Gilden D, Kubisiak J, Rubenstein L,Shah H. Delivery of ipratropium and albuterolcombination therapy for chronic obstructive pul-monary disease: effectiveness of a two-in-oneinhaler versus separate inhalers. Am J Manag Care.2002;8:902–11.

46. Battisti WP, Wager E, Baltzer L, et al. Good publi-cation practice for communicating company-spon-sored medical research: GPP3. Ann Intern Med.2015;163:461–4.

47. Singh D, Fabbri LM, Vezzoli S, Petruzzelli S, Papi A.Extrafine triple therapy delays COPD clinicallyimportant deterioration vs ICS/LABA, LAMA, orLABA/LAMA. Int J Chron Obstruct Pulmon Dis.2019;14:531–46.

Adv Ther (2020) 37:2956–2975 2975