Spontaneous Breathing Trials With T-Piece or Pressure ...

Once a patient is deemed ready to breathe spontaneously, a screening test, called a spontaneous breathing trial (SBT), is usually performed, ... Skiptomaincontent ReviewArticleReviews SpontaneousBreathingTrialsWithT-PieceorPressureSupportVentilation JoséAugustoSPellegrini,RafaelBMoraes,JuçaraGMaccari,RoselainePdeOliveira,AugustoSavi,RodrigoARibeiro,KarenEABurnsandCassianoTeixeira RespiratoryCareDecember2016,61(12)1693-1703;DOI:https://doi.org/10.4187/respcare.04816 JoséAugustoSPellegriniDivisionofCriticalCare,HospitalMoinhosdeVento(HMV),PortoAlegre,RS,Brazil.DivisionofCriticalCare,HospitaldeClínicasdePortoAlegre(HCPA),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteForcorrespondence: [email protected]RafaelBMoraesDivisionofCriticalCare,HospitaldeClínicasdePortoAlegre(HCPA),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteJuçaraGMaccariDivisionofCriticalCare,HospitalMoinhosdeVento(HMV),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteRoselainePdeOliveiraDivisionofCriticalCare,HospitalMoinhosdeVento(HMV),PortoAlegre,RS,Brazil.MedicalSchool,UniversidadeFederaldeCiênciasdaSaúdedePortoAlegre(UFCSPA),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteAugustoSaviDivisionofCriticalCare,HospitalMoinhosdeVento(HMV),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteRodrigoARibeiroInstituteforEducationandResearch,HMV,GraduatePrograminEpidemiology,UniversidadeFederaldoRioGrandedoSul(UFRGS),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteKarenEABurnsInterdepartmentalDivisionofCriticalCareMedicine,St.Michael'sHospitalandtheUniversityofTorontoandtheLiKaShingKnowledgeInstitute,Toronto,Canada.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissiteCassianoTeixeiraDivisionofCriticalCare,HospitalMoinhosdeVento(HMV),PortoAlegre,RS,Brazil.MedicalSchool,UniversidadeFederaldeCiênciasdaSaúdedePortoAlegre(UFCSPA),PortoAlegre,RS,Brazil.FindthisauthoronGoogleScholarFindthisauthoronPubMedSearchforthisauthoronthissite ArticleFigures&DataReferencesInfo&MetricsPDF AbstractSpontaneousbreathingtrials(SBTs)areamongthemostcommonlyemployedtechniquestofacilitateweaningfrommechanicalventilation.ThepreferredSBTtechnique,however,isstillunclear.ToclarifythepreferableSBT(T-pieceorpressuresupportventilation[PSV]),weconductedthissystematicreview.WethensearchedtheMEDLINE,EMBASE,SciELO,GoogleScholar,CINAHL,ClinicalTrials.gov,andCochraneCENTRALdatabasesthroughJune2015,withoutlanguagerestrictions.WeincludedrandomizedcontrolledtrialsinvolvingadultsubjectsbeingweanedfrommechanicalventilationcomparingT-piecewithPSVandreporting(1)weaningfailure,(2)re-intubationrate,(3)ICUmortality,or(4)weaningduration.Anticipatingclinicalheterogeneityamongtheincludedstudies,wecomparedprespecifiedsubgroups:(1)simple,difficult,orprolongedweaningand(2)subjectswithCOPD.WesummarizedthequalityofevidenceforinterventioneffectsusingtheGRADE(GradingofRecommendationsAssessment,Development,andEvaluation)methodology.Weidentified3,674potentiallyrelevantstudiesandreviewed23papersinfull.Twelvestudies(2,161subjects)metourinclusioncriteria.Overall,theevidencewasofverylowtolowquality.SBTtechniquedidnotinfluenceweaningsuccess(riskratio1.23[0.94–1.61]),ICUmortality(riskratio1.11[0.80–1.54]),orre-intubationrate(riskratio1.21[0.90–1.63]).PrespecifiedsubgroupanalysissuggestedthatPSVmightbesuperiortoT-piecewithregardtoweaningsuccessforsimple-to-weansubjects(riskratio1.44[1.11–1.86]).Fortheprolonged-weaningsubgroup,however,T-piecewasassociatedwithashorterweaningduration(weightedmeandifference−3.08[−5.24to−0.92]d).Inconclusion,low-qualityevidenceisavailableconcerningthistopic.PSVmaybeassociatedwithlowerweaningfailureratesinthesimple-to-weansubgroup.Incontrast,inprolonged-weaningsubjects,T-piecemayberelatedtoashorterweaningduration,althoughthisisathighriskofbias.Furtherstudyofthedifficult-to-weanandCOPDsubgroupsisrequired.weaningmechanicalventilationcriticalcareIntroductionSuccessfulweaningofpatientsfrommechanicalventilationconstitutesoneofthemostchallengingtasksforICUpractitioners.Timelyidentificationofpatientswhoarecapableofspontaneousbreathingcanshortenthemechanicalventilationdurationandpotentiallyreducemechanicalventilation-relatedcomplications.1–5Onceapatientisdeemedreadytobreathespontaneously,ascreeningtest,calledaspontaneousbreathingtrial(SBT),isusuallyperformed,althoughtheliteratureremainsconflictedonthissubject.6–10AnSBTistypicallyperformedbydisconnectingthepatientfromtheventilatorandattachingaT-piecetotheendotrachealtube.11Someclinicians,however,prefertouselowlevelsofpressuresupportventilation(PSV),orautomatictubecompensation.8Switchingfromcontinuousmandatoryventilationtospontaneousbreathingcandecreaseleft-ventricularperformanceandunmasklatentleft-ventricularheartfailure.ConcernsexistregardingthepotentialforSBTfailureratestobehigherwithT-pieceSBTsthanwithlowlevelsofPSV,possiblybecauseoftheincreasedexpenditureofrespiratorymuscleenergy12andcardiogenicpulmonaryedemasecondarytotheMüllermaneuver.13AlthoughPSVmaybealessdemandingSBTwithregardtorespiratorymuscleeffortandhydrostatichomeostasis,especiallywiththeadditionofPEEPtopreventthedevelopmentofleft-ventricularheartfailure,itmayalsodulltheclinicalpictureofintolerancecomparedwiththatofunassistedT-pieceSBTs.13Manytrialshavepreviouslyassessedthisquestion,althoughheterogeneousmethodologicalaspectsandconflictingresultslimitadequateevidenceappraisal.Previousmeta-analyseshavebeenconductedinthisfieldbuthavenotdirectlycomparedSBTsorusedupdatedinformation.14ObjectivesOurobjectivewastoclarifythepreferredSBTtechnique(T-pieceorlowlevelsofPSV)forcriticallyillpatientsweaningfrommechanicalventilationaccordingtoICUmortality,re-intubationrates,weaningfailure,andweaningduration.MethodsWeconductedasystematicreviewbasedonstandardmethodsandreportedourfindingsinaccordancewiththePreferredReportingItemsforSystematicReviewsandMeta-Analysesstatement.15DataSourcesandSearchesWeaimedtoidentifyallrandomizedcontrolledtrials(RCTs)assessingtheefficacyandoutcomesofT-piececomparedwithPSVtrialsinadultpatientsweaningfrominvasivemechanicalventilation.WeconductedelectronicsearchesoftheMEDLINE,EMBASE,CochraneCENTRAL,CINAHL,SciELO,GoogleScholar,andClinicalTrials.govdatabasesforstudiesactivelyrecruitingpatients.AlldatabasesweresearchedfromtheirinceptiontoJune2015.OurMEDLINEsearchincludedthefollowingtermsandkeywords:(“weaning”OR“VentilatorWeaning”[MeSH])AND(“MechanicalVentilation”OR“Respiration,Artificial”[MeSH])AND(“spontaneousbreathingtrial”OR“T-piece”OR“ttube”),usingtheRobinsonandDickersinRCTfilterforPubMed.16TheelectronicsearchstrategyappliedstandardfiltersfortheidentificationofRCTsfromeachdatabase.Wescreenedthereferencelistsofretrievedpublicationsforpotentiallyeligibletrials.Wedidnotapplylanguagerestrictions.StudySelectionWerestrictedouranalysistoRCTsaimingtolimitpotentialsourcesofbias.Weexcludedcrossovertrialsandquasi-randomizedtrials.Regardlessofspecificweaningprotocols,includedtrialshadtocomparebetweenT-pieceandPSVforconductingSBTs.WeconsideredT-pieceSBTtobetheprocedureoftemporarilydisconnectingapatientfromtheventilatorwhilemaintaininganexternaloxygensupply,commonlybyusingaT-piececonnectedtotheendotrachealtube.PSVwasconsideredtobeanSBTwhenemployedinasystematicfashion,followingapredefinedprotocolspecificallydesignedtoidentifypatientsforextubationor,inthecaseoftracheostomizedpatients,fordefinitiveremovalfrommechanicalventilation.OutcomeMeasuresTheoutcomesassessedincluded(1)ICUmortality,(2)rateofre-intubationwithin48hfollowingextubation,(3)weaningfailureprecludingextubation,and(4)weaningduration.Weusedauthors'definitionsforthepost-randomizationweaningduration.DataExtractionandQualityAssessmentTwoindependentreviewers(JPandRM)screenedthetitlesandabstractsofretrievedcitationsandthefulltextsofpotentiallyeligiblestudiestoidentifytrialsthatmetourinclusioncriteria.Datafromeachpotentiallyrelevanttrialwereindependentlyextractedbythereviewersusingapredefineddataextractionform.AccordingtotheCochraneriskofbiastool,weappraisedtheadequacyofrandomsequencegeneration,thereportingofallocationconcealment,theblindingofparticipantsandoutcomeassessments,andthedescriptionsoflossestofollow-upandexclusions;westillassessedadherencetotheintention-to-treatprinciple.Wesolveddisagreementsbyconsultingathirdreviewer(CT)whenneeded.DataSynthesisandAnalysisQualitativeAnalysis.WeusedanarrativesummaryapproachtoqualitativelydescribethestudycharacteristicsandvariationsinqualityindicatorsandtoconsiderhowthesefactorsaffectedourunderstandingoftheoutcomesoftheincludedRCTs.QuantitativeAnalysis.WeusedtheCochraneCollaborationguidelinestoconductourmeta-analysis.17AllstatisticalanalyseswereperformedwithReviewManagerversion5.3(NordicCochraneCenter,Copenhagen,Denmark),theCochraneCollaboration'ssoftwareforpreparingandupdatingCochranesystematicreviews.Weexpressedthepooledeffectsestimatesforbinaryandcontinuousvariablesusingriskratiosandweightedmeandifferenceswith95%CIs.WetestedforheterogeneitybetweenstudiesusingtheCochranQandI2tests.Wepredefinedstatisticalheterogeneityasbeinglow,intermediate,orhigh,correlatedtoI2statistics<25%,from25to50%,or>50%,respectively.17Meta-analyseswithrandom-effectsmodelswereemployedforalloutcomes,duetoanticipatedclinicalheterogeneityintermsofpatientpopulations.Weattemptedtoidentifyclinicalfactorsaspotentialsourcesofheterogeneity,assessingforprespecifiedsubgroups,including(1)weaningdifficulty(simple,difficult,orprolongedweaning)and(2)COPD(vssubjectswithoutCOPD).Toassessthepotentialpublicationbiasfromsmallstudyeffects,weconstructedfunnelplotsdisplayingthelogriskratioonthehorizontalaxisandtheSEofthelogriskratioontheverticalaxis.WeemployedEgger'stesttoevaluatetheriskforpublicationbias.WesummarizedthequalityofevidenceaccordingtotheGradingofRecommendationsAssessment,Development,andEvaluation(GRADE)guidelines18andreportedthemusingtheGRADEproweb-basedtool.ResultsStudySelectionOurinitialelectronicsearchidentified3,674abstracts.Ofthese,weexcluded3,651becausetheydidnotdescribeRCTs,didnotevaluateweaningtechniques,wereduplicatereferences,orwerenotrelevant.Weretrieved23studiesforamoredetailed,full-textanalysis,andweexcluded11ofthesestudies19–29(Fig.1).Wethenidentified12suitablestudiescomprising2,161subjects.Bothreviewerscompletelyagreedonthefinalselectionofincludedstudies.Wealsoidentified2ongoingRCTs,includingonefromourgroup,fromtheClinicalTrials.govdatabase(Pellegrini,http://clinicaltrials.gov/show/NCT01464567;Agarwal,http://clinicaltrials.gov/show/NCT00911378AccessedAugust15,2016).DownloadfigureOpeninnewtabDownloadpowerpointFig.1.Flowchart.StudyDescriptionTheincludedstudieswerepublishedbetween1994and2015andwerefrom10countries.Eightofthe12includedstudiesweresingle-centerstudies.30–37Onestudywaspublishedonlyinabstractform,37andfulldetailsofthestudywerenotavailabletotheauthorsofthisreview.Table1summarizesthecomponentsoftheriskofbiasassessment.Duetothenatureoftheinterventionbeingstudied,allstudieswereunblindedwithregardtothepatientsenrolledandtheoutcomesassessed.Only3studiesspecifiedadequaterandomsequencegeneration,and7studiesdidnotreportadherencetotheintention-to-treatprinciple.OurqualitativeanalysisofkeystudycharacteristicsissummarizedinTable2.Viewthistable:ViewinlineViewpopupDownloadpowerpointTable1.RiskofBiasAssessmentViewthistable:ViewinlineViewpopupDownloadpowerpointTable2.QualitativeAnalysisofKeyStudyCharacteristicsOutcomesAssessedAllbut2trials31includedinthestudyreportedweaningfailureoutcome,whichwasdefinedasfailuretoextubatethesubjectimmediatelyfollowingtheSBT.Allbutone36study,apartfromthoseassessingonlytracheostomizedsubjects,reported48-hre-intubationrates.EightstudiesreportedICUmortality.32–35,37–39,41StudyProtocolsBrochardetal38compared3differentstrategiesforgradualweaningfrommechanicalventilation.BeyondT-pieceandPSV,theauthorsrandomlyassignedsubjectstoathirdgroupusinggradualtitrationofsynchronizedintermittentmandatoryventilation.T-pieceSBTswereperformedupto8times/d,progressingfrom5to120mininprogressivesteps.InsubjectsassignedtothePSVgroup,thePSVlevelwassystematicallyadjusted2times/dtomaintainabreathingfrequencybetween20and30breaths/minindecrementalstepsof2–4cmH2O.Estebanetal40comparedonce-a-dayT-pieceSBTswith3othermethods,includingintermittentmandatoryventilation,intermittenttrialsofspontaneousbreathing(conductedtwoofmoretimesadayifpossible),inadditiontoPSV.ThestudybyVitaccaetal41differedfromtheotherincludedstudiesbyenrollingonlydifficult-to-weantracheostomizedsubjectswithCOPDwhorequiredmechanicalventilationforatleast15d.Thistrialwasconductedin3long-termweaningunits,withsubjectstransferredfrom24ICUsafterarangeof15–39donmechanicalventilation.Theauthorsalsocomparedtheirresultswithhistoricalcontrolsoranuncontrolledclinicalpractice.Jubranetal34assessedasimilarpopulation.Subjectswhorequiredmechanicalventilationfor>21dinalong-termweaningunitwererandomlyassignedtounassistedbreathingthroughatracheostomycollarortoprogressivereductionsinPSVbasedontheirbreathingfrequencies.Sixstudiesincludedsimple-to-weansubjects,accordingtopreviouslypublisheddefinitions.2,30–32,35,36,39Threestudiesassesseddifficult-to-weansubjects,33,38,40and2studiesincludedprolonged-weaningandtracheostomizedsubjects.34,41Twoadditionalstudiesevaluatedpostoperativesubjects.35,37InthestudiesthatpredefinedtheSBTduration,a120-mintrialwasmostcommonlyemployed(7studies).ThreeauthorsreportedprogressivelyincreasingdurationofSBTsbasedonsubjects'tolerance.34,38,41Althoughincludedinvarioustrials,subjectswithCOPDrepresentedonlyasmallfractionofthestudypopulation.Only2oftheincludedstudiesspecificallyassessedsubjectswithCOPD33,41;oneofthesealsospecificallyenrolledprolonged-weaningsubjects.41EvidenceSynthesisT-pieceSBTswereassociatedwithariskratio(95%CI)of1.11(0.80–1.54)forICUmortalityand1.21(0.90–1.63)forthe48-hre-intubationrate(Figs.2and3).TheevidencefromtrialsaddressingtheseoutcomeswasconsideredverylowtolowqualitybasedontheGRADEapproach(Table3).Studylimitations,inconsistency,andimprecisioncontributedtodowngradingtheoverallqualityofevidenceinthepooledRCTs.DownloadfigureOpeninnewtabDownloadpowerpointFig.2.ICUmortality.M-H=Maentel-Haentzel;PSV=pressuresupportventilation.DownloadfigureOpeninnewtabDownloadpowerpointFig.3.48-hreintubationrate.M-H=Maentel-Haentzel;PSV=pressuresupportventilation.Viewthistable:ViewinlineViewpopupDownloadpowerpointTable3.GradeEvidenceProfileForweaningfailure,wefoundariskratioof1.23(0.94–1.61)(Fig.4)withmoderatetohighheterogeneity(I2=48%).Whenevaluatingpotentialsourcesofclinicalheterogeneity,weexcludedprolonged-weaningstudiesfromouranalysis(8studiesremaining;1,237subjects)andnotedariskratioof1.47(1.17–1.84)favoringPSVwithregardtoweaningsuccess.TheI2statisticforthisanalysiswas0%,suggestingthatprolonged-weaningstudiesrepresentanimportantsourceofclinicalheterogeneity.DownloadfigureOpeninnewtabDownloadpowerpointFig.4.Weaningfailure.M-H=Maentel-Haentzel;PSV=pressuresupportventilation.SubgroupAnalysesWeaningDifficulty.Sevenstudies(1,600subjects)focusedonsimpleweaning,whichwasdefinedassuccessonthefirstSBTintheabsenceofpreviousweaningfailure.PSVinthispopulationwasassociatedwithbetteroutcomesrelatedtoweaningsuccess(riskratio=1.44,1.11–1.86;I2=0%)butnotwithlowerre-intubationorICUmortalityrates.Threestudies(197subjects)specificallyassesseddifficult-to-weansubjects.Inthissubgroup,significantdifferencesinclinicaloutcomeswerenotfoundbetweenthealternativeSBTtechniques.Only2studies,comprising364subjects,focusedonprolonged-weaningsubjects.WhereastheSBTtechniquehadnoeffectonweaningfailurerateorICUmortality,T-piecewasassociatedwithashorterweaningduration(weightedmeandifference=−3.08(−5.24to−0.92)dofweaning)thanthatassociatedwithPSV(Fig.5).TheseaforementionedoutcomeshadanI2=0%inthissubgroup.DownloadfigureOpeninnewtabDownloadpowerpointFig.5.ForestplotcomparingT-pieceandpressuresupportventilation(PSV)forweaningduration.M-H=Maentel-Haentzel.COPDSubjects.Inthe12includedstudies,weidentified338subjectswithCOPD,although3studiesdidnotspecificallyreportthiscondition.Betweenthe2studiesenrollingexclusivelysubjectswithCOPD,weidentifiedremarkableclinicalheterogeneitypreventingpooling.WhereasthestudyofMatićetal33includeddifficult-to-weansubjectswithCOPD(definedasonefailedweaningattempt),theVitaccastudy41includedtracheostomizedsubjectsventilatedfor≥15d.TheriskratiosintheMatićandVitaccastudiesforweaningfailurewithT-pieceSBTswere1.61(0.82–3.16)and0.86(0.33–2.21),respectively,andriskratiosforICUmortalitywere2.06(0.41–10.47)and0.67(0.12–3.67),respectively.Weperformedfunnel-plotanalysisforeachoutcomeanddidnotidentifypublicationbias.ThefunnelplotforweaningfailureisshowninFigure6.Egger'stestdidnotsuggestpublicationbias(P=.37).DownloadfigureOpeninnewtabDownloadpowerpointFig.6.Funnelplotforweaningfailureoutcome.DiscussionSummaryofEvidenceT-pieceandPSVare2ofthemostcommonlyusedtechniqueswhenconductingSBTsinclinicalpracticetoday.Nevertheless,existingevidencedirectlycomparingthese2approachesissparse,heterogeneous,andofpooroverallquality.Smallstudypopulationswithloweventrates,variabilityamongtheappliedSBTtechniques,andremarkablydifferentpopulationslimitthepoolingandadequateinterpretationofevidence.T-pieceandPSVtechniqueshavetheoreticalsingularitiesthatmayinfluencebedsidejudgmentwhenchoosingoneSBTtechniqueoveranother.WhenusingT-pieceSBTs,onemightbelookingforspecificityandthusmightproceedwithextubationonlyforthosepatientsabletotoleratethehemodynamicperturbationsofthisdisturbingtest.13,42Also,previousstudieshaveshownthatthepostextubationworkofbreathingcouldbemorecloselyparalleledbyunassistedbreathing(asinaT-piecetrial)thanbyalow-pressuresupporttrial.43Bycontrast,Ezingeardetal44demonstratedthatmoresubjectscouldbesuccessfullyextubatedafteraPSVSBT,includingsomesubjectswhopreviouslyfailedaT-pieceSBT.Thesefindingsaresupportedbymoderate-qualityevidencewithregardtosimple-to-weansubjects,forwhomPSVmightbeassociatedwithreducedweaningfailurerates,notadverselyinfluencingre-intubationrates.Incontrast,low-qualityevidencesuggeststhatprolonged-weaningsubjects31,37appeartobenefitfromT-pieceSBTsintermsofweaningduration.Inthesesubjects,progressivestepstowardpredeterminedreductionsinPSVaccordingtothesubject'stolerancemayprolongthedurationofmechanicalventilation,potentiallyincreasingtheriskformechanicalventilation-relatedcomplications.However,weobservedthatthefewstudiesincludingsubjectswhoexperiencedsucheventsfoundthattheSBTtechniquehasnoinfluenceonmortalityinthissubgroup.PatientswithCOPDrepresentagrowingpopulationworldwide,remainingasoneofthemostprevalentconditionsleadingtorespiratoryfailureandconsequentmechanicalventilation.45–47Theseindividualsrepresentsomeofthemostchallenginggroupstoweanfrommechanicalventilation48;paradoxically,thispopulationisunderrepresentedinRCTs.TwogroupsofauthorsevaluatedsubjectswithCOPDexclusively,buttheirstudiesincludedmarkedlydifferentprofiles.Onegroupofauthors33enrolleddifficult-to-weansubjectswithCOPD,andtheother41focusedontracheostomized,prolonged-weaningsubjectsinlong-termweaningunits.Recognizingthatclinicalheterogeneitywouldhindertheinterpretationoffindings,wedecidednottopooltheseresults.ArelevantaspectthatshouldbekeptinmindconcernstheconsiderationofSBTsasaninterventionratherthanadiagnostictesttryingtoidentifypatientswhoarepotentialcandidatesforextubation,predictingtoleranceofunassistedbreathing.Inthelatterapproach,questionsfocusondiagnosticaccuracy,andthenweaningfailureorre-intubationratesareimportantendpointsfordescribingweaningtrialsensitivityorspecificity.Nevertheless,theassessmentofdiagnosticpropertiesofSBTsforpredictingsuccessfulextubationisnotstraightforward.Extubationfailureratesarewidelyreportedasbeingaround15–20%,whichmakesspecificityofthetrialforpredictingsuccessfulextubationonly80–85%.Ontheoppositeside,testsensitivity(theproportionofpatientsabletotolerateextubationdespitefailingtheweaningtest)isdifficulttoevaluatebecausepatientswhofailaweaningtestareusuallynotextubated.Furthermore,criteriaforterminationofaweaningtrialandevendefinitionsoftestfailureareessentiallysubjectiveandclinician-dependent,potentiallybiasingoutcomesbeyondthetestitself.Accordingly,wedecidedtostayinlinewithpreviousstudiesandassesstheclinicalimpactofSBTsasaninterventioninimportantoutcomes,beyonditsdiagnosticroleinpredictingpatienttolerancetomechanicalventilationdiscontinuation.Overall,ourresultsareconsistentwiththoseofaCochranereview14andareapplicabletothegeneralweaningpopulationencounteredbycliniciansinclinicalpractice.Ourreview,however,addsimportantadditionalinformationfrom4publishedRCTs,increasingthesizeoftheincludedpopulation(2,161subjectsherevs1,208subjectspreviouslyanalyzed).Inaddition,wedefinedanapriorisubgroupanalysisaimedatidentifyingdifferenteffectsofthealternativeSBTtechniquesbasedonweaningdifficultyandthepresenceofCOPD.Verylowtolowoverallqualityofevidencestronglylimitsdefinitivefindingsinthisfield.StrengthsandLimitationsWeconductedasystematicsearchofseveraldatabaseswithoutlanguagerestrictionstoidentifyallRCTscomparingT-pieceandPSVSBTtechniquesinweaningsubjects.Weemployedstandardizedtechniquestoassessriskofbiasandoverallqualityofevidence.Ourreviewhasseverallimitationsthatreducethestrengthofinferencesthatcanbemade.First,qualityassessmentpermitsclassifyingtheevidenceasverylowtolowquality.Althoughsomeaspectsofbiasassessmentarenotrelevantinthisarea(eg,blindingofpatientsandinvestigatorsinnecessarilyunblindedtrials),others,suchassequencegenerationandallocationconcealment,revealmethodologicalissuesthatmayimpactstudyfindings.Imprecisionofavailabledatawasanimportantsourceofdowngradingofevidenceformanyoutcomes.Second,weidentifiedimportantclinicalheterogeneityamongstudiesthathinderedthepoolingofestimatesandlimitedthegeneralizabilityofourfindings.Someaspectsrelatedtomechanicalventilationsettings(differentpressuresupportlevelsandprotocols,adjunctiveuseofPEEP)certainlycontributetoconflictingresults.Thiscouldbeconsideredoneofthemostimportantissuesinthisreview.Third,subgroupanalysisshouldbeinterpretedwithcaution,accordinglytostudypopulationsandoutcomesreported.Finally,thedifficult-to-weansubjectandCOPDsubjectsubgroupsremainscarcelystudied,limitingconclusionsintheseareas.ConclusionsThequalityofavailableevidenceprecludesdefinitiveconclusionsaboutassessedoutcomes.Low-qualityevidencesuggeststhatPSVSBTsmayresultinlowerweaningfailureratesinsimple-to-weansubjectsbutdonotaffectre-intubationratesorotherimportantoutcomes.Conversely,inprolonged-weaningsubjects,aT-piecemayreducetheweaningdurationcomparedwithPSVSBTs.FuturetrialsshouldcompareSBTtechniquesindifficult-to-weanandCOPDsubjects.FootnotesCorrespondence:JoséAugustoSantosPellegriniMD,RuaRamiroBarcelos,910,BlocoD,90035-001-PortoAlegre,RS,Brazil.E-mail:joseaugusto.pellegrini{at}gmail.com.Theauthorshavedisclosednoconflictsofinterest.Copyright©2016byDaedalusEnterprisesReferences1.↵KlompasM,AndersonD,TrickW,BabcockH,KerlinMP,LiL,etal.Thepreventabilityofventilator-associatedevents:theCDCPreventionEpicentersWakeUpandBreatheCollaborative.AmJRespirCritCareMed2015;191(3):292–301.OpenUrlCrossRefPubMed2.↵BolesJM,BionJ,ConnorsA,HerridgeM,MarshB,MelotC,etal.Weaningfrommechanicalventilation.EurRespirJ2007;29(5):1033–1056.OpenUrlAbstract/FREEFullText3.PeñuelasÓ,ThilleAW,EstebanA.Discontinuationofventilatorysupport:newsolutionstoolddilemmas.CurrOpinCritCare2015;21(1):74–81.OpenUrl4.BalasMC,VasilevskisEE,OlsenKM,SchmidKK,ShostromV,CohenMZ,etal.Effectivenessandsafetyoftheawakeningandbreathingcoordination,deliriummonitoring/management,andearlyexercise/mobilitybundle.CritCareMed2014;42(5):1024–1036.OpenUrlCrossRefPubMed5.↵TrogrlićZ,vanderJagtM,BakkerJ,BalasMC,ElyEW,vanderVoortPH,IstaE.Asystematicreviewofimplementationstrategiesforassessment,prevention,andmanagementofICUdeliriumandtheireffectonclinicaloutcomes.CritCare2015;19:157.OpenUrlCrossRefPubMed6.↵TaniosMA,NevinsML,HendraKP,CardinalP,AllanJE,NaumovaEN,EpsteinSK.Arandomized,controlledtrialoftheroleofweaningpredictorsinclinicaldecisionmaking.CritCareMed2006;34(10):2530–2535.OpenUrlCrossRefPubMed7.TeixeiraC,ZimermannTeixeiraPJ,HohërJA,deLeonPP,BrodtSF,daSivaMoreiraJ.Serialmeasurementsoff/VTcanpredictextubationfailureinpatientswithf/VT