1-2015-Long-acting muscarinic antagonist + long-acting beta agonist

REVIEW

Long-acting muscarinic antagonist +long-acting beta agonist versus long-acting beta agonist +inhaled corticosteroid for

COPD:A systematic review and meta-analysis

N OBUYUKI HORITA,1,2N AOKI MIYAZAWA,1K OJI TOMARU,1M IYO INOUE 1AND T AKESHI KANEKO 2

1

Department of Respiratory Medicine,Saiseikai Yokohamashi Nanbu Hospital,and 2Department of Pulmonology,Yokohama

City University Graduate School of Medicine,Yokohama,Japan

ABSTRACT

Some trials have been conducted to compare long-acting muscarinic antagonist (LAMA)+long-act-ing beta agonist (LABA)versus LABA +inhaled corticosteroids (ICS)for chronic obstructive pulmo-nary disease (COPD),but no meta-analysis were reported.

Two investigators independently searched for eli-gible articles using the PubMed,Web of Science and Cochrane databases.Articles in authors’reference ?les were also regarded as candidates.The eligibility criteria for the current meta-analysis were original trials written in English comparing the impact of LAMA +LABA and LABA +ICS for COPD patients.A pooled value for the continuous value was calculated using the genetic inverse variance method for mean difference.Incidence of events was evaluated using the odds ratio (OR).Minimal clinically important differ-ence were 50mL for forced expiratory volume in 1s (FEV 1),four points for St George Respiratory Question-naire (SGRQ)and one point for transition dyspnoea index (TDI).

We included seven randomized controlled trials and one cross-over trial with follow-up period of 6–https://www.wendangku.net/doc/459223627.html,pared with LABA +ICS,LAMA +LABA led to signi?cantly greater improvements of trough FEV 1by 71(95%CI:48–95)mL,TDI by 0.38points (95%CI:0.17–0.58),less exacerbations with an OR of 0.77(95%CI:0.62–0.96)and less pneumonia with an OR of 0.28(95%CI:0.12–0.68).Frequencies of any adverse event,serious adverse event,adverse event leading to discon-tinuation,all-cause death and change of total score of SGRQ were not different in both https://www.wendangku.net/doc/459223627.html,MA +LABA might be a better option for treating COPD than LABA +ICS.

Key words:adverse effect,dyspnoea,exacerbation,forced

expiratory volume,pneumonia.

Abbreviations:AE,adverse event;95%CI,95%con?dence interval;COPD,chronic obstructive pulmonary disease;FEV1,forced expiratory volume in 1s;ICS,inhaled corticosteroids;LABA,long-acting beta agonist;LAMA,long-acting muscarinic antagonist;MCID,minimal clinically important difference;RCT,randomized controlled trial;OR,odds ratio;SFC,salmeterol/?uticasone combination;SGRQ,St George Respira-tory Questionnaire;TDI,transition dyspnoea index;UMEC/VI,combined umeclidinium +vilanterol;QVA149,combined indacaterol +glycopyrronium bromide.

INTRODUCTION

Chronic obstructive pulmonary disease (COPD)is a progressive in?ammatory disease characterized by air?ow limitation and systematic in?ammation,for which inhaled medications play an essential part of disease control.1Long-acting muscarinic antagonist (LAMA),long-acting beta agonist (LABA)and inhaled corticosteroids (ICS),which are administrated via an inhaler device,are key medications for COPD treatment.1Treatment with a combination of two or three inhaled medications belonging to different pharmacological classes can provide a greater bene?t compared with a single medication https://www.wendangku.net/doc/459223627.html,e of combined medications in one device has the additional advantage of improving patient adherence to the treatment.Currently,we have two types of combined medication:LAMA +LABA and LABA +ICS.The former includes combined indacaterol +glycoppyronium bromide (QVA149,Ultibro)and combined umeclidinium +vilanterol (UMEC/VI,Anoro);and the latter includes the salmeterol/?uticasone combination (SFC,Adoair,Seretide,Advair).2–5

The current guidelines recommend the use of ICS only for a patient with group C and D COPD:severe to very severe air?ow limitation,≥2exacerbations per year,and/or ≥1with hospitalization for exacerbation.However,prescription rates of ICS and combined LABA +ICS are very high.6Among the combined inhaler medications,SFC has been prescribed the

Correspondence:Nobuyuki Horita,Department of Pulmonology,Yokohama City University Graduate School of Medicine,3-9Fukuura,Kanazawa-ku,Yokohama 236-0004,Japan.Email:nobuyuki_horita@yahoo.co.jp

Con?ict of Interest:KT received grants and/or lecture fees from GlaxoSmithKline,Boehringer Ingelheim,P?zer,Meiji and Novartis within the last 3years.

Received 30December 2014;invited to revised 10April 2015;revised:13April 2015;accepted 15June 2015(Associate Editor:Neil Eves).

Article ?rst published online:03August 2015

bs_bs_banner

most frequently.This is partly because SFC is the oldest combination drug,and it is partly because many randomized controlled trials(RCT)have indi-cated that SFC can decrease exacerbation of COPD, improve the quality of life and prevent the yearly decline of forced expiratory volume in1s(FEV1).4,5In addition to SFC,QVA149and UMEC/VI are other novel attractive combined medications consisting of two key bronchodilating medications.2,3,7,8

Thus,it is interesting to compare the effect of LAMA+LABA and LABA+ICS.Some trials have been conducted to compare LAMA+LABA and LABA+ICS for COPD treatment,9–17though,no meta-analysis were reported.We therefore performed the current systematic review and meta-analysis to provide updated information comparing LAMA+LABA and LABA+ICS.

METHODS

Study search

Institutional review board approval and patient consent were not required because of the review nature of this study.

Two investigators independently searched for eli-gible articles using the PubMed,Web of Science and Cochrane databases as of October2014.The search formula used for PubMed was shown in Supplemen-tary Appendix S1.

Articles in authors’reference?les were also regarded as candidates.

The eligibility criteria for the current meta-analysis were original RCT written in English comparing the impact of LAMA+LABA and LABA+ICS for COPD patients.Outcomes were to include one or more of the following:trough FEV1,post-dose FEV1,24-h weighted mean FEV1,total score of St George Respira-tory Questionnaire(SGRQ),transition dyspnoea index(TDI),any adverse event(AE),serious AE,AE leading to discontinuation,exacerbation,pneumonia and all-cause death.All of the following were allowed: parallel group studies,crossover studies,blinded studies and open-label studies.Duplicate use of the same data was evaluated carefully.The quality of an eligible study was evaluated using Chalmer’s scale,18 wherein100and0indicates the best and the poorest quality,respectively.

When an accessed original report had not provided suf?cient data,the authors of the current manuscript tried to have a contact with authors of the original report.

Outcomes and statistics

A continuous variable was evaluated as a change from the baseline to the end of observation.A value in the LAMA+LABA arm minus a value in the LABA+ICS arm was shown.A pooled value for the continuous value was calculated using the genetic inverse vari-ance method for mean difference.19

Incidence of event was evaluated using the odds ratio(OR).The OR was calculated using odds for the LABA+ICS arm as a reference.Thus,OR<1indicated that less frequent events were observed in the LAMA+LABA arm.

The heterogeneity of original studies was evaluated with the I2statistics,whereby I2=0%indicates no het-erogeneity,0%

50%≤I2<75%indicates moderate heterogeneity and 75%≤I2indicates strong heterogeneity.20We had planned to use a?xed model when I2<50%and a random model when50%≤I2.19When two or more of trials compared the same medications,a subgroup for sensitivity analysis and between subgroup differences was evaluated.We interpreted results from subgroup analysis with caution as these results were not pri-marily focused in pre-speci?ed protocol and were derived from a small sample size.

To distinguish the clinically meaningful/ meaningless change,we use cut-off values,clinically important differences(MCID):50mL for FEV1,four points for SGRQ,one point for TDI.21–23

All analysis was performed in Review Manager ver 5.3(Cochrane Collaboration,Oxford,UK).

RESULTS

Study search

Of101articles that met the preliminary criteria,we found eight eligible RCT9–16(Fig.1).A pharmaceutical company and an author of a conference abstract were contacted by us and provided data.11–13,17This confer-ence abstract17was about two RCT reported by other two documents.11,12Eight eligible RCT included seven double-blinded trials,one open label trial,seven par-allel group trials and one crossover trial(Table1).The number of randomized patients ranged from46to 744.The follow-up period ranged from six to26 weeks.The quality score ranged from32to70with a median of55,which suggested the study quality was generally acceptable.One study received the lowest score of32due to the open-label design of the trial. Seven out of the eight RCT had substantial con?icts of interest:three were sponsored by GlaxoSmithKline, two were sponsored by Novartis Pharma,one was sponsored by both Boehringer Ingelheim and P?zer, and one was supported by Boehringer Ingelheim for statistical analysis and the manuscript editing. Seven out of eight studies excluded cases with recent or ongoing exacerbations.All studies included patients with%predicted FEV1both≥50%and<50%. The treatment regimens were as follows:QVA149 (110/50μg once a day)versus SFC(50/500μg twice a day)for two trials,UMEC/VI(62.5/25μg once a day) versus SFC(50/250μg twice a day)for two trials, UMEC/VI(62.5/25μg once a day)versus SFC(50/ 500μg twice a day)for one trial,tiotropium(18μg once a day)+indacaterol(150μg once a day)versus SFC(50/250μg twice a day)for one trial,tiotropium (18μg once a day)+salmeterol(50μg twice a day) versus SFC(50/500μg twice a day)for one trial,and tiotoropium(18μg once a day)+formoterol(12μg twice a day)versus salmeterol(50μg twice a day)+?uticasone(500μg twice a day)for one trial

N Horita et al.

1154

Figure 1Flow chart for study search.ICS,inhaled corticosteroids;LABA,long-acting beta agonist;LAMA,long-acting muscarinic antagonist;n ,number of reports;RCT,randomized controlled trial.

Table 1Summary of included randomized controlled trials

Study

Regimen

Design

n

Follow-up period(week)

Severity

Excluding exacerbation

Quality score

ILLUMINATE 20129

QVA149SFC500Double blind Parallel

5232640–80%Ex (1year)

69LANTERN 201410

QVA149SFC500Double blind Parallel

7442630–80%mMRC ≥2≥2Ex,Ex (6weeks),Admission (1year)58DB2114930201311

UMEC/VI SFC250Double blind Parallel

7101230–70%Ex (1year)52DB2114951201412

UMEC/VI SFC500Double blind Parallel

7101230–70%Ex (1year)52DB2116134201313UMEC/VI SFC250

Double blind Parallel 7101230–70%Ex (1year)52Hoshino 201414

Tio +Ind150SFC250Open label Parallel

461630–80%Ex at entry 32Magnussen 201215Tio +Sal SFC500Double blind Crossover 3448×2

<65%Not excluded 70Rabe 200816

Tio +For12Sal +Flu500

Double blind Parallel

605

6

<65%

No oral steroid (6weeks)

59

Quality score,evaluated using Chalmer’scale (100is the best).Severity,forced expiratory volume in 1s was in the range of presented percentage of the predicted value.

Ex,exacerbation;Flu500,?uticasone (500μg twice a day);For12,formoterol (12μg twice a day);Ind150,indacaterol (150μg once a day);mMRC,modi?ed Medical Research Council dyspnoea scale;n :number of randomized patients;QVA149,combined indacaterol +glycopyrronium bromide (110/50μg once a day);Sal,salmeterol (50μg twice a day);SFC250,salmeterol/?uticasone combination (50/250μg twice a day);SFC500,salmeterol/?uticasone combination (50/500μg twice a day);Tio,tiotropium (18μg once a day);UMEC/VI,combined umeclidinium +vilanterol (62.5/25μg once a day).COPD inhaler medication:A meta-analysis 1155

(Table1).Most studies included patients with%pre-dicted FEV1of30–80%(Table1).

Meta-analysis

Trough FEV1

In each study,trough FEV1was de?ned as FEV115–60 min before inhalation,or simply pre-dose FEV1.A random-model meta-analysis using data from4648 cases(cases in a crossover trial were double counted) in seven RCT indicated that pooled trough FEV1 increase from the baseline was larger in the LAMA+LABA arm by71mL(95%con?dence interval (CI):48–95mL)(>MCID:50mL).Meta-analysis with whole RCT was associated with moderate heteroge-neity(I2=68%),and test for medication-oriented subgroup had very strong heterogeneity(I2=87.8%). However,subgroup analysis presented no or little het-erogeneity in each subgroup.These?nding suggested the heterogeneity among whole RCT were due to dif-ference of medications.Pooled trough FEV1increases from baseline were87mL,90mL and31mL for QVA versus SFC subgroup,UMEC/VI versus SFC subgroup and other medications subgroup,respectively(Fig.2). Total score of St George Respiratory Questionnaire

A random-model meta-analysis using data from3382 cases in?ve RCT indicated that pooled total score of SGRQ improvement(decrease)from baseline was slightly larger in LAMA+LABA arm by?1.54points (95%CI:?3.31to0.23.I2:75%),but reached neither MCID of four points nor statistical signi?cance(Sup-plementary Figure S1).

Transition dyspnoea index

A?xed-model meta-analysis using data from2666 cases in four RCT indicated that pooled TDI score improvement(increase)from baseline was signi?-cantly larger in the LAMA+LABA arm by0.38points (95%CI:0.17to0.58.I2:49%),but was smaller than MCID of one point(Supplementary Figure S1).

Any adverse event

A?xed-model meta-analysis using data from4648 cases(cases in a crossover trial were double counted) in seven RCT suggested that any AE was observed in similar frequencies in both arms with an OR of0.90 (95%CI:0.79–1.01.I2:0%)(Fig.2).

Serious adverse event

A?xed-model meta-analysis using data from4691 cases(cases in a crossover trial were double counted) in eight RCT indicated that serious AE was observed in similar frequencies in both arms with an OR of 0.84(95%CI:0.62–1.16.I2:22%)(Supplementary Figure S1).

Adverse event leading to discontinuation

A?xed-model meta-analysis using data from3425 cases in six RCT indicated that AE leading to discon-tinuation was observed in similar frequencies in both arms with an OR of0.76(95%CI:0.53–1.09.I2:0%) (Supplementary Figure S1).

Exacerbation

One study did not present the number of exacerba-tion,but presented the number of COPD worsening, which we regard as COPD exacerbation.A?xed-model meta-analysis using data from2529cases (cases in a crossover trial were double counted)in four RCT indicated that exacerbation was less fre-quently observed in the LAMA+LABA arm with an OR of0.77(95%CI:0.62–0.96).Although among whole RCT analysis,there was little heterogeneity with I2of 7%),test for medication-oriented subgroup had strong heterogeneity(I2=62.0%).QVA149seemed to be related to less exacerbation compared to SFC (Fig.2).

Pneumonia

A?xed-model meta-analysis using data from2666 cases in four RCT indicated that pneumonia occurred less frequently in LAMA+LABA arm with a very small OR of0.28without heterogeneity(95%CI:0.12–0.68. I2:0%)(Fig.2).

All-cause death

Only12deaths were observed throughout eight RCT. A?xed-model meta-analysis using data from4691 cases in eight RCT indicated that all-cause death was observed in the same frequency in two arms,with an OR of1.00(95%CI:0.37–2.67I2:0%)(Supplementary Figure S1).

Sensitivity analysis

We additionally conducted sensitivity analyses excluding a cross-over RCT by Magnussen.These sen-sitivity analyses yielded similar results from analyses including the cross-over RCT as below,which sug-gested that results of our meta-analyses was robust. Pooled trough FEV1increase from the baseline was larger in the LAMA+LABA arm by78mL(95%CI: 56–101mL,P<0.001).Any AE and serious AE observed were of similar frequencies in both arms with OR0.89(95%CI:0.78–1.02.P=0.08)and0.78 (95%CI:0.56–1.09.P=0.15),respectively.Exacerba-tion occurred less frequently in LAMA+LABA arm with an OR of0.71(95%CI:0.56–0.91,P=0.005). DISCUSSION

The current meta-analysis with eight RCT suggests that,compared to LABA+ICS,LAMA+LABA pro-vided larger trough FEV1improvement by71mL. Especially,QVA149provided larger improvement compared to SFC.This is a reasonable result given the strong and long-lasting dual bronchodilating effect by QVA149and UMEC/VI.7,8,24As we adopted MCID for FEV1of50mL,the difference of71mL was judged clinically important.However,considering that some researcher proposed MCID of100mL or larger,we

N Horita et al.

1156

F i g u r e 2F o r r e s t p l o t s t o c o m p a r e l o n g -a c t i n g m u s c a r i n i c a n t a g o n i s t (L A M A )+l o n g -a c t i n g b e t a a g o n i s t (L A B A )a n d L A B A +i n h a l e d c o r t i c o s t e r o i d s (I C S ).I V ,i n v e r s e v a r i a n c e ;M -H ,M a n t e l –H a e n s z e l ;S E ,s t a n d a r d e r r o r .

COPD inhaler medication:A meta-analysis

1157

may have better interpreted that a pooled difference of FEV1of71mL was of marginal clinical meaningful. Less exacerbations were observed in LAMA+LABA arm with an OR of0.77.Although most clinicians choose SFC to reduce exacerbation,1,25our result showed that LAMA+LABA may be a preferable choice to avoid https://www.wendangku.net/doc/459223627.html,MA+LABA was also associated with less pneumonias with an OR of0.28, which may be partly because ICS increases pneumo-nia due to its immunosuppressive effect on the airway.26

From the beginning,the indication of ICS for COPD has been questionable.6Both COPD and asthma are associated with chronic in?ammation of the lung; however,there are differences in the type of in?am-matory cells and mediators involved.While ICS is clearly indicated for asthma in which immunoglobu-lin E and eosinophils play an important role,27ICS is not as effective on neutrophil and macrophage,which are the predominant in?ammatory cells observed in the respiratory tract of COPD cases.1Recently,the importance of ICS treatment for asthma–COPD overlap syndrome has been emphasized.28Even though some previous trials have shown a slightly favourable impact of ICS on COPD,this may be a result of the substantial bene?t of ICS in treating the overlap syndrome.Another doubtful?nding is that some trials have indicated that ICS decreases the exacerbations of COPD but that ICS increases pneu-monia.25,26Considering the similar concept of exacer-bation of COPD and pneumonia,these observations cannot be explained in an integrated fashion.Triple therapy comprised of LAMA+LABA+ICS is often selected as the treatment option for advanced COPD. However,a meta-analysis suggested adding ICS to the dual-bronchodilator therapy is of little use.29 Despite the availability of guidelines,1some epide-miological studies suggested and warned that ICS is over prescribed for COPD patients.Price et al.ana-lysed24957patients in UK and found that in patients without concomitant asthma,53.7%of the total COPD population and50.2%of the GOLD Stage2 subset were receiving ICS.30Price et al.concluded that most COPD patients receive ICS irrespective of severity of air?ow limitation,asthma diagnosis and exacerbation history.30Drivenes et al.performed a cross-sectional study with149Norwegian patients with COPD treated by general practitioner and revealed that55.6%were treated with a combina-tion of LABA/ICS,which was concluded as overprescribing of LABA/ICS.31White et al.conducted an epidemiological study with3537COPD patients at41London general practices.Over-prescription of ICS in GOLD stage I or II(n=403,38%)and in GOLD III or IV without exacerbations(n=231,33.6%)were con?rmed.32

One limitation of the current meta-analysis is the limited number of included RCT for some compari-son.Another limitation is about inclusion criteria for this analysis.We did not set criteria for follow-up period,and no study observed patients for more than a half year.As COPD is a long-term condition,it is desirable to observe for longer duration.Similarly, crossover trials are not always the preferred methods in conducting COPD trials because carry-over effect may diminish the true impact of an intervention.We also included one open label trial,but is not also a desirable trial design.There is always a trade-off between the quality of inclusion criteria and the number of included studies.Therefore,we selected a currently feasible list of inclusion criteria.The favour-able results of LAMA+LABA were largely driven from the QVA149versus SFC comparison.Thus,internal validity was not fully ensured.Strong heterogeneity among studies and con?ict of interest of the included original studies also downgrade the overall credibility of this analysis.

In conclusion,the current meta-analysis indicated that,compared to LABA+ICS,LAMA+LABA may be a better option for treating COPD.Given the limita-tions of this study,we have to interpret result from this meta-analysis modestly.

Acknowledgements

We would like to thank GlaxoSmithKline for providing data for DB2114930,DB2114951and DB2116134.We would also thank Dr James F.Donohue,University of North Carolina for providing data for DB2114930and DB2114951.

REFERENCES

1Global Initiative for Chronic Obstructive Lung Disease(GOLD) The Global Strategy for the Diagnosis,Management and Preven-tion of COPD,2014.[Accessed20on October20th,2014.]Avail-able from URL:https://www.wendangku.net/doc/459223627.html,/

2Frampton JE.QVA149(indacaterol/glycopyrronium?xed-dose combination):a review of its use in patients with chronic obstructive pulmonary disease.Drugs2014;74:465–88.

3Schachter EN.Indacaterol/glycopyrronium bromide?xed-dose combination for the treatment of COPD.Drugs Today(Barc.) 2013;49:437–46.

4Nannini LJ,Poole P,Milan SJ,Holmes R,Normansell R.

Combined corticosteroid and long-acting beta2-agonist in one inhaler versus placebo for chronic obstructive pulmonary disease.Cochrane Database Syst.Rev.2013;11:CD003794.

5Hanania NA.The impact of inhaled corticosteroid and long-acting beta-agonist combination therapy on outcomes in COPD.

Pulm.Pharmacol.Ther.2008;21:540–50.

6Suissa S,Barnes PJ.Inhaled corticosteroids in COPD:the case against.Eur.Respir.J.2009;34:13–16.

7Malerba M,Morjaria JB,Radaeli A.Differential pharmacology and clinical utility of emerging combination treatments in the management of COPD–role of umeclidinium/vilanterol.Int J Chron Obstruct Pulmon Dis.2014;9:687–95.

8Scott LJ,Hair P.Umeclidinium/Vilanterol:?rst global approval.

Drugs2014;74:389–95.

9Vogelmeier CF,Bateman ED,Pallante J,Alagappan VK,D’Andrea P,Chen H,Banerji D.Ef?cacy and safety of once-daily QVA149 compared with twice-daily salmeterol-?uticasone in patients with chronic obstructive pulmonary disease(ILLUMINATE):a randomised,double-blind,parallel group https://www.wendangku.net/doc/459223627.html,ncet Respir.

Med.2013;1:51–60.

10Zhong N,Wang X,Zhou N,Zhang F,Patalano F,Humphries M, Wang L,Banerji D.Ef?cacy and safety of once-daily QVA149 compared with twice-daily salmeterol/?uticasone combination (SFC)in patients with COPD:the LANTERN study.Int.J.Chron.

Obstruct.Pulmon.Dis.2015;10:1015–26.

11GlaxoSmithKline.A study to compare the ef?cacy and safety of umeclidinium/vilanterol and?uticasone propionate/salmeterol in subjects with chronic obstructive pulmonary disease(COPD):

N Horita et al.

1158

(NCT01817764/DB2114930).2014.[Accessed27Dec2014.] Available from URL:https://https://www.wendangku.net/doc/459223627.html,/ct2/show/ NCT01817764

12GlaxoSmithKline.A study to compare the ef?cacy and safety of umeclidinium/vilanterol with?uticasone propionate/ salmeterol in subjects with chronic obstructive pulmonary disease(COPD):(NCT01879410/DB2114951).2014.[Accessed27 Dec2014.]Available from URL:https://https://www.wendangku.net/doc/459223627.html,/ct2/ show/NCT01879410.

13GlaxoSmithKline.A study to evaluate the ef?cacy and safety of umeclidinium bromide/vilanterol compared with?uticasone propionate/salmeterol over12weeks in subjects with chronic obstructive pulmonary disease(COPD)(NCT01822899/ DB2116134).2014.[Accessed27Dec2014.]Available from URL: https://https://www.wendangku.net/doc/459223627.html,/ct2/show/NCT01822899

14Hoshino M,Ohtawa J,Akitsu https://www.wendangku.net/doc/459223627.html,parison of airway dimen-sions with once daily tiotropium plus indacaterol versus twice daily Advair in chronic obstructive pulmonary disease.Pulm.

Pharmacol.Ther.2014;14:pii:S1094–5539,00091-1.European Respiratory Society,abstract700090;Session281;Date:Septem-ber82014.

15Magnussen H,Paggiaro P,Schmidt H,Kesten S,Metzdorf N, Maltais F.Effect of combination treatment on lung volumes and exercise endurance time in COPD.Respir.Med.2012;106:1413–

20.

16Rabe KF,Timmer W,Sagkriotis A,Viel https://www.wendangku.net/doc/459223627.html,parison of a com-bination of tiotropium plus formoterol to salmeterol plus ?uticasone in moderate COPD.Chest2008;134:255–62.

17Donohue JF,Worsley S,Zhu CQ,Hardaker L,Church A.Improve-ments in lung function with umeclidinium/vilanterol versus ?uticasone propionate/salmeterol in patients with moderate-to-severe COPD and infrequent exacerbations.Respir.Med.2015;

109:870–81.

18Chalmers TC,Smith H Jr,Blackburn B,Silverman B,Schroeder B, Reitman D,Ambroz A.A method for assessing the quality of a randomized control trial.Control.Clin.Trials1981;2:31–49.

May.

19Higgins PJ,Green S.Cochrane Handbook for Systematic Reviews of Interventions(Version5.1.0).The Cochrane Collaboration, London.2011.[Accessed20Oct2014.](Available from URL: https://www.wendangku.net/doc/459223627.html,/front_page.htm).

20Higgins JP,Thompson SG,Deeks JJ,Altman DG.Measuring inconsistency in meta-analyses.BMJ2003;327:557–60.

21American Thoracic Society.Minimal Clinically Signi?cant Difference.2007.Available from URL:http://qol.thoracic .org/sections/measurement-properties/minimal-clinically

-signi?cant-difference.html

22Jones PW.St.George’s Respiratory Questionnaire:MCID.COPD 2005;2(Suppl.1):75–9.23Mahler DA,Witek TJ Jr.The MCID of the transition dyspnea index is a total score of one unit.COPD2005;2:99–103.

24Horita N,Kaneko T.Role of combined indacaterol and glycopyrronium bromide(QVA149)for the treatment of COPD in Japan.Int.J.Chron.Obstruct.Pulmon.Dis.2015;10:813–22.

25Calverley PM,Anderson JA,Celli B,Ferguson GT,Jenkins C, Jones PW,Yates JC,Vestbo J.TORCH investigators.Salmeterol and?uticasone propionate and survival in chronic obstructive pulmonary disease.N.Engl.J.Med.2007;356:775–89.

26Yang IA,Clarke MS,Sim EH,Fong KM.Inhaled corticosteroids for stable chronic obstructive pulmonary disease.Cochrane Database Syst.Rev.2012;(7):CD002991.

27Global Initiative for Asthma(GINA)Global Strategy For Asthma Management and Prevention.2014.[Accessed1Nov2014.]Avail-able from URL:https://www.wendangku.net/doc/459223627.html,/local/uploads/?les/ GINA_Report_2014_Aug12.pdf

28Global Initiative for Asthma.Diagnosis of Diseases of Chronic Air?ow Limitation:Asthma COPD and Asthma-COPD Overlap Syndrome(ACOS).2014.[Accessed1Nov2014.]Avail-able from URL:https://www.wendangku.net/doc/459223627.html,/local/uploads/?les/ AsthmaCOPDOverlap.pdf

29Horita N,Kaneko T.Triple therapy vs.dual bronchodilator therapy for chronic obstructive pulmonary disease:is it worth the cost?Respir.Investig.2015;53:173–5.

30Price D,West D,Brusselle G,Gruffydd-Jones K,Jones R, Miravitlles M,Rossi A,Hutton C,Ashton VL,Stewart R et al.Man-agement of COPD in the UK primary-care setting:an analysis of real-life prescribing patterns.Int J Chron Obstruct Pulmon Dis.

2014;9:889–904.

31Drivenes E,Ostrem A,Melbye H.Predictors of ICS/LABA pre-scribing in COPD patients:a study from general practice.BMC Fam.Pract.2014;15:42.

32White P,Thornton H,Pinnock H,Georgopoulou S,Booth HP.

Overtreatment of COPD with inhaled corticosteroids–implica-tions for safety and costs:cross-sectional observational study.

PLoS ONE2013;8:e75221.

Supplementary Information

Additional Supplementary Information can be accessed via the html version of this article at the publisher’s website.

Appendix S1Search formula used for PubMed.

Figure S1Forrest plots to compare long-acting muscarinic antagonist(LAMA)+long-acting beta agonist(LABA)and LABA+inhaled corticosteroids(ICS).IV,inverse variance.M–H, Mantel–Haenszel;SE,standard error.

COPD inhaler medication:A meta-analysis1159