Functional outcomes associated with varying levels of targeted temperature management after out-of-hospital cardiac arrest ? An INTCAR2 registry analysis

Clinical paper

Functional outcomes associated with varying

levels of targeted temperature management after out-of-hospital cardiac arrest — An INTCAR2 registry analysis

Jesper Johnsson

*, Josefine Wahlstro¨m

, Josef Dankiewicz

, Martin Annborn

, Sachin Agarwal

, Allison Dupont

, Sune Forsberg

, Hans Friberg

, Robert Hand

, Karen G. Hirsch

, Teresa May

, John A. McPherson

, Michael R Mooney

,

Nainesh Patel

, Richard R. Riker

, Pascal Stammet

, Eldar Søreide

, David B. Seder

, Niklas Nielsen

aDepartmentofAnaesthesiologyandIntensiveCare,HelsingborgHospital,Helsingborg,Sweden

bDepartmentofClinicalSciences,LundUniversity,Lund,Sweden

cDepartmentofCardiology,Ska˚neUniversityHospital,Lund,Sweden

dDepartmentofNeurology,ColumbiaUniversityMedicalCenter,NewYorkCity,UnitedStates

eDepartmentofCardiology,EasternGeorgia,UnitedStates

fDepartmentofIntensiveCare,Norrta¨ljeHospital,CenterforResuscitationScience,KarolinskaInstitute,Sweden

gDepartmentofClinicalSciences,LundUniversity,IntensiveandPerioperativeCare,Ska˚neUniversityHospital,Malmo¨,Sweden

hDepartmentofMedicalServices,EasternMaineMedicalCenter,UnitedStates

iDepartmentofNeurology,StanfordUniversity,UnitedStates

jDepartmentofCriticalCareServices,MaineMedicalCenter,Portland,ME,UnitedStates

kVanderbiltUniversityMedicalCenter,Nashville,UnitedStates

lMinneapolisHeartInstitute,AbbottNorth-WesternHospital,UnitedStates

mDepartmentofCardiology,LehighValleyHealthNetwork,PA,UnitedStates

nMedicalandHealthDepartment,NationalFireandRescueCorps,Luxembourg

oCriticalCareandAnaesthesiologyResearchGroup,StavangerUniversityHospital,Norway

pDepartmentofClinicalMedicine,UniversityofBergen,Bergen,Norway

Abstract

Introduction:Targetedtemperature management (TTM) afterout-of-hospitalcardiac arrest(OHCA)hasbeenrecommendedininternational guidelinessince2005.TheTTM-trialpublishedin2013showednodifferenceinsurvivalorneurologicaloutcomeforpatientsrandomisedto33Cor 36C,andmanyhospitalshavechangedpractice.TheoptimalutilizationofTTMisstilldebated.Thisstudyaimedtoanalyseifadifferencein temperaturegoalwasassociatedwithoutcomeinanunselectedinternationalregistrypopulation.

Methods:Thisisaretrospectiveobservationalstudybasedonaprospectiveregistry—theInternationalCardiacArrestRegistry2.Patientswere categorizedasreceivingTTMinthelowerrangeat32 34C(TTM-low)orat35 37C(TTM-high).Primaryoutcomewasgoodfunctionalstatus definedascerebralperformancecategory(CPC)of1 2athospitaldischargeandsecondaryoutcomewasadverseeventsrelatedtoTTM.Alogistic

* Correspondingauthorat:DepartmentofAnaesthesiologyandIntensiveCare,HelsingborgHospital,CharlotteYlensGata10,SE-25187,Helsingborg, Sweden.

E-mailaddress:[email protected](J.Johnsson).

https://doi.org/10.1016/j.resuscitation.2019.10.020

0300-9572/©2019TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/

licenses/by-nc-nd/4.0/).ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Availableonlineatwww.sciencedirect.com

Resuscitation

j our na lho me pa g e :ww w. e l s e v i e r. c om/ l o ca t e / re s usc i ta t i on

regressionmodelwascreatedtoevaluatetheindependenteffectoftemperaturebycorrectingforclinicalanddemographicfactorsassociatedwith outcome.

Results:Of1710patientsincluded,1242(72,6%)receivedTTM-lowand468(27,4%)TTM-high.InpatientsreceivingTTM-low,31.3%survivedwith goodoutcomecomparedto28.8%intheTTM-highgroup.Therewasnosignificantassociationbetweentemperatureandoutcome(p=0.352).In analysesadjustedforbaselinedifferencestheORforagoodoutcomewithTTM-lowwas1.27,95%CI(0.94 1.73).Haemodynamicinstabilityleadingto discontinuationofTTMwasmorecommoninTTM-low.

Conclusions:Nosignificantdifferenceinfunctionaloutcomeathospitaldischargewasfoundinpatientsreceivinglower-versushighertargeted temperaturemanagement.

Keywords:Cardiacarrest,Out-of-hospital,Outcome,Targetedtemperaturemanagement,TTM

Introduction

Theuseoftargetedtemperaturemanagement(TTM)asanintervention tomitigatesecondaryneurologicinjuryincomatosesurvivorsofcardiac arresthasbeenwidelyadoptedduringthelast15yearsdespitelowto very-lowoverallqualityofevidence.^1,2TheTTM-trialpublishedin2013 compared a target temperature of 33C 36C in out-of-hospital cardiacarrest (OHCA)patientsand didnotdemonstrateabenefit regardingsurvivalorneurologicaloutcome.³Thistrialwas,however, differentcomparedtoearliertrialsinthatbothinterventiongroupswere tightlytemperaturecontrolledandkeptattemperaturebelownormal, avoidingthenaturaltemperaturetrajectoryforcardiacarrestpatients;

hencethetrialcomparedmildhypothermiatoverymildhypothermia.^4,5 Additional subgroup analyses and observational data support the neutralresultoftheTTM-trial^{6 10}andsince2013manycentershave changedtheirstandardpracticetreatmentstrategyaimingforatarget temperatureof36C.

Somecentershavecontinuedtousetraditionalinducedhypother- mia(32 34C)whereassomedonotuse,orhaveabandonedTTM, despite the updated European Resuscitation Council (ERC)- and American Heart Association (AHA)-guidelines from 2015,strongly recommendingTTMat32 36CforadultsurvivorsofOHCAwithan initial shockable rhythm who remain unresponsive after return of spontaneous circulation (ROSC).^1,2 Thus, there is a substantial internationalvariationofclinicalpracticewithdifferentapproachesto TTM.^11,12Recently,resultsfromlargeintensivecaredatabaseshave confirmedachangeintheuseofTTMafterOHCA;fewerpatients receiveTTMandmorepatientsexperiencefeverduringtheintensive carestay.^12,13However,anyimpactonoverallsurvivalorneurological functionhasbeendifficulttodistinguish.Tendenciestowardsworse outcomehave beenreportedwiththese changes,thoughinconsistently linkedtochangesinTTMpractices.¹³

Inthisstudyweaimedtoseeifdifferencesintargettemperature affectedfunctionaloutcomesinaninternationalobservationalregistry of OHCA-patients where baseline variables allow for adjusted analyses.

Methods

TheInternationalCardiacArrestRegistry2

INTCAR2isamultinational,internet-basedregistryofcardiacarrest patients treated in an intensive care unit (ICU) setting.

INTCAR2 received data from 25 centers in the United States, Sweden,NorwayandLuxembourg.Theregistrywas startedasa continuationofINTCAR1andtheHypothermiaNetworkRegistry.¹⁴

Itpredominantlyencompassesaprospectivelyregisteredsample ofconsecutivepatientsmostofwhomweretreatedwithtemperature management, andincludesdetailsaboutpresumedcauses,treat- mentandoutcomesforpatientsaftercardiacarrestatalllocations admittedtointensivecare.

Patients

Thepatients inour studywereOHCA-patients treated atcenters reportingtoINTCAR2between2008and2017(startandenddatesof INTCAR2). Patients registered before 2013 were excluded to minimizetreatmentbiasduetothechangeintreatmentstrategyof OHCApatientsfollowingpublicationoftheTTM-trial.³

InclusioncriteriawereOHCA-arrestpatientsofanycauseofarrest, 18yearsofage,stableROSC,notrespondingtoverbalcommandsat admission and being treated in an ICU-setting with temperature management. Exclusion criteria werearrest in the EDor location missing,missingoutcomedataormissingtemperatureallocation.

Each participating center treated patients according to local protocols,includingchoiceofcoolingdevicesandcoolingmethods.

TheethicalreviewboardinLund,Swedenapprovedtheregistry (272/2007)andlocalethicalapprovalwasgrantedasperregulations of each participating hospital. Information about the study was providedtopatientswhoregainedconsciousnessortonextofkin,if requiredbylegalstatuteineachcountry.

Data

INTCAR2-data were derived from ambulance charts, admission journals,ICUobservationchartsandmedicalrecordsfromhospitals andrehabilitationcenters.Pre-hospitaldataweredefinedaccordingto the Utstein guidelines¹⁵ and in-hospital data according to the extendedUtsteinguidelinesforreportingpost-resuscitationcare.¹⁶

Comorbiditieswereregisterediftheywerepharmacologicallyor previouslysurgicallytreated,orsubjecttocontinuousmonitoringat thetimeofthecardiacarrest.TimetoROSCwasdefinedastimefrom collapse until return ofspontaneous circulation, leadingto stable circulationwithouttheneedforcardiopulmonaryresuscitation(CPR) for atleast20min.Temperature managementwas definedasan active attempt to keep the patient’s body temperature within a prescribedtargetrange.TTMat32 34CwasdefinedasTTM-low andTTMat35 37CasTTM-high.AdverseeventsduringICUcare wererecordedaccordingtoapredefinedprotocol.

Outcome

Primary outcome was survivalwith goodneurological functionat hospitaldischarge,usingtheCerebralPerformanceCategory(CPC)

scalewhereCPC1=goodcerebralperformancewithnormalfunction orminordisability;CPC2=moderatecerebraldisability,independent in activities of daily life; CPC 3=severe cerebral disability and dependentonothersfordailyactivities;CPC4=apatientincomaora vegetative state; and CPC 5=dead.¹⁷ The CPC scale was dichotomizedintogood(CPC1and2)andpoor(CPC3 5)outcome accordingtotheUtsteinguidelines.^18,19

In asubset of patients, no outcome datawere registeredat hospitaldischargebuthadlongterm(180days)follow-updata.In thesecases,weusedthefollow-upoutcomeasahospitaldischarge outcome-substitute in the analysis (last observation carried backwards). The primary outcome was reported for all patients accordingto TTM-group. We also performed subgroup analysis using the prespecified subgroups defined in the TTM-trial: age (above/below65years),sex(male/female),initialrhythmshockable (yes/no),timetoROSC(above/below25min)andcirculatoryshock onarrivalinhospital(yes/no).²⁰

Thesecondary outcome was adverse events related to TTM duringICUcare:pneumonia(definedasthreeofthefollowingfour criteria:progressiveornewinfiltratesonchestX-ray(mandatory), feverabove38Cin thefirst 72hofadmission,leukocytosis and purulent mucusin endotracheal tube;majorbleeding (defined as cerebralbleedingorbleedingrequiringtransfusion);haemodynamic instabilityleadingtodiscontinuationofTTM;severesepsisandseptic shockdefinedaccordingtothecriteriaoftheAmericanCollegeof ChestPhysiciansandSocietyofCriticalCareMedicine²¹leadingto discontinuationofTTM;andseizuresbasedonclinicaldetectionand diagnosisduringTTM.

Statisticalanalysis

Continuous variables are presented as mean one standard deviation if normally distributed and as median and interquartile rangeifnon-normallydistributed.

Binaryandcategoricalvariablesarepresentedasnumbersand percentages.CategoricaldatawerecomparedusingChi-Squaretest, continuousnormallydistributeddatawerecomparedusingStudent’s t-test and non-normally distributed data by the Wilcoxon Mann Whitneytest.Aunivariatelogisticregressionwasperformedand presentedasoddsratios(OR)with 95%confidenceintervals(CI) indicatingtheassociationofthevariablewithagoodoutcomeandOR- values>1indicatingafavourableassociation.

A multivariate analysis was also performed using logistic regressionwithadjustmentforimportantcovariableswithapotential to influence outcome after cardiac arrest including age, sex, comorbidities, bystander-CPR, arrest characteristics, circulatory shock on admission and urgent angiography prior to hospital discharge.Someofthesevariableswerenotcompleteinthedataset but due to an overall low number of missing values (<5%) no imputationwasperformed.²²

Aforestplotwascreatedassessinginteractionofage(aboveor below65years),sex,timetoROSC(aboveorbelow25min),initial rhythm (shockable or non-shockable) and circulatory shock on admissiontoinvestigatewhetheranyofthesegroupswouldsignal apositiveassociationtoeitherTTM-highorTTM-low.Finally,adverse eventsduringthepatients’ICUstaywerecomparedbetweenthetwo temperature groups.A p-value<0.05 wasconsidered statistically significantandalltestsweretwo-tailed.Rwasusedforstatistical analysis(RCoreTeam,2013).R:Alanguageandenvironmentfor statisticalcomputing.RFoundationforStatisticalComputing,Vienna, Austria.URL(http://www.R-project.org/).

Results

BetweenOctober2008andNovember 2017,3252cardiacarrest patients were registered in the INTCAR2 database. Of these, 1710wereeligibleforthefinalanalysisafterexcludingpatientswith age <18 years (n=23),arrestin-hospital, in theEDorunknown (n=868),missingdataonoutcomeortargetedtemperature(n=177) and registeredbefore 2013(n=474). Ofthe1710patients,1059 (61,9%)wasregisteredintheUnitedStates,427(24,9%)inSweden, 142(8,3%)inNorwayand82(4,8%)inLuxembourg(Fig.1).The patientswere groupedaccordingtoprescribedtemperature treat- ment,including1242(72.6%)patientstreatedwithTTM-lowand468 (27.4%)withTTM-high. ThedistributionofTTM-low vs.TTM-high patientswasnotevenlydistributedbetweenparticipatingcountries.

PatientsintheTTM-lowgroupwerepredominatelyenteredfromthe UnitedStatesandNorway,whilepatientsintheTTM-highgroupwere fromSwedenandLuxembourg(Fig.1).

Baselinecharacteristicsforthetwotreatmentgroupsareshownin Table1.ThereweremoremalepatientsinTTM-high(n=358,76.5%) compared to TTM-low (n=829, 66.7%) and patients in the low temperaturegroupwereyoungerwithameanageof59.2(15.8)

Fig.1–Targetedtemperaturebycountry.

ThemajorityofpatientsreceivingTTM-highdidsoinSweden,withLuxembourgcontributingsomepatients.The UnitedStatesalmostexclusivelyusesTTM-low.TTM-lowdenotes32 34CandTTM-highdenotes35 37C.

TTM:targetedtemperaturemanagement.

compared to 63.7 (14.5) years. Patients in TTM-low had more comorbidities in general compared to TTM-high. There was no significant difference regarding frequency of witnessed arrest betweenthetwotemperaturegroupswhereasbystander-CPRwas morecommonintheTTM-high(n=278,60.4%vs.n=645,52.4%).

ArrestwithEmergency Medical Service(EMS)presentwasmore common in TTM-low (n=174, 14.1% vs. n=38, 8.3%). The percentage of patients with a shockable rhythm did not differ significantlybetweengroupswhereastimetoROSCwassignificantly longerinTTM-high(34min[IQR24 53]vs.29min[IQR19 48]).

More urgent angiography was performed in the TTM-high group (n=231,51%vs.n=431,38%)andpost-arrestshockonadmission wasmorecommoninTTM-low(n=562,48.4%vs.n=180,39.1%).

TheICUlengthofstaywasshorterfortheTTM-highgroup(4days [IQR2.00 7.75]vs. 5days [IQR3.00 9.00],p<0.001)whereas hospitallengthofstaydidnotdiffersignificantlybetweengroups(7 days[IQR3.00 16.00]forTTM-highvs.7days[IQR3.00 13.00]for TTM-low,p=0.15.

Outcome

Thenumberofpatientswithgoodfunctionaloutcome(CPC1-2)was 389of1242(31,3%)inTTM-lowand135of468(28,8%)inTTM-high.

Mortality(CPC5)wasalsosimilar,59.2%(735of1242)inTTM-low and61.8%inTTM-high(289of468)(Fig.2).AChi-squaretestfor temperaturevs.outcomehadap-valueof0.352.

The univariate analysis showed no statistically significant association between a low temperature and a good functional outcome(OR=1.12,95%CI0.89 1.42,p=0.32),confirmedinthe multivariate analysis (OR =1.27, 95% CI 0.94 1.73, p=0.11) (Table2).Amongcovariables,thepresenceofashockablerhythm had the strongest multivariate association with a good outcome (OR=4.39,95%CI3.23 6.01,p<0.001).

Forthepredefinedsubgroupanalyses,inpatientswithfemalesex andpresenceofcirculatoryshockonhospitaladmission,TTM-high wasassociatedwithagoodoutcome(Fig.3).

Secondaryoutcome

Regarding adverse events during the ICU stay, haemodynamic instabilityleadingtodiscontinuedTTMwasmorecommoninTTM-low (n=58,4.9%vs.n=8,1.7%,p<0.001)andpneumoniawassimilarly common inboth groups(n=435, 38,4%in TTM-low andn=170, 37.1%inTTM-high,p=0.67)(Table3).Therewerenostatistically significantdifferencesinthefrequencyofadverseeventsregarding majorbleeding(n=88,7.8%inTTM-lowvs.n=30,6.6%inTTM-high, p=0.47),sepsis(n=3,0.3%inTTM-lowvs.n=0,0%inTTM-high, p=0.66)orseizures(n=98,8.5%inTTM-lowvs.n=39,8.4%inTTM- low,p=1.00).

Discussion

Inthislargeretrospective,observationalregistrystudyweinvestigat- edwhethertheresultsfromtheTTM-trialcouldbedemonstratedin Table1–Baselinecharacteristicsstratifiedintoinhigh-andlowtargetedtemperaturegroups.

TTM-low^a(n=1242) TTM-high^b(n=468) p-value

Age(years)meanSD 59.515.8 63.714.5 <0.001

Malesex(%) 829(66.7) 358(76.5) <0.001

Previouschronicheartfailure(%) 198(15.9) 89(19.0) 0.149

Previousmyocardialinfarction(%) 233(18.8) 71(15.2) 0.097

Previoushypertenison(%) 589(47.4) 173(37.0) <0.001

Previousinsulindependentdiabetes(%) 142(11.4) 51(10.9) 0.821

Previousnon-insulindependentdiabetes(%) 170(13.7) 45(9.6) 0.029

PreviousCOPD(%) 215(17.3) 53(11.3) 0.003

Previousdementiaorcognitiveimpairment(%) 51(4.1) 13(2.8) 0.251

Witnessedcardiacarrest(%) 928(75.6) 364(79.8) 0.077

BystanderCPR(%) 0.001

-Yes(%) 645(52.4) 278(60.4)

-No(%) 411(33.4) 144(31.3)

-ArrestwithEMSpresent(%) 174(14.1) 38(8.3)

VT/VForAED-advisedshockablerhythm(%) 618(50.0) 248(53.6) 0.210

TimetoROSCinminutes(IQR)^c 29(19 48) 34(24 53) 0.006

UrgentAngiography(%) 431(37.5) 231(50.5) <0.001

Shockonadmission(%)^d 562(48.4) 180(39.1) 0.001

ICUlengthofstayindays(IQR) 5(3.00 9.00) 4(2.00 7.75) <0.001

Hospitallengthofstayindays(IQR) 7(3.00 13.00) 7(3.00 16.00) 0.15

Dataarepresentedasmean(SD),n(%)ormedian(IQR).ndenotesthenumberofcaseswithvaliddata.Ap-valueof<0.05wasconsideredsignificant.SD:

standarddeviation;IQR:interquartilerange;TTM:targetedtemperaturemanagement;COPD:chronicobstructivepulmonarydisease;CPR:cardio-pulmonary resuscitation;EMS:EmergencyMedicalService;VT:ventriculartachycardia;VF:ventricularfibrillation;AED:automatedexternaldefibrillator;ROSC:returnof spontaneouscirculation;ICU:intensivecareunit.

aTTM-lowdenotes32 34C.

bTTM-highdenotes35 37C.

cIfunwitnessed,timeiscalculatedfromemergencycall.

dShockonadmissionisdefinedassystolicbloodpressureoflessthan90mmHgformorethan30minorend-organhypoperfusionunlessvasoactivesare administered.

OHCApatientsincludedintheINTCAR2-registrycontainingcardiac arrestdatawherebaselinevariablesallowforadjustedanalyses.Our analysesshowednostatisticallysignificantdifferenceinfunctional outcomeathospitaldischargebetweenpatientstreatedwithTTM-low

(32 34C)orTTM-high(35 37C)ineitherunadjustedoradjusted analyses.

AlthoughthecrudenumbersforgoodoutcomebetweentheTTM- groupswerestrikinglysimilar,themultivariableanalysisrevealeda Fig.2–CPCdistributionandcomparisonbetweenthelow-andhightargetedtemperaturegroups.

Valuesarepercentagesofthetotalamountofpatientsinthatgroup.CPC:cerebralperformancecategory;CPC1,good cerebralperformance,mighthavemildneurologicalorpsychologicaldeficit.CPC2,moderatecerebraldisability.Able toworkinshelteredenvironmentandenoughfunctionforindependentactivitiesofdailylife.CPC3,severecerebral disability,conscious,dependantonotherpeoplefordailysupport(awidespectrumofcerebralfunction).CPC4,coma orvegetativestate.CPC5,braindead.TTM-lowdenotes32 34CandTTM-highdenotes35 37C.

TTM:targetedtemperaturemanagement.

Table2–Univariateandmultivariatelogisticregressionanalysisofbaselinefactorsandtheirassociationwith outcome.

Univariateanalysis Multivariateanalysis

OddsRatio(95%CI) p-Value OddsRatio(95%CI) p-Value

TTM-low^a 1.12(0.89 1.42) 0.32 1.27(0.94 1.73) 0.11

Age(peryear) 0.97(0.97 0.98) <0.001 0.97(0.96 0.98) <0.001

Malesex 1.98(1.56 2.53) <0.001 1.33(0.97 1.83) 0.08

Previouschronicheartfailure 0.67(0.50 0.90) 0.008 1.01(0.68 1.47) 0.97

Previousmyocardialinfarction 1.00(0.76 1.30) 0.983 1.10(0.76 1.59) 0.60

Previoushypertenison 0.61(0.49 0.75) <0.001 0.78(0.58 1.04) 0.09

Previousinsulindependentdiabetes 0.35(0.23 0.52) <0.001 0.45(0.26 0.75) <0.001

Previousnon-insulindependentdiabetes 0.67(0.48 0.93) 0.019 0.88(0.57 1.36) 0.57

PreviousCOPD 0.30(0.21 0.43) <0.001 0.46(0.27 0.74) <0.001

Previousdementiaorcognitiveimpairment 0.18(0.06 0.42) <0.001 0.23(0.07 0.64) 0.01

Witnessedcardiacarrest 1.96(1.50 2.58) <0.001 1.80(1.26 2.58) <0.001

VT/VForAED-advisedshockablerhythm 6.31(4.96 8.08) <0.001 4.39(3.23 6.01) <0.001

TimetoROSC(perminute)^b 0.98(0.98 0.99) <0.001 0.99(0.98 0.99) <0.001

Urgentangiography 2.74(2.20 3.41) <0.001 1.60(1.21 2.13) <0.001

Shockonadmission^c 0.45(0.36 0.56) <0.001 0.51(0.39 0.67) <0.001

BystanderCPR

-No Ref Ref Ref Ref

-Yes 2.20(1.73 2.82) <0.001 1.43(1.05 1.95) 0.02

-ArrestwithEMSpresent 1.06(0.72 1.54) 1.000 1.47(0.90 2.37) 0.12

Oddsratiosforgoodneurologicaloutcomeforthegroupinentiretywhereavalueof>1indicateseachfactor’sbeneficialinfluenceonoutcome.Ap-valueof

<0.05wasconsideredsignificant.Inthemultivariatemodeladjustmentforpotentialconfoundingfactorspreviouslyknowntoinfluenceoutcomeafterout-of- hospitalcardiacarrest(OHCA)suchasage,gender,co-morbidities,arrestcharacteristics,angiographyandshockonadmissionwasmade.CI:confidence Interval;TTM:targetedtemperaturemanagement;COPD:chronicobstructivepulmonarydisease;VT:ventriculartachycardia;VF:ventricularfibrillation;AED:

automatedexternaldefibrillator;ROSC:returnofspontaneouscirculation;CPR:cardiopulmonaryresuscitation;EMS:EmergencyMedicalService.

aTTM-lowdenotes32 34C.

bIfunwitnessedarrest,timeiscalculatedfromemergencycall.

cShockonadmissionisdefinedasasystolicbloodpressureoflessthan90mmHgformorethan30minorend-organhypoperfusionunlessvasoactivesare administered.

tendency towards a more favourable outcome in TTM-low after adjustmentforpotentialconfoundingfactorspreviouslyknowntobe associatedwithoutcome.^{23 25}Similarconcernswereraisedinprior observationalstudies.^12,13Althoughcomplexmediationanalysisof datafrom45935patientsinastudyfromBradleyetal.¹³suggested inconsistency regarding the role of target temperature in theses outcomes,thelackofrandomisationandhighpotentialforbiasand confoundingsuggestsgreatcautionwheninterpretingtheseresults.²⁶ Similarly,ourresultsmustbeinterpretedwithcaution,andpotential benefitofTTM-low maybe worthexploringinfurtherrandomised clinicaltrials.

Theoverallincidenceofadverseeventswaslowinbothgroups, howeverpneumoniawasthemorecommonandoccurredwithsimilar frequency in both temperature groups. The high incidence of pneumonia during post-cardiac arrest care is described in other OHCAcohorts.^3,27,28MoreTTM-lowpatientshadTTMdiscontinued duetohemodynamicinstability,andtherateofTTMdiscontinuationin ourstudywashigherthanreportedintheTTM-trial.³Thismightreflect

agreatertendencytoaborttemperaturetreatmentinunstableand deteriorating patientsifnotbeingpartofaresearchtrialprotocol.

Interestingly,thesignalfromtheTTM-trialthatpatientsincirculatory shock on hospital arrivaltendedto have abetteroutcome when treatedwithTTMat36C,assuggestedbyAnnbornetal.,⁶wasalso evidentinourcohort.Additionally, subgroupanalysissuggestsan associationbetweenagoodoutcomeandwomentreatedwithTTM- high,whichwasnotseenintheTTM-trial,thoughthepointestimate wasinthesamedirection.³

Ourtreatmentgroupsdifferedinbaselinecharacteristicssuchas age,sex,comorbidities,arrestcharacteristics,pre-hospitalcircum- stances,cardiacinterventionsandshockonadmission,allvariables significantly associated with outcome after cardiac arrest. These differencesmayreflectgeographic,demographicandpolicy-related orpatient-selectionfactorsspecifictotreatingphysicians.InSweden, themeanageatarrestishigher,malepatientssufferingcardiacarrest outnumberfemalepatients,shockablerhythmsaremorecommon andthefrequencyofbystander-CPRismuchhighercomparedtothe Fig.3–Oddsratioofoutcomeaccordingtosubgroups.

Theforestplotshowstheoddsratioforfivepredefinedsubgroupsinregardtowhetherthesesubgroupswerefavoured byalow-orahightargetedtemperatureathospitaldischarge.Thehorizontalbarsrepresent95%confidenceintervals (CI).p-valuesareforthetestsofsubgroupheterogeneity(testsofinteractions)andap-valueof<0.05wasconsidered significant.ForunwitnessedcardiacarreststhetimetoROSCwascalculatedfromtimeofemergencycall.Shockon admission is defined as a systolic blood pressure of less than 90mmHg for more than 30min or end-organ hypoperfusionunlessvasoactivesareadministered.Lowtargeteddenotes32 34C(TTM-low)andHightargeted denotes35 37C(TTM-high).

Table3–Adverseeventsfortotalsampledichotomizedinlow-andhightargetedtemperaturegroups.

n TTM-low^a(%) TTM-high^b(%) p-Value

SignsofseizureduringTTM 1616 98(8.5) 39(8.4) 1.00

PneumoniaClinicalorMicrobialdiagnosis^c 1590 435(38.4) 170(37.1) 0.67

Majorbleeding^d 1591 88(7.8) 30(6.6) 0.47

TTMdiscontinued-Haemodynamicinstability 1649 58(4.9) 8(1.7) <0.001

TTMdiscontinued-Severesepsis/septicshock^e 1649 3(0.3) 0(0) 0.66

SecondaryoutcomeinthestudywasadverseeventsduringtheICUstay.Dataarepresentedasn(%)andndenotesthenumberofcaseswithvaliddata.The eventswerecomparedusingChi-squareandap-valueof<0.05wasconsideredsignificant.TTM:targetedtemperaturemanagement.

aTTM-lowdenotes32 34C.

bTTM-highdenotes35 37C.

cPneumoniaisdefinedas3ofthefollowingcriteria:progressiveornewinfiltratesonchestX-ray(mandatory),feverabove38C,leucocytosisandpurulentmucus intube.

dMajorbleedingisdefinedascerebralbleedingorbleedingrequiringtransfusion.

eSeveresepsis/septicshockisdefinedaccordingtothecriteriaoftheAmericanCollegeofChestPhysicianandSocietyofCriticalCareMedicine.

United States.^29,30 The marked difference in baseline variables betweentheUnitedStatesandEuropemightindicatethepresenceof other unidentifiedandunmeasured factors thatdiffer, resultingin considerableresidualconfounding.Avalidatedcardiacarrest-specific severity scoring model could facilitate the comparison between groupswithdifferentbaselinecharacteristics.

There are a number of limitations to this study. This was a retrospectivestudyofprospectivelycollectedregistrydataandthe samplesize wasdetermined byconvenience. No auditorformal qualitycontrolwasperformed,makingerroneousdataandmisinter- pretedentriesintheINTCAR2databasepossible.Thegeneralizability ofourfindingsmaybelimited,asourresultsreflectstandardsinhighly specializedOHCA-centersusingTTM.Hospitalcharacteristicsare associatedwithOHCAoutcome,favoringcenterswith24-hcardiac interventional services.^31,32 Recent studies have shown that the variation in outcome after cardiac arrest may be influenced by variationsinwithdrawaloflifesustainingtherapy(WLST)strategies andin-hospitalmanagementdifferences.^33,34

Hospitaldischargemaynotbeanidealoutcomeassessmenttime point,sincefunctionaloutcomemayevolveaftercardiacarrest,and timeof discharge variedconsiderably.³⁵ TheTTM-trial, however, showedthatthedifferenceinneurologicalfunctionbetweenhospital survivaland180-daysurvivalwaslimited.³

Our sample-size differed between TTM-low and TTM-high (Table1)duetothefactthatthemajorityofINTCAR2-patientswere registeredintheUnitedStateswheretreatmentat33Cwasmore commonintheparticipatingsites.Thereversesituationwaspresent forpatientsincludedinSweden(Fig.1)wheretemperaturecontrolat 36ChasbecomestandardcareaftertheTTM-trial.Thisdifferencein treatmentstrategiesindifferentcountriesmightrepresentabiaswhen analysingdatafromaninternationalmulticenterregistry.Therefore, patients registered before 2013 were excluded to minimize any treatmentbiasfollowingthepublicationoftheTTM-trial.

Duringthefive-yearinclusionperiod,changesmayhaveoccurred incardiacarrestcare,includingstandardizedintensivecarebundles andmoreearlycardiacintervention.Advancedpre-hospitalcarehas alsoevolvedandbothavailabilityofpublicdefibrillatorsandlayperson awarenessofcardiacarrestandbystander-CPRmayhaveincreased.

In addition, fewer patients presented with shockable rhythms.³⁶ Finally,thelackofinternationalstandardizedprocessesforprognos- ticationandWLSTincardiacarrestpatientsmayhaveinfluenced outcomeinthesepatients.

Strengthsincludealargemultinationalperspective,aprospective registry, well established cardiac arrest centers, well defined covariables important for adjustment of treatment effects and consecutivelyenteredpatientswhichmaybetterreflectreal-world practicesthanclinicaltrialsdo.

Whiletheoverallmortalityfromcardiacarrestremainshigh,the prognosis for unconscious OHCA patients with initial shockable rhythmsandROSCadmittedtotheICUareimproving,asmorethan halfwillsurvivewithagoodfunctionaloutcome.⁸

Controllingbodytemperatureisapotential treatmentthatmay preventsecondarybraindamagebuttheprecisemechanismsarestill unknown. Optimalpost-cardiac arrestcare remainscontroversial, includingwhichtemperaturetotarget,howlongtodelivertemperature control,theoptimalwayofrewarmingandwhetherdifferenttarget temperatures are appropriate for different patients.^{37 39} Overall qualityofevidenceforthistherapyisloworverylow,andfurther studies are necessary to determine benefits and risksrelatedto temperature management.¹ TheTTM2-trial (NCT02908308)isan

ongoinginternational,multicenter,parallelgroup,investigatorinitiat- ed,superioritytrialinwhich1900OHCApatientswillberandomisedto a targeted temperature of 33C or to normothermia with early treatmentoffever(37.8C).⁴⁰

Conclusions

ThislargeinternationalregistrystudyofOHCApatientsrevealedno significantdifferenceinoutcomebetweenpatientstreatedwithTTM- loworTTM-high,supportingthefindingsfromtheTTM-trial.When adjustingforconfoundingfactors,themultivariateanalysisindicateda non-significanttendencytowardsbetterfunctionaloutcomewithTTM- low. This was, however, associated with more hemodynamic instability and discontinuation of TTMtherapy. Limitations in the currentevidencesupportlargerrandomisedtrialstobetterestablish thepotentialbenefitsandharmsofspecificapproachestoTTMafter OHCA.

Conflict of interest

Theauthorsdeclare thattheyhaveno conflictofinterestwiththe contentsofthisarticle.

Acknowledgements

DrJesperJohnssonhasreceivedindependentresearchgrantsto fund research time from Stig and Ragna Gorthon’s Foundation, ThelmaZoega’sFoundation,VOFoUSkånevårdSund,theEuropean RegionalDevelopmentFundthroughtheInterregIVAOKSprogram and government funding of clinical research within the Swedish NationalHealthServices(ALF).Nocommercialfundingwasreceived.

Thefundingorganizationsdidnothaveanyaccesstothedata,nordid theyhaveanyinfluenceondataanalysisorinterpretation.

Appendix A. Supplementary data

Supplementarymaterialrelatedtothisarticlecanbefound,intheonline version,atdoi:https://doi.org/10.1016/j.resuscitation.2019.10.020.

REFERENCES

1.NolanJP,SoarJ,CariouA,etal.EuropeanResuscitationCouncil andEuropeanSocietyofIntensiveCareMedicineguidelinesfor post-resuscitationcare2015:section5oftheEuropean ResuscitationCouncilguidelinesforresuscitation2015.

Resuscitation2015;95:202 22.

2.CallawayCW,DonninoMW,FinkEL,etal.Part8:post-cardiacarrest care:2015AmericanHeartAssociationguidelinesupdatefor cardiopulmonaryresuscitationandemergencycardiovascularcare.

Circulation2015;132:S465 82.

3.NielsenN,WetterslevJ,CronbergT,etal.Targetedtemperature managementat33degreesCversus36degreesCaftercardiac arrest.NEnglJMed2013;369:2197 206.

4.BernardSA,GrayTW,BuistMD,etal.Treatmentofcomatose survivorsofout-of-hospitalcardiacarrestwithinducedhypothermia.N EnglJMed2002;346:557 63.

5.HypothermiaafterCardiacArrestStudyG.Mildtherapeutic hypothermiatoimprovetheneurologicoutcomeaftercardiacarrest.N EnglJMed2002;346:549 56.

6.AnnbornM,Bro-JeppesenJ,NielsenN,etal.Theassociationof targetedtemperaturemanagementat33and36degreesCwith outcomeinpatientswithmoderateshockonadmissionafterout-of- hospitalcardiacarrest:aposthocanalysisoftheTargetTemperature Managementtrial.IntensiveCareMed2014;40:1210 9.

7.Winther-JensenM,KjaergaardJ,WanscherM,etal.Nodifferencein mortalitybetweenmenandwomenafterout-of-hospitalcardiacarrest.

Resuscitation2015;96:78 84.

8.CronbergT,LiljaG,HornJ,etal.Neurologicfunctionandhealth- relatedqualityoflifeinpatientsfollowingtargetedtemperature managementat33degreesCvs36degreesCafterout-of-hospital cardiacarrest:arandomizedclinicaltrial.JAMANeurol 2015;72:634 41.

9.FrydlandM,KjaergaardJ,ErlingeD,etal.Targettemperature managementof33degreesCand36degreesCinpatientswithout-of- hospitalcardiacarrestwithinitialnon-shockablerhythm—aTTMsub- study.Resuscitation2015;89:142 8.

10.Winther-JensenM,PellisT,KuiperM,etal.Mortalityandneurological outcomeintheelderlyaftertargettemperaturemanagementforout-of- hospitalcardiacarrest.Resuscitation2015;91:92 8.

11.DeyeN,VincentF,MichelP,etal.Changesincardiacarrestpatients’

temperaturemanagementafterthe2013“TTM”trial:resultsfroman internationalsurvey.AnnIntensiveCare2016;6:4.

12.SalterR,BaileyM,BellomoR,etal.Changesintemperature managementof cardiac arrestpatients followingpublicationofthetarget temperaturemanagementtrial.CritCareMed2018;46:1722 30.

13.BradleySM,LiuW,McNallyB,etal.Temporaltrendsintheuseof therapeutichypothermiaforout-of-hospitalcardiacarrest.JAMA NetworkOpen2018;1:e184511.

14.NielsenN,HovdenesJ,NilssonF,etal.Outcome,timingandadverse eventsintherapeutichypothermiaafterout-of-hospitalcardiacarrest.

ActaAnaesthesiolScand2009;53:926 34.

15.JacobsI,NadkarniV,BahrJ,etal.Cardiacarrestandcardiopulmonary resuscitationoutcomereports:updateandsimplificationoftheUtstein templatesforresuscitationregistries.Astatementforhealthcare professionalsfromataskforceoftheinternationalliaisoncommittee onresuscitation(AmericanHeartAssociation,European

ResuscitationCouncil,AustralianResuscitationCouncil,New ZealandResuscitationCouncil,HeartandStrokeFoundationof Canada,InterAmericanHeartFoundation,ResuscitationCouncilof SouthernAfrica).Resuscitation2004;63:233 49.

16.LanghelleA,NolanJ,HerlitzJ,etal.Recommendedguidelinesfor reviewing,reporting,andconductingresearchonpost-resuscitation care:theUtsteinstyle.Resuscitation2005;66:271 83.

17.PhelpsR,DumasF,MaynardC,SilverJ,ReaT.Cerebralperformance categoryandlong-termprognosisfollowingout-of-hospitalcardiac arrest.CritCareMed2013;41:1252 7.

18.PerkinsGD,JacobsIG,NadkarniVM,etal.Cardiacarrestand cardiopulmonaryresuscitationoutcomereports:updateoftheUtstein ResuscitationRegistryTemplatesforOut-of-HospitalCardiacArrest:

astatementforhealthcareprofessionalsfromataskforceofthe InternationalLiaisonCommitteeonResuscitation(AmericanHeart Association,EuropeanResuscitationCouncil,AustralianandNew ZealandCouncilonResuscitation,HeartandStrokeFoundationof Canada,InterAmericanHeartFoundation,ResuscitationCouncilof SouthernAfrica,ResuscitationCouncilofAsia);andtheAmerican HeartAssociationEmergencyCardiovascularCareCommitteeand theCouncilonCardiopulmonary,CriticalCare,Perioperativeand Resuscitation.Circulation2015;132:1286 300.

19.BlondinNA,GreerDM.Neurologicprognosisincardiacarrestpatients treatedwiththerapeutichypothermia.Neurologist2011;17:241 8.

20.NielsenN,WetterslevJ,al-SubaieN,etal.Targettemperature managementafterout-of-hospitalcardiacarrest—arandomized,

parallel-group,assessor-blindedclinicaltrial—rationaleanddesign.

AmHeartJ2012;163:541 8.

21.BoneRC,BalkRA,CerraFB,etal.Definitionsforsepsisandorgan failureandguidelinesfortheuseofinnovativetherapiesinsepsis.The ACCP/SCCMConsensusConferenceCommittee.AmericanCollege ofChestPhysicians/SocietyofCriticalCareMedicine.Chest 1992;101:1644 55.

22.JakobsenJC,GluudC,WetterslevJ,WinkelP.Whenandhowshould multipleimputationbeusedforhandlingmissingdatainrandomised clinicaltrials—apracticalguidewithflowcharts.BMCMedRes Methodol2017;17:162.

23.KarlssonV,DankiewiczJ,NielsenN,etal.Associationofgenderto outcomeafterout-of-hospitalcardiacarrest—areportfromthe InternationalCardiacArrestRegistry.CritCare2015;19:182.

24.GeriG,DumasF,BougouinW,etal.Immediatepercutaneous coronaryinterventionisassociatedwithimprovedshort-andlong-term survivalafterout-of-hospitalcardiacarrest.CircCardiovascInterv 20158:.

25.TermanSW,ShieldsTA,HumeB,SilbergleitR.Theinfluenceofage andchronicmedicalconditionsonneurologicaloutcomesinoutof hospitalcardiacarrest.Resuscitation2015;89:169 76.

26.LeeH,HerbertRD,McAuleyJH.Mediationanalysis.JAMA 2019;321:697 8.

27.DankiewiczJ,NielsenN,LinderA,etal.Infectiouscomplicationsafter out-of-hospitalcardiacarrest—acomparisonbetweentwotarget temperatures.Resuscitation2017;113:70 6.

28.GagnonDJ,NielsenN,FraserGL,etal.Prophylacticantibioticsare associatedwithalowerincidenceofpneumoniaincardiacarrest survivorstreatedwithtargetedtemperaturemanagement.

Resuscitation2015;92:154 9.

29.HerlitzJ.Svenskahjärt-lungräddningsregistretÅrsrapport2017.

2017.

30.KimLK,LooserP,SwaminathanRV,etal.Sex-baseddisparitiesin incidence,treatment,andoutcomesofcardiacarrestintheUnited States,2003 2012.JAmHeartAssoc20165:.

31.StubD,SmithK,BrayJE,BernardS,DuffySJ,KayeDM.Hospital characteristicsareassociatedwithpatientoutcomesfollowingout-of- hospitalcardiacarrest.Heart2011;97:1489 94.

32.SchoberA,SterzF,LaggnerAN,etal.Admissionofout-of-hospital cardiacarrestvictimstoahighvolumecardiacarrestcenterislinkedto improvedoutcome.Resuscitation2016;106:42 8.

33.MayTL,LaryCW,RikerRR,etal.Variabilityinfunctionaloutcomeand treatmentpracticesbytreatmentcenterafterout-of-hospitalcardiac arrest:analysisofInternationalCardiacArrestRegistry.Intensive CareMed2019;45:1176.

34.MayTL,RuthazerR,RikerRR,etal.Earlywithdrawaloflifesupport afterresuscitationfromcardiacarrestiscommonandmayresultin additionaldeaths.Resuscitation2019;139:308 13.

35.ArrichJ,ZeinerA,SterzF,etal.Factorsassociatedwithachangein functionaloutcomebetweenonemonthandsixmonthsaftercardiac arrest:aretrospectivecohortstudy.Resuscitation2009;80:876 80.

36.YehRW,SidneyS,ChandraM,SorelM,SelbyJV,GoAS.Population trendsintheincidenceandoutcomesofacutemyocardialinfarction.N EnglJMed2010;362:2155 65.

37.KapinosG,BeckerLB.TheAmericanAcademyofNeurologyaffirms therevivalofcoolingfortherevived.Neurology2017;88:2076 7.

38.BrayJE,StubD,BloomJE,etal.Changingtargettemperaturefrom 33degreesCto36degreesCintheICUmanagementofout-of- hospitalcardiacarrest:abeforeandafterstudy.Resuscitation 2017;113:39 43.

39.PoldermanKH,VaronJ.Weshouldnotabandontherapeuticcooling aftercardiacarrest.CritCare2014;18:130.

40.DankiewiczJ,CronbergT,LiljaG,etal.Targetedhypothermia versustargetedNormothermiaafterout-of-hospitalcardiacarrest (TTM2):arandomizedclinicaltrial-rationaleanddesign.AmHeartJ 2019;217:23 31.