COPD : a stepwise or a hit hard approach?

www.revportpneumol.org

SPECIAL

ARTICLE

COPD:

A

stepwise

or

a

hit

hard

approach?

A.J.

Ferreira

_,

_A.

_Reis

_,

_N.

_Marc

_¸al

_,

_P.

_Pinto

_,

_C.

_Bárbara

_,

_on

_behalf

_of

_the

_GI

DPOC-Grupo

de

Interesse

na

Doenc

¸a

Pulmonar

Obstrutiva

Crónica

a_{PulmonologyDepartment,CentroHospitalarUniversitáriodeCoimbra,Portugal} b_{FacultyofMedicine,UniversityofCoimbra,Portugal}

c_{PulmonologyDepartment,CentroHospitalarTondela-Viseu,EPE,Portugal} d_{PulmonologyDepartment,HospitaldeVilaFrancadeXira,Portugal} e_{ChestDepartment,CentroHospitalarLisboaNorte,Lisbon,Portugal}

f_{EnvironmentalHealthInstitute(ISAMB),FacultyofMedicine,UniversityofLisbon,Portugal}

Received22August2015;accepted27December2015 Availableonline27February2016

KEYWORDS COPD; Stepwise; Hithard; Step-up; ICSwithdrawal; Bronchodilators; ICS

Abstract CurrentguidelinesdifferslightlyontherecommendationsfortreatmentofChronic ObstructivePulmonaryDisease(COPD)patients,andalthoughtherearesomeundisputed recom-mendations,thereisstilldebateregardingthemanagementofCOPD.Oneofthehindrancesto decidingwhichtherapeuticapproachtochooseislatediagnosisormisdiagnosisofCOPD.After aproperdiagnosisisachievedandseverityassessed,thechoicebetweenastepwiseor‘‘hit hard’’approachhastobemade.ForGOLDApatientsthestepwiseapproachisrecommended, whilstforB,CandDpatientsthisremainsdebatable.Moreover,inpatientsforwhominhaled corticosteroids(ICS)arerecommended,astep-upor‘‘hithard’’approachwithtripletherapy willdependonthepatient’scharacteristicsand,forpatientswhoarebeingover-treatedwith ICS,ICSwithdrawalshouldbeperformed,inordertooptimizetherapyandreduceexcessive medications.

Thispaperdiscussesandproposesstepwise,‘‘hithard’’,step-upandICSwithdrawal ther-apeutic approaches for COPD patients based on their GOLD group. We conclude that all approacheshavebenefits,andonlyacarefulpatientselectionwilldeterminewhichapproach isbetter,andwhichpatientswillbenefitthemostfromeachapproach.

©2016SociedadePortuguesadePneumologia.PublishedbyElsevierEspa˜na,S.L.U.Thisisan openaccessarticleundertheCCBY-NC-NDlicense( http://creativecommons.org/licenses/by-nc-nd/4.0/).

∗_{Correspondingauthor.}

E-mailaddresses:[email protected],[email protected](C.Bárbara).

http://dx.doi.org/10.1016/j.rppnen.2015.12.012

2173-5115/©2016SociedadePortuguesadePneumologia.PublishedbyElsevierEspa˜na,S.L.U.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

Currentguidelines differslightly ontherecommendations for treatment of Chronic Obstructive Pulmonary Disease (COPD)patients,mainlybecausepatientstratificationisnot consensual across guidelines.1---5 _{Although there are some}

undisputed recommendations, such as smoking cessation, physicalactivityprograms,andinfluenzaandpneumococcal vaccination, there is still debate regarding the manage-ment of COPD.6---12 _{The therapeutic approachproposedby}

the GlobalInitiative for Chronic ObstructiveLung Disease (GOLD), and based solely on the GOLD classification of COPD,2 _{is not entirely satisfactory, given the variability}

withinGOLDgroups,namelyregardinghospitalizationsand mortality.13_{However,therapyhastobebasedonsome}

clas-sificationsystem, and the GOLD classification is themost widelyaccepted,evenwithitscaveats.

One of the hindrances to deciding which therapeutic approach to choose is late diagnosis or misdiagnosis of COPD.Patientswhoarenotdiagnosed atthe earlystages of the disease cannot receive the early treatment which hasbeen showntobebeneficial.6,11,14 _{On theotherhand,}

patientsmisdiagnosedwithasthmaorAsthma-COPDoverlap syndrome (ACOS), will be overtreated with inhaled corti-costeroids (ICS), andare likelytosee no improvementin theirsymptomburden.Infact,tworecentanalysesshowed that,incurrent clinicalpractice,ICSarebeingprescribed inappropriately,15,16_{andthatthousandsofpatientsmaybe}

overtreated.

After a proper diagnosis is achieved, and severity assessed,thechoiceforastepwiseor‘‘hithard’’approach has to be made, and if for GOLD A patients the step-wise approachis recommended,2 _{for B,C and D patients}

this remains debatable.17 _{The argumentfor the stepwise}

approach is to not overtreat patients, but some patients may benefit from a ‘‘hit hard’’ approach, with the aim of maximal bronchodilation.12,13,18---20 _{In patients who will}

benefit from dual bronchodilation, a long-acting mus-carinic antagonist/long-acting beta-agonist (LAMA/LABA) fixed-dose combination is advantageous.11,12,21---24 _{Also, in}

patients for whom ICS is recommended, a step-up or ‘‘hit hard’’ approach with triple therapy will depend on the patient’s characteristics.2,4,5,17 _{For patients who are}

beingovertreatedwithICS,ICSwithdrawal shouldbe per-formed,inordertooptimizetherapyandreduceexcessive medications.21,22,25---28_{However,thisraisesanotherquestion:}

howtodecidewhenapatientisbeingovertreated? There arecurrentlynoreliableoraccuratebiomarkersofresponse totherapyanddiseaseprogression,sothedecision concern-ingICSwithdrawalmustbebasedontheavailableobjective testsandsubjectiveinstruments.2

ResultsfromarecentUKPrimaryCareSetting retrospec-tivestudyshowedthat,24monthsafterCOPDdiagnosisand prescriptionofinitialtherapy,severaltreatmentstrategies areused:switchin medication,stepwise, step-upandICS withdrawal,28_{suggestingthathereisanunmetclinicalneed}

torefinetherapybeyondGOLDandotherinternationaland nationalguidelines.

This paper discusses and proposes stepwise and ‘‘hit hard’’therapeuticapproachesfor COPDpatientsbasedon theirGOLDgroup.Analternativetreatmentapproach,based onphenotypes, isaddressed elsewhere.29 _{We suggest two}

subgroupsforGOLDAandGOLDBpatients,withdifferent therapeuticapproaches.Finally,weconcludethat,inCOPD, therapyshouldbetailoredtothepatient,takinginto consid-erationco-morbidities, presenceof hyperinflation,history ofchronic bronchitis,levelsof physicalactivity,and each individualpatientcharacteristics.

GOLD

A

patients

ItisdifficulttoidentifyasymptomaticGOLDApatientswith noexacerbations,giventhat theyhave noreasontoseek medicalhelp.Spirometricscreeningof asymptomatic indi-vidualsisnotsupportedbyevidence,althoughinindividuals over40 yearsold andwitha smokinghistory of>10 pack years,spirometrymaybeperformedwiththeaimofearly diagnosis.1 _{Indeed, some of these patients are identified}

duringscreenings,butmanyofthosewhoarenoteligiblefor screening(e.g.,non-smokers),mayremainundiagnosed.11

Also,thesepatientstendtounderestimatetheirsymptoms andadapttheirdailyactivitiesbyexerciseself-limitation,1

hence reporting to be asymptomatic. These unidentified patientscannotreceivetheearlytreatment,whichhasbeen showntobebeneficial.6,11,14

IdentificationofGOLDApatients

Besides screening, these patients aremainly identified in four situations: (a) in clinical visits for other causes or complaints;(b) whentheyaresubjected totests for non-respiratory reasons; (c) in the emergency room due to an acuteepisode; or (d) during pre-surgery testing. Once identified, it is imperative not to lose these patients to follow-up, as they will eventually evolve to other GOLD groupandtherapywillhavetobeadjusted.Acorrect diag-nosis is of the utmost importance, sinceit leads to both undertreatmentandovertreatment(e.g.,COPDdiagnosed asasthma).

Wesuggestanactivecase-findingapproachforthe iden-tificationofGOLDApatients.Wefurtherproposethatthese patients are flagged whenever they are diagnosed and, thereafter, that they aremanaged by their general prac-titioner,inclosecooperationwithapulmonologist.

RecommendedtherapeuticapproachforGOLDA patients

Given that these patients are often excluded from Ran-domizedClinicalTrials(RCTs),therearenosystematicdata availableonwhichtherapyshouldbeusedorhowtheywill respond.6_{Shouldtheybetreated?When?Withwhich}

med-ication and how? How will they progress with or without therapy?AstudybasedontheECLIPSEcohortshowedthat, at 3 yearsfollow-up, 57% of patients initially assigned to GOLDAremainedintheAgroup,whilsttheremaining43% progressedtootherGOLDgroups.13 _{Basedonthisstudy,all}

GOLDApatientsshouldreceivetreatment.

Current guidelines generally recommend for these patientssmokingcessation,physicalactivityprograms,and influenza and pneumococcal vaccination. Also, the use of a short-acting beta-agonist (SABA) or a short-acting

muscarinicantagonist(SAMA)asneededis mostly consen-sual, although LABA or LAMA may be used as alternative therapies.1---5_{Someauthorsspeculatethatearlyintervention}

with long-acting bronchodilators may improve patient-reportedoutcomes.11 _{However,onemajorissuewiththese}

patientsiscompliancetolong-termdrugs,6_andpatient

edu-cationisfundamentalindelayingCOPDevolution.

In GOLD A patients, both phenotype3,6,30 _and

co-morbidities1---5_{shouldbetakenintoaccountwhenchoosing}

betweenaLABAor aLAMA,withtheaimof achievingthe bestoutcomesanddelayingdiseaseprogression.BothLAMAs andLABAshave shown similarprofiles regardingFEV1 and dyspnea improvement, exercise tolerance, exacerbations reductionand safety.21,23,24,26,31---40 _{However, thereis some}

evidence that LAMAsmay delay lung function decline34,35

anddecreaseall-causemortality,34_andmaybemore

effec-tivethanLABAsinpreventingexacerbations.41_Despitethis,

andalthough LABAs donot seem toinfluence mortality,42

twostudies showed superiority in providing better symp-tomaticimprovementthanaLAMA.43,44 _{Thesedatasuggest}

thatLAMAsshouldbepreferredfor patientsat higherrisk ofexacerbations, whilstLABAswould bebetter for symp-tomatic control, although it depends on the individual clinicalresponseofeachpatient.IfaLABAischosen, inda-caterolmaybepreferabletoothercommerciallyavailable LABAssince,besideshavingalltheabovementioned advan-tagesofLABAs,itistheonlyonce-dailyLABA withstudies designedtoinvestigateexacerbations45,46_{andprovedableto}

reducethem,althoughinmoderatetoseverepatients.39,40

However,evidenceisstilltooscarcetoproposea recom-mendationbetweenaLABAandaLAMA.Also,thefinancial aspectshould notbe disregardedwhen choosingbetween a LABA or a LAMA, as some patients may not be able to affordsometherapiesandrespondwelltomoreaffordable alternatives.Dualbronchodilationisnotanoptionforthese patients.

Finally,giventheheterogeneityofCOPD,evenpatients classifiedasbelongingtotheAgroupshowalargevariability. WeproposethatgroupApatientsneedtobesub-divided intotwogroups,AX1andAX2,andthetherapeuticapproach shouldbebasedonthissubdivision---Table1.

Wefurtherrecommendthat,ifaLABA ischosen, inda-caterol may be preferable as it also reduces the rate of exacerbations.

We also recommend that therapy should be tailored to the patient, taking into consideration co-morbidities,

Table1 ProposeddivisionofGOLDApatientsintwo sub-groupsandrespectivetherapeuticapproaches.

Sub-groupcharacteristics Therapeuticapproach AX1:FEV1>80%;

noworseningofFEV1in annualassessment

SABAorSAMAonlySOS

AX2:50%<FEV1<80%; and/orworseningofFEV1in annualassessment

LABAorLAMA

FEV1---forcedexpiratoryvolumein1second;SABA---short act-ing␤2-agonist;SAMA---shortactingmuscarinicantagonist;LABA ---longacting␤2-agonist;LAMA---longactingmuscarinic antag-onist.

presence of hyperinflation, history of chronic bronchitis, levelsofphysicalactivityandadverseeffectsofeachdrug. Finally,allCOPDpatientsshouldbeofferedacompletely freesmokingcessationprogram,includingconsultationsand therapy.

GOLD

B

patients

Currentguidelinesgenerallyrecommendforthesepatients smokingcessation,physicalactivityprograms,influenzaand pneumococcalvaccination,andpulmonaryrehabilitation.1---5

All guidelinesagreethat bronchodilators arethe baseline therapy for all stages of COPD, but the choice of which bronchodilatortouseis lefttothephysician.12 _Also,most

guidelinesgenerallyrecommendastepwiseapproach, and dual bronchodilationonly when one bronchodilator is not sufficient to provide satisfactory symptom relief.12 _GOLD

recommends LAMA or LABA asthe firstchoice medication and LAMA+LABA asthe alternative choice.2 _{Ifindeed the}

choiceisastepwiseapproach,thenwhichlong-acting bron-chodilator should be used,a LABA or a LAMA? As already discussed above, evidenceis stilltoo scarce topropose a recommendation.

On the other hand, Agustiand Fabbri defend the‘‘hit hard’’ approach with dual bronchodilation for GOLD B patients,17 _{and there are several arguments in favor}

of this approach. Group B and C patients show a simi-lar risk of all-cause mortality,13 _{suggesting that a more}

aggressive treatment approach should be used in these patients. Also, many patients receiving long-acting bron-chodilatormonotherapy continuetoexperiencesignificant symptoms,18 _{and dual bronchodilation provides better}

symptomatic relief,12,23,24 _{improves FEV1 in patients with}

moderate-to-severe COPD,23,24,47 _{and improves health}

status.23_{Moreover,reductionofhyperinflation,asachieved}

with maximal (dual) bronchodilation, increases exercise tolerance,12,31 _{and higher levels of physical activity are}

associated with a better functional status48 _{and reduced}

riskofhospitalizationsandmortality,49_{evenatlevelsaslow}

as the equivalent to walking or cycling 2h/week.50 _Also,

it has been reported that objectively measured physical activityis thestrongestpredictorofall-causemortalityin patientswithCOPD.51 _{Therefore,itcanbespeculatedthat}

dual bronchodilation will have both short-and long-term beneficial effects in COPD patients. In patients who will benefit from dual bronchodilation,LAMA/LABA fixed-dose combinationsareexpectedtobecomethenewstandardin COPDtreatment.9_{InGOLD2or3patients,withorwithout}

ahistoryofexacerbations,dualbronchodilationwith once-daily indacaterol/glycopyrronium (IND/GLY) has clinically meaningful improvements in symptomatic parameters versus asalmeterol/fluticasonecombination (SFC)21,22 _and

tiotropium,23 _{and is superior totreatment with its}

mono-components, indacateroland glycopyrronium,23 _suggesting

theexistenceofsynergisticactivitybetweentheLABAand theLAMA.11 _{IND/GLYalsoprovidedsuperiorimprovements}

in patient-reported dyspnea and lung function versus placebo and tiotropium.24 _{Moreover, indacaterol and}

gly-copyrroniumshowaveryfastandlong-lasting(about24h) relaxation of airway smooth muscle.12 _{There are several}

potential advantages and beneficial effects of having a combinationofLABA+LAMAonthesamedevice.52

Table2 ProposeddivisionofGOLDBpatientsintwo sub-groupsandrespectivetherapeuticapproaches.

Sub-group characteristics Therapeuticapproach BX1:mMRC=2;AND FEV1>70%; ANDnocardiovascular co-morbidities a)ifnotmedicated, initiateLABAorLAMA

BX2:mMRC>2;OR FEV1<70%;

ORwithcardiovascular co-morbidities

LABA+LAMA(‘‘hithard’’ approach)+rehabilitation withexercisetraining mMRC---modifiedMedicalResearchCouncildyspneascale;FEV1 ---forcedexpiratoryvolumein1second;LABA---longacting_␤2 -agonist;LAMA---long-actingmuscarinicantagonist.

Finally,giventheheterogeneityofCOPD,evenpatients classifiedasbelongingtotheBgroupshowlargevariability. WeproposethatgroupBpatientsneedtobesub-divided intotwogroups,BX1andBX2,andthetherapeuticapproach shouldbebasedonthissubdivision---Table2.

We further recommend that, when the treatment of choice is dual bronchodilation, a combination of LABA+LAMAonthesamedeviceispreferable.

GOLD

C

and

D

patients

In addition to all general recommendations for GOLD B patients,1---5_{guidelinessuggest:LAMAand/orLABA,withor}

without ICS,1,2,4 _{withICS recommended for patients with}

frequentexacerbationsthatarenotadequatelycontrolled by long-acting bronchodilators;2,4 _{a stepwise or step-up}

approach, or immediatetriple therapy,depending on fre-quency of exacerbations;5 _{norecommendation for ICS on}

the non-exacerbator patient phenotype.3 _{Agusti & Fabbri}

proposeastepwiseorstep-upapproachdependingon dys-pnea and risk of exacerbations, respectively.17 _{A recent}

analysis showed that, in current clinical practice, how-ever, group D patients are more frequently prescribed tripletherapy,regardlessofpulmonaryfunctionandriskof exacerbations,16 _{which is contrary towhat the guidelines}

recommend.

Regardingdualbronchodilation,severalrandomized clin-icaltrialsinGOLD2to4patients,withorwithoutahistory of exacerbations, showed that dual bronchodilation with IND/GLYimprovessymptoms,21---24,37_{andpreventsmoderate}

tosevereCOPDexacerbations.37_IntheSHINE,23_BLAZE24_and

SPARK37 _{studies, patients already onICS therapy at}

base-line maintained the ICS during the study,whereas in the ILLUMINATE21 _andLANTERN22 _{studiespatients onICS}

ther-apybeforestudystartunderwentawashoutperiod.Taken together,thesedatasuggestthatdualbronchodilationisan appropriatetreatmentoptionforpatientswithsevereand verysevereCOPD.

CandDsubgroups

TherapeuticoptionshavetoconsiderthethreeCandD sub-groups:

Table 3 Proposed therapeutic approach for C and D patients.

C/DSub-group Therapeuticapproach

C1 LAMA

D1 LAMA+LABA

C2 LABA+ICS

D2 LABA+ICS+LAMA

C3,D3 LABA+ICS+LAMA

C1,D1---patientsathigh riskduetopoorfunction;C2,D2 ---patientsathighriskduetoexacerbations;C3,D3---patientsat highriskduetobothpoorfunctionandexacerbations;LABA ---long acting_␤2-agonist;LAMA---long-actingmuscarinic antago-nist;ICS-inhaledcorticosteroid.

- C1,D1(highriskduetopoorfunction) - C2,D2(highriskduetoexacerbations)

- C3,D3(highriskduetobothpoorfunctionand exacerba-tions)

We propose thatthe therapeutic approachis based on thesesubgroups---Table3.Whenthetreatmentofchoiceis dualbronchodilation,acombinationofLABA+LAMAonthe samedeviceispreferable.ForC1andD1patients,a step-wiseor‘‘hithard’’approachshouldbedecideddepending onsymptoms,withthe‘‘hithard’’approachrecommended in strongly symptomatic patients. For C2, D2, C3 and D3 patients,a step-up approachor immediatetriple therapy shouldbedecideddependingonthe frequencyof exacer-bations. We suggest the patient to be re-assessed every 3 months during a one year period after initiation of ICS (spirometry,theModifiedMedicalResearchCouncilDyspnea Scale(mMRC),theCOPDAssessmentTest(CAT), inflamma-tion,symptoms).Wefurtherrecommendthat,iftherewere noexacerbationsduring12months,COPDwasstable,and assessedparametersarewithintheexpectedrange, with-drawalofICScouldbeconsidered.

ICS

Although ICS are not indicated for patients without exacerbations,2---5,17 _{ICS/LABA or even triple therapy is}

widelyprescribedinreal-lifemanagementofCOPD,evenin patientswithmildormoderateCOPDseverity.BothGeneral Practitionersandspecialists in respiratorymedicine often use triple therapy even for patients who are not suffer-ingfrom severe COPD.8 _{Although the reasons for this are}

unclear,wespeculatethatthisismainlyduetothe gener-alizedideathatapatienttakingICSwillbemorecontrolled thanapatientwhoisnotonICStherapy,andwillnot exac-erbateordecompensate.Thisisnottrue.Infact,patients whodonotneedICStherapywillbeovermedicated,willnot benefitfromtripletherapy,andwillsufferallthepossible adverseeventsofICS,namelypneumonia.

A recent review from UK general practice showed that, in 2009, patients in all GOLD stages were receiv-ing triple therapy.15 _{The question of whether the less}

severepatientswerefrequentexacerbators,andthusbeing treatedaccordingtothecurrentguidelines,remained unan-swered.Anotherrecentanalysisconfirmedthat,incurrent clinical practice, ICS are indeed used inappropriately.16

These reports indicate that ICS prescription does not fol-lowthe current guidelinesandthousands of patientsmay beovertreated.

WithdrawalfromICS

Ifpatientsareindeedbeingovertreated,thentheywillnot onlybesubjectedtotheICS’numeroussideeffectsbutthey willalsonotbenefitfromtheICS,renderingtherisk/benefit ratio of the ICS too high to justify its use. A systematic review from 2011 on trials with withdrawal of ICS found noevidencethatwithdrawingpatients fromICSin routine practiceledtoimportantdeteriorationofoutcomes,namely frequencyofexacerbationsandexercisetolerance.Onlyone oftheincludedtrialsreportedasignificantdeclineinlung function.53 _{The ILLUMINATE}21 _{and the LANTERN}22 _studies

showedthat,inGOLD2and3patientswithorwithoutone moderateorsevere exacerbationintheprevious year,ICS canbesafelywithdrawn,andonce-daily IND/GLYprovided significant, sustained, and clinically meaningful improve-mentsinlungfunctionversustwice-dailySFC.Inaddition, IND/GLYprovidedsignificant symptomaticbenefitandwas superiortoSFC inachieving bronchodilationandreducing therateofexacerbations.AnotherstudyinwhichICS ther-apy was either switched or withdrawn, in patients with moderateto severe COPD,showed that ICS canbe safely discontinued,andtheadditionoffluticasone---salmeterolto tiotropiummayimprovelungfunctionanddecrease hospi-talizations,butdoesnotaffectratesofexacerbations.26_The

OPTIMOstudy,27 _{a real-lifestudy,showed thatICSmaybe}

withdrawn,providedappropriatetherapywith bronchodila-tors is maintained, but the WISDOM study54 _{showed that}

ICSwithdrawalhadnoeffectonexacerbationsbutledtoa decreaseinpulmonaryfunction.Thedifferentresultsfrom theOPTIMOandWISDOMstudiesmaylieinonesimplefact: inthe OPTIMO27 _{study patientshad an averageFEV1≈71%}

predicted,beingprobablynon-exacerbatorGOLDBpatients, andthereforedidnotneedICS,whilsttheWISDOM54 _study

patientshadmuchmoresevereairwaylimitation(FEV1≈34% predicted,GOLD3and4),probablyfittingthecriteriafor ICS use, and were thus affected by withdrawal. Another explanationfor theseresultsliesinthedesignofthe WIS-DOM study, where even patients whohad never been on ICS,receivedan ICS-containingtriple therapyfor 6weeks beforerandomization.Moreover,theOPTIMOstudy explic-itlyexcludespatientswitha‘‘historyofasthma’’,whilstthe WISDOMstudyexcludespatientswitha‘‘currentdiagnosis ofasthma’’,whichmakesitpossibleforpatientswithACOS tobeincludedintheWISDOMstudy.

A very recent UK Primary Care Setting retrospective studyshowsthat,innewlydiagnosedCOPDpatients, with-drawalfromICSiscommonafter12or24monthsoftherapy initiation.28 _{Unfortunately,and althoughincluded patients}

wereclassifiedasGOLD1to4,treatmentapproachesused werenotstratifiedbyGOLDstage.

Regardlessallavailableevidence,thereisageneralized concernamong physicians that, if a patientis withdrawn fromICS, that patient will decompensate or exacerbate. Even ifthepatient isre-assessed and does nothave indi-cationsfor ICS, the reluctance towithdraw remains. This concernstemsmainlyfromtheresultsoftheTORCH19,20,42

trial,whichconvincedthecommunityofphysiciansthatICS improves severaloutcomes,slows diseaseprogressionand reduces moderate-to-severeexacerbations, eveniffailing to show a beneficial effect of ICS on all-cause mortality rates.However,inlightofmorerecentdata,aspresented above,thereisnoevidence-basedreasonforsuchaconcern. Another hindrance to deciding whether to withdraw fromICSornotismisdiagnosis,e.g.COPDmisdiagnosedas asthma,orsuspicionoftheACOSphenotype.Theonlyway toovercomethishindranceistoattainaproperdiagnosis.

Finally, patients prefer to be withdrawn from ICS due to the general perception that corticosteroids are ‘‘dangerous’’ --- and indeed they can be, if not properly prescribed.

WerecommendthatAandBpatientswithout exacerba-tions,whoareovertreated withICS,shouldbewithdrawn from ICS in order to optimizetherapy and reduce exces-sivemedications,providedtheyarenotACOS,andkepton closesurveillance.CandDpatientswithoutexacerbations shouldalsobewithdrawnfromICS,withre-assessments rec-ommendedevery3monthsduringaoneyearperiod.

HowtowithdrawfromICS?

The two options are withdrawal of ICS with or without tapering off. The UK Primary Care Setting retrospective study does not specify how withdrawal from ICS was achieved.28 _{BoththeILLUMINATEstudy}21_{andtheLANTERN}

study22 _{mention a washout period of up to 7 days, but}

also do not specify if this period was with or without tapering off.Aaronetal.alsodonotspecifyhowICSwas discontinued.26 _{TheOPTIMOstudy}27 _{doesnotmentionhow}

ICS was withdrawn, but the WISDOMstudy54 _{does specify}

thatICSwastaperedoffovera12weekperiod.

Giventhelackofconcretedata,wecanonlyspeculate thatifapatientdoes notneedICS,then thereis noneed totaperoff.Inaddition,asitisanon-systemicmedication, thereisnoscientificrationalefortaperingoff.

Werecommendthatpatientswithnoexacerbations dur-ing12monthsandwhoarestableshouldbewithdrawnfrom ICS. Taperingoff is notnecessary andICS shouldbe with-drawnin asinglestep.Wesuggest GOLDAandBpatients bere-assessedevery6months,andGoldCandD patients every 3 months, duringa one year period afterICS with-drawal (spirometry, mMRC, inflammation, symptoms). We furtherrecommendthat,inCandDpatients,iftherewere exacerbations during 6 months, or if pulmonary function consistentlydecreases,thereshouldbeastep-upwithICS.

WhichpatientswillbenefitfromICS?

There is a need for biomarkers of response to therapy anddiseaseprogression, sothatthe momentof therapeu-ticadjustments canbedeterminedin atimelyway. Some biomarkers of disease activity and progression have been proposed,55,56 _{but much more research needs tobe done}

beforethese areclinicallyapplicable, and canguide per-sonalizedmanagementofCOPDpatients.Perhapsthemost promisingavailablemarkerissputumeosinophilia,andmost recentlyalsobloodeosinophilia.57 _{COPDpatientswith}

andsystemiccorticosteroids.10,58---61_{Asforcirculatinglevels}

ofC-reactive protein(CRP), resultsfromdifferentstudies arecontroversial,andCRPmaynotbeasuitablebiomarker inCOPD,duetoitslowspecificityandhighvariability.58

We recognize that there are currently no accurate biomarkerstoguidetherapyinCOPD.WeproposethatCAT shouldbeusedinallconsultations,inordertomonitor ther-apeuticresponse,givenitisapredictorofexacerbations.

ICSclass,doseandpneumonia

TheconsensusisthatICSuseincreasestheriskofpneumonia inpatientswithCOPD,1,2,4,20---22,36,62_{andevenarecentICS,}

fluticasonefuroate,hasbeenassociatedwithanincreased pneumonia risk and deaths from pneumonia in COPD patients.62---64_{However,somestudiesfindaverysmall}

differ-enceintheriskofpneumoniawithICS.21,22,26_These

dissimi-larresultsmaystemfromthefactthatsomestudiesaretoo shortfor patientstodeveloppneumonia,and/orthat sev-eralICSclassesanddosesareusedindifferentstudies.Both fluticasoneandbudesonidehavebeenreportedtoincrease the risk of pneumonia,64---67 _{but results concerning a dose}

effect range fromno difference between fluticasone and budesonide,65_{tofluticasonebeingassociatedwithahigher,}

dose dependent risk, when compared tobudesonide,66 _to

fluticasone not being dose dependent, while budesonide shows asignificantdifferencebetweenthe twocommonly used doses.67 _{The general dose-effect of ICS on}

pneumo-nia risk has been confirmed in a large USA retrospective cohortstudy,buttheuseofseveralICSclassesprecludesany conclusionregardingspecificICSclass-relateddose-effect.68

ResultsconcerningthetrueeffectofdifferentICSclasses and doses on the risk of pneumonia remain inconclusive. Morestudiesareneededtoallowanevaluationofwhether differentclassesofICSareassociatedwithdifferent pneu-moniariskinCOPDpatients.

Given thecontroversysurroundingICS therapyin COPD patients,wesuggestthatawithdrawalorstep-upapproach shouldtake into accounttheabove recommendationsbut betailored to thepatient,taking intoconsideration each individualpatientcharacteristics.

Conclusions

Both stepwise and ‘‘hit hard’’ approaches have bene-fits.Onlyacarefulpatientselection willdeterminewhich approachisbetter,andwhichpatientswillbenefitthemost fromeachapproach.InCOPD,therapyshouldbetailoredto thepatient,takingintoconsiderationco-morbidities, pres-enceofhyperinflation,historyofchronicbronchitis,levels ofphysicalactivity,andeachindividualpatient characteris-tics.

Ethical

responsibilities

Protection of human and animal subjects.The authors declarethatnoexperimentswereperformedonhumansor animalsforthisstudy.

Confidentialityofdata.Theauthorsdeclarethatnopatient dataappearinthisarticle.

Right to privacy and informed consent.The authors declarethatnopatientdataappearinthisarticle.

Conflict

of

interest

The authors declare collaborating and receiving fees frompharmaceuticalcompaniesotherthanNovartiseither through participation in advisory board or consultancy meetings, congress symposia, clinical trial conduct or investigator-initiatedtrials.

Role

of

funding

source

Fundingfor this paperwasprovidedby NovartisPortugal. Fundingwasusedtoaccessallnecessaryscientific bibliog-raphyandcovermeetingexpenses.NovartisPortugalhadno roleinthecollection,analysisandinterpretationofdata,in thewriting ofthepaperandinthedecisiontosubmitthe paperforpublication.

Acknowledgements

The authorswishtothank NovartisPortugal for the fund-ingfor thispaper,which wasusedtoaccessall necessary scientificbibliographyandcovermeetingexpenses.

References

1.Direcc¸ãoGeraldeSaúde.Normanr028/2011-Diagnóstico e TratamentodaDoenc¸aPulmonarObstrutivaCrónica.In: Portu-gal.2013.

2.GlobalInitiativeforChronicObstructiveLungDisease.Global Strategy for the Diagnosis, Management and Prevention of ChronicObstructivePulmonaryDisease(Revised2015);2015.

3.Miravitlles M, Soler-CatalunaJJ, CalleM, MolinaJ, Almagro P,QuintanoJA,etal.SpanishguidelineforCOPD(GesEPOC). Update2014.ArchBronconeumol.2014;50Suppl.1:1---16.

4.National Institute for Health and Care Excellence. Chronic obstructive pulmonary disease: Management of chronic obstructive pulmonary disease in adultsin primary and sec-ondary care(partialupdate);2010. http://www.nice.org.uk/ guidance/CG101

5.O’DonnellDE,AaronS,BourbeauJ,HernandezP,MarciniukDD, BalterM,etal.CanadianThoracicSocietyrecommendationsfor managementofchronicobstructivepulmonarydisease---2007 update.CanRespirJ.2007;14Suppl.B:5B---32B.

6.Rabe KF,WedzichaJA.Controversiesintreatmentofchronic obstructivepulmonarydisease.Lancet.2011;378:1038---47.

7.CazzolaM,SegretiA,RoglianiP.Comparativeeffectivenessof drugsforchronicobstructivepulmonarydisease.DrugsToday (Barc).2012;48:785---94.

8.CazzolaM,MateraMG.Triplecombinationsinchronic obstruc-tivepulmonarydisease---isthreebetterthantwo?ExpertOpin Pharmacother.2014;15:2475---8.

9.de Miguel-Diez J, Jimenez-Garcia R. Considerations for new dual-actingbronchodilatortreatments forchronicobstructive pulmonarydisease.ExpertOpinInvestigDrugs.2014;23:453---6.

10.LeighR,PizzichiniMM,MorrisMM,MaltaisF,HargreaveFE, Pizzi-chiniE.StableCOPD:predictingbenefitfromhigh-doseinhaled corticosteroidtreatment.EurRespirJ.2006;27:964---71.

11.PatalanoF,BanerjiD,D’AndreaP,FogelR,AltmanP,Colthorpe P.AddressingunmetneedsinthetreatmentofCOPD.EurRespir Rev.2014;23:333---44.

12.NardiniS,CamiciottoliG,LociceroS,MaselliR,PasquaF, Pas-salacqua G, et al. COPD: maximization of bronchodilation. MultidiscipRespirMed.2014;9:50.

13.Agusti A, Edwards LD, Celli B, Macnee W, Calverley PM, MullerovaH,etal.Characteristics,stabilityandoutcomesof the2011GOLDCOPDgroupsintheECLIPSEcohort.EurRespir J.2013;42:636---46.

14.DecramerM,CooperCB.TreatmentofCOPD:thesoonerthe better?Thorax.2010;65:837---41.

15.JamesGD,DonaldsonGC,WedzichaJA,NazarethI.Trendsin managementandoutcomesofCOPDpatientsinprimarycare, 2000---2009:aretrospectivecohortstudy.NPJPrimCareRespir Med.2014;24:14015.

16.VestboJ,VogelmeierC,SmallM,HigginsV.Understandingthe GOLD2011Strategyasappliedtoareal-worldCOPDpopulation. RespirMed.2014;108:729---36.

17.AgustiA,FabbriLM.InhaledsteroidsinCOPD:whenshouldthey beused?LancetRespirMed.2014;2:869---71.

18.DransfieldMT,BaileyW,CraterG,EmmettA,O’DellDM,Yawn B. Diseaseseverity and symptoms among patientsreceiving monotherapyforCOPD.PrimCareRespirJ.2011;20:46---53.

19.CelliBR, ThomasNE,AndersonJA, FergusonGT,JenkinsCR, JonesPW,etal.Effectofpharmacotherapyonrateofdecline of lung function in chronic obstructive pulmonary disease: results from the TORCH study. Am J Respir Crit Care Med. 2008;178:332---8.

20.JenkinsCR,JonesPW,CalverleyPM,CelliB,AndersonJA, Fer-gusonGT,etal.Efficacyofsalmeterol/fluticasonepropionateby GOLDstageofchronicobstructivepulmonarydisease:analysis fromtherandomised,placebo-controlledTORCHstudy.Respir Res.2009;10:59.

21.VogelmeierCF,BatemanED,PallanteJ,AlagappanVK,D’Andrea P,ChenH,etal.Efficacyandsafetyofonce-dailyQVA149 com-paredwithtwice-dailysalmeterol-fluticasoneinpatientswith chronic obstructive pulmonary disease (ILLUMINATE): a ran-domised,double-blind,parallelgroupstudy.LancetRespirMed. 2013;1:51---60.

22.ZhongN,WangC,ZhouX,ZhangN,PatalanoF,HumphriesM. Efficacyandsafetyofonce-dailyQVA149comparedwith twice-dailysalmeterol/fluticasonecombination(SFC)inpatientswith COPD:theLANTERNstudy.In:PosterpresentedattheEuropean RespiratorySocietyAnnualCongress.2014.

23.Bateman ED, Ferguson GT, BarnesN, Gallagher N, Green Y, HenleyM,etal.DualbronchodilationwithQVA149versus sin-gle bronchodilator therapy: the SHINE study. Eur Respir J. 2013;42:1484---94.

24.MahlerDA,DecramerM,D’UrzoA,WorthH,WhiteT,Alagappan VK,etal.DualbronchodilationwithQVA149reduces patient-reported dyspnoea in COPD: theBLAZE study. Eur RespirJ. 2014;43:1599---609.

25.De Coster DA,Jones M. Tailoringof corticosteroidsin COPD management.CurrRespirCareRep.2014;3:121---32.

26.AaronSD, Vandemheen KL,FergussonD,MaltaisF,Bourbeau J,GoldsteinR,etal.Tiotropiumincombinationwithplacebo, salmeterol,orfluticasone-salmeterolfortreatmentofchronic obstructivepulmonarydisease:arandomizedtrial.AnnIntern Med.2007;146:545---55.

27.RossiA,GuerrieroM,CorradoA.Withdrawalofinhaled corticos-teroidscanbesafeinCOPDpatientsatlowriskofexacerbation: areal-lifestudyontheappropriatenessoftreatmentin moder-ateCOPDpatients(OPTIMO).RespirRes.2014;15:77.

28.Wurst KE, Punekar YS, Shukla A. Treatmentevolution after COPD diagnosis in the UK primary care setting. PLOS ONE. 2014;9:e105296.

29.FragosoE,AndréS,Boleo-ToméJP,AreiasV,MunháJ,Cardoso J.UnderstandingCOPD:avisiononphenotypes,comorbidities andtreatmentapproach;2016[acceptedforpublicationinRev PortPneumol].

30.Hurst JR, Vestbo J, Anzueto A, Locantore N, Mullerova H, Tal-Singer R, et al. Susceptibilityto exacerbation inchronic obstructive pulmonary disease. N Engl J Med. 2010;363: 1128---38.

31.BeehKM,KornS,BeierJ,JadayelD,HenleyM,D’AndreaP,etal. Effect ofQVA149on lungvolumesand exercise tolerance in COPDpatients:theBRIGHTstudy.RespirMed.2014;108:584---92.

32.DecramerM,CelliB,KestenS,LystigT,Mehra S,TashkinDP. Effect oftiotropiumon outcomesin patientswithmoderate chronic obstructive pulmonary disease (UPLIFT): a prespeci-fiedsubgroupanalysisofarandomisedcontrolledtrial.Lancet. 2009;374:1171---8.

33.Johansson G, Lindberg A, Romberg K, Nordstrom L, Gerken F,RoquetA.Bronchodilatorefficacyoftiotropiuminpatients withmild to moderateCOPD. Prim CareRespir J. 2008;17: 169---75.

34.TashkinDP,CelliB,SennS,BurkhartD,KestenS,MenjogeS, etal.A4-yeartrialoftiotropiuminchronicobstructive pul-monarydisease.NEnglJMed.2008;359:1543---54.

35.TroostersT, Celli B,Lystig T, KestenS,Mehra S, Tashkin DP, etal.TiotropiumasafirstmaintenancedruginCOPD:secondary analysisoftheUPLIFTtrial.EurRespirJ.2010;36:65---73.

36.WedzichaJA, Calverley PM,Seemungal TA, Hagan G, Ansari Z, Stockley RA. The prevention of chronic obstructive pulmonary disease exacerbations by salmeterol/fluticasone propionateortiotropiumbromide.AmJRespirCritCareMed. 2008;177:19---26.

37.WedzichaJA,DecramerM,Ficker JH,NiewoehnerDE, Sand-strom T, Taylor AF, et al. Analysis of chronic obstructive pulmonarydiseaseexacerbationswiththedualbronchodilator QVA149comparedwithglycopyrroniumandtiotropium(SPARK): arandomised,double-blind,parallel-groupstudy.LancetRespir Med.2013;1:199---209.

38.RocheN,ChanezP.BronchodilatorcombinationsforCOPD:real hopesoranewPandora’sbox?EurRespirJ.2013;42:1441---5.

39.WedzichaJA,BuhlR,LawrenceD,YoungD.Monotherapywith indacaterol once daily reducesthe rate of exacerbations in patientswithmoderate-to-severeCOPD:post-hocpooled anal-ysisof6monthsdatafromthreelargephaseIIItrials.Respir Med.2015;109:105---11.

40.DecramerML,Chapman KR, DahlR, FrithP,DevouassouxG, Fritscher C,et al. Once-daily indacaterol versus tiotropium forpatientswithseverechronicobstructivepulmonarydisease (INVIGORATE): a randomised, blinded, parallel-group study. LancetRespirMed.2013;1:524---33.

41.Vogelmeier C, Hederer B, Glaab T, Schmidt H, Rutten-van MolkenMP, BeehKM,et al.Tiotropiumversussalmeterolfor the prevention of exacerbations of COPD. N Engl J Med. 2011;364:1093---103.

42.CalverleyPM,Anderson JA, CelliB, FergusonGT, JenkinsC, JonesPW,etal.Salmeterolandfluticasonepropionateand sur-vivalinchronicobstructivepulmonarydisease.NEngl JMed. 2007;356:775---89.

43.Buhl R, Dunn LJ, Disdier C, Lassen C, Amos C, Henley M, etal. Blinded 12-weekcomparison ofonce-dailyindacaterol andtiotropiuminCOPD.EurRespirJ.2011;38:797---803.

44.Donohue JF, Fogarty C,Lotvall J, MahlerDA, Worth H, Yor-gancioglu A, et al. Once-daily bronchodilators for chronic obstructivepulmonarydisease:indacaterolversustiotropium. AmJRespirCritCareMed.2010;182:155---62.

45.FergusonGT, FeldmanGJ, Hofbauer P, HamiltonA, Allen L, KorduckiL,etal.Efficacyandsafetyofolodateroloncedaily delivered via Respimat(R) in patients withGOLD 2-4 COPD: resultsfromtworeplicate48-weekstudies.IntJChronObstruct PulmonDis.2014;9:629---45.

46.ZuWallackR,AllenL,HernandezG,TingN,AbrahamsR. Effi-cacy and safety of combining olodaterol Respimat((R)) and tiotropiumHandiHaler((R)) inpatientswith COPD:resultsof

tworandomized,double-blind,active-controlledstudies.IntJ ChronObstructPulmonDis.2014;9:1133---44.

47.MahlerDA,D’UrzoA,BatemanED,OzkanSA,WhiteT,PeckittC, etal.Concurrentuseofindacaterolplustiotropiuminpatients withCOPD provides superiorbronchodilation compared with tiotropiumalone:arandomised,double-blindcomparison. Tho-rax.2012;67:781---8.

48.Garcia-AymerichJ, SerraI, GomezFP,Farrero E,Balcells E, RodriguezDA,etal.Physicalactivityandclinicalandfunctional statusinCOPD.Chest.2009;136:62---70.

49.Garcia-AymerichJ,LangeP,SerraI,SchnohrP,AntoJM. Time-dependentconfoundinginthestudyoftheeffectsofregular physicalactivityinchronicobstructivepulmonarydisease:an applicationofthemarginalstructural model.AnnEpidemiol. 2008;18:775---83.

50.Garcia-AymerichJ,LangeP,BenetM,SchnohrP,AntoJM. Reg-ularphysicalactivityreduceshospitaladmissionandmortality inchronicobstructivepulmonary disease:apopulationbased cohortstudy.Thorax.2006;61:772---8.

51.WaschkiB,KirstenA,HolzO,MullerKC,MeyerT,WatzH,etal. Physicalactivityisthestrongestpredictorofall-cause mortal-ityinpatientswithCOPD:a prospectivecohortstudy.Chest. 2011;140:331---42.

52.Ferreira J, Drummond M, Pires N, Reis G, Alves C, Robalo-Cordeiro C. Optimal treatment sequence in COPD: can a consensusbefound?RevPortPneumol.2016;22:39---49.

53.NadeemNJ,TaylorSJ,EldridgeSM.Withdrawalofinhaled cor-ticosteroidsinindividualswithCOPD---asystematicreviewand commentontrialmethodology.RespirRes.2011;12:107.

54.MagnussenH,DisseB,Rodriguez-RoisinR,KirstenA,WatzH, Tetzlaff K, et al. Withdrawal of inhaled glucocorticoids and exacerbationsofCOPD.NEnglJMed.2014;371:1285---94.

55.ShawJG,VaughanA,DentAG,O’HarePE,GohF,BowmanRV, et al. Biomarkers of progression of chronic obstructive pul-monarydisease(COPD).JThoracDis.2014;6:1532---47.

56.JiJ, von Scheele I, BergstromJ, Billing B, DahlenB, Lantz AS, et al. Compartment differences of inflammatory activ-ity in chronic obstructive pulmonary disease. Respir Res. 2014;15:104.

57.PascoeS, LocantoreN, DransfieldMT, BarnesNC,Pavord ID. Bloodeosinophil counts, exacerbations,and response to the additionofinhaledfluticasonefuroatetovilanterolinpatients with chronic obstructive pulmonary disease: a secondary

analysisofdatafromtwoparallelrandomisedcontrolledtrials. LancetRespirMed.2015;3:435---42.

58.Lock-JohanssonS, Vestbo J,SorensenGL.Surfactantprotein D, Clubcell protein16, Pulmonary and activation-regulated chemokine. C-reactive protein, and Fibrinogen biomarker variation in chronic obstructive lung disease. Respir Res. 2014;15:147.

59.Siva R, Green RH, Brightling CE, Shelley M, Hargadon B, McKennaS,etal.Eosinophilicairwayinflammationand exac-erbationsofCOPD:arandomisedcontrolledtrial.EurRespirJ. 2007;29:906---13.

60.BrightlingCE,MonteiroW,WardR,ParkerD,MorganMD, Ward-lawAJ,etal.Sputumeosinophiliaandshort-termresponseto prednisoloneinchronicobstructivepulmonarydisease:a ran-domisedcontrolledtrial.Lancet.2000;356:1480---5.

61.BrightlingCE,McKennaS,HargadonB,BirringS,GreenR,Siva R,etal.Sputumeosinophiliaandtheshorttermresponseto inhaledmometasoneinchronicobstructivepulmonarydisease. Thorax.2005;60:193---8.

62.ErnstP,SaadN,SuissaS.InhaledcorticosteroidsinCOPD:the clinicalevidence.EurRespirJ.2015;45:525---37.

63.DransfieldMT,BourbeauJ,JonesPW,HananiaNA,MahlerDA, Vestbo J, et al. Once-daily inhaled fluticasone furoate and vilanterolversusvilanterolonlyforpreventionofexacerbations of COPD: two replicate double-blind, parallel-group, ran-domisedcontrolledtrials.LancetRespirMed.2013;1:210---23.

64.McKeage K. Fluticasone furoate/vilanterol: a review of its use in chronic obstructive pulmonary disease. Drugs. 2014;74:1509---22.

65.NanniniLJ,LassersonTJ,PooleP.Combinedcorticosteroidand long-acting beta(2)-agonist inone inhaler versus long-acting beta(2)-agonistsforchronicobstructivepulmonarydisease.CDS Rev.2012;9:CD006829.

66.Suissa S, Patenaude V, Lapi F, Ernst P. Inhaled corticos-teroids in COPD and the riskof serious pneumonia. Thorax. 2013;68:1029---36.

67.Kew KM, Seniukovich A. Inhaled steroids and risk of pneu-monia for chronic obstructive pulmonary disease. CDS Rev. 2014;3:CD010115.

68.YawnBP,LiY,TianH,ZhangJ,ArconaS,KahlerKH.Inhaled cor-ticosteroiduseinpatientswithchronicobstructivepulmonary diseaseandtheriskofpneumonia:aretrospectiveclaimsdata analysis.IntJChronObstructPulmonDis.2013;8:295---304.