Adherence to guidelines is associated with improved survival following in-hospital cardiac arrest

(1)

Clinical

paper

Adherence

to

guidelines

is

associated

with

improved

survival

following

in-hospital

cardiac

arrest

Fredrik

Hessulf

a,b,

*

,

Johan

Herlitz

b,c

,

Araz

Rawshani

b

,

Solveig

Aune

d

,

Johan

Israelsson

e,f,g

,

Marie-Louise

So¨dersved Ka¨llestedt

h

,

Per

Nordberg

i

,

Peter

Lundgren

a,b,j

,

Johan

Engdahl

b,k

a

DepartmentofAnaesthesiologyandIntensiveCareMedicine,HallandHospital,SE-30185Halmstad,Sweden

b_Department_of_Molecular_and_Clinical_Medicine,_Institution_of_Medicine,_Sahlgrenska_Academy,_University_of_Gothenburg,_Gothenburg,_Sweden c_PreHospen _Centre_of_Prehospital_Research;_Academy_of_Caring_Science,_Welfare_and_Work_Life,_University_of_Bora˚s,_SE-501₉₀_Bora˚s,

Sweden

d_Unit_for_{EMS-coordination,}_Provider_Governance_and_{Coordination,}_Head_Office,_Region_Va¨stra_Go¨taland,_Sweden e

DepartmentofInternalMedicine,DivisionofCardiology,KalmarCountyHospital,Kalmar,Sweden

f

FacultyofHealthandLifeSciences,LinnaeusUniversity,Kalmar,Sweden

g_Department_of_Medical_and_Health_Sciences,_Division_of_Nursing_Science,_Linko¨ping_University,_Linko¨ping,_Sweden h_Centre_for_Clinical_Research,_Uppsala_University,_{Va¨stera˚s,}_Sweden

i_Karolinska_Institute,_Institution_for_Clinical_Research_and_Education,_South_Hospital,_Stockholm,_Sweden j

DepartmentofCardiology,SahlgrenskaUniversityHospital,Gothenburg,Sweden

k

KarolinskaInstitutet,DepartmentofClinicalSciences,DanderydHospital,DivisionofCardiovascularMedicine,Stockholm,Sweden

Abstract

Background:Mostresuscitationguidelineshaverecommendationsregardingmaximumdelaytimesfromcollapsetocallingfortherescueteamand initiationof treatmentfollowingcardiac arrest.Theaim ofthestudywasto investigatetheassociation betweenadherence toguidelines for cardiopulmonaryresuscitation(CPR)afterin-hospitalcardiacarrest(IHCA)andsurvivalwithafocusondelaytotreatment.

Methods:WeusedtheSwedishRegistryforCPRtostudy3212patientswithashockablerhythmand9113patientswithnon-shockablerhythmfrom January1,2008toDecember31,2017.Adultpatientsolderthanorequalto18yearswithawitnessedIHCAwhereresuscitationwasinitiatedwere included.Weassessedtrendsinadherencetoguidelinesandtheirassociationswith30-daysurvivalandneurologicalfunction.Adherencetoguidelines wasdefinedasfollows:timefromcollapsetocallingfortherescueteamandCPRwithin1minfornon-shockablerhythms.Forshockablerhythms, adherencewasdefinedasthetimefromcollapsetocallingfortherescueteamandCPRwithin1minanddefibrillationwithin3min.

Results:Inpatientswithashockablerhythm,the30-daysurvivalforthosetreatedaccordingtoguidelineswas66.1%,ascomparedto46.5%among thosenottreatedaccordingtoguidelinesononeormoreparameters,adjustedoddsratio1.84(95%CI1.52 2.22).Amongpatientswitha non-shockablerhythmthe30-daysurvivalforthosetreatedaccordingtoguidelineswas22.8%,ascomparedto16.0%amongthosenottreatedaccordingto guidelinesononeormoreparameters,adjustedoddsratio1.43(95%CI1.24 1.65).Neurologicalfunction(cerebralperformancecategory1 2) amongsurvivorswasbetteramongpatientstreatedinaccordancewithguidelinesforbothshockable(95.7%vs91.1%,<0.001)andnon-shockable rhythms(91.0%vs85.5%,p<0.008).AdherencetotheSwedishguidelinesforCPRincreasedslightly2008 2017.

Abbreviations:CA,cardiacarrest;CPR,cardiopulmonaryresuscitation;IHCA,in-hospitalcardiacarrest;ROSC,returnofspontaneouscirculation;VF/ pVT,ventricularfibrillation/pulselessventriculartachycardia;PEA,pulselesselectricalactivity;SRC,SwedishResuscitationCouncil;CAT,cardiacarrest team;ICU/CCU,intensivecareunit/coronarycareunit;AED,automatedexternaldefibrillator;OR,oddsratio.

* Correspondingauthorat:DepartmentofAnaesthesiologyandIntensiveCareMedicine,HalmstadHospital,SE-30185Halmstad,Sweden. E-mailaddress:fredrik.hessulf@regionhalland.se(F.Hessulf).

https://doi.org/10.1016/j.resuscitation.2020.07.009

Received1May2020;Receivedinrevisedform22June2020;Accepted6July2020

0300-9572/©2020TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/ by/4.0/).ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).

Available

online

at

www.sciencedirect.com

Resuscitation

(2)

Conclusions:Adherencetoguidelineswasassociatedwithincreasedprobabilityofsurvivalandimprovedneurologicalfunctioninpatientswitha shockableandnon-shockablerhythm,respectively.Increasedadherencetoguidelinescouldincreasecardiacarrestsurvival.

Keywords:In-hospitalcardiacarrest,Cardiopulmonaryresuscitation,Chainofsurvival,Guidelines

Background

Survivalfollowingin-hospitalcardiacarrest(IHCA)hasincreasedin recentyears1butremainslowat15 30%.1,2Multiplefactorsinfluence thechanceofsurvival,andsomeofthemostcrucialaspectshave been summarized in the concept of the chain of survival: early recognition and call for help,early and efficient cardiopulmonary resuscitation(CPR),swiftdefibrillationandeffectivepostresuscitation care.3Mostguidelineshaverecommendationsregardingtheseinitial andcrucialstepstohelpclinicians takethebestpossibleactions duringcardiacarrest(CA)treatment.

TheSwedishResuscitationCouncil(SRC)hasissuedguidelines for thetreatment ofIHCA basedon theEuropean Resuscitation Council (ERC)/American Heart Association (AHA) guidelines.4 In additiontotheERC/AHArecommendationsthatpatientsfoundina shockablerhythmshouldbedefibrillatedwithin3mintheSwedish guidelinesstatethatthedelayfromCAtocallingforthecardiacarrest team(CAT)andtostartingofCPRshouldnotexceed1min.In2017, callingfortherescueteamandCPRwasperformedwithin1minand defibrillationwasmadewithin3minin80 90%ofIHCAsinSweden.5 Theassociationsbetweenanearlycallfortherescueteam,the swiftinitiationofCPR,immediatedefibrillationandsurvivalfollowing CAarewellestablished.6 8_However,_few_studies_have_specifically

investigated whether adherence to resuscitation guidelines is associatedwithincreasedchanceofsurvival.

The present study is based on the Swedish Registry for

Cardiopulmonary Resuscitation (SRCPR). The aim of the study was to investigate whether adherence to current resuscitation guidelines was associated with an increased chance of 30-day survival, neurological function, andwhetheradherence increased from2008to2017.Secondly,wequantifyhowlargethepotentialfor improvement in survival could be by comparing survival among patients treated according to guidelines versus notaccording to guidelines.

Methods

Studypopulationandsetting

This is a retrospective registry-based cohort study of all IHCAs registeredintheSRCPRfrom1January2008to31December2017. All of Sweden's seventy-three hospitals with CATs currently participateintheregistry.Bytheendof2017,70hospitalscontributed toreportingdata.TheregistrywasdesignedtocomplywiththeUtstein styleofreportingIHCAresearch.9Theregistryisanationalquality registry and publishes an annual report presenting nationwide characteristicsandtrends.

TheSRCPRcontainsvariablesincludingage,gender,witnessed status,causeofalarm,locationofcardiacarrest,electrocardiogram (ECG)monitoring,timetocall,timetoCPR,timetodefibrillation,initial rhythm,treatmentcharacteristicsandshort-termoutcome,as-wellas 30-daysurvivalandneurologicfunctionpre-arrestandatdischarge

(measuredbyCerebralPerformanceCategories,CPC-score).The studycomplieswiththeDeclarationofHelsinkiandwasapprovedby theregionalethicalboardinGothenburg,Sweden(referenceno 349-1620160516).

ValidationoftheSRCPR

ForreviewandvalidationoftheSRCPR,pleaserefertoSupplement S1.

Inclusioncriteria

‘Thestudyincludedallpatientsaged18yearsandolderwhosuffered awitnessedIHCA,between1January2008and31December2017 and in who resuscitationwas started. Furtherrequirementswere completedataoninitialrhythm,timetocalltheCATandtimetofirst shock’. For a review of included and excluded patients see supplementalmaterial.

Definitionofexposuregroups

Patientsweremanagedaccordingtoguidelinesifthefollowingcriteria werefulfilled:timefromCAtocallingforthecardiacarrestteam(CAT) and to the start of CPR did not exceed 1min and the time to defibrillationdidnotexceed3min.Thelastcriteriononlyappliedto patients who presented with a shockable rhythm. Patients were categorizedaseitherfulfillingornotfulfillingthesecriteria.Patients weredividedinto thosepresentingwithshockablerhythm or non-shockablerhythms.

Statisticalanalysis

Baselinecharacteristicsarepresentedusingmeans,mediansand proportions,withappropriatemeasurementsofdispersion. Hypothe-sistestswerenotcomputedforbaselinefeatures.10_We_assessed

trendsinadherencetoguidelines.Thiswasdonebyexaminingthe proportionofpatientstreatedaccordingtoguidelinesfrom2008to 2017.Similarly,weassessedtheproportionofpatientssurvivingto30 days (2008 2017)inrelation toadherencetoguidelines. Logistic regressionwasperformedforeachgrouptoevaluateannualchanges in 30-day survival. Differences in survival trends, in relation to adherencetoguidelines,wereevaluatedusinganinteractionterm betweencalendaryearandadherencetoguidelines.

Theassociationbetweenguidelineadherenceand30-daysurvival wasultimatelyevaluatedusinglogisticregression,withadjustmentfor covariates ofclinicalimportanceordisplayingvaryingdistributions betweenthegroupsatbaseline.

Weconstructedgradientboostingmodelstoestimatethemost importantpredictorsor30-daysurvivalamongpatientssufferingan IHCA with a shockable and non-shockable initial rhythm. The measurements of strength ofan associationin gradient boosting are denotedasrelativeinfluence.Gradientboostingisamachine learningtechniquethatcanbeusedforregressionanalysis.Ituses machinelearningtechniquesandovercomessomeofthelimitations

(3)

ofconventionallogisticregressionmodels(suchaslinearity)andcan discovernonlinearassociationsandhigh-orderinteractions.11

Wealsocomputedoddsratios(ORs)intheentirepopulationby

imputing missing data using Multiple Imputation by Chained

Equations(MICE)algorithm.12Onedatasetwasimputedandwe comparedtheobtainedORswiththoseobtainedinthecompletecase set.MICEisamethodofdealingwithmissingdata,andevaluatesifthe lossofdataislikelytohaveaffectedtheobservedresults.p-Values <0.05 were considered statistically significant. Standard mean differenceisthedifferencebetweenthemeansforthetwogroups divided bytheir standard deviation. Values below 0.1(10%) are consideredinconsequential.

WeusedR(https://r-project.org)forallanalyses.

Missingdata

Forreviewofmissingdataandacomparisonofcaseswithcomplete andincompletedata,pleaserefertoSupplementS1 S7.

Results

Baselinecharacteristics

Atotalof23,460IHCApatientswereenteredontheSRCPRbetween 1January2008and31December2017,ofwhich12,325patientsmet

thestudycriteria.Ofthese3212hadashockablerhythmand9113a non-shockablerhythm.

Shockablerhythms

Therewasnodifferenceinthemeanagesofpatientswithashockable rhythmregardlessofwhethertreatmentwasaccordingtoguidelines ornot(70.6v71.9years,respectively)(Table1).Thedistributionof gender was similar. Those who were not treated according to guidelinesweretwiceasfrequentlylocatedinregularhospitalwards, whereasthosewhoweretreatedaccordingtoguidelinesweremore frequentlylocatedinacoronarycareunit(CCU),emergencyroom (ER),intensivecareunit(ICU),operatingroom(OR),or catheteriza-tionlaboratory.ECGmonitoringwasinuseamong80.8%ofpatients whoweretreatedaccordingtoguidelines,ascomparedwith56.0%of patients whowere. Defibrillationwas morecommon inthegroup treatedaccordingtoguidelines.Allcoexistingconditionsapartfrom previousmyocardialinfarctionweremorecommoninpatientswho werenottreatedaccordingtoguidelines.

Non-shockablerhythms

Therewasnodifferenceinthemeanagesofpatientswitha non-shockablerhythmregardlessofwhethertreatmentwasaccordingto guidelines ornot(73.3 v 74.1years respectively) (Table 2). The distribution of gender was similarin relation to guideline status. Patients whowerenottreatedaccordingtoguidelinesweremore frequently located in regular wards. ECG monitoring was more

Table1–Baselinecharacteristicsofpatientswithashockablerhythm.

Non-adherencetoguidelines Adherencetoguidelines p SMD

n 970 2242 Age(mean(SD)) 71.93(12.26) 70.61(12.51) 0.006 0.106 Gender=female(%) 312(32.2) 680(30.3) 0.321 0.040 Location(%) <0.001 0.609 Regularward 419(43.2) 469(20.9) CCU 162(16.7) 763(34.0) Catheterizationlaboratory 79(8.1) 317(14.1) ER 119(12.3) 331(14.8) ICU 82(8.5) 215(9.6) OR 14(1.4) 20(0.9) Otherward 47(4.8) 48(2.1)

Outpatientward,lab,radiology 48(4.9) 79(3.5)

ECGmonitoring(%) 543(56.3) 1812(81.4) <0.001 0.563

Non-academichospital(%) 571(58.9) 1459(65.1) 0.001 0.128

NoCPRbeforeCATarrival(%) 77(8.3) 96(4.3) <0.001 0.163

Notdefibrillated(%) 37(3.8) 1(0.0) <0.001 0.277 Heartfailure(%) 393(44.6) 850(40.4) 0.039 0.084 Diabetes(%) 260(28.1) 549(25.1) 0.096 0.067 Respiratoryinsufficiency(%) 134(14.6) 278(12.9) 0.222 0.050 Myocardialinfarction(%) 299(33.3) 715(33.5) 0.916 0.006 Cancer(%) 152(16.7) 264(12.1) 0.001 0.129 Stroke(%) 103(11.2) 207(9.5) 0.170 0.056

MinutestoCPR(median[IQR]) 1.00[0.00,2.00] 0.00[0.00,0.00] <0.001 0.529 Minutestocall(median[IQR]) 2.00[1.00,2.00] 0.00[0.00,1.00] <0.001 0.530 Minutestodefibrillation(median[IQR]) 4.00[2.00,6.00] 1.00[1.00,2.00] <0.001 0.077 SMD:thestandardizedmeandifferenceisthedifferencebetweenthemeansforthetwogroupsdividedbytheirstandarddeviation.Valuesbelow0.1(10%)are consideredinconsequential(i.e.,nodifferencebetweenthegroups).CCU:cardiaccareunit;ED:emergencydepartment;ICU:intensivecareunit;OR:operation room;CPR:cardiopulmonaryresuscitation;CAT:cardiacarrestteam.

(4)

Table2_–Baselinecharacteristicsofpatientswithanon-shockablerhythm.

n 2131 6982 Age(mean(sd)) 74.05(12.82) 73.31(13.03) 0.021 0.058 Gender=female(%) 898(42.1) 2872(41.1) 0.424 0.020 Location(%) <0.001 0.405 Regularward 1319(61.9) 3193(45.7) Catheterizationlaboratory 74(3.5) 341(4.9) CCU 231(10.8) 1337(19.1) ER 159(7.5) 904(12.9) ICU 108(5.1) 581(8.3) OR 40(1.9) 147(2.1) Otherward 77(3.6) 129(1.8)

Outpatientward,lab,radiology 123(5.8) 350(5.0)

ECGmonitored(%) 813(38.6) 3892(56.2) <0.001 0.357

Non-academichospital(%) 1235(58.0) 4526(64.8) <0.001 0.141

NoCPRbeforealarm-grouparrival(%) 205(9.8) 360(5.2) <0.001 0.175

Notdefibrillated(%) 1835(87.2) 6069(87.9) 0.365 0.023 Heartfailure(%) 672(34.2) 2197(34.0) 0.890 0.004 Diabetes(%) 535(25.9) 1904(27.9) 0.074 0.046 Respiratoryinsufficiency(%) 523(25.7) 1582(23.6) 0.054 0.049 Myocardialinfarction(%) 446(22.2) 1502(22.6) 0.680 0.011 Cancer(%) 465(22.7) 1344(20.0) 0.008 0.066 Stroke(%) 295(14.4) 823(12.1) 0.008 0.066

MinutestoCPR(median[IQR]) 1.00[0.00,2.00] 0.00[0.00,0.00] <0.001 0.152 Minutestocall(median[IQR]) 2.00[2.00,3.00] 0.00[0.00,1.00] <0.001 0.162 SMD:thestandardizedmeandifferenceisthedifferencebetweenthemeansforthetwogroupsdividedbytheirstandarddeviation.Valuesbelow0.1(10%)are consideredinconsequential(i.e.,nodifferencebetweenthegroups).CCU:cardiaccareunit;ED:emergencydepartment;ICU:intensivecareunit;OR:operation room;CPR:cardiopulmonaryresuscitation;CAT:cardiacarrestteam.

Table3_–30-DaysurvivalandCPC-scoreamongpatientswithashockablerhythm.

n 970 2242 Survived=yes(%) 451(46.5) 1483(66.1) <0.001 0.404 CPC1(%) 296(75.1) 1112(85.0) <0.001 CPC2 63(16.0) 140(10.7) CPC3 27(6.9) 44(3.4) CPC4 5(1.3) 10(0.8) CPC5 3(0.8) 3(0.2)

SMD:thestandardizedmeandifferenceisthedifferencebetweenthemeansforthetwogroupsdividedbytheirstandarddeviation.Valuesbelow0.1(10%)are consideredinconsequential(i.e.,nodifferencebetweenthegroups).CPC:cerebralperformancecategory.

Table4_–30-DaysurvivalandCPC-scoreamongpatientswithanon-shockablerhythm.

n 2131 6982 Survived=yes(%) 341(16.0) 1590(22.8) <0.001 0.172 CPC1(%) 191(67.5) 974(74.3) <0.008 CPC2 51(18.0) 219(16.7) CPC3 32(11.3) 97(7.4) CPC4 9(3.2) 15(1.1) CPC5 0(0.0) 6(0.5)

SMD:thestandardizedmeandifferenceisthedifferencebetweenthemeansforthetwogroupsdividedbytheirstandarddeviation.Valuesbelow0.1(10%)are consideredinconsequential(i.e.,nodifferencebetweenthegroups).CPC:cerebralperformancecategory.

(5)

commoninpatientstreatedaccordingtoguidelines(55.7%vs38.2%). Therewerenomarkeddifferencesincoexistingconditionsinrelation toguidelinestatus

Adherencetoguidelinesand30-daysurvival

Theover-all 30-daysurvival ratewas 31.4% (n=12,325) Among survivors>90%hadaCPC-scoreof1 2.Inpatientswithashockable rhythm,comparingtreatmentaccordingtoguidelinestowithout,

30-daysurvivalwas66.1%v46.5%(OR1.88,95%CI1.56 2.26).In patientswithanon-shockablerhythm,comparingtreatmentaccording toguidelinestowithout,30-daysurvivalwas22.8%v16.0%(OR1.43,

95% CI 1.25 1.67). Neurological function (CPC 1 2) among

survivors was better among patients treated in accordance with guidelinesforbothshockable(95.7%vs91.1%)andnon-shockable rhythms(91.0%vs85.5%)(Tables3and4).Stratificationofpatients showedthat,amongpatientswithashockablerhythm,womenand patients >65 years had the greatest benefit of treatment in

Table5_–ComparisonofadjustedORsfor30-daysurvivalbetweencompletecasesandcaseswithimputeddata.

Completecases Imputeddata

Initialrhythm Variable OR LL UL OR LL UL

Shockable Overall <1.838 1.524 2.217 1.732 1.496 2.004

Shockable Men 1.686 1.344 2.115 1.606 1.345 1.917

Shockable Women 2.245 1.594 3.17 1.988 1.532 2.582

Shockable Age<65years 1.333 0.886 1.995 1.204 0.884 1.632

Shockable Age65yearsorolder 2.022 1.632 2.507 1.923 1.625 2.276

Non-shockable Overall 1.427 1.237 1.651 1.334 1.182 1.507

Non-shockable Men 1.417 1.176 1.714 1.348 1.151 1.583

Non-shockable Women 1.449 1.159 1.822 1.323 1.097 1.601

Non-shockable Age<65years 1.208 0.925 1.586 1.204 0.96 1.514

Non-shockable Age65yearsorolder 1.53 1.289 1.823 1.394 1.208 1.613

UL=upperlimit,LL=lowerlimit.Theadjustedoddsratio(95%CI)for30-daysurvivalamongpatientstreatedaccordingtoguidelinesamongshockableand non-shockablerhythmsstratifiedbysexandage,completecasesandimputeddata.ORsobtainedintheimputeddatasetwereinlinewiththoseobtainedinthe completedataset.

Figure1_–(A)ForestplotwiththeadjustedORsfor30-daysurvivalamongpatientswithashockablerhythmformultiple

variablesincludingage,sex,wardtype,monitoring,academicvsnon-academichospitalandmultiplecomorbidities.

(B)ForestplotwiththeadjustedORsfor30-daysurvivalamongpatientswithanon-shockablerhythmformultiple

(6)

accordancewithguidelines(adjustedOR2.25(CI1.59 3.17and2.02 (CI 1.63 2.51), respectively). For non-shockable rhythms, again womenandpatients>65yearshadthegreatestbenefitoftreatmentin accordancewithguidelines(adjustedOR1.45(CI1.16 1.82and1.53 (CI1.29 1.82))buttherelativebenefitwas smallercompared to shockablerhythms(Table5aandb).ORsobtainedintheimputed datasetwereinlinewiththoseobtainedinthecompletecasedataset (Table5).

Independentpredictorsofsurvival

After adjusting for covariates, multiple factors including age, comorbidities, location of cardiac arrest and hospital type were associatedwithsurvival(Fig.1AandB).Therelativeinfluenceofthe most important predictors of survival is shown in Supplement Figs.S8aandb.

Trendsinadherencetoguidelinesandsurvival

Adherencetoguidelinesincreasedfrom68.5%in2008to69.7%in 2017for patientswith shockablerhythms. Forpatients with non-shockablerhythmsadherencetoguidelinesincreasedfrom73.1%in 2008to78.6%in2017.PleaseseeFig.2A D.

Patientswithashockablerhythm:Thelineartrendindicatedan annual0.73%increaseintheprobabilityofsurvivalamongpatients whoweretreatedaccordingtoguidelines.Thecorrespondingfigure forpatientsnottreatedaccordingtoguidelineswas0.81%.

Patientswithanon-shockablerhythm:Thelineartrendindicated anannual0.45%increaseintheprobabilityofsurvivalamongpatients whoweretreatedaccordingtoguidelines.Thecorrespondingfigure forpatientsnottreatedaccordingtoguidelineswas1.51%.

Discussion

Inthislargepopulation-basedstudyofIHCAinSweden,wefoundthat adherencetocurrentresuscitationguidelineswasassociatedwithan increased chance of 30-day survival and improved neurological function. We also found that adherence to guidelines increased slightlyinthestudyperiod2008 2017.

Adherencetoguidelinesandsurvival

Our results show that adherence to guidelines increased the probabilityofsurvivalat30daysandneurologicalfunctionamongall patients,regardlessofinitialrhythm.Adherencetoguidelineswas thesinglemostimportantmodifiablefactor(amongpatientswitha

shockable rhythm) and second most important factor (among

patients with a non-shockablerhythm) asshown in the relative importanceplots.Althoughotherfactorscaninfluenceoutcome(for exampleage,comorbidities,location)theymaynotbeamenableto being influenced. However, adherence to guidelines along with monitoring of patients is modifiable factors and thereby has a potentialforimprovement.Itisworthnotingthatsurvivalincreased among all patients but theabsolute difference between patients treatedandnottreatedaccordingtoguidelinesremainedsubstantial overtime.Themechanismbehindthegeneralincreaseinsurvival seen overtime, regardless of adherence to guidelines, remains unknown.Itispossiblethatseveralfactorsincludingqualityofchest compressions, improved attitude towards CPR and selection of cases for successful resuscitation contribute.In this study, it is importanttostressthatchangesinsurvivalovertimewasnotthe majortargettoaddress.

(7)

Patientstreatedaccordingtoguidelinesconstitute70 75%ofall patientsandtheyhadamean30-daysurvivalratethatwas19.1and 6.8 percentage points higher than that of patients not treated accordingtoguidelinesforshockableandnon-shockablerhythms, respectively.Itisincorrecttoassumethatmovingpatientsfromthe “nottreated to guidelines”category to the“treated to guidelines” categorywouldincreasetheirchanceofsurvivalby6.7 19.1%,but theabsolutedifferencebetweenthetwogroupsindicatethatthereare substantial gains to be made by increasing adherence. McEvoy et al.13 _showed _that _adherence _to _the ₂₀₀₅_ACLS _protocol_was

associatedwithanincreasedchanceofROSC,whiledivergencefrom protocol(wrongaction,wrongtimingoromission)wassignificantly associated with a reduced chance of ROSC. In a similar study Honarmandetal.14_studied₁₆₀_{resuscitation}_events_following_IHCA

andfoundthatalargernumberofdeviationsfromACLSguidelines wasassociatedwithareducedchanceofROSCbuttherewasno associationwith survivalto hospital discharge. Anderson etal.15 showedthat,onahospitallevel,greateradherencetoprespecified processof care measures was associatedwith survivalrate and

neurologicoutcome.Interestingly,ourresultsshowedthatitwasthe patients with a shockable rhythm that benefited the most from adherencetoguidelines(ORforsurvival1.88v1.43).Thismayreflect thefactthatsomeofthepatientswerepossiblydefibrillatedwithinless than2min,forexamplethoseintheCCUorcathlab.Ithaspreviously beenshownthatsurvivaldecreasessignificantlywhenthetimefrom collapsetodefibrillationexceeds2min.8

Barrierstoadherencetoguidelines

WhatarethebarrierstodetectallCAsimmediately,callingforhelpand initiatingCPRwithin1minanddefibrillationwithin3min?Intheory,the identificationofthedeterioratingpatientatincreasedriskofCAshould reducetheriskofCAanddifferenttrack-and-triggersystemshave beensuggested,buttodatetheyhavefailedtoprospectivelyshowa survivalbenefit.16_The_time_from_collapse_to_the_recognition_of_CA,

callingfortherescueteamandwitnessedstatushaspreviouslybeen showntobeindependentlyassociatedwithsurvival.2Inthepresent study,onlywitnessedIHCAswereincludedintheanalysis.

ECG-Figure2_–Adherencetoguidelinesandsurvival.(AandC)Theproportionofpatientstreatedaccordingtoguidelines

overtime(2008 2017)forshockable(A)andnon-shockable(C)rhythms.(BandD)Thesurvivalrateovertime(2008

2017)forshockableandnon-shockablerhythmscomparingpatientstreatedaccordingtoguidelineswiththosenot

(8)

monitoredIHCAshouldbydefinitionberecognizedimmediatelyand thetimetocallshouldbeminimized.Thenumberofstaffandtheir individual and collective knowledge of cardiac arrest care vary betweenwards andhospitals andthiscouldexplainsome ofthe variation seen in delays from collapse to action. This is further emphasizedbythefactthatsurvivaldecreasesoutsideofficehours possiblyduetolessstaffonthehospitalwards.Ithaspreviouslybeen shownthatCAsonspecificwardswithhighlevelsofexpertiseinthe CAfield,suchastheCCU,haveanincreasedchanceofsurvival.2Our resultsshowthatadherencetoguidelineswasgenerallybetteron high-resourcewardssuchastheICU/CCU/cathlabascomparedto regularwards.Toourknowledge,thishasnotbeenreportedbefore.It hasbeenshownpreviously5_that_the_time_intervals_from_collapse_to₁₎

callingfortherescueteam,2)startingCPRand3)defibrillationwas significantlylongeringeneralwardscomparedwiththeICU/CCU/cath lab/OR.Surprisingly,adherencetoguidelineswasloweratacademic hospitalscomparedwithnon-academichospitals,perhapsreflecting logisticalbarriersto initiationof treatment.Even so,survivalwas higher amongpatientstreated atacademic hospitals witha non-shockablerhythm(andnotstatisticallydifferentforpatientswitha shockablerhythm).

When it comes to the barriers to initiatingCPR, the patients location(inapatientroomvsan elevator)andposition,alack of adequateequipment(CPRboard),alackofeducationareafewofthe reasonswhyCPRmightnotbeinitiatedimmediately.

Defibrillation is the appropriate treatment in patients with a shockablerhythm,butin20%ofcasesthereisadelayofmorethan 3min.5_A_lack_of_AEDs_at_strategic_locations_inside_the_hospital_and

inadequate training in how to operate an AED are two possible explanationsofthisshortcoming.

Trendsinadherence2008 2017

Adherencetoguidelineswasslightlyhigherforpatientswitha non-shockablerhythm(mean76.3%)comparedwithshockablerhythms (mean69.2%),adherencetoguidelinesincreasedslightlyin2008 2017.PreviousstudieshaveshownthatACLStrainingimproves hospital employees’ theoretical knowledge of CPR,17 improves practicalskillswhenassessed inCAsimulationtraining18_and_the

presenceofACLS-trainedinstructorsisassociatedwithanincreased chanceofROSCand1-yearsurvival.19_Although_never_shown,_it_is

possible that improved ACLS training increases the likelihood of adherence to CPR guidelines, and thiswould support continued emphasisonCPRtraining.

Limitations

PatientsfoundinashockablerhythmintheCCU,ICUandthecathlab aresometimessuccessfullydefibrillatedwithoutcallingfortherescue team.Wethereforeassumethattheremaybeunder-reportingofIHCA andanunderestimationoftheimportanceofadherencetoguidelines sincemanyCAsinabovelocationsaresuccessfullyresuscitated.

Thepatientcohortwasdividedaccordingtotheinitialrhythm.The registry does not contain information on rhythm change during resuscitation,afactorknowntohaveaneffectonoutcome.

Furthermore, thecurrent studyonly investigated theeffect of guidelinescompliancespecificallydelaysintimetocall,startofCPR anddefibrillation.Otherfactors,e.g.,qualityofchestcompressions, whichalsoinfluenceoutcomeswerenotaddressed.

Finally,itislikelythatthereissomevariationintheaccuracyofthe estimatedtimeintervalsfromcollapsetocallingfortherescueteam, theinitiationofCPRandtosomedegreetodefibrillation,andthis shouldbeacknowledged.

Conclusion

Increasedadherencetoguidelineshasthepotentialtofurtherimprove IHCA survival as demonstrated by the substantial difference in survivalratesbetweenpatientstreatedandnottreatedaccordingto guidelinesandtheindependentassociationwith30-daysurvivaland improvedneurologicalfunction.ContinuedfocusonCPReducation and the retentionof acquired skills are suggested as means of achieving further improvements.Considering thelargeamountof resourcesspentondevelopingnationalandinternationalguidelines, more evaluation oftheir impacton qualityofcareand survivalis needed.

Authors

’

contributions

Fredrik Hessulf: Conceptualization, Formal analysis, Funding acquisition,Methodology,Projectadministration,Validation, Visuali-zation,Writing originaldraft,Writing review&editing.

JohanHerlitz:Conceptualization,Datacuration,Formalanalysis, Funding acquisition,Methodology,Projectadministration, Supervi-sion,Validation,Visualization,Writing review&editing.

Araz Rawshani: Conceptualization, Data curation, Formal

analysis,Fundingacquisition,Methodology,Projectadministration, Supervision,Validation,Visualization,Writing review&editing.

Solveig Aune: Conceptualization, Methodology, Visualization, Writing review&editing.

JohanIsraelsson:Methodology,Visualization,Writing review &editing.

Marie-Louise So¨ dersved-Ka¨llestedt: Methodology, Visualiza-tion,Writing review&editing.

PerNordberg:Methodology,Visualization,Writing review& editing.

Peter Lundgren: Conceptualization, Data curation, Formal analysis,Fundingacquisition,Methodology,Projectadministration, Supervision,Validation,Visualization,Writing review&editing.

Johan Engdahl: Conceptualization, Data curation, Formal

analysis,Fundingacquisition,Methodology,Projectadministration, Supervision,Validation,Visualization,Writing review&editing.

Ethical

approval

and

consent

to

participate

ThestudywasapprovedbytheregionalethicalboardinGothenburg, Sweden,whichwaivedtheneedforinformedconsentdueto the retrospectivedesignofthestudy.

Availability

of

data

and

material

Thedatasetsusedand/oranalyzedduringthecurrentstudywillbe availablefromthecorrespondingauthorinresponsetoreasonable requests.

(9)

Conflict

of

interest

Theauthorsdeclarethattherearenoconflictsofinterest.

Acknowledgements

Fredrik Hessulf received research grants from Region Halland ResearchandDevelopmentandtheSwedishResuscitationCouncil. JohanHerlitzreceivedfundingfromtheLaerdalFoundationandthe SwedishHeartFoundation.Thestudywasalsofinancedbygrants fromtheSwedishstateundertheagreementbetweentheSwedish government and the County Council, the ALF agreement (grant

ALFGBG-716901).ArazRawshani wassupportedbytheSwedish

ResearchCouncil(grant2019-02019).JohanEngdahlwassupported bytheStockholmCountyCouncil(clinicalresearchappointment).

Appendix

A. Supplementary

data

Supplementarymaterialrelatedtothisarticlecanbefound,inthe onlineversion,atdoi:10.1016/j.resuscitation.2020.07.009.

REFERENCES

1.GirotraS,NallamothuBK,SpertusJA,LiY,KrumholzHM,ChanPS. Trendsinsurvivalafterin-hospitalcardiacarrest.NEnglJMed 2012;367:1912 20.

2.HessulfF,KarlssonT,LundgrenP,etal.Factorsofimportanceto 30-daysurvivalafterin-hospitalcardiacarrestinSweden a population-basedregisterstudyofmorethan18,000cases.IntJCardiol 2018;255:237 42.

3.SpearpointKG,McLeanCP,ZidemanDA.Earlydefibrillationandthe chainofsurvivalin‘in-hospital’adultcardiacarrest;minutescount. Resuscitation2000;44:165 9.

4.SoarJ,NolanJP,BottigerBW,etal.EuropeanResuscitationCouncil GuidelinesforResuscitation2015:Section3.Adultadvancedlife support.Resuscitation2015;95:100 47.

5.VelasquezT,MackeyG,LuskJ,etal.ESICMLIVES2016:partthree. IntensiveCareMedExp2016;4:28.

6.MartensPR,MullieA,CalleP,VanHoeyweghenR.Influenceon outcomeaftercardiacarrestoftimeelapsedbetweencallforhelpand

startofbystanderbasicCPR.TheBelgianCerebralResuscitation StudyGroup.Resuscitation1993;25:227 34.

7.Hasselqvist-AxI,RivaG,HerlitzJ,etal.Earlycardiopulmonary resuscitationinout-of-hospitalcardiacarrest.NEnglJMed 2015;372:2307 15.

8.ChanPS,KrumholzHM,NicholG,NallamothuBK.Delayedtime todefibrillationafterin-hospitalcardiacarrest.NEnglJMed 2008;358:9 17.

9.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.

10.WassersteinRL,LazarNA.TheASA'sstatementonp-values:context, process,andpurpose.AmStat2016;70:129 33.

11.ZhangZ,ZhaoY,CanesA,SteinbergD,LyashevskaO.Predictive analyticswithgradientboostinginclinicalmedicine.AnnTranslMed 2019;7:152.

12.vanBuurenS.Multipleimputationofdiscreteandcontinuous databyfullyconditionalspecification.StatMethodsMedRes 2007;16:219 42.

13.McEvoyMD,FieldLC,MooreHE,SmalleyJC,NietertPJ,Scarbrough SH.TheeffectofadherencetoACLSprotocolsonsurvivalofeventin thesettingofin-hospitalcardiacarrest.Resuscitation2014;85:82 7.

14.HonarmandK,MephamC,AinsworthC,KhalidZ.Adherenceto advancedcardiovascularlifesupport(ACLS)guidelinesduring in-hospitalcardiacarrestisassociatedwithimprovedoutcomes. Resuscitation2018;129:76 81.

15.AndersonML,NicholG,DaiD,etal.Associationbetweenhospital processcompositeperformanceandpatientoutcomesafter in-hospitalcardiacarrestcare.JAMACardiol2016;1:37 45.

16.LyonsPG,EdelsonDP,ChurpekMM.Rapidresponsesystems. Resuscitation2018;128:191 7.

17.KallestedtML,RosenbladA,LeppertJ,HerlitzJ,EnlundM.Hospital employees’theoreticalknowledgeonwhattodoinanin-hospital cardiacarrest.ScandJTraumaResuscEmergMed2010;18:43.

18.MakinenM,AuneS,Niemi-MurolaL,etal.AssessmentofCPR-Dskills ofnursesinGoteborg,SwedenandEspoo,Finland:teaching leadershipmakesadifference.Resuscitation2007;72:264 9.

19.MorettiMA,CesarLA,NusbacherA,KernKB,TimermanS,Ramires JA.Advancedcardiaclifesupporttrainingimproveslong-termsurvival fromin-hospitalcardiacarrest.Resuscitation2007;72:458 65.