Comparison of the TIMI, GRACE, PAMI and CADILLAC risk ...

The GRACE score is based on a large registry of patients across the entire spectrum of coronary syndromes and is designed to determine all-cause ... BrowseSubjectAreas ? ClickthroughthePLOStaxonomytofindarticlesinyourfield. FormoreinformationaboutPLOSSubjectAreas,click here. Article Authors Metrics Comments MediaCoverage ReaderComments Figures Figures AbstractAcutecoronarysyndrome(ACS)patientswithdiabeteshavesignificantlyworsecardiovascularoutcomesthanthosewithoutdiabetes.ThisstudyaimedtocomparetheperformanceofTheThrombolysisInMyocardialInfarction(TIMI),GlobalRegistryofAcuteCoronaryEvents(GRACE),PrimaryAngioplastyinMyocardialInfarction(PAMI),andControlledAbciximabandDeviceInvestigationtoLowerLateAngioplastyComplications(CADILLAC)riskscoresinpredictinglong-termcardiovascularoutcomesindiabeticpatientswithST-segmentelevationmyocardialinfarction(STEMI).FromtheAcuteCoronarySyndrome-DiabetesMellitusRegistryoftheTaiwanSocietyofCardiology,patientswithSTEMIwereincluded.TheTIMI,GRACE,PAMI,andCADILLACriskscoreswerecalculated.Thediscriminativepotentialofriskscoreswasanalyzedusingtheareaunderthereceiver-operatingcharacteristicscurve(AUC).Inthe455patientsincluded,allfourriskscoresystemsdemonstratedpredictiveaccuracyfor6-,12-and24-monthmortalitywithAUCvaluesof0.67–0.82.TheCADILLACscorehadthebestdiscriminativeaccuracy,withanAUCof0.8207(p<0.0001),0.8210(p<0.0001),and0.8192(p<0.0001)for6-,12-,and24-monthmortality,respectively.Italsohadthebestpredictivevalueforbleedingandacuterenalfailure,withanAUCof0.7919(p<0.05)and0.9764(p<0.0001),respectively.PatientswithCADILLACriskscores>8hadpoorer2-yearsurvivalthanthosewithlowerscores(log-rankp<0.0001).Inconclusion,theCADILLACriskscoreismoreeffectivethanotherriskscoresinpredicting6-month,1-year,and2-yearall-causemortalityindiabeticpatientswithSTEMI.Italsohadthebestpredictivevalueforin-hospitalbleedingandacuterenalfailure. Citation:KaoY-T,HsiehY-C,HsuC-Y,HuangC-Y,HsiehM-H,LinY-K,etal.(2020)ComparisonoftheTIMI,GRACE,PAMIandCADILLACriskscoresforpredictionoflong-termcardiovascularoutcomesinTaiwanesediabeticpatientswithST-segmentelevationmyocardialinfarction:FromtheregistryoftheTaiwanSocietyofCardiology.PLoSONE15(2): e0229186. https://doi.org/10.1371/journal.pone.0229186Editor:YoshiakiTaniyama,OsakaUniversityGraduateSchoolofMedicine,JAPANReceived:May9,2019;Accepted:February2,2020;Published:February13,2020Copyright:©2020Kaoetal.ThisisanopenaccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.DataAvailability:AllrelevantdataarewithinthemanuscriptanditsSupportingInformationfiles.Funding:TheresearchwasfundedbytheTaiwanSocietyofCardiology(www.tsoc.org.tw)andTaipeiMedicalUniversityWan-FangHospital(https://www.wanfang.gov.tw/)(105swf06)inthecollectionofdatadonebyJSY.Thefundershadnoroleinstudydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript.Competinginterests:Theauthorshavedeclaredthatnocompetinginterestsexist. IntroductionDiabetesmellitus(DM)isassociatedwithpooroutcomesinpatientswithcoronaryarterydisease(CAD)[1].InTaiwan,acutecoronarysyndrome(ACS)patientswithdiabeteshadsignificantlyworseoutcomesthanthosewithoutDM,includingall-causedeathandcombinedresultsfordeath,re-infarction,andstroke[2].ToimproveACS-relatedmortalityandmorbidityinTaiwan,theAcuteCoronarySyndrome-DiabetesMellitusRegistryoftheTaiwanSocietyofCardiology(TSOCACS-DMRegistry)wasestablishedtoassessthequalityofcareforACSpatientswithDM.ThisstudywasconductedtodetermineaccurateriskstratificationinthemanagementofACSpatientswithDM.Severalriskscoreshavebeendevelopedinthelast20yearstostratifypatientshospitalizedwithACS[3–8].ThemostwidelyusedriskscoreistheThrombolysisInMyocardialInfarction(TIMI)algorithm,whichissimpletocalculateandisderivedfromselectedclinical-trialcohorts.ForST-segmentelevationmyocardialinfarction(STEMI)patients,theTIMIscoreisbasedoneightclinicalindicatorsavailableuponadmission,withscoresrangingfrom0to14.ThesecondmostusedscoreistheGlobalRegistryofAcuteCoronaryEvents(GRACE)riskmodel,whichuseseightvariablesandisapplicabletotheentirespectrumofACS.ThePrimaryAngioplastyinMyocardialInfarction(PAMI)scoreisbasedonclinicalandelectrocardiographiccharacteristics.ThePAMIriskscore,witharangeof0to15points,wasfoundtobeastrongpredictoroflatemortalityinSTEMIpatientundergoingprimarypercutaneouscoronaryintervention(PCI)[8].Finally,theControlledAbciximabandDeviceInvestigationtoLowerLateAngioplastyComplications(CADILLAC)riskscoreincorporatesthemeasurementofbaselineleftventricular(LV)function.ItisthesinglemostpowerfulpredictorofsurvivalinACSpatients[3].ForpatientswithSTEMIundergoingPCI,TIMI,PAMI,orCADILLACriskscoresallprovideimportantprognosticinformationandenableaccurateidentificationofhigh-riskpatients[4].Furthermore,TIMIandGRACEriskscorespredict5-yearall-causemortalitywellinpatientswithSTEMItreatedwithprimaryPCI[5].Table1featuresthecomponentsoftheserisk-scoringmodels.However,theseriskscoresweredevelopedbyenrollingpatientsmostlyfromWesterncountries.Althoughtheseriskscoreshavebeenexternallyvalidatedinthegeneralpopulationforpredictingall-causedeathandre-myocardialinfarctionfromtheshorttermtoa1-yearfollow-upperiod,thereislimiteddataontheabilityoftheseriskscoresystemstopredictlong-termcardiovasculareventsandin-hospitaloutcomes,includingacuterenalfailureorbleeding,inspecificpopulationssuchasDMpatients.TheaimsofthispresentprospectiveobservationalstudyweretocomparetheprognosticvalueoffourriskscoresintheriskstratificationofTaiwanesediabeticpatientswithSTEMI,andtoexaminewhethertheseriskscorescouldbeappliedtopredicteithershort-termin-hospitaloutcomesorfuturecardiovasculareventsuptotwoyearsafterSTEMI. Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageTable1.Riskscoringmodelsandtheircomponents. https://doi.org/10.1371/journal.pone.0229186.t001 MethodsThestudycompliedwiththeDeclarationofHelsinkiandwasapprovedbytheTaipeiMedicalUniversity-JointInstitutionalReviewBoard(EthicsReference:201312017).Writteninformedconsentwasobtainedfromallstudyparticipants.AllpatientswereparticipantsintheTSOCACS-DMRegistry.Thisisaprospective,nationwide,multicenter,non-interventional,observationalclinicalregistry–basedstudy,thedetailedrecruitmentproceduresofwhichhavebeenpublished[9].Inbrief,theinclusioncriteriaincludedpatients1)whowereadmittedtothehospitalwithACSwithintheprevious30days;2)withahistoryoftype2DMornewly-diagnosedDMdefinedaccordingtoWorldHealthOrganization(WHO)criteria;3)aged≥20years;and4)whoagreedtoprovideinformedconsent.TheexclusioncriteriaincludedACSaccompaniedwithorprecipitatedbysignificantcomorbiditysuchasseveregastrointestinalbleeding,trauma,peri-operativeorperi-proceduralmyocardialinfarction(MI),orparticipationinaninvestigationaldrugtrial.Intotal,1,534ACSpatientswithDM,including455STEMIpatients,750non-ST-segmentelevationmyocardialinfarction(NSTEMI)patients,and329unstableangina(UA)patients,wereregisteredbetweenJanuary2013andDecember2015.ApprovalforuseoftheTSOCACS-DMRegistrywasacquiredfromtheInstitutionalReviewBoardofeachparticipatinghospital.Allsubjectscompletedsignedinformedconsentandpermissiontorecordfollow-upoutcomes. Patientswithtype2DMwerediagnosedaccordingtothecriteriaoftheAmericanDiabetesAssociationandtheWHO.Thosewhohadalreadytakenoralhypoglycemicagent(s),hadhemoglobinA1Clevelsof6.5%orhigher,orfastingplasmaglucose126mg/dLorhigher,or2-hourpost-prandialbloodsugar200mg/dLorhigher,wereconsideredtobeDMpatients[10]. ACSreferstoaspectrumofconditionscompatiblewithacutemyocardialischemiaand/orinfarctionthatareusuallyduetoanabruptreductionincoronarybloodflow[11].PatientswithACSandelevatedcardiacbiomarkervaluesarediagnosedwithMI.Forthesakeofimmediatetreatmentstrategiessuchasreperfusiontherapy,itisusualpracticetodesignateMIinpatientswithchestdiscomfortorotherischemicsymptomswhodevelopSTelevationintwocontiguousleadsasSTEMI[12]. Alldata,includingdemographiccharacteristics,medicaltherapy,laboratorytests,andinvasivemeasurement,includingquantitativecoronaryanalysisandTIMIflowgradeassessmentafterPCI,werecollectedbyphysiciansandstudynurses.Medicationsortreatmentsuponadmission,durationofhospitalization,andstatusatdischargewerealsocollected.Alldataweresubmittedelectronicallytoacentrallaboratoryandauditedforqualityassurance. Echocardiographicassessmentwascarriedout3–5daysafterMIonset.LeftventricularejectionfractionwasestimatedprimarilyusingthebiplaneSimpson’sformulawithapicaltwo-andfour-chamberviews. Theprimaryendpointofinterestwasall-causemortalityat6months,12months,and24months.Thesecondaryendpointsincludedin-hospitalrecurrentnon-fatalMI,TIMImajor/minorbleeding[13],new-onsetcardiogenicshock,andacuterenalfailure.Acuterenalfailureisdefinedasathree-foldincreaseofserumcreatinineordecreaseinglomerularfiltrationrateof>75%oraurineoutputof<0.3mL/kgperhourfor>24hoursoranuriafor>12hours[14].Allrecordswerecollectedfrommedicalrecordsbywell-trainedstudynurses. Numericaldataarepresentedasthemean±thestandarddeviation(SD)ormedianwithinterquartilerangewhilecategoricalvariablesareshownasfrequencywithpercentage.Thediscriminativepotentialofriskscoreswasperformedusingtheareaunderthereceiver-operatingcharacteristicscurve(AUC)[15].StatisticallysignificantdifferencesbetweenAUCswereexaminedusingDeLong’stest.CalibrationwasevaluatedwithHosmer-Lemeshowgoodness-of-fitX2estimatesusingdeciles[16].Theoptimalcut-offthresholdsweredeterminedbyusingthehighestYoudenindex.The2-yearsurvivalprobabilityofeachriskscoreswereestimatedusingtheKaplan-Meiermethodandexaminedbylog-ranktests.AllanalyseswereperformedusingSASsoftwareversion9.4(SASInstituteInc.,Cary,NC,USA)andSTATAsoftwareversion15.0(STATACorpLP,CollegeStation,TX,USA). ResultsTable2revealsthebaselinecharacteristicsandmedicaltherapyuponhospitaladmissionofthe455STEMIpatientsstudied.Amongthese,theaverageagewas61.5±11.9yearsand78%weremale.Morethan70%oftheseSTEMIpatientshadahistoryofhypertension.About15%werenewlydiagnosedwithDMduringthiscourseofhospitalization.MorethanhalfoftheSTEMIpatientshadKillipclassIseverity(57.4%).Uponhospitaladmission,28.8%oftheseSTEMIpatientsusedinsulin.Resultsoftheremaininglaboratorytests,includingcreatinekinase,glycatedhemoglobin,lipidlevel,andinvasiveprocedures,arepresentedinTable3.Morethanhalf(53.8%)ofpatientshadTIMIflow0uponhospitaladmission.Selectivecoronaryangiographyshowedthree-vesseldiseasein145(31.9%)patients. Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageTable2.Baselinecharacteristicsofpatientsandmedicaltherapyuponhospitaladmission. https://doi.org/10.1371/journal.pone.0229186.t002 Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageTable3.Characteristicsoflaboratorytestsandinvasiveprocedures. https://doi.org/10.1371/journal.pone.0229186.t003TheAUCofeachriskscoreforprimaryandsecondaryendpointsareshowninTable4.TheCADILLACriskscorehadthebestdiscriminativeaccuracy,withanAUCof0.8207(p<0.0001),0.8210(p<0.0001),and0.8192(p<0.0001)for6-,12-,and24-monthmortality,respectively.Calibrationofeachriskscorewasperformedbycomparingpredictedprobabilitieswith6-,12-,and24-monthmortalityestimates,andallmodelshadanadequategoodness-of-fit(S1Fig,S2Fig,S3Fig,andS4Fig).TheAUCcalculatedforeachofriskscoremodelsformortalityat24monthsoffollowupareshowninFig1ThehighestperformanceoftheCADILLACriskscorewasobserved.Inaddition,theCADILLACriskscorealsohadthebestpredictivevalueforin-hospitalbleedingandacuterenalfailure,withanAUCof0.7919(p<0.05)and0.9764(p<0.0001),respectively.Asforin-hospitalrepeatedMI,theGRACEriskscorehadthehighestpredictiveaccuracywithanAUCof0.9288(p<0.05).Similarly,theGRACEscorewasthebestpredictivetoolfornewonsetcardiogenicshock,withanAUCof0.8648(p<0.0001). Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageFig1.PredictiveaccuracyofPAMI,TIMI,CADILLACandGRACEscoringmodelsfor2-yearmortality.Accordingtothetertilesofeachriskscore,patientswiththehighertertile(T3)ofeachriskscorehadunfavorable2-yearsurvivalthanthosewithmiddletertile(T2)andlowertertile(T1)ofeachriskscore(Fig2).However,therewasnodeathsubjectsinT1andT2oftheCADILLACscoregroup.Wefurtherfoundthebestcut-offpointfortheCADILLACriskscore(8points)byusingYouden’sindex(Table5).SincetheCADILLACriskscorehadthebestpredictiveaccuracyformortality,weusedittoestimatethesurvivalrateat2years,asshowninFig3.PatientswithCADILLACriskscores>8hadpoorer2-yearsurvivalthanthosewithriskscores≤8(bothlog-rankp<0.0001). https://doi.org/10.1371/journal.pone.0229186.g001 Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageFig2.Kaplan-MeieroverallsurvivalcurveforpatientswithSTEMIstratifiedbytertilesofTIMI,PAMI,CADILLACandGRACEriskscores.Thecut-offpointsofthesetertiles(T1-T3)are≤2,3,and≥4forTIMI;≤3,4–5,and≥6forPAMI;≤3,4–5,and≥6CADILLAC;≤115,116–137,and≥138forGRACE,respectively. https://doi.org/10.1371/journal.pone.0229186.g002 Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageFig3.ObservedsurvivalbyControlledAbciximabandDeviceInvestigationtoLowerLateAngioplastyComplications(CADILLAC)scores.Two-yearmortalityofpatientswithCADILLACscores≤8and>8(forboth,log-ranktestp<0.0001). https://doi.org/10.1371/journal.pone.0229186.g003 Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageTable4.TheAUCoffourscoringmodelsforprimaryandsecondaryendpoints. https://doi.org/10.1371/journal.pone.0229186.t004 Download: PPTPowerPointslidePNGlargerimageTIFForiginalimageTable5.TheAUC,sensitivity,specificity,Youden’sindex,andcut-offpointofeachriskscore. https://doi.org/10.1371/journal.pone.0229186.t005 DiscussionInthecurrentstudy,4riskstratificationmodels(TIMI,GRACE,PAMI,CADILLAC)werecomparedinTaiwanesediabeticpatientsdiagnosedwithSTEMIaccordingtotheACSguidelinesoftheTSOC.Toourbestknowledge,thepresentstudyisthefirsttodemonstratetherelevantdiscriminatoryabilityofthesefourriskscoresformortalityandclinicaloutcomesattimepointsuptotwoyearsinacohortofdiabeticpatientswithSTEMI.Thefourmodelshadgoodpredictivevalueinestimating2-yearmortality,althoughtheAUCswereslightlydifferent.AUCvaluesobtainedfromourdatabaseforpredictingone-yearmortalitywereevenbetterthanthosedescribedbytheoriginalauthors[4,5].Furthermore,theCADILLACriskscorehadthebestpredictivevalueforbleedingandacuterenalfailureandthehighestprognosticaccuracyformortalityateachobservedtimepoint,including6-month,1-year,and2-year.OurstudydemonstratedthattheCADILLACriskscoreisthebesttoolforpredictionlong-termmortalityinTaiwanesediabeticpatientsdiagnosedwithSTEMI,accordingtothenationwidereal-worldregistry. Thepredictiveaccuracywas0.82at1-yearfollow-upfortheCADILLACriskscoreinourdataset,consistentwiththefactthatthescorewasoriginallydevelopedtodetermine1-yearsurvival.Inthepresentstudy,theCADILLACriskscorehadahigherprognosticaccuracythanthe0.741-yearmortalitypredictioninthestudybyKozieradzkaetal[5].Additionally,theAUCremainedunchangedandretainedaverygoodpredictivepowerfor2-yearmortality.ItperformedbetterthaninthederivationandvalidationsetsoftheCADILLACrandomizedclinicaltrial,inwhichtheprognosticaccuracywas0.79[4]. SomefactorsmayexplainthebetterprognosticpoweroftheCADILLACriskscoreinthisstudy.First,theCADILLACriskscorewasdevelopedbasedon1-yearsurvivalanalysis.Itistheonlyriskscoringmodelwhichtakesintoconsiderationejectionfractionandthree-vesseldisease.InaKoreanclinicalregistry,LVdysfunction,poorTIMIflowafterPCI,andmulti-vesseldiseasewereassociatedwithlongtermmajorcardiovasculareventsafterMI[17].ThepresenceofLVdysfunctionassessedbybaselineleftventriculographyinpatientswhoundergoPCIisapowerfulpredictorofearlyandlate(3-year)mortality[18].Inthisregistry,98%oftheSTEMIpopulationreceivedcoronaryangiogram.Detailedassessmentofculpritlesions,TIMIflowandLVfunctionwerecompletedaccordingly.Theseclinicaldatamayprovideadditionalprognosticrelevantinformationforthestudypopulation.ThoseSTEMIpatientswithLVdysfunctionmaydevelopcardiorenalsyndrome,reflectinganabruptworseningofcardiacfunctionleadingtoacutekidneyinjury[19].Second,elevatedserumcreatinineandanemiahadbeenprovedtobetheindependentbaselinepredictorstopredictbleedinginpatientswithacutecoronarysyndrome[20].TheCADILLACriskscoreistheonlysystemthatconsidersanemiaandcreatinine(renalinsufficiency)inthese4riskstratificationmodels.Therefore,itshouldbemoresensitiveinpredictingin-hospitalbleedingandacuterenalfailure. TheGRACEscoreisbasedonalargeregistryofpatientsacrosstheentirespectrumofcoronarysyndromesandisdesignedtodetermineall-causemortalityat6months[21,22].Thepoorperformance(0.67)oftheGRACEscoreinpredictingall-causedeathat6monthsofourdatasetwasexpected.ThepooraccuracycanbeexplainedbytherelativelylowernumberofKillipIII/IVpatients(around20%).However,theGRACEscoredidhavethehighestprognosticaccuracyforsecondaryendpointsregardingrecurrentMIandnew-onsetcardiogenicshock(Table4).Theexactreasonforthisaccuracyisnotentirelyclear.Itmaybeduetothepriorepisodeofcardiacarrestandtheincreasednumberofcardiacmarkers,botharethecomponentsofGRACEscore,indicatingmyocardialdamageinprogressandcausingfurthereventssuchascardiogenicshockandrecurrentMI.Theother3riskscoringmodels(TIMI,PAMI,andCADILLAC)donotincludethecomponentofcardiacarrestorincreasedcardiacmarkers.Therefore,GRACEscoremightbemoresensitiveinpredictingin-hospitalrecurrentMIandnew-onsetcardiogenicshock. Becauseofguideline-directedmedicaltherapyandinterventions,in-hospitaland1-yearmortalityratesforpatientswithSTEMIhavesignificantlydecreased[23–27].However,DMisstillanindependentpredictorof3-yearmortalityand3-yearmajoradversecardiacevents[28].Therefore,itisimportanttousebestpracticesguidelinestomanagediabeticSTEMIpatients.Cliniciansshouldalsoemphasizeevidence-basedmedicaltherapiesandavailablereperfusiontherapy.ForthosediabeticSTEMIpatientswithhigherCADILLACscores(>8),strictadherencetooptimalmedicaltreatmentismandatory. Ourworkhadthreemainlimitations.Firstly,ourdatasetincluded15%newlydiagnosedDMpatients.Wecouldnotrecordtheexactdiagnosisyearofknowndiabeticpatientsinourcohortbecauseoflimitedinformation.Nevertheless,therewerenosignificantdifferencesinadverseeventsbetweennew-diagnosedandknowndiabeticpatients[28].Secondly,thestudycohortwasrelativelysmallandlimitedindiabeticpopulation.However,thesepatientswerefollowedupprospectivelyandthedatathoroughlyanalyzed.Finally,missingdatapreventedourapplicationinourstudycohortofnewerriskscoringmodelssuchastheSyntaxscore.Theangiography-basedscoringmodelwascreatedforpredictinglong-termmajoradversecardiaceventswhentreatingseverecoronaryarterydiseasesuchasmulti-vesseldiseaseorleftmaincoronaryarteryinvolvement.Fully31.9%ofpatientsofourdatasethadthree-vesseldisease.Suchpatientsshouldbeevaluatedforpotentialrevascularization. ConclusionsSeveralriskscoringmodelsshowedahighpredictivevaluetoestimate1-yearmortalityinTaiwanesediabeticSTEMIpatients.Amongthem,theCADILLACsystemwassuperioratpredicting6-month,1-year,and2-yearmortality.WeshouldespeciallymonitorpatientswithhigherCADILLACscores(>8).Strictadherencetomedicaltherapyguidelinesandintensivecardiovascularriskfactormodificationshouldbeencouraged. SupportinginformationS1Fig.CalibrationplotfortheTIMIscorefor2-year,1-year,and6-monthdeath.2-yearHosmerandLemeshowgoodness-of-fittestp=0.54001-yearHosmerandLemeshowgoodness-of-fittestp=0.39916-monthHosmerandLemeshowgoodness-of-fittestp=0.4618Abbreviation:TIMI,ThrombolysisInMyocardialInfarction. https://doi.org/10.1371/journal.pone.0229186.s001(DOCX) S2Fig.CalibrationplotforthePAMIscorefor2-year,1-year,and6-monthdeath.2-yearHosmerandLemeshowgoodness-of-fittestp=0.67861-yearHosmerandLemeshowgoodness-of-fittestp=0.64226-monthHosmerandLemeshowgoodness-of-fittestp=0.2774Abbreviation:PAMI,PrimaryAngioplastyinMyocardialInfarction. https://doi.org/10.1371/journal.pone.0229186.s002(DOCX) S3Fig.CalibrationplotfortheCADILLACscorefor2-year,1-year,and6-monthdeath.2-yearHosmerandLemeshowgoodness-of-fittestp=0.54161-yearHosmerandLemeshowgoodness-of-fittestp=0.65056-monthHosmerandLemeshowgoodness-of-fittestp=0.8121Abbreviation:CADILLAC,ControlledAbciximabandDeviceInvestigationtoLowerLateAngioplastyComplications. https://doi.org/10.1371/journal.pone.0229186.s003(DOCX) S4Fig.CalibrationplotfortheGRACEscorefor2-year,1-year,and6-monthdeath.2-yearHosmerandLemeshowgoodness-of-fittestp=0.63691-yearHosmerandLemeshowgoodness-of-fittestp=0.46776-monthHosmerandLemeshowgoodness-of-fittestp=0.7567Abbreviation:GRACE,GlobalRegistryofAcuteCoronaryEvents. https://doi.org/10.1371/journal.pone.0229186.s004(DOCX) Acknowledgments OnbehalfoftheTSOCACSDMRegistryStudyGroup:Wei-HsianYin,Chih-ChengWu,Shih-HungChan,Yen-WenWu,Kuo-YungWang,Juey-JenHwang,Kuan-ChengChang,Wen-CholVoon,I-ChangHsieh,Kou-GiShyu,JunTedChong,Wei-ShiangLin,Chih-NengHsu,Kwo-ChangUeng,Chih-PingHsia,JongShiuanYeh,Ju-ChiLiuGuang-YuanMar,Jhih-YuanShih,Jen-YuanKuo,Hsuan-MingTsao,Wei-KungTseng,Cheng-HsuYANG,Chao-ChienChang,Jeng-FengLin,Chern-EnChiang,andMeng-HuanLei. References1. LeonBM,MaddoxTM.Diabetesandcardiovasculardisease:Epidemiology,biologicalmechanisms,treatmentrecommendationsandfutureresearch.WorldJDiabetes.2015;6:1246–1258.pmid:26468341 ViewArticle PubMed/NCBI GoogleScholar 2. WeiCC,ShyuKG,ChengJJ,LoHM,ChiuCZ.Diabetesandadversecardiovascularoutcomesinpatientswithacutecoronarysyndrome–DatafromTaiwan’sAcuteCoronarySyndromeFullSpectrumDataRegistry.ActaCardiolSin.2016;32:31–38.pmid:27171367 ViewArticle PubMed/NCBI GoogleScholar 3. HalkinA,SinghM,NikolskyE,GrinesCL,TchengJE,GarciaE,etal.Predictionofmortalityafterprimarypercutaneouscoronaryinterventionforacutemyocardialinfarction:TheCADILLACriskscore.JAmCollCardiol.2005;45:1397–1405.pmid:15862409 ViewArticle PubMed/NCBI GoogleScholar 4. LittnerovaS,KalaP,JarkovskyJ,KubkovaL,PrymusovaK,KubenaP,etal.GRACEscoreamongsixriskscoringsystems(CADILLAC,PAMI,TIMI,DynamicTIMI,Zwolle)demonstratedthebestpredictivevalueforpredictionoflong-termmortalityinpatientswithST-elevationmyocardialinfarction.PLoSOne.2015;10:e0123215.pmid:25893501 ViewArticle PubMed/NCBI GoogleScholar 5. KozieradzkaA,KamińskiKA,MaciorkowskaD,OlszewskaM,DobrzyckiS,NowakK,etal.GRACE,TIMI,ZwolleandCADILLACriskscores—dotheypredict5-yearoutcomesafterST-elevationmyocardialinfarctiontreatedinvasively?IntJCardiol.2011;148:70–75.pmid:19923016 ViewArticle PubMed/NCBI GoogleScholar 6. AminST,MorrowDA,BraunwaldE,SloanS,ContantC,MurphyS,etal.DynamicTIMIriskscoreforSTEMI.JAmHeartAssoc.2013;2:e003269.pmid:23525425 ViewArticle PubMed/NCBI GoogleScholar 7. BrkovicV,DobricM,BeleslinB,GigaV,VukcevicV,StojkovicS,etal.AdditiveprognosticvalueoftheSYNTAXscoreoverGRACE,TIMI,ZWOLLE,CADILLACandPAMIriskscoresinpatientswithacuteST-segmentelevationmyocardialinfarctiontreatedbyprimarypercutaneouscoronaryintervention.IntJCardiovascImaging.2013;29:1215–1228.pmid:23443340 ViewArticle PubMed/NCBI GoogleScholar 8. ScruthEA,ChengE,Worrall-CarterL.RiskscorecomparisonofoutcomesinpatientspresentingwithST-elevationmyocardialinfarctiontreatedwithpercutaneouscoronaryintervention.EurJCardiovascNurs.2013;12:330–336.pmid:22694809 ViewArticle PubMed/NCBI GoogleScholar 9. ChenKC,YinWH,WuCC,ChanSH,WuYW,YangWangK,etal.In-hospitalimplementationofevidence-basedmedicationsisassociatedwithimprovedsurvivalindiabeticpatientswithacutecoronarysyndrome–datafromTSOCACS-DMregistry.ActaCardiolSin.2018;34:211–223.pmid:29844642 ViewArticle PubMed/NCBI GoogleScholar 10. AmericanDiabetesAssociation.Diagnosisandclassificationofdiabetesmellitus.DiabetesCare.2014;37Suppl1:81–90.pmid:23959568 ViewArticle PubMed/NCBI GoogleScholar 11. AmsterdamEA,WengerNK,BrindisRG,CaseyDEJr,GaniatsTG,HolmesDRJr,etal.2014AHA/ACCguidelineforthemanagementofpatientswithnon-ST-elevationacutecoronarysyndromes:AreportoftheAmericanCollegeofCardiology/AmericanHeartAssociationtaskforceonpracticeguidelines.JAmCollCardiol.2014;64:e139–228.pmid:25260718 ViewArticle PubMed/NCBI GoogleScholar 12. ThygesenK,AlpertJS,JaffeAS,SimoonsML,ChaitmanBR,WhiteHD,etal.Thirduniversaldefinitionofmyocardialinfarction.Circulation.2012;126:2020–2035.pmid:22923432 ViewArticle PubMed/NCBI GoogleScholar 13. CannonCP,BattlerA,BrindisRG,CoxJL,EllisSG,EveryNR,etal.AmericanCollegeofCardiologykeydataelementsanddefinitionsformeasuringtheclinicalmanagementandoutcomesofpatientswithacutecoronarysyndrome.AreportoftheAmericanCollegeofCardiologyTaskForceonClinicalDataStandards(AcuteCoronarySyndromeWritingCommittee).JAmCollCardiol.2001;38:2114–2130pmid:11738323 ViewArticle PubMed/NCBI GoogleScholar 14. BellomoR,RoncoC,KellumJA,MehtaRL,PalevskyP,theADQIworkgroup:Acuterenalfailure–definition,outcomemeasures,animalmodels,fluidtherapyandinformationtechnologyneeds:theSecondInternationalConsensusConferenceoftheAcuteDialysisQualityInitiative(ADQI)Group.CritCare.2004;8:R204–212.pmid:15312219 ViewArticle PubMed/NCBI GoogleScholar 15. HanleyJA,McNeilBJ.Amethodofcomparingtheareasunderreceiveroperatingcharacteristiccurvesderivedfromthesamecases.Radiology.1983;148:839–843.pmid:6878708 ViewArticle PubMed/NCBI GoogleScholar 16. LemeshowS,HosmerDWJr.Areviewofgoodnessoffitstatisticsforuseinthedevelopmentoflogisticregressionmodels.AmJEpidemiol,1982;115:92–106.pmid:7055134 ViewArticle PubMed/NCBI GoogleScholar 17. ParkHW,KangMG,KimK,KohJS,ParkJR,JeongYH,etal.Long-termprognosisandclinicalcharacteristicsofpatientswithnewlydiagnoseddiabetesmellitusdetectedafterfirstacutemyocardialinfarction:FromKAMIR-NIHregistry.KoreanCircJ.2018;48:134–147.pmid:29441746 ViewArticle PubMed/NCBI GoogleScholar 18. DaneaultB,GénéreuxP,KirtaneAJ,WitzenbichlerB,GuagliumiG,ParadisJM,etal.Comparisonofthree-yearoutcomesafterprimarypercutaneouscoronaryinterventioninpatientswithleftventricularejectionfraction<40%versus≥40%(fromtheHORIZONS-AMItrial).AmJCardiol.2013;111:12–20.pmid:23040595 ViewArticle PubMed/NCBI GoogleScholar 19. RoncoC,HaapioM,HouseAA,AnavekarN,BellomoR.Cardiorenalsyndrome.JAmCollCardiol2008;52:1527–1539.pmid:19007588 ViewArticle PubMed/NCBI GoogleScholar 20. MehranR,PocockSJ,NikolskyE,ClaytonT,DangasGD,KirtaneAJ,etal.Ariskscoretopredictbleedinginpatientswithacutecoronarysyndrome.JAmCollCardiol.2010;55:2556–2566.pmid:20513595 ViewArticle PubMed/NCBI GoogleScholar 21. EagleKA,LimMJ,DabbousOH,PieperKS,GoldbergRJ,VandeWerfF,etal.Avalidatedpredictionmodelforallformsofacutecoronarysyndrome:Estimatingtheriskof6-monthpostdischargedeathinaninternationalregistry.JAMA.2004;291:2727–2733.pmid:15187054 ViewArticle PubMed/NCBI GoogleScholar 22. ElLev,KornowskiR,Vaknin-AssaH,PorterA,TeplitskyI,Ben-DorI,etal.ComparisonofthepredictivevalueoffourdifferentriskscoresforoutcomesofpatientswithST-elevationacutemyocardialinfarctionundergoingprimarypercutaneouscoronaryintervention.AmJCardiol.2008;102:6–11.pmid:18572028 ViewArticle PubMed/NCBI GoogleScholar 23. FoxKA,StegPG,EagleKA,GoodmanSG,AndersonFAJr,GrangerCB,etal.Declineinratesofdeathandheartfailureinacutecoronarysyndromes,1999–2006.JAMA.2007;297:1892–1900.pmid:17473299 ViewArticle PubMed/NCBI GoogleScholar 24. MandelzweigL,BattlerA,BoykoV,BuenoH,DanchinN,FilippatosG,etal.ThesecondEuroHeartSurveyonacutecoronarysyndromes:Characteristics,treatment,andoutcomeofpatientswithACSinEuropeandtheMediterraneanBasinin2004.EurHeartJ.2006;27:2285–2293.pmid:16908490 ViewArticle PubMed/NCBI GoogleScholar 25. McManusDD,GoreJ,YarzebskiJ,SpencerF,LessardD,GoldbergRJ.Recenttrendsintheincidence,treatment,andoutcomesofpatientswithSTEMIandNSTEMI.AmJMed.2011;124:40–47.pmid:21187184 ViewArticle PubMed/NCBI GoogleScholar 26. RoeMT,MessengerJC,WeintraubWS,CannonCP,FonarowGC,DaiD,etal.Treatments,trends,andoutcomesofacutemyocardialinfarctionandpercutaneouscoronaryintervention.JAmCollCardiol.2010;56:254–263.pmid:20633817 ViewArticle PubMed/NCBI GoogleScholar 27. JernbergT,JohansonP,HeldC,SvennbladB,LindbäckJ,WallentinL.Associationbetweenadoptionofevidence-basedtreatmentandsurvivalforpatientswithST-elevationmyocardialinfarction.JAMA.2011;305:1677–1684.pmid:21521849 ViewArticle PubMed/NCBI GoogleScholar 28. ErteltK,BrenerSJ,MehranR,Ben-YehudaO,McAndrewT,StoneGW.ComparisonofoutcomesandprognosisofpatientswithversuswithoutnewlydiagnoseddiabetesmellitusafterprimarypercutaneouscoronaryinterventionforST-elevationmyocardialinfarction(theHORIZONS-AMIStudy).AmJCardiol.2017;119:1917–1923.pmid:28427734 ViewArticle PubMed/NCBI GoogleScholar DownloadPDF   Citation XML Print Printarticle Reprints Share Reddit Facebook LinkedIn Mendeley Twitter Email   Advertisement SubjectAreas? FormoreinformationaboutPLOSSubjectAreas,click here. Wewantyourfeedback.DotheseSubjectAreasmakesenseforthisarticle?ClickthetargetnexttotheincorrectSubjectAreaandletusknow.Thanksforyourhelp! Myocardialinfarction   IstheSubjectArea"Myocardialinfarction"applicabletothisarticle? Yes No Thanksforyourfeedback. Cardiovasculardiseaserisk   IstheSubjectArea"Cardiovasculardiseaserisk"applicabletothisarticle? Yes No Thanksforyourfeedback. Diabetesmellitus   IstheSubjectArea"Diabetesmellitus"applicabletothisarticle? Yes No Thanksforyourfeedback. Coronaryangioplasty   IstheSubjectArea"Coronaryangioplasty"applicabletothisarticle? Yes No Thanksforyourfeedback. Medicalriskfactors   IstheSubjectArea"Medicalriskfactors"applicabletothisarticle? Yes No Thanksforyourfeedback. Angioplasty   IstheSubjectArea"Angioplasty"applicabletothisarticle? Yes No Thanksforyourfeedback. Coronaryheartdisease   IstheSubjectArea"Coronaryheartdisease"applicabletothisarticle? Yes No Thanksforyourfeedback. Hemorrhage   IstheSubjectArea"Hemorrhage"applicabletothisarticle? Yes No Thanksforyourfeedback.