Trials directly comparing alternative spontaneous breathing ...

After identification, patients may undergo a spontaneous breathing trial (SBT) to assess their ability to breathe spontaneously with minimal or ... Skiptomaincontent Advertisement SearchallBMCarticles Search DownloadPDF Research OpenAccess Published:01June2017 Trialsdirectlycomparingalternativespontaneousbreathingtrialtechniques:asystematicreviewandmeta-analysis KarenE.A.Burns1,2,3,IbrahimSoliman1,2,NeillK.J.Adhikari2,4,AmerZwein1,2,JessicaT.Y.Wong5,CarolinaGomez-Builes1,2,JoseAugustoPellegrini6,7,LuChen1,2,NuttapolRittayamai1,2,MichaelSklar1,2,LaurentJ.Brochard1,2&JanO.Friedrich1,2  CriticalCare volume 21,Article number: 127(2017) Citethisarticle 10kAccesses 31Citations 49Altmetric Metricsdetails AbstractBackgroundTheeffectofalternativespontaneousbreathingtrial(SBT)techniquesonextubationsuccessandotherclinicallyimportantoutcomesisuncertain.MethodsWesearchedMEDLINE,EMBASE,CENTRAL,CINAHL,Evidence-BasedMedicineReviews,OvidHealthStar,proceedingsoffiveconferences(1990–2016),andreferencelistsforrandomizedtrialscomparingSBTtechniquesinintubatedadultsorchildren.PrimaryoutcomeswereinitialSBTsuccess,extubationsuccess,orreintubation.Tworeviewersindependentlyscreenedcitations,assessedtrialquality,andabstracteddata.ResultsWeidentified31trials(n = 3541patients).Moderate-qualityevidenceshowedthatpatientsundergoingpressuresupport(PS)comparedwithT-pieceSBTs(ninetrials,n = 1901)wereaslikelytopassaninitialSBT(riskratio(RR)1.00,95%confidenceinterval(CI)0.89–1.11;I 2 = 77%)butmorelikelytobeultimatelyextubatedsuccessfully(RR1.06,95%CI1.02–1.10;11trials,n = 1904;I 2 = 0%).ExclusionofonetrialwithinconsistentresultsforSBTandextubationoutcomessuggestedthatPS(vsT-piece)SBTsalsoimprovedinitialSBTsuccess(RR1.06,95%CI1.01–1.12;I 2 = 0%).Limiteddatasuggestthatautomatictubecompensationpluscontinuouspositiveairwaypressure(CPAP)vsCPAPaloneorPSincreaseSBTbutnotextubationsuccess.ConclusionsPatientsundergoingPS(vsT-piece)SBTsappeartobe6%(95%CI2–10%)morelikelytobeextubatedsuccessfullyand,iftheresultsofanoutliertrialareexcluded,6%(95%CI1–12%)morelikelytopassanSBT.FuturetrialsshouldinvestigatepatientsforwhomSBTandextubationoutcomesareuncertainandcomparetechniquesthatmaximizedifferencesinsupport. BackgroundWeaningaccountsforapproximately40%ofthetimespentonmechanicalventilation[1,2].Comparedwithnonprotocolizedcare,randomizedcontrolledtrials(RCTs)andasystematicreviewindicatethatweaningprotocolsreducethedurationofmechanicalventilation,weaningtime,andintensivecareunit(ICU)lengthofstay(LOS)[3,4].Afteridentification,patientsmayundergoaspontaneousbreathingtrial(SBT)toassesstheirabilitytobreathespontaneouslywithminimalornosupport.CliniciansconductSBTstofacilitatedecision-makingregardingtimelyextubationandtominimizepatients’exposuretoinvasiveventilation.Inmakingdecisions,clinicians‘tradeoff’therisksassociatedwithdelayedextubationandthoseassociatedwithaprematurefailedattemptatextubation.SeveraltechniquesarecommonlyusedtoconductSBTs,includingpressuresupport(PS)withorwithoutpositiveend-expiratorypressure(PEEP),continuouspositiveairwaypressure(CPAP),automatictubecompensation(ATC),andtheT-piece.WhereassomeSBTtechniquesdeliverpressureduringinspirationtoovercomeendotrachealtuberesistance(e.g.,PS,ATC),othertechniquesaimtoimproverespiratorymechanicsorcardiacfunction(e.g.,CPAP)andmayoverestimatepatients’abilitytobreatheautonomouslyafterextubation[5].Conversely,theT-pieceprovidesnosupport,isperceivedtoincreaseworkofbreathing(WOB),andmayunderestimatepatients’abilitytobreathespontaneouslyafterextubation[5].ThemostrecentAmericanCollegeofChestPhysicians/AmericanThoracicSocietyClinicalPracticeGuidelines[6]supportthatanSBTisthemajordiagnostictesttodeterminewhetherpatientscanbeextubatedbutgiveonlyaconditional(weak)recommendationthatPSSBTsshouldbeusedastheinitialSBT,basedonlimiteddata(threeorfourincludedRCTs).ACochranereviewofninetrialscomparedPSandT-piece‘weaning’inadultsandfoundnonsignificantdifferencesbetweentechniquesonweaningsuccess,pneumonia,reintubation,ICUmortality,andLOS.Inasubgroupanalysis(fourtrials,n = 940)theauthorsnotedthatpatientsweresignificantlymorelikelytopassaPSSBTvsaT-pieceSBT(riskratio(RR)1.09,95%confidenceinterval(CI)1.02–1.17)[7].ThisreviewdidnotdirectlycomparealternativeSBTtechniquesandwaslimitedtofullpublicationsofadults.Atpresent,noSBTtechniquehasbeenshowntobesuperiortoanother.WesoughttosummarizetheRCTevidencedirectlycomparingallalternativeSBTtechniquesinvolvingcriticallyilladultsandchildrenoninitialSBTsuccess,extubationsuccess,reintubationrate(primaryoutcomes),andotherimportantoutcomes.MethodsDatasourcesWesearchedMEDLINE(1966–February2017),EMBASE(1980–February2017),theCochraneCentralRegisterofControlledTrials(CENTRAL,February2017),CINAHL(1982–February2017),evidence-basedmedicinereviews,andOvidHealthStar(1999–February2017)toidentifypotentiallyeligibletrialsusingdatabase-specificsearchstrategieswithoutlanguagerestrictions.WeusedtheoptimallysensitivesearchstrategiesofTheCochraneCollaborationforMEDLINEandEMBASE[8,9,10].Twoauthors(KEAB,JOF)independentlyscreenedcitationtitlesandabstractsandevaluatedfull-textversionsofpotentiallyrelevanttrials.Fiveauthorshand-searchedconferenceproceedingsoffivescientificmeetingsfrom1990–2016:EuropeanSocietyofIntensiveCareMedicine,AmericanCollegeofChestPhysicians(except1999–2002,unavailable),AmericanThoracicSociety,InternationalSymposiumofIntensiveCareandEmergencyMedicine,andSocietyofCriticalCareMedicine(including2017forthelatter).Ethicsapprovalwasnotrequired.StudyselectionWeincludedrandomizedorquasi-randomizedtrialscomparingtwoormoreSBTtechniquesincriticallyilladultsorchildrenreportingatleastoneofinitialSBTorextubationoutcome(successorfailure),reintubation,timetoextubationorsuccessfulextubation,timetofirstsuccessfulSBT,mortality,ventilator-associatedpneumonia(VAP),totaldurationofventilation,ICUorhospitalLOS,postextubationuseofnoninvasiveventilation(NIV),oradverseeventsusingauthors’definitions.Weexcludedtrialsthatevaluated:neonatalortracheostomizedpatients;SBTsaspartofaweaningstrategy;automatedSBTs(e.g.,SmartCare™,Intellivent®);NIVvscontinuedinvasiveventilation;andSBTconductvsnoSBT.Twoauthors(KEAB,JOF)independentlyselectedtrialsmeetinginclusioncriteria,andanotherauthor(LJB)adjudicateddifferences.DataextractionandqualityassessmentTwounblindedauthors(KEAB,JOF)abstracteddataregardingthestudyriskofbias(RoB)(randomization,allocationconcealment,blindedoutcomesassessment,completenessoffollow-up,selectiveoutcomesreporting,stoppingearlyforbenefit)andrecordedoutcomes,usingauthors’definitionsforreportedoutcomes,onastandardizedform[11].WeevaluatedRoB(yes,unclear,no)foreachdomain.Disagreementswereresolvedbyconsensusandarbitrationwithathirdauthor(LJB).DatasynthesisWepooleddataacrossstudiesusingrandomeffectsmodels.WederivedsummaryestimatesofRRandmeandifference(MD)with95%CIforbinaryandcontinuousoutcomes,respectively,usingReviewManager5.3(CochraneCollaboration,Oxford,UK)[12].Wepooled‘initialSBTsuccess’intrialsthatconductedmorethanoneSBT.WeevaluatedstatisticalheterogeneityforeachoutcomeusingtheI 2measurewiththresholdsof0–40%(mightnotbeimportant),30–60%(moderate),50–90%(substantial),and>75%(considerable)[13,14].Wesummarizedtrialsbasedonthetechniquescompared(e.g.,T-piecevsother)(Additionalfile1).Weplannedsubgroupanalysestocomparetheeffectsofdifferenttechniquesonprimaryoutcomesinperioperativevsnonperioperativetrialsandbasedon:durationofventilationatrandomization(nonperioperativetrials);thesupportprovidedduringSBTs;thetypeoflungdisease;andmethodologicquality(low/moderatevshighRoB).Weassessedforsubgroupdifferencesusingthechi-squaretest[15].WeusedtheGradingofRecommendationsAssessment,DevelopmentandEvaluation(GRADE)systemtoassessthequalityofthebodyofevidenceassociatedwiththeprimaryoutcomesandsignificantsecondaryoutcomes[16].Weassessedforpublicationbiaswhenatleast10trialswereidentified[17].Awrittenprotocolwasusedtoguidethereviewprocess.ResultsTrialidentificationWeidentified4218uniquecitations.Weassessed187articlesforeligibilityandexcluded156studies(Fig. 1).Thirty-onetrials[18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48]reportingon3541patientsmetourinclusioncriteriaincludingfivetrialscomparingthreeSBTtechniques[23,24,29,30,48].Twotrials[38,46]appearedtobepublished,atleastinpart,induplicate[49,50].ThemostcommoncomparisonswereT-piecevsPS(13trials),T-piecevsCPAP(ninetrials),PSvsATC(threetrials),andCPAPvsATC(threetrials).Fourtrials[21,34,38,40]werenotpublishedinEnglish.Sixtrials[25,27,33,36,45,47]werepublishedasabstracts,ofwhichtwo[36,47]providedpartialorfull-textmanuscripts.Ninetrials[19,20,21,23,24,25,44,45,47]evaluatedperioperativepopulations(sixcardiacsurgical[19,20,21,24,25,44]andthreeothersurgical[23,45,47]).Threetrials[28,43,46]evaluatedpediatricpatients.Fig.1Identificationoftrialsincludedinthemeta-analysis.SBTspontaneousbreathingtrial,PEEPpositiveend-expiratorypressureFullsizeimage QualityassessmentWejudgedrandomizationandallocationconcealmenttobeatlowRoBin16(52%)trialsand17(55%)trials,respectively.Onequasi-randomizedtrialallocatedpatientsbasedonevenorodddays[34].Notrialevaluatedoutcomesinablindedmanner.Wejudged15(48%)trialstohavecompleteoutcomesreporting.Eighteen(58%)trialsconductedanintention-to-treatanalysisand26(84%)trialsdidnotstopearlyforbenefit.Overalltrialqualitywasmoderate(Additionalfile2:FigureS1).PrimaryoutcomesInitialSBTsuccessSeventeenT-piece,12CPAP,eightATC,and13PStrialsdirectlycomparedoneSBTtechniquewithanotherandreportedinitialSBTsuccess(Table 1).ComparedwithT-pieceSBTs,moderate-qualityevidencesupportsthatpatientsundergoingPSSBTswerenotmorelikelytopassanSBT(RR1.00,95%CI0.89–1.11;p = 1.0;ninetrials,n = 1901)withconsiderableheterogeneity(I 2 = 77%)(Table 2,Fig. 2,Additionalfile3:FigureS2).Table1CharacteristicsofincludedtrialsFullsizetable Table2Summaryoffindings:PSvsT-pieceSBTsonSBTandextubationsuccessFullsizetable Fig.2ForestplotcomparingeffectofSBTtechnique(PSvsT-piece)onSBTsuccess.Thepooledriskratiowith95%confidenceinterval(CI)wascalculatedusingarandomeffectsmodel.Weightreferstothecontributionofeachstudytotheoverallestimateoftreatmenteffect.PSpressuresupportFullsizeimage Low-qualityevidencefromthreetrials(n = 247)suggeststhatpatientsweresignificantlymorelikelytopassanSBTwithATC + CPAPcomparedwithCPAPalone(RR1.12,95%CI1.04–1.22;p = 0.005,I 2 = 0%).Similarly,low-qualityevidencefromthreetrials(n = 276)showedthatpatientsweresignificantlymorelikelytopassanSBTwithATC + CPAPcomparedwithPS(RR1.10,95%CI1.01–1.20;p = 0.02,I 2 = 0%)(Table 3).Table3SummaryestimatesofeffectforcomparisonsofATCvsothertechniquesonSBTsuccessFullsizetable ExtubationsuccessSeventeenT-piece,eightCPAP,eightATC,and14PStrialscomparedoneSBTtechniquewithanotherandreportedextubationsuccess(Table 1).Moderate-qualityevidencesupportsthatpatientsundergoingPScomparedwithT-pieceSBTsweresignificantlymorelikelytobeextubatedsuccessfully(RR1.06,95%CI1.02–1.10;p = 0.007;11trials,n = 1904;I 2 = 0%)(Table 2,Fig. 3,Additionalfile4:FigureS3).Fig.3ForestplotcomparingeffectofSBTtechnique(PSvsT-piece)onextubationsuccess.Thepooledriskratiowith95%confidenceinterval(CI)wascalculatedusingarandomeffectsmodel.Weightreferstothecontributionofeachstudytotheoverallestimateoftreatmenteffect.RRriskratio,PSpressuresupportFullsizeimage ReintubationrateFourteenT-piece,nineCPAP,sevenATC,and13PStrialscomparingoneSBTtechniquewithanotherreportedthereintubationrateandfoundnostatisticallysignificantdifferencesbetweentechniques(Table 1,Additionalfile5:FigureS4).SecondaryoutcomesTherewasnoeffectofoneSBTtechniquevsanotheronICUmortality(sevenT-piece,threeCPAP,andfivePStrials),hospitalmortality(fourT-pieceandfourPStrials),orthemostprotractedmortalitymeasure(10 T-piece,fourCPAP,andsevenPStrials).Notrialreportedtimetoextubationortimetosuccessfulextubation.Meta-analysisofthreetrialscomparingATC + CPAPwithCPAPalonefoundnodifferenceinNIVuseafterextubation(RR0.53,95%CI0.27–1.06;p = 0.07,I 2 = 0%).Sensitivity,subgroup,andpost-hocanalysesExclusionofasinglequasi-randomizedtrialcomparingPSvsT-pieceSBTs[34]didnotchangethesignificantincreaseinextubationsuccessfavoringPSSBTs(RR1.05,95%CI1.01–1.10;p = 0.02,I 2 = 0%).Meta-analysesofPSvsT-pieceSBTsshowedbenefitinsevennonperioperativetrials(n = 1273;RR1.07,95%CI1.01–1.13;p = 0.02,I 2 = 94%)(high-qualityevidence)comparedwithtwoperioperativetrials(n = 548;RR0.86,95%CI0.61–1.22;p = 0.41,I 2 = 0%)(low-qualityevidence);however,aninteractiontestshowednodifferencebetweenthesesummaryestimates(p = 0.23)(Additionalfile6:TableS1,Additionalfile7:FigureS5).Subgroupanalysesbasedondurationofventilationamongnonperioperativetrialswasnotfeasiblegivensimilarreporteddurationsofventilation.Subgroupanalysescomparingmorevslessinspiratorysupportandtypeoflungdiseasewerenotsignificantforcommonlyreportedcomparisonsofalternativetechniques.Asensitivityanalysiswasnotpossibleduetotheabsenceofblindedoutcomesassessmentacrosstrials.Inspectionofafunnelplotfor11trialscomparingPSwithT-pieceSBTsonextubationsuccessdidnotsuggestpublicationbias.Weconductedapost-hocanalysisthatexcludedasingle,surgicaltrial[47]whichenrolledsurgicalpatients,waspublishedinabstractformonly,andhadinternallyinconsistentresults(i.e.,lowerinitialSBTsuccessratebuthigherextubationsuccessrateforPSvsT-pieceSBTs).Whenthistrialwasexcluded,meta-analysesshowedthatmorepatientspassedaninitialPS(vsT-piece)SBT(RR1.06,95%CI1.01–1.12;p = 0.03)withoutheterogeneity(I 2 = 0%)andweresimilarlyextubatedsuccessfully(RR1.06,95%CI1.01–1.12;p = 0.03,I 2 = 0%).ExcludingeachoftheotherRCTsinafullleave-one-study-outanalysisdidnotresultinsignificantchangestothepooledeffectestimateforSBTsuccess(pooledRRrange,0.98–1.01)orheterogeneity(I 2range,77–80%)suggestingthatonlyasingletrial[47]wasanoutlierforthisoutcome.Inasimilaranalysisforextubationsuccess,thepooledRRrangedfrom1.05to1.07andremainedstatisticallysignificantwithnoheterogeneity(I 2 = 0%)regardlessofwhichRCTwasremoved,suggestingthatnotrialwasanoutlierordisproportionatelyinfluencedtheextubationsuccesssummaryestimate.DiscussionWeidentified31trialsofoverallmoderatequalityreportingon3541patients.Moderate-qualityevidencesupportedthatSBTsuccessratesweresimilarwithPSandT-piece,withsubstantialheterogeneity.However,post-hocexclusionofanunpublishedtrial[47]withinconsistentresultseliminatedtheheterogeneityandshowedthatSBTsuccesswas6%morelikelywithPSSBTs.Meta-analysisalsoshoweda6%higherprobabilityofsuccessfulextubationfollowingPS(vsT-piece)SBTs,withnoheterogeneity,irrespectiveofthistrial’sinclusion[47].Althougha6%relativeimprovementinprobabilityofsuccessfulextubationmayappearsmall,itcorrespondstoaclinicallyimportantnumberneededtotreatof22(95%CI13–67)whenthebaselineriskofextubationsuccessis75%(Table 2).Low-qualityevidencefromthreetrialsindicatedthatpatientswere12%morelikelytopassanSBTwithATC + CPAP/PEEPcomparedwithCPAPandwere10%morelikelytopassanSBTwithATC + CPAP/PEEPcomparedwithPS,althoughextubationsuccessratesweresimilar.WefoundnodifferencesbetweenalternativeSBTtechniquesonreintubationrateormortality,althoughCIswerewide.SubgroupanalysissuggestedbeneficialeffectsofPSvsT-pieceSBTsonSBTsuccessinsevennonperioperativetrials(high-qualityevidence)comparedwithtwoperioperativetrials(low-qualityevidence),buttheRRswerenotstatisticallydissimilar.MosttrialsdirectlycomparedT-piecewithPSSBTs(13trials)andT-piecewithCPAPSBTs(ninetrials).VeryfewtrialsassessedalternativeSBTtechniquesinchildren.Mosttrials(n = 22)wereconductedinpatientswhowerenonperioperativeandforwhomextubationdecisionsareconsideredmorechallenging.Inpoolingoutcomes,wenotedthatfiveofsixcardiacsurgerytrialsreporteda100%SBTsuccessrateinbotharmsandthreesurgicaltrialsreporteda100%extubationsuccessrateinbotharms.ThesefindingssuggestthatthemostimportantquestioninpostoperativepatientswithahighpretestprobabilityofSBTandextubationsuccessmaybewhetheranSBTisnecessaryandthatquestionsregardingthebestSBTtechniquetousemaybemostrelevanttopatientsatindeterminateorlowpretestprobabilityofsuccess.OursystematicreviewdiffersfromtwopreviousreviewsbydirectlycomparingSBTtechniquesandexcludingtrialsevaluatingSBTtechniquesasonecomponentofaweaningstrategy[7,51].Moreover,wehand-searchedconferenceproceedingsspanning26 years,wherefeasible,andincludedpediatrictrials.ComparedwiththeCochranereviewofninetrials[7],weincludednineadditionaltrials(onepediatrictrial[28],fouradulttrials[34,38,42,44],twoabstracts[27,47],andtwothree-armtrials[30,48])comparingT-pieceandPSSBTsandexcludedfourweaningtrials[52,53,54,55].Contrarytotheirfindings,wefoundthatpatientswereonlymorelikelytopassaPS(vsT-piece)SBTafterexclusionofasingleoutliertrial[47]butweresignificantlymorelikelytobeextubatedsuccessfully.Thisfindingremainedsignificantafterexclusionofasinglepediatrictrial[28].Comparedwitharecentmeta-analysisof12trials[51],weincludedfiveadditionaltrials(onepediatrictrial[28],threeadulttrials[34,42,48]includingathree-armtrial[48],andanabstract[27])andexcludedfourtrialsinvolvingweaningortracheostomizedpatients[52,53,55,56].Similartotheirreview,wefoundthattheSBTtechniquedidnotinfluenceratesofweaningsuccess,mortality,orreintubation.ConsiderabledebateexistsregardingtheSBTtechniquethatbestsimulatespatient’sWOBafterextubation.AnSBTapproximatesthepatient’sabilitytobreathespontaneously,butitisanimperfecttestbecauseitcannottakeintoconsiderationfactors(e.g.,upperairwayresistance,respiratorymusclefatigue,cardiacdecompensation,accumulationofsecretions)thatmayoccurafterextubation.ThereareseveralphysiologicalreasonsfortheclinicalobservationthatPSSBTsmayleadtomoresuccessfulinitialSBTsandextubations.Byovercomingaportionofthepressuregradientacrosstheendotrachealtube,lowlevelsofPSorCPAPprovideminimalbutpotentiallyimportantsupportduringanSBT.AsystematicreviewoftheeffectofdifferentSBTsonphysiologicaloutcomesfoundthatmetricsofpatienteffort(WOB(n = 142)andpressure–timeproduct(PTP)(n = 129))weresignificantlyhigherduringT-piecevsPSSBTs,withsignificantheterogeneity(I 2 ≥ 75%).ThesemetricsduringT-pieceSBTswerealsomorecomparablewiththepostextubationperiodcomparedwithPSSBTs,althoughsamplesizesweresmall(n = 77forWOB;n = 52forPTP)andheterogeneitywasmoderate(I 2 = 67%and62%,respectively)[57].Mostpatients,especiallythosewithhighpretestprobabilityofsuccesswhorepresentthemajorityofpatientssubmittedtoSBTs[6,58],canbeextubatedeasilyafteraninitialSBT[58].However,T-pieceSBTsmaybeappropriateinselectedpatients(e.g.severeleftventriculardysfunction,neuromuscularweakness,difficultairway)whencliniciansareuncertainregardingtheirabilitytobreatheontheirownandwhentheyprioritizealowfalsepositiverateforpassinganSBTandbeingextubatedsuccessfullyduetotherisksassociatedwithextubationfailure[59].Conversely,whenT-piece(vsPS)SBTsareusedinpatientswithahighlikelihoodofextubationsuccess,theymayinduceahighfalsenegativerate.ComparedwithT-pieceSBTs,ourreviewmaysuggestthatPSSBTsfacilitateextubationdecision-making.EvenifPSSBTsunderestimatepostextubationWOB,theirsuccessfulcompletionmayoffsetclinicianreluctancetoextubate,resultinginmoretimelyandsuccessfulextubationwithoutincreasedreintubation[60,61].ThismaybetheprimaryreasonwhyPSSBTsresultinbothhigherSBTandextubationsuccessrates;otherwiseonemightexpectatestthatunderestimatespostextubationWOBtoyieldahigherSBTsuccessratefollowedbyahigherreintubationrate.However,reintubationmayberelatednotonlytoSBTtechniqueandoutcomebutalsotoextubationreadinessandnewandconcomitantillnesses.Furthermore,comparedwithT-pieceSBTs,PSSBTsdonotrequirethatcliniciansassembleaT-piececircuitordisconnectthepatientfromtheventilatororitsalarms,andpermitapplicationofPEEPthatreducesthepotentialforlossoflungaerationimmediatelypriortoextubation.AlthoughpassinganinitialSBTisanimportantoutcome,patientsmayundergoserialSBTsbeforeextubation,andstakeholdersprioritizebeingextubatedsuccessfully[62].Severaladditionalfindingswarrantcommentary.First,fewtrialsreporteduseofdailyscreeningorthecriteriausedtoidentifySBTcandidatesandassessextubationreadiness.Second,wenotedwidevariationininitialSBTsuccessratesacrosstrialscomparingPS(range,54.3–100.0%)withT-piece(range,53.0–100.0%)SBTsandcomparingATC + CPAP(range,64.7–96.7%)withPS(range,52.6–86.0%)SBTs,and,similarly,broadvariationinextubationsuccessratescomparingPS(range,60.0–100.0%)withT-piece(range,50.0–100.0%)SBTs.Conversely,wenotedhigherSBTsuccessratesinthreetrialscomparingATCandCPAP(range,93.3–96.6%)withCPAPalone(range,80.0–86.7%).WhileSBTandextubationsummaryestimatesdifferedquantitativelyacrosscomparisons,qualitativelythedirectionofeffectfavoredSBTsconductedwithinspiratorysupport.Third,onlyeighttrials[25,28,29,30,34,37,38,43]specifiedadditionofCPAP(orPEEP)toPSduringSBTs.Fourth,trialswerepredominantlyofmoderatequality.Finally,althoughweincludedpediatrictrials,weonlyidentifiedthreesuchtrials.Becauseeachpediatrictrialcompareddifferenttechniques(T-piecevsPS + PEEP[28],ATC/CPAPvsPS/CPAP[43],andCPAPvsT-piece[46])ourabilitytocompareoutcomesinchildrenvsadultswaslimited.Resultsforthepediatrictrialswerenotdifferentfromthepooledresultsoftheadulttrialsforanyoftheoutcomes(interactionpvaluesallnonsignificant;resultsnotshown).OurreviewisthefirsttodirectlycomparealternativeSBTtechniquesandwasstrengthenedbyanextensivesearch,duplicatecitationscreeninganddataabstraction,useofrandomeffectsmodelstopooldata,andconductofprespecifiedsubgroupanalyses.Ourreviewalsohaslimitations.Includedtrialswerepredominantlyofmoderatequalitywithnotrialconductingblindedoutcomesassessmentandsummaryestimateswerelimitedbyvariableoutcomesreportingandunclearprospectivefollow-up.StatisticalnoisecouldbeminimizedifSBTtechniqueswereappliedseriallyuntilextubationandifextubationwasrestrictedtopatientswhopassedanSBT.Onlyfivetrials,allcomparingPSwithT-pieceSBTs,reportedconductingSBTsdaily[31,35],dailyupto3 days[29,48],orforanundisclosedtime[38].Despitesubgroupanalyses,wecannotfullyelucidatetheimpactofpretestprobabilityofSBTorextubationsuccessontheeffectofSBTaspatientsatintermediateorhighlikelihoodofSBTandextubationsuccessdominatedtheanalysis.Theimplicationsofourfindingsforpatientswithlowpretestprobabilityremainuncertain.Finally,ourabilitytoassesstheimpactofotherfactors(e.g.,pre-randomizationdurationofventilation,typeofICU,presenceofdedicatedrespiratorycarepersonnel)insubgroupanalyseswasconstrainedbysmallnumbersoftrialsandvariableandlimitedreporting.Furthermore,SBTdurationswerevariable.TwoRCTsrandomizingpatientsto30vs120 minutesdurationofT-piece[63]orPS[64]SBTsfoundnonsignificantdifferencesinSBTandextubationsuccessrates.Inaddition,averyrecentmulticenterRCTpublishedonlyinabstractform[65]suggeststhatreventilatingpatientsfor1 hourafterasuccessfulSBTmayincreasesuccessfulextubationrates.ConclusionPatientsundergoingPSvsT-pieceSBTsappeartobe6%(95%CI2–10%,)morelikelytobeextubatedsuccessfully,and,iftheresultsofanoutliertrialareexcluded,are6%(95%CI1–12%)morelikelytopassanSBT.PSSBTswerenotassociatedwithanincreasedriskofreintubationormortality,butCIswerewide. AbbreviationsATC: Automatictubecompensation CI: Confidenceinterval CPAP: Continuouspositiveairwaypressure GRADE: GradingofRecommendationsAssessment,DevelopmentandEvaluation ICU: Intensivecareunit LOS: Lengthofstay MD: Meandifference NIV: Noninvasiveventilation PEEP: Positiveend-expiratorypressure PS: Pressuresupport RCT: Randomizedcontrolledtrial RoB: Riskofbias RR: Riskratio SBT: Spontaneousbreathingtrial VAP: Ventilator-associatedpneumonia WOB: Workofbreathing References1.EstebanA,AliaI,IbanezJ,etal.Modesofmechanicalventilationandweaning.AnationalsurveyofSpanishhospitals.TheSpanishLungFailureCollaborativeGroup.Chest.1994;106:1188–93.CAS  Article  PubMed  GoogleScholar  2.EstebanA,AnzuetoA,FrutosF,etal.Characteristicsandoutcomesinadultpatientsreceivingmechanicalventilation:a28-dayinternationalstudy.JAMA.2002;287(3):345–55.Article  PubMed  GoogleScholar  3.ElyEW,BakerAM,DunaganDP,etal.Effectonthedurationofmechanicalventilationofidentifyingpatientscapableofbreathingspontaneously.NEnglJMed.1996;335:1864–9.CAS  Article  PubMed  GoogleScholar  4.BlackwoodB,BurnsK,CardwellC,etal.Useofweaningprotocolsforreducingdurationofmechanicalventilationincriticallyilladultpatients:anupdatedCochranesystematicreviewandmeta-analysis.CochraneDatabaseSystRev.2014;11:CD006904. GoogleScholar  5.KuhlenR,MaxM,DembinskiR,etal.Breathingpatternandworkloadduringautomatictubecompensation,pressuresupportandT-piecetrialsinweaningpatients.EurJAnaesthesiol.2003;20(1):10–6.CAS  Article  PubMed  GoogleScholar  6.OuletteDR,etal.Liberationfrommechanicalventilationincriticallyilladults:anOfficialAmericanCollegeofChestPhysicians/AmericanThoracicSocietyClinicalPracticeGuideline.Inspiratorypressureaugmentationduringspontaneousbreathingtrials,protocolsminimizingsedation,andnoninvasiveventilationimmediatelyafterextubation.Chest.2017;151(1):166–80.Article  GoogleScholar  7.LadeiraMT,VitalFM,AndrioloRB,etal.PressuresupportversusT-tubeforweaningfrommechanicalventilationinadults.CochraneDatabaseSystRev.2014;5:CD006056. GoogleScholar  8.DickersinK,SchererR,LefebvreC.Identifyingrelevantstudiesforsystematicreviews.BMJ.1994;309(6964):1286–91.CAS  Article  PubMed  PubMedCentral  GoogleScholar  9.LefebvreC,ManheimerE,GlanvilleJ.Chapter6:Searchingforstudies.In:CochraneHandbookforSystematicReviewsofInterventionsVersion5.0.1(updatedSeptember2008).TheCochraneCollaboration.2011.www.cochrane-handbook.org.Accessed16May2017.10.RobinsonKA,DickersinK.DevelopmentofahighlysensitivesearchstrategyfortheretrievalofreportsofcontrolledtrialsusingPubMed.IntJEpidemiol.2002;31(1):150–3.Article  PubMed  GoogleScholar  11.HigginsJPT,GreenS,editors.CochraneHandbookforSystematicReviewsofInterventionsVersion5.0.0.TheCochraneCollaboration.2008.http://handbook.cochrane.org.Accessed16May2017.12.Reviewmanger(RevMan)Version5.3.0.TheCochraneCollaboration.Copenhagen:TheNordicCochraneCentre;2011.13.HigginsJP,ThompsonSG.Quantifyingheterogeneityinameta-analysis.StatMed.2002;21:1539–58.Article  PubMed  GoogleScholar  14.HigginsJPT,GreenS(editors).CochraneHandbookforSystematicReviewsofInterventionsVersion5.1.0(updatedMarch2011).TheCochraneCollaboration.2011.http://handbook.cochrane.org.Accessed16May2017.15.BorensteinM,HigginsJP.Meta-analysisandsubgroups.PrevSci.2013;14:134–43.Article  PubMed  GoogleScholar  16.GuyattG,OxmanAD,KunzR,fortheGRADEWorkingGroup,etal.Whatis“QualityofEvidence”andwhyisitimportanttoclinicians?BMJ.2008;336:995–8.Article  PubMed  PubMedCentral  GoogleScholar  17.EggerM,DaveySmithG,SchneiderM,etal.Biasinmeta-analysisdetectedbyasimple,graphicaltest.BMJ(ClinicalResearchEd).1997;315(7109):629–34.CAS  Article  GoogleScholar  18.FeeleyTW,SaumarezR,KlickJM,etal.Positiveend-expiratorypressureinweaningpatientsfromcontrolledventilation.Lancet.1975;2(7938):725–8.CAS  Article  PubMed  GoogleScholar  19.HastingsPR,BushnellL,SkillmanJJ,etal.CardiorespiratorydynamicsduringweaningwithIMVversusspontaneousventilationingood-riskcardiac-surveypatients.Anesthesiology.1980;53:429–31.CAS  PubMed  GoogleScholar  20.PrakashO,MeijS,VanDerBordenB.Spontaneousventilationtestvsintermittentmandatoryventilation.Chest.1982;81:403–6.CAS  Article  PubMed  GoogleScholar  21.KollerW,SpissC,FinaU,DumaS.EntwohnungnachpostoperativerBeatmungKardiochirurgischerPatienten,CPAPversusZEEP[German].DerAnesthesist.1983;32:483–7.CAS  GoogleScholar  22.JonesDP,ByrneP,MorganC,etal.Positiveend-expiratorypressurevs.T-piece.Extubationaftermechanicalventilation.Chest.1991;100:1655–59.CAS  Article  PubMed  GoogleScholar  23.AbalosA,LeibowitzAB,DistafancoD,etal.Myocardialischemiaduringtheweaningperiod.CritCare.1992;3:32–6. GoogleScholar  24.BaileyCR,JonesRM,KelleherAA.Theroleofcontinuouspositiveairwaypressureduringweaningfrommechanicalventilationincardiacsurgicalpatients.Anaesthesia.1995;50:677–81.CAS  Article  PubMed  GoogleScholar  25.SchincoMA,WhitmanGJR,WeimanDS,etal.Pressuresupportventilationincombinationwithcontinuouspositiveairwaypressureisabetterweaningtrialinthepostcoronarybypasspatient.CritCareMed.1995;23:A236.Article  GoogleScholar  26.EstebanA,AliaI,GordoF,etal.ExtubationoutcomeafterspontaneousbreathingtrialswithT-tubeorpressuresupportventilation.AmJRespirCritCareMed.1997;56:459–65.Article  GoogleScholar  27.HolandaMA,RochaEM,BandeiraRM,etal.Pressuresupportventilation(PSV)versusT-tubeaspre-extubationspontaneousbreathingtrials(SBT).AmJRespirCritCareMed.2000;161:A559. GoogleScholar  28.FariasJA,RettaA,OlazarriF,etal.Acomparisonoftwomethodstoperformabreathingtrialbeforeextubationinpediatricintensivecarepatients.IntensiveCareMed.2001;27:1649–54.CAS  Article  PubMed  GoogleScholar  29.HaberthurC,MolsG,ElsasserS,etal.Extubationafterbreathingtrialswithautomatictubecompensation,T-tube,orpressuresupportventilation.ActaAnaesthesiolScand.2002;46:973–9.CAS  Article  PubMed  GoogleScholar  30.KoksalGM,SayilganC,SenO,etal.Theeffectsofdifferentweaningmodesontheendocrinestressresponse.CritCare.2004;8:R31–4.Article  PubMed  GoogleScholar  31.MaticI,Majeric-KoglerV.ComparisonofpressuresupportandT-tubeweaningfrommechanicalventilation:randomizedprospectivestudy.CroatMedJ.2004;45:162–6.PubMed  GoogleScholar  32.CohenJD,ShapiroM,GrozovskiE,etal.Extubationoutcomefollowingaspontaneousbreathingtrialwithautomatictubecompensationversuscontinuouspositiveairwaypressure.CritCareMed.2006;34:682–6.Article  PubMed  GoogleScholar  33.LiangCY,ChenCW,LinWJ,etal.Aprospective,randomizedcomparisonofautomatedtubecompensation(ATC)andT-pieceintheweaningofmechanicallyventilatedpatients.AmJRespirCritCareMed.2006;173:A41. GoogleScholar  34.ColomboT,BoldriniAF,JulianoSRR,etal.Implementation,assessmentandcomparisonoftheT-tubeandpressure-supportweaningprotocolsappliedtotheintensivecareunitpatientswhohadreceivedmechanicalventilationformorethan48 hours[Portugese].RevBrasTerIntensiva.2007;19:31–7.Article  PubMed  GoogleScholar  35.MaticI,DanicD,Majeric-KoglerV,etal.Chronicobstructivepulmonarydiseaseandweaningofdifficult-to-weanpatientsfrommechanicalventilation:randomizedprospectivestudy.CroatMedJ.2007;48:51–8.PubMed  PubMedCentral  GoogleScholar  36.FayedAM,ElFekyIL.Addingautomatictubecompensationtocontinuouspositiveairwaypressureinweaningpatientswithchronicobstructivepulmonarydisease,isitworthtrying?CritCareMed.2008;36:A137. GoogleScholar  37.CohenJ,ShapiroM,GrozovkiE,etal.Predictionofextubationoutcome:arandomized,controlledtrialwithautomatictubecompensationvs.pressuresupportventilation.CritCare.2009;13:R21.Article  PubMed  PubMedCentral  GoogleScholar  38.ZhangB,QinYZ.Aclinicalstudyofrapid-shallowbreathingindexinspontaneousbreathingtrialwithpressuresupportventilationandT-piece[Chinese].ChineseCritCareMed.2009;21:397–401. GoogleScholar  39.Figueroa-CasasJ,MontoyaR,ArzabalaA,etal.Comparisonbetweenautomatictubecompensationandcontinuouspositiveairwaypressureduringspontaneousbreathingtrials.RespirCare.2010;55:549–54.PubMed  GoogleScholar  40.Molina-SaldarriagaFJ,Fonseca-RuizNJ,CastroC,etal.Studyofspontaneousbreathinginpatientswithchronicobstructivepulmonarydisease:continuouspositiveairwaypressure(CPAP)versusT-Tube[Portuguese].MedIntensiva.2010;34:453–8.CAS  Article  PubMed  GoogleScholar  41.CekmanN,ErdemliO.ThecomparisonoftheeffectsofT-pieceandCPAPonhemodynamicparameters,arterialbloodgasesandsuccessofweaning.BratislLekListy.2011;112:512–16. GoogleScholar  42.VatsN,SinghJ,KairaS.ExtubationoutcomeafterspontaneousbreathingtrialswithT-tubeorpressuresupportventilation.IndianJPhysiotherOccupTher.2012;6:86–9. GoogleScholar  43.El-beleidyASE,KhattabAAE,El-SherbiniSA,etal.Automatictubecompensationversuspressuresupportventilationandextubationoutcomeinchildren;arandomizedcontrolledstudy.ISRNPediatr.2013;871376:1–6.Article  GoogleScholar  44.LourencoIS,FrancoAM,BassettoS,etal.PressuresupportventilationversusspontaneousbreathingwithT-tubeforinterruptingtheventilationaftercardiacoperations.RevBrasCirCardiovasc.2013;28:455–61.Article  PubMed  PubMedCentral  GoogleScholar  45.SherifAA,AtalaahHA.Predictionofweaningoutcome,feasibilityofautomatictubecompensation(ATC)forweaningofchronicobstructivepulmonarydisease(COPD)patientsfrommechanicalventilation.AnesthAnalg.2013;116:S-108. GoogleScholar  46.BilanN,GanjiS.ComparisonofCPAPwithhumidifier,blender,andT-pieceontheoutcomeofweaninginpatientswithneurologicdisorders.IranJChildNeurol.2015;9:42–5.PubMed  PubMedCentral  GoogleScholar  47.ChittawatanaratK,OrrapinS,OrrapinS.AnopenlabelrandomizedcontroltrialbetweenlowpressuresupportandT-piecemethodfordiscontinuationfrommechanicalventilationandextubationingeneralsurgicalintensivecareunits.CritCare.2015;19:P268.Article  PubMedCentral  GoogleScholar  48.TeixeiraSN,OsakuEF,CostaCRLM,etal.Comparisonofproportionalassistventilationplus,T-tubeventilationandpressuresupportventilationasspontaneousbreathingtrialsforextubation:arandomizedstudy.RespirCare.2015;60:1527–35.Article  PubMed  GoogleScholar  49.ZhangB,QinYZ.ComparisonofpressuresupportventilationandT-pieceindeterminingrapidshallowbreathingindexinspontaneousbreathingtrials.AmJMedSci.2014;348:300–5.Article  PubMed  GoogleScholar  50.BilanN,GangiS.Weaningfromventilatorandeffectofblender-humidifieronoutcome.IntJPediatr.2014;2Suppl6:39–45. GoogleScholar  51.PellegriniJAS,MoraesRB,MaccariJG,etal.SpontaneousbreathingtrialswithT-pieceorpressure-supportventilation:asystematicreviewandmeta-analysisofRCTs.RespirCare.2016;61(12):1693–703.Article  PubMed  GoogleScholar  52.BrochardL,RaussA,BenitoS,etal.Comparisonofthreemethodsofgradualwithdrawalfromventilatorysupportduringweaningfrommechanicalventilation.AmJRespirCritCareMed.1994;150:896–903.CAS  Article  PubMed  GoogleScholar  53.EstebanA,FrutosF,TobinMJ,etal.Acomparisonoffourmethodsofweaningpatientsfrommechanicalventilation.NEnglJMed.1995;332:345–50.CAS  Article  PubMed  GoogleScholar  54.KohY,HongSB,LimCM,etal.Effectofanadditional1-hour T-piecetrialonweaningoutcomeatminimalpressuresupport.JCritCare.2000;15:41–5.CAS  Article  PubMed  GoogleScholar  55.VitaccaM,VianelloA,ColomboD,etal.Comparisonoftwomethodsforweaningpatientswithchronicobstructivepulmonarydiseaserequiringmechanicalventilationformorethan15 days.AmJRespirCritCareMed.2001;164:225–30.CAS  Article  PubMed  GoogleScholar  56.JubranA,GrantBJ,DuffnerLA,etal.Effectofpressuresupportvsunassistedbreathingthroughatracheostomycollaronweaningdurationinpatientsrequiringprolongedmechanicalventilation:arandomizedtrial.JAMA.2013;309(7):671–7.CAS  Article  PubMed  PubMedCentral  GoogleScholar  57.SklarMC,BurnsK,RittayamaiN,etal.Efforttobreathewithvariousspontaneousbreathingtrialtechniques:asystematicreviewandphysiologicalmeta-analysis.AmJRespirCritCareMed.2016.[Epubaheadofprint].58.TobinMJ,JubranA.Variableperformanceofweaning-predictortests:roleofBayes’theoremandspectrumandtest-referralbias.IntensiveCareMed.2006;32:2002–12.Article  PubMed  GoogleScholar  59.ThilleAW,RichardJC,BrochardL.Thedecisiontoextubateintheintensivecareunit.AmJRespirCritCareMed.2013;187:1294–302.Article  PubMed  GoogleScholar  60.AfessaB,HogansL,MurphyR.Predicting3-dayand7-dayoutcomesofweaningfrommechanicalventilation.Chest.1999;116:456–61.CAS  Article  PubMed  GoogleScholar  61.StroetzRW,HubmayrRD.Tidalvolumemaintenanceduringweaningwithpressuresupport.AmJRespirCritCareMed.1995;152:1034–40.CAS  Article  PubMed  GoogleScholar  62.BurnsKEA,KarottaiyamvelilJacobS,AguirreV,etal.Stakeholderengagementintrialdesign:asurveyofcitizen’spreferenceforoutcomesandtreatmentoptionsduringweaning.AnnAmThoracSoc.2016;13(11):1962–8.Article  PubMed  GoogleScholar  63.EstebanA,AliaI,TobinMJ,GilA,GordoF,VallverduI,BlanchL,BonetA,VazquezA,dePabloR,TorresA,dealCalMA,MaciasS,fortheSpanishLungFailureCollaborativeGroup.Effectofspontaneousbreathingtrialdurationonoutcomeofattemptstodiscontinuemechanicalventilation.AJRCCM.1999;159:512–8.CAS  GoogleScholar  64.PerrenA,DomenighettiG,MauriS,GeniniF,VizzardiN.Protocol-directedweaningfrommechanicalventilation:clinicaloutcomeinpatientsrandomizedfora30-minor120-mintrialwithpressuresupportventilation.IntensCareMed.2002;28:1058–63.Article  GoogleScholar  65.FernandezMM,FernandezR,MagretM,González-CastroA,BouzaMT,IbañezM,GarcíaC,BalerdiB,MasA,ArauzoV,AñónJM,RuisF,FerreresJ,TomásR,AlabertM,TizónAI,AltabaS,LlamasN.Reconnectiontomechanicalventilationforonehourafterasuccessfulspontaneousbreathingtrialreducesextubationfailureandreintubationincriticallyillpatients:amulticenterrandomizedcontrolledtrial.IntensCareMedExp.2016;4Suppl1:A470. GoogleScholar  DownloadreferencesAcknowledgementsThismanuscriptwaspresentedinabstractformattheannualassemblyoftheAmericanThoracicSociety(2016).TheauthorswishtothankMrDavidLightfoot(librarian,St.Michael’sHospital)forhisassistanceinconductingtheoriginalandupdatedliteraturesearches. Funding Nospecificfundingwasreceivedforthisstudy.Atthetimeofconductingthissystematicreview,KEABandJOFheldClinicianScientist—Phase2AwardsfromtheCanadianInstitutesofHealthResearch,andKEABalsoheldanOntarioMinistryofResearchandInnovationEarlyResearcherAward.LJBholdstheKeenanChairinCriticalCareandAcuteRespiratoryFailure(St.Michael’sHospital,Toronto,Canada). Availabilityofdataandmaterial KEABandJOFhadfullaccesstoallofthedatainthestudy,takeresponsibilityfortheintegrityofthedataandtheaccuracyofthedataanalysis,andwillmakeavailabletheRevManfilesusedtoconducttheanalyses. Authors’contributions KEABandJOFconductedtheliteraturesearches,selectedstudiesmeetinginclusioncriteria,extracteddataandassessedstudyquality,andconductedriskofbiasassessments.KEAB,JOF,IS,JAP,LC,NR,andMSscreenedabstracts.JAPreviewedselectedstudiespublishedinforeignlanguages,extracteddataandassessedstudyquality,andassistedwithadjudicatingwhetherstudiesmetinclusioncriteria.JOFalsoreviewedselectedforeignlanguagepublications.IS,AZ,NKJA,JTYW,andCG-Bsearchedconferenceproceedings.ISandKEABadjudicatedabstractsforinclusion.LJBadjudicateddisagreementsbetweenreviewers(studyselectionandquality).KEAB,JOF,NKJA,JAP,andLJBpreparedtheinitialandsubsequentdraftsofthemanuscript,andintegratedcommentsintorevisedversionsofthemanuscript.JOFdouble-checkeddataentry.Allauthorsrevisedandapprovedthefinalversionofthemanuscript. Competinginterests Allauthorsdeclare:(1)nosupportfromanyorganizationforthesubmittedwork;(2)nofinancialrelationshipswithanyorganizationsthatmighthaveaninterestinthesubmittedworkintheprevious3 years;and(3)nootherrelationshiporactivitiesthatcouldappeartohaveinfluencethesubmittedwork. LJB’slaboratoryreceivedequipmentand/orresearchgrantsfromMaquet(researchonNAVA),Covidien(PAV),Philips(sleep),FisherPaykel(highflow),AirLiquide(CPR),andGeneralElectric(lungvolume);andLJBreceivedlecturefeesfromCovidienandMaquet. Consentforpublication Notapplicable. Ethicsapprovalandconsenttoparticipate Notapplicable. Sourcesofdirectfinancialsupport None. Publisher’sNote SpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. AuthorinformationAffiliationsStMichael’sHospitalandtheKeenanResearchCentre/LiKaShingKnowledgeInstitute,30BondStreet,Toronto,ON,M5B1W8,CanadaKarenE.A.Burns, IbrahimSoliman, AmerZwein, CarolinaGomez-Builes, LuChen, NuttapolRittayamai, MichaelSklar, LaurentJ.Brochard & JanO.FriedrichInterdepartmentalDivisionofCriticalCare,UniversityofToronto,Toronto,ON,CanadaKarenE.A.Burns, IbrahimSoliman, NeillK.J.Adhikari, AmerZwein, CarolinaGomez-Builes, LuChen, NuttapolRittayamai, MichaelSklar, LaurentJ.Brochard & JanO.FriedrichDepartmentofClinicalEpidemiologyandBiostatistics,McMasterUniversity,Hamilton,ON,CanadaKarenE.A.BurnsDepartmentofCriticalCareMedicineandSunnybrookResearchInstitute,SunnybrookHealthSciencesCentre,Toronto,ON,CanadaNeillK.J.AdhikariDepartmentofPublicHealth,UniversityofToronto,Toronto,ON,CanadaJessicaT.Y.WongDivisionofCriticalCareofMoinhosdeVentoHospital,PortoAlegre,BrazilJoseAugustoPellegriniDivisionofCriticalCareofHospitaldeClínicasdePortoAlegre,PortoAlegre,BrazilJoseAugustoPellegriniAuthorsKarenE.A.BurnsViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarIbrahimSolimanViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarNeillK.J.AdhikariViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarAmerZweinViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJessicaT.Y.WongViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarCarolinaGomez-BuilesViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJoseAugustoPellegriniViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarLuChenViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarNuttapolRittayamaiViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarMichaelSklarViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarLaurentJ.BrochardViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJanO.FriedrichViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarCorrespondingauthorCorrespondenceto JanO.Friedrich.Additionalfiles Additionalfile1:Additionalmethods,results,anddiscussion.(DOCX15 kb)Additionalfile2:FigureS1.Riskofbiasfortheincludedtrials.(DOCX30 kb)Additionalfile3:FigureS2.ForestplotcomparingPSvsothertechniquesonSBTsuccess.(DOCX41 kb)Additionalfile4:FigureS3.ForestplotcomparingPSvsothertechniquesonextubationsuccess.(DOCX42 kb)Additionalfile5:FigureS4.ForestplotcomparingSBTtechnique(PSvsothertechnique)onreintubation.Thepooledriskratiowith95%CIwascalculatedusingarandomeffectsmodel.Weightreferstothecontributionofeachstudytotheoverallestimateoftreatmenteffect.(DOCX40 kb)Additionalfile6:TableS1.SummaryoffindingsforPSvsT-pieceSBTsonSBTsuccessbasedonpretestprobability.(DOCX16 kb)Additionalfile7:FigureS5.Subgroupanalysis:forestplotcomparingPSvsT-pieceonSBTsuccessbasedonpretestprobability.(DOCX30 kb)Rightsandpermissions OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.0InternationalLicense(http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.TheCreativeCommonsPublicDomainDedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated. ReprintsandPermissionsAboutthisarticleCitethisarticleBurns,K.E.A.,Soliman,I.,Adhikari,N.K.J.etal.Trialsdirectlycomparingalternativespontaneousbreathingtrialtechniques:asystematicreviewandmeta-analysis. CritCare21,127(2017).https://doi.org/10.1186/s13054-017-1698-xDownloadcitationReceived:03February2017Accepted:02May2017Published:01June2017DOI:https://doi.org/10.1186/s13054-017-1698-xSharethisarticleAnyoneyousharethefollowinglinkwithwillbeabletoreadthiscontent:GetshareablelinkSorry,ashareablelinkisnotcurrentlyavailableforthisarticle.Copytoclipboard ProvidedbytheSpringerNatureSharedItcontent-sharinginitiative KeywordsWeaningSpontaneousbreathingtrialExtubationSystematicreviewMeta-analysisExtubationoutcome DownloadPDF Advertisement CriticalCare ISSN:1364-8535 Contactus Submissionenquiries:AccesshereandclickContactUs Generalenquiries:[email protected]