全球急性冠脉综合征注册研究原文(GRACE研究)

TheGlobalRegistryofAcuteCoronaryEvents,1999to2009eGRACEKAAFox,1KAEagle,2JMGore,3PhGSteg,4FAAnderson,3fortheGRACEandGRACE2Investigators5ABSTRACTTheaimofGRACEwastoprovidealargemultinationalregistryofthefullspectrumofpatientswithacutecoronarysyndromes(ACS)inordertodefinepatientcharacteristicsandoutcomesandderivepredictiveriskscores.Thestudywasdesignedandadministeredbyanindependentsteeringcommittee;dataanalyseswereperformedundertheguidanceofthesteeringcommitteeattheCenterforOutcomesResearchoftheUniversityofMassachusetts.Regularfeedbackregardinglocal,regionalandinternationalguidelineandperformancemeasureswasprovidedtoindividualhospitalsandclustersofhospitals.Regionalandinternationalbenchmarkdatawereavailabletoallsites.MainGRACEinvolved123hospitalsin14countriesinNorthandSouthAmerica,Europe,AustraliaandNewZealand.GRACE2(ExpandedGRACE)comprised154hospitalsinEurope,NorthandSouthAmerica,Asia,AustralasiaandChina.Continuousrecruitmentandfollow-uptookplacebetween1999and2009.Thefirst10-20patientspersite(dependingonhospitalsize)wereenrolledeachmonth,resultingintherecruitmentof102341patients,whowerecategorizedashavingST-segmentelevationmyocardialinfarction,non-ST-elevationmyocardialinfarctionorunstableangina.Standardizedcasereportforms(datafaxorelectronic)werecompletedbytrainedstudycoordinators,andincludedfieldsrelatingtodemographicfactors,comorbidconditions,treatmentsandin-hospitalandpost-discharge(6-month)events.Bloodsampling,geneticanalysesandlonger-termfollow-upwereundertakeninGRACEsubstudies.Prospectiveindividualpatientfollow-upwascarriedout.Allsiteswereauditedlocally;10%ofindividualpatientrecordswereauditedina2-yearcycle.Lessthan1%of20keybaselinefields,andlessthan1%ofdischargediagnosisanddischargestatusdata,weremissing.Six-monthfollow-upwas85%complete.Publicationsandriskscoresareavailableatficanalyseswereconsidered,incompetition,byanindependentpublicationscommittee.INTRODUCTIONTheGlobalRegistryofAcuteCoronaryEvents(GRACE)programmewasestablished,denovo,in1999toresolvemajoruncertaintiesintowhatconstitutesanacutecoronarysyndrome(ACS),todefinehowpatientswithanACSaretreated,andtocharacterisetheiroutcomes.Thisisadynamicprocessandthiscontinuousdecade-longstudyhasprovidedatemporalreflectionofpracticebetween1999and2009.Thisapproachdiffersfromcross-sectional‘snapshot’surveys,whichprovidedataonlyatspecifictimepoints.TheGRACEpublicationshavedescribedthespectrumofpatientswithsuspectedACS(includingST-segmentelevationmyocardialinfarction(STEMI),non-ST-elevationmyocardialinfarctionandACSwithoutbiomarkerrelease),theirriskpredictorsandtheirin-hospitaland6-monthoutcomes.GRACEaimedtonarrowthe‘gap’betweenevidenceandclinicalpractice.Byprovidingfeedbackwithareferencestandardofrobustregionalandinternationaldataeachquarter,acliniciancouldindexlocalhospitalpracticetolargerdatasetsandidentifyopportunitiestoimprovepractice.GRACEcomplementsinforma-tionfromrandomisedtrialsinselectedpopulations:itdefineshowpracticeisappliedinalarge‘real-world’reflectionofthefullspectrumofacutecoronarydisease.SeverallargeobservationalstudieshavebeenconductedinpatientswithACS(table1),10e17andtheyvaryintheextenttowhichtheycomplywithproposedqualitystandardsofdesign,reportingandqualityassurance.18ThemostcriticalissueiswhetheraregistryreflectsthefullspectrumofACS,ratherthanaselectedpopulation(eg,thosetreatedininterventionalcentresorpatientsiden-tifiedonlyfromcardiaccareunits).GRACEMETHODSGRACEwasdesignedtoreflectanunselectedpopulationofpatientswithACS,irrespectiveofgeographicalregion.Atotalof123hospitalslocatedin14countriesinNorthandSouthAmerica,Europe,AustraliaandNewZealandhavecontrib-uteddatatothisobservationalcohortstudy.Allparticipatingcountriesandhospitalclusterswereestablishedattheoutset.Toavoidsiteselectionbiasclusterswererequiredtoincludeacompletespec-trumofhospitalsthatadmitpatientswithACS(withinageographicalregion).Thiswasvalidatedforeachregion.Toavoidinclusionbiasthefirst10e20patients(dependingonhospitalsize)admittedwithsuspectedACSineachcalendarmonthwere‘tracked’,irrespectiveoftheireventualhospitallocation(includingcardiacunits,medicalunits,careoftheelderlyandintensivecareunits).Thecyclicauditprogrammeofallsites(two-yearcyclewith10%ofallpatientsauditedbyaseniorGRACEcoordinatorvisitingeachcluster)wasdesignedtominimisetheriskofinclusionbias.GRACEemployedlocaltraining,rigorousqualitycontrolandauditofparticipatingcentres.Inthe‘warmpursuit’design,thetrackingofpatientsafterarrivalintheemergencydepartmentensuresthatpatientscaredforoutsidecardiacunits(eg,careoftheAdditionaldataarepublishedonlineonly.Toviewthesefilespleasevisitthejournalonline()1CentreforCardiovascularScience,UniversityofEdinburgh,Edinburgh,UK2UniversityofMichiganCardiovascularCenter,AnnArbor,Michigan,USA3CenterforOutcomesResearch,UniversityofMassachusettsMedicalSchool,Worcester,Massachusetts,USA4INSERMU-698,Universite´Paris7,AP-HP,CentreHospitalierBichat-ClaudeBernard,Paris,France5SeeappendixforfulllistofGRACEandGRACE2investigatorsandcoordinatorsC