搜索
2021重点-PLOSONEAMI行PCI术前干预的meta(南方医)NicorandilinPatientswithAcuteMyocardialInfarctionUndergoingPrimaryPercutaneousCoronaryIntervention:ASystematicReviewandMeta-AnalysisMuliWu1?,ZhengHuang1*?,HaojunXie2,ZhongjiangZhou11DepartmentofCardiology,NanfangHospital,SouthernMedicalUniversity,Guangzhou,Guangdong,China,2DepartmentofRespiratoryandCriticalCareMedicine,NanfangHospital,SouthernMedicalUniversity,Guangzhou,Guangdong,ChinaIntroductionEarlyreperfusionoftotallyoccludedcoronaryarteriesreducesinfarctsize,cardiacmortalityrates,andin-hospitalevents[1,2].However,somepatientscontinuetohavedeterioratingcardiacfunctionandbadprognoses,thoughttobeduetoreperfusioninjury.Reperfusioninjuryprobablyrepresentsmyocytecelldeathduetoreperfusion,noreflow,myocardialstunningandreperfusionarrhythmias[3].Nicorandil,ahybridofanadenosinetriphosphate(ATP)-sensitivepotassiumchannelopenerandnitrates,wasusedasanadjunctivetherapywithprimarypercutaneouscoronaryintervention(PCI)foracutemyocardialinfarction(AMI).Themechanismsforthesalutaryactionsofnicorandilhavebeenpostulated,includinganti-freeradicalandneutrophilmodulatingproperties[4,5],vasodilatationofsmallcoronaryarteries[6],andmimickingofischaemicpreconditioning[7].Butrelevantclinicaltrialsshowedcontroversialresultsonwhethernicorandilhadpotentialtoimprovecoronaryarteryreflowandventricularfunction[8-10].Inaddition,rareeventsofclinicaloutcomeswerereportedbecauseofmanytrialswithlimitedsamplesize.Thusweperformedasystematicreviewofrandomizedcontrolledtrials(RCTs)toinvestigatetheeffectofnicorandilpriortoreperfusiontherapyoncardioprotectionandclinicaloutcomesinpatientswithAMI.MethodsDatasourcesandsearchesWeidentifiedallpublishedstudies,includingfull-textandabstract,whichcomparedtheeffectofnicorandilbyintravenousand/orintracoronaryadministrationwithcontrol(placeboornonicorandiltreatment)priortoreperfusionbyprimaryPCIinAMIpatients.SearcheswereperformedusingCochraneCentralRegisterofControlledTrials(CENTRAL)(TheCochraneLibraryIssue12,2021),Pubmed(1966toDecember2021),(1974toNovember2021)andISImeetingandproceedings(previoustoNovember2021)electronicdatabases.Thesearchstrategywasdevelopedwithoutlanguageandpublicationrestrictionandusedthemedicalsubjectheadingsandtextwords,suchas“nicorandil”,“myocardialinfarction”,and“randomizedcontrolledtrials”.Wehandsearchedthesupplementsofseveninternationalcorejournalsonthecardiovascularfield,andmanuallyscannedthereferencelistsofalleligiblearticlesandrelevantmeta-analyses.WealsosearchedGoogleScholar,TCTMDand(seeTableS1).StudiesselectionTworeviewers(M.W.andH.X.)performedstudyselectionindependently,withdisagreementssolvedthroughdiscussionandbytheopinionofathirdreviewer(Z.H.)ifnecessary.Studieswereconsideredpotentiallyeligibleforthissystematicreviewiftheymetthefollowingcriteria:(1)RCTsaboutpatientswhosufferedfromAMIandperformedprimaryPCI,(2)nicorandilwasadministeredpriortoreperfusionbyintravenousand/orintracoronaryandcomparedwithcontrol(placeboornonicorandiltreatment),and(3)thestudiesincludedatleastoneofthefollowinginterestingoutcomes:thrombolysisinmyocardialinfarction(TIMI)flowgradeafterPCI,TIMIframecountafterPCI,leftventricularejectionfraction(LVEF),peakcreatinekinase(CK)valueandclinicaloutcomesincludingcardiacdeath,ventriculartachycardia(VT)orfibrillation(VF),orcongestiveheartfailure(CHF)DataextractionandqualityassessmentTworeviewers(M.W.andH.X.)independentlyundertookthedataextractionandthequalityassessment.Thedisagreementswouldbesolvedthroughdiscussionandbytheopinionofathirdreviewer(Z.H.)ifnecessary.TheriskofbiaswasassessedaccordingtotheguidanceintheCochraneHandbookversion5.1.0[11].Weratedtheriskofselectionbiasbyassessingrandomizationandallocationconcealment,ofperformancebiasbyassessingblindingofparticipantsandpersonnel.TheriskofdetectionbiasandattritionbiaswereratedbyassessingblindingofoutcomeassessmentfortheoutcomeTIMIflowgradeandLVEF,respectivelyandselectionbiasfortheoutcomeTIMIflowgradeonly.Wetriedtocontacttheauthorsbyemailforobtaininginformation,iftheirarticlesdidnotreporttheinformationindetail.Forstudiesthatwerereportedin1publications,weextracteddatafromthemostcompletepublicationandusedotherpublicationsassupplements.StatisticalanalysisTheverifieddatawereanalyzedusingReview-Managersoftware(RevMan,version5.1.6forWindows).Wedeterminedpooledmeandifference(MD)andcorresponding95%confidenceintervals(CI)forcontinuousdata,andpooledriskratios(RRs)fordichotomousdata.Ifstatisticallysignificantdifferenceexistedintheendpointswithdichotomousdata,wewouldcalculatethenumberneededtotreat(NNT)fromthepooledriskdifferenceforeachendpoint.TheappropriatenessofpoolingdataacrossstudieswasassessedwiththeuseoftheCochranQandtheI2testforheterogeneity.Datawerepooledbyuseofafixed-effects(FE)model(Mantel-Haenszelmethod)ifI2≤50%.IfI250%,wetriedtofindclinicalheterogeneityacrossstudiesfirstandaddresseditbysensitivityanalysesorsubgroupanalyses,andifsignifica