Cardiac pathology 6 months after hospitalization for COVID-19 and association with the acute disease severity

(1)

Cardiac pathology 6 months after

hospitalization for COVID-19 and association with the acute disease severity

PederL.Myhre,MD,PhDâ^,^b^,^∗,SiriL.Heck,MD,PhD^b^,^c^,^∗,JuliaB.Skranes,MDâ^,^b,ChristianPrebensen,MD,PhD^b^,^d, ChristineM.Jonassen,PhDê,TrygveBerge,MD,PhD^b^,^f,AlbulenaMecinaj,MDâ^,^b,WoldegabrielMelles,MSc^c, GunnarEinvik,MD,PhD^b^,^g,CharlotteB.Ingul,MD,PhD^h,ArnljotTveit,MD,PhD^b^,^f,JanErik Berdal,MD,PhD^b^,^d, Helge Røsjø,MD,PhD^b^,ⁱ,MagnusN.Lyngbakken,MD,PhDâ^,^b,andTorbjørnOmland,MD,PhD,MPHâ^,^bLorenskog, Norway;Oslo,Norway;Gralum,Norway;Trondheim,Norway

Background

Coronavirusdisease2019(COVID-19)maycausemyocardialinjuryandmyocarditis,andreportsof persistentcardiacpathologyafterCOVID-19haveraisedconcernsoflong-termcardiacconsequences.Weaimedtoassess thepresenceofabnormalcardiovascularresonanceimaging(CMR)ﬁndingsinpatientsrecoveredfrommoderate-to-severe COVID-19,anditsassociationwithmarkersofdiseaseseverityintheacutephase.

Methods

Fifty-eight(49%)survivorsfromtheprospectiveCOVIDMECHstudy,underwentCMRmedian175[IQR105- 217] daysafter COVID-19 hospitalization.AbnormalCMRwas deﬁnedas leftventricularejection fraction(LVEF)<50%

or myocardial scarby late gadolinium enhancement. CMRindiceswere comparedto healthycontrols(n = 32),andto circulatingbiomarkersmeasuredduringtheindexhospitalization.

Results

AbnormalCMRwaspresentin12(21%)patients,ofwhom3wereclassifiedwithmajorpathology(scarand LVEF<50%orLVEF<40%).Therewasnodifferenceintheneedofmechanicalventilation,lengthofhospitalstay,andvital signs betweenpatients withvswithout abnormalCMR after6 months.Severe acute respiratorysyndrome coronavirus 2 viremiaandconcentrationsofinflammatorybiomarkersduringtheindexhospitalizationwerenotassociatedwithpersistent CMRpathology.CardiactroponinTandN-terminalpro-B-typenatriureticpeptideconcentrationsonadmission,werehigher inpatientswithCMRpathology,buttheseassociationswerenotsignificantafteradjustingfordemographicsandestablished cardiovasculardisease.

Conclusions

CMRpathology6monthsaftermoderate-to-severeCOVID-19was presentin21%of patientsanddid notcorrelatewithseverityofthe disease.CardiovascularbiomarkersduringCOVID-19werehigherinpatientswithCMR pathology,butwithnosigniﬁcantassociationafteradjustingforconfounders.

Trial Registration

COVIDMECHStudyClinicalTrials.govIdentiﬁer:NCT04314232(AmHeartJ2021;242:61–

70.)

Keywords:COVID-19;cardiacmagneticresonanceimaging;CMR;biomarkers;troponin;NT-proBNP

Fromthe^aDepartmentofCardiology,DivisionofMedicine,AkershusUniversityHospi- tal,Lørenskog,Norway,^bInstituteofClinicalMedicine,FacultyofMedicine,Universityof Oslo,Oslo,Norway,^cDepartmentofDiagnosticImaging,AkershusUniversityHospital, Lørenskog,Norway,^dDepartmentofInfectiousDiseases,DivisionofMedicine,Akershus UniversityHospital,Lørenskog,Norway,^eCenterforLaboratoryMedicine,ØstfoldHos- pitalTrust,Grålum,Norway,^fDepartmentofMedicalResearch,BærumHospital,Vestre VikenHospitalTrust,Norway,^gDepartmentofPulmonology,DivisionofMedicine,Ak- ershusUniversityHospital,Lørenskog,Norway,^hDepartmentofCirculationandMed- icalImaging,NorwegianUniversityofTechnologyandScience,Trondheim,Norway, iDivisionofResearchandInnovation,AkershusUniversityHospital,Lørenskog,Norway

∗Theseauthorscontributedequally.

Abbreviations:COVID-19,Coronavirusdisease2019;SARS-CoV-2,Severeacuterespiratory syndromecoronavirus2;LGE,LateGadoliniumEnhancement;IL-6,Interleukin-6;ICU,In- tensivecareunit;NEWS,NationalEarlyWarningScore;CRP,C-reactiveprotein;PCT,Pro- calcitonin;cTnT,CardiactroponinT;NT-proBNP,N-terminalpro-B-typenatriureticpeptide;eGFR,Estimatedglomerularfiltrationrate.

SubmittedApril18,2021;acceptedAugust2,2021

Reprintrequests:TorbjørnOmland,MD,PhD,DepartmentofCardiology,AkershusUni- versityHospital,Sykehusveien27,Lørenskog1478,Norway.

Cardiac involvement in coronavirus disease 2019 (COVID-19)iscommon.Patientswithunderlyingcardio- vascular disease (CVD) are at increased risk of severe disease and cardiac complications. Arrhythmias, acute coronarysyndromeand heartfailurerelated eventsare known to occur in patients hospitalized with COVID- 19,^1-3inadditionto the lessfrequent fulminant severe acute respiratory syndrome coronavirus 2 (SARS-COV-

E-mailaddress:torbjorn.omland@medisin.uio.no. 0002-8703

This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)

https://doi.org/10.1016/j.ahj.2021.08.001

(2)

2)myocarditis.⁴Myocardialinjury,reflectedbyelevated concentrationsofcardiactroponins,isfrequent butthe prevalencedependsonthebaselineriskofthepopula- tionandclinicalsetting.⁵Alargeproportionofhospital- izedCOVID-19patientswithelevatedcardiactroponins hasbeenreportedtohaveelevatedcardiovascularmag- neticimaging(CMR)measurementsofT1,extracellular volume or late gadoliniumenhancement (LGE) with a non-ischemicpatternintheacutephase.⁶Retrospective studieshave suggested that cardiac troponinand natri- ureticpeptidesaremarkersofriskinCOVID-19.²^,⁷This is howeverless clearinprospective studieswith unse- lectedpatients.⁸

Persistentsymptoms,particularlyfatigueanddyspnea, arecommonafter COVID-19.⁹Myocardialinflammation and ventricular dysfunction determined by CMR were reported in 78% of patients who recovered from pre- dominantlymild-moderateCOVID-19infection,irrespec- tiveofcardiacsymptoms.¹⁰Instudiesofcollegeathletes withCOVID-19,thepresenceofCMRfindingsconsistent withmyocarditisvariessignificantlyfrom1.5%to15%.%

11-13 The presence ofabnormal CMRfindings in patients recovered from moderate-to-severe COVID-19, and the association with cardiac biomarker concentrations and SARS-CoV-2viremiaintheacutephaseisunknown.

Methods

Studydesignandparticipants

COVID MECH (NCT04314232) was a prospective, observational study consecutively enrolling unselected adult patients hospitalized with laboratory-confirmed COVID-19. The study was conducted at Akershus Uni- versityHospitalinNorwaybetween March18 andMay 4,2020.¹⁴ParticipantswereclassifiedasICUpatientsif they were admitted to theICU and received intensive care treatment(mechanicalventilation) for>24 hours.

HistoryofCVD,pulmonarydisease,hypertensionanddi- abeteswasrecordedfromtheelectronicmedicalrecords and after interviewingthepatients atdiscretion ofthe treatingphysicians.NationalEarlyWarningScoreisaval- idated toolused for detection and response to clinical deteriorationinadultpatients.Itiscalculatedfrom6vi- talsigns,withlowriskmeasurementsyielding0points, andabnormalvaluesgivingupto3pointsperitem,with amaximumscoreof20.¹⁵

Afterdischarge, patientswere invitedbymailortele- phonetoafollow-upstudy.Of128participantsincluded intheCOVIDMECHbiobankstudy,therewere118sur- vivors attime ofthefollow-up study,and 102wereel- igibleforparticipation. Sixteenpatientswere excluded due to cognitive impairment, major language barriers, beingstillhospitalizedattimeofinitiation offollow-up study, residing outside the hospital catchment area or withunavailablecontactinformation.Oftheeligiblepa- tientswhowereinvitedtothefollow-upstudy,63(62%)

werescheduledforCMRexamination,while28didnot wanttoparticipateand11didnotreplytotheinvitation.

FourpatientsabortedtheCMRexaminationduetoclaus- trophobia andone did not attendthe scheduled CMR, leaving58patientswithavailableCMRwhomakeupthe populationinthecurrentstudy(Suppl.Figure1).

Patients were compared to 32 healthy participants fromtheprospective,population-based, age-cohortAk- ershus Cardiac Examination 1950 Study.¹⁶The con- trolswerenormotensive, non-obesenon-smokers with- outknowndiabetesorcardiovasculardisease.

The COVID MECH, COVID CMR and Akershus Car- diac Examination 1950 studies were approved by the Regional Ethics Committee (#20/02873; #20/05884;

#2011/1475)andbytheinstitutionalDataProtectionOf- ficer(#117589;#148701;#12_093).Noextramuralfund- ingwasusedtosupportthiswork.Theauthorsaresolely responsibleforthedesignandconductofthisstudy,all studyanalyses,thedraftingandeditingofthepaperand itsfinalcontents.

CMRprotocol

TheCMRswereconductedatAkershusUniversityHos- pitalbetweenJune24andNovember182020ona1.5 MRIscanner(Achieva;PhilipsMedicalSystems,Best,The Netherlands).Short-axis,steady-state-free precessionse- quenceswere acquiredincontiguous8mm short axis slicesforassessingventricularvolumesandejectionfrac- tion.T2STIRimages wereacquiredin10mmslicesin a singlemidventricular short axisviewand 1four and 1 left ventricular 2 chamber views. Two-dimensional, phasesensitiveinversionrecoveryLGEimagingincon- tiguous10mm short-axisslicescoveringtheventricles and3long-axisviewsfor assessing myocardialscarring was performed starting 10 minutes after injection of 0.15mmol/kggadoteratemeglumine(ClariscanGé,GE Healthcare).MyocardialT1andT2mapping sequences for the assessment of diffuse myocardial fibrosis and edema were acquired in single 10 mm midventricular short-axisslices.T1mapswereacquiredbeforeand 15 minutes after contrast administration using MOLLI se- quenceswith5s(3s)3sand4s(1s)3s(1s)2smapping schemes, respectively. A gradient-spin echo sequence was used for T2mapping. T1 and T2 maps were gen- eratedondedicatedsoftware(cvi42,v5.11.4,CircleCar- diovascularInc,Calgary,Canada).Bloodhematocrit for calculationoftheextracellularvolumefractionwasmea- suredatthetimeofCMRexamination.

CMRassessment

Assessment of ventricular volumes and EF and mass wasperformedoncvi42accordingtoSCMRguidelines.¹⁷ Trabeculations and papillarymuscles were included in theLV volumes.The presenceof scarwas assessed on LGEsequencesbysemiautomaticsignalintensitythresh- olding5standarddeviationsaboveremotemyocardium

(3)

Figure1

Title:Non-ischemicleftventricular(LV)scarbylategadoliniumenhancementimaging,Caption:Lategadoliniumimagesdemonstratingnon- ischemicscarsinA,thebasalinferolateralLVofa54-year-oldmale209daysafterhospitalizationforCOVID-19;B,thebasoseptalLV ofa60yearold male175daysafterhospitalizationforCOVID-19;andC, thebasolateralLVofa50yearoldmale202daysafter hospitalizationforCOVID-19.

andcategorizedasischemicornon-ischemic.Presenceof LGE(myocardialscar)orleftventricularEF(LVEF)<50%

weredefinedasabnormalCMR.Thesewerefurtherclas- sifiedintomajorpathology(scarandLVEF<50%orLVEF

<40%) and minor pathology (scar and LVEF ≥50% or LVEF40%-49%).

T2 STIR images were visually assessed for focal myocardial and pericardial edema. Myocardial T1 and T2 relaxationtimesweremeasuredbyconservatively plac- ingregionsofinterestinthemidventricularseptum.Ar- easofLGEandsignificantartifactswereexcludedfrom themeasurements.ECVfractionwascalculatedasprevi- ouslydescribed.¹⁸Myocardialfeaturetrackingstrainana- lyzeswasperformedoncvi42.Leftventricularlongitudi- nalstrainwasassessedin3longaxisviews,andcircum- ferentialstrainin3shortaxisslices(basal,midventricular,andapical).Examinationswithpersistinginadequate trackingafterupto2timescontourcorrectionwereex- cludedfromanalysis.

Laboratoryanalysis

Blood samples were collected at admission and on target day 3 (day 2-5 accepted) during hospitalization andstoredat-80°Cinastudy-specificbiobankpending analysis.Measurements ofinterleukin-6(IL-6), procalcitonin,ferritin,cardiactroponinT(cTnT)andN-terminal pro-B-type natriuretic peptide (NT-proBNP) were per- formedbytheElecsysimmunoassayontheCobase801 platform(RocheDiagnostics,Rotkreuz,Switzerland).C- reactive protein was measured as part of clinical rou- tine. Five patients had missing biobank samples, and for these cTnT, NT-proBNP and ferritinwere recorded fromtheclinicalroutinemeasurements, whileIL-6and procalcitoninare reportedasmissing. SARS-CoV-2RNA in plasma(viremia) was detected byreverse transcrip-

tionreal-timepolymerasechainreactiononaQuantStu- dio 7system (Thermo Fisher Scientific, Waltham,Mas- sachusetts,USA).Detailsofthelaboratoryanalysishave beenreportedpreviously.¹⁴

Statisticalanalyses

Values are reported as N (%), median (quartile 1 to quartile 3) or mean ± SD, as appropriate. Cate- gorical and continuous variables were compared us- ingthechi-square test forbinary variables,ANOVA for parametriccontinuousvariables,andtheKruskal-Wallis testfornon-parametriccontinuousvariables.Changein biomarkerconcentrationsfromhospitaladmissiontoday 3werecalculatedbysubtractionandcomparedbynon- parametric tests. To account for possible confounders betweenbiomarkerconcentrationsandCMRpathology, we performed multivariable logistic and linear regres- sionmodelsthatwereadjustedforage,sex,raceandes- tablishedCVD(selectedapriori),usinglog-transformed biomarker concentrations. We also adjusted for time from index hospitalization to CMR examination in ad- ditionalmodels.All statisticalanalyses were performed usingStataSoftware(version16,StataCorp.,CollegeSta- tion,TX,USA).A2-sidedP-valueof<.05wasconsidered statisticallysignificant.

Results

Baselinecharacteristics

ThetimefromhospitaladmissiontotheCMRexamina- tionwasmedian175(IQR105-217)days(range75-246).

The58 patientswithavailable CMRwere agedmedian 56(Q1-Q350-70) years,30(56%)weremale, 30(57%) were Caucasian and median BMI was 27.2 (24.2-29.4) kg/m². EstablishedCVD was present in5 (9%), hyper-

(4)

tension in 12 (21%), diabetes mellitus in 6 (11%) and chronic kidneydisease in2(4%) patients. The median length ofstay attheindexhospitalization was7(4-11) days and 11 (19%) were treated withmechanical ventilation in the ICU. Hydroxychloroquine was given to 23(40%)patients,whilenopatientswere treatedwith high dose corticosteroids or convalescent plasma. On admission for acute COVID-19,69% reported dyspnea, 25%chestpainand56%fatigue.AftertheacuteCOVID- 19infection55%reportedpersistentdyspnea,4%chest painand 64%fatigue.Cardiac arrhythmiawasreported in2patientsduringtheacuteCOVID-19.Screeningwith 24-hourECG monitoringpost-COVID-19 revealed1pa- tient with paroxysmal atrial fibrillation and 4 patients with short episodes of self-limiting non-sustained ven- triculartachycardia.Therewasnosignificantdifference withrespecttodemographics,comorbidities,vitalsigns onadmission,lengthofstayorICU treatmentbetween thestudypopulationandotherparticipantsinthetotal COVIDMECHcohort(Suppl.Table1).

CardiacpathologyonCMR

Byprotocol,the2patientswithchronickidneydisease didnotreceivecontrastandwereaccordinglynoteval- uated for myocardialscar by LGE. Both these patients hadnootherpathologyonCMR(comparableLVandRV structureandfunction,T1,T2and straintothosewith availableLGE),andwereaccordinglyclassifiedwithnor- malCMR.Inaddition,1patienthadunevaluableLGEse- quencesduetoseveremotionartifacts.Thispatienthad reducedLVEF,andwasaccordinglyclassifiedwithabnor- malCMR.

In total, 12 (21%) patients were classified with abnormal CMR. Among these, 3 patients had major my- ocardialpathology:1withbothmyocardialscar(acom- binedischemic/non-ischemicscarof6.5%scarvolume) andreducedLVEF(38%):1withLVEF37%andnoscar;

and1withLVEF39%andunavailableLGE-measurements duetounevaluableLGE. Theremaining9patientswith abnormal CMR were classified with minor pathology on CMR. One patient had a combined ischemic/non- ischemic scar of 3.0 % scar volume, the other 8 had non-ischemicscars(meanscarvolume2.0±1.1%;range 0.7%-4.2%)andLVEF≥50%(mean57±6%,range50%- 69%)(Figure1).TherewerenodifferencesinnativeT1 orT2valuesbetweenpatientswithandwithoutmyocar- dialpathologybyconventionalCMRfindings.Pericardial enhancementwasnotidentifiedinanyofthepatients.

TheCMRmeasurementswerecomparedto32healthy controls using the same CMR equipment, method for analysisandanalyst.In additiontobeingfreeofcardio- vascularcomorbidities and establishedrisk factors,the healthycontrolswereolderandmorefrequentlyfemale andofwhiteracecomparedtotheCOVIDCMRpatients (Suppl.Table2).Therewerenosignificantdifferencesin LVEF,RVEF,nativeT1,nativeT2,extracellularvolume,LV

strainandmyocardialscarbetweentheCOVIDpatients andhealthycontrols,althoughtherewasaborderlinesig- nificanthigherT1inCOVIDpatients(mean1006±31 msvs993±29ms,P=.05;Suppl.Table3).

CMRpathologybyclinicalcharacteristicsand diseaseseverityattheindexhospitalization

Patients with abnormal vs normal CMR were older, withmore prevalent CVD and chronic pulmonary disease (Table I). Vital signs at admission of the index hospitalization were comparable in patients with and withoutabnormal CMR, apartfrom lowertemperature in those with abnormal CMR. Disease severity scor- ing, length of hospital stay and the proportion of pa- tientsrequiringmechanicalventilationintheICUwere also comparable in patients with and without abnor- malCMR(Figure2).Clinical characteristics,comorbidities,vitalsignsduringtheindexhospitalizationandhos- pitaloutcome stratified by the presence of minor and majorpathology on CMR is presented in Suppl. Table 4.Therewerenodifferencesinpatientreportedsymp- tomsduringacuteCOVID-19andpost-COVID-19bypres- ence of CMR pathology (Suppl. Table 5). Cardiac ar- rhythmiasduring acute-COVID-19were more common inpatients withCMR pathology after 6 months, while therewerenodifferencesinprevalenceofarrhythmias fromscreeningpost-COVID-19.Therewasnodifference intreatmentwithhydroxychloroquinebetweenpatients withandwithoutabnormalCMR(50%vs37%,P=.41, respectively).

Eleven(19%)ofpatientshadsevereCOVID-19requir- ingmechanicalventilation inthe ICU.Only 1ofthese hadabnormalCMR(non-ischemicscarandLVEF≥50%).

TableIIdisplays measurements ofLV, RV andLAstruc- tureandfunction,myocardialscar,T1,T2andstrainmea- surements in patients withsevere COVID-19 requiring mechanicalventilationinthe ICUand inpatients with moderate COVID-19treated in medical wards. Overall, therewere nosignificant differences inCMR measure- mentsbetweenthe2groups6monthsafterhospitaliza- tion(TableII).

CMRpathologybycardiovascularandinﬂammatory biomarkersmeasuredduringtheindex

hospitalization

cTnTandNT-proBNPconcentrationsonadmissionfor theindexhospitalizationweremedian8(IQR4-13)ng/L and97(IQR35-195)ng/L,respectively.Elevatedconcen- trations of cTnT (≥14 ng/L) was present in 16 (28%) and NT-proBNP (≥250 ng/L) in 20 (35%) at any time pointduringthe hospitalstay. SARS-CoV-2viremiawas detectablein19(36%)ofpatientsduringtheindexhos- pitalization.ComparedtopatientswithnormalCMRaf- ter6months,patientswithCMR pathologyhadhigher admissionconcentrationsofcTnT(median[IQR]13[11- 25]ng/Lvs7[4-11]ng/L,P=.003)andNT-proBNP(357

(5)

TableI. Baselinecharacteristics,vitalsignsonadmissionandhospitaloutcomefortheacuteCOVID-19hospitalization,stratiﬁedbythe presenceofpathologyoncardiacmagneticresonanceimaging(CMR)6monthslater

NormalCMR CMRpathology P-value

n=46 n=12

Age,years 54[46,70] 68[57,78] .030

Malesex 26(56.5%) 8(66.7%) .53

Whiterace 23(50.0%) 8(66.7%) .30

Bodymassindex,kg/m² 27.7[24.2,29.4] 25.8[24.0,29.4] .33

Obesity 11(23.9%) 3(25.0%) .94

DiabetesMellitus 6(13.0%) 0(0.0%) .19

Hypertension 9(20.0%) 3(25.0%) .71

Cardiovasculardisease 2(4.3%) 3(25.0%) .023

Chronicpulmonarydisease 1(2.2%) 3(25.0%) .005

Chronickidneydisease 2(4.3%) 0(0.0%) .46

Currentsmoking 0(0.0%) 1(8.3%) .05

Temperature,°C 38.2[37.5,39.0] 37.3[36.8,38.3] .033

Heartrate,/min 90[77,98] 86[73,89] .24

Systolicbloodpressure,mmHg 129[120,140] 131[125,138] .76

Oxygensaturation,% 95[93,96] 94[93,95] .43

NEWS-score 5[3,7] 4[3,5] .24

Lengthofstay(d) 8[4,12] 6[4,9] .42

ICUadmission 10(21.7%) 1(8.3%) .29

ICU, intensive care unit; NEWS, National Early Warning Score.

TableII. Cardiacmagneticresonanceimagingmeasurementsofleftventricular(LV)andrightventricular(RV)structureandfunction, myocardialscarbylategadoliniumenhancement,T1andT2inCOVID-19patientsrequiringmechanicalventilationattheICUandin patientstreatedatthemedicalwards

Medicalwardn=47 ICUn=11 P-value

Myocardialscar 9(20.5%) 1(9.1%) .38

Scarvolume,% 2.7±1.8 1.9 .69

LVenddiastolicvolumeindexed,ml/m² 74.6±13.7 79.9±14.3 .26

LVendsystolicvolumeindexed,ml/m² 30.8±9.3 33.8±6.9 .32

LVstrokevolume,ml 87.4±20.4 95.1±30.0 .31

LVejectionfraction,% 59.0±7.8 57.6±5.4 .58

LVmassindexed,g/m² 48.6±10.6 48.9±9.1 .93

LVcircumferentialstrainshortaxis,% -18.7±3.4 -19.1±1.8 .69

LVlongitudinalstrainlongaxis,% -16.3±2.2 -16.4±1.9 .89

LAvolumeindexed,ml/m² 34.1±13.3 32.1±9.8 .67

RVenddiastolicvolumeindexed,ml/m² 72.9±13.1 79.6±15.7 .15

RVendsystolicvolumeindexed,ml/m² 31.1±7.9 34.8±7.1 .16

RVstrokevolume,ml 83.3±18.3 92.6±29.7 .19

RVejectionfraction,% 57.5±6.7 56.2±4.0 .52

Extracellularvolume,% 25.0±3.0 23.8±1.1 .49

NativeT1,ms 1010±31 989±25 .05

NativeT2,ms 51.5±2.9 52.1±2.3 .32

LA, left atrial; LV, left ventricular; RV, right ventricular.

[88-616]ng/Lvs97[26-156]ng/L,P=.013)(Figure2).

For cTnT, these differences were attenuated when adjusting for demographics (age, sex and race) and CVD (P=.12),whiletheassociationpersistedinmultivariable modelsforNT-proBNP(P=.03).ForNT-proBNPthisdif- ferencewasdrivenbypatientswithmajorCMRpathol- ogy who hadparticularly high concentrations(median 665 (IQR 487-15461) ng/L) while patients with minor andnoCMRpathologyhadcomparableconcentrations:

median 109(IQR 86-449)vs92(IQR 26-156)ng/L,P=

.11(Suppl. Table6). Patientswith elevatedNT-proBNP duringhospitalizationalsohadsignificantlyhigherCMR markersofmyocardialedema(T2)andlowerLVejection fraction,but these associations were attenuated when adjusting for demographics and CVD (Suppl. Table 7), also when analyzingadmission NT-proBNPas a continuous variable (P= .16). Elevated cTnT was associated with higher T2 values and larger LV and LA volumes, and the association with T2 persisted also inadjusted models(P=.039;Suppl.Table8),butnotwhenanalyz-

(6)

Figure2

Title:ClinicalvariablesandbiomarkerconcentrationsduringhospitalizationforCOVID-19inpatientswithandwithoutpathologyonCMR after6months,Caption:ProportionofpatientswithandwithoutpathologyonCMRwithneedforintensivecareunit(ICU)treatment,Severe AcuteRespiratorySyndromeCoronavirus2(SARS-CoV-2)viremia,NationalEarlyWarningScore(NEWS)andlengthofhospitalstayduring admissionforCOVID-19,andconcentrationsofcardiactroponinT(cTnT),N-terminalpro-B-typenatriureticpeptide(NT-proBNP),C-reactive protein(CRP)andferritinmeasured atadmissionandchangetoday3in patients.CMRpathologywasdeﬁnedasmyocardialscaror reducedleftventricularejectionfraction.Thewhiskersrepresentquartile1toquartile3forcontinuousvariables.

ingcTnTasacontinuousvariable(P=.62).CMRmark- ers reflecting fibrosis (native T1 and ECVfraction), LV mass,LV function(includingstrain measurements) and RV function were overall comparable in patients with andwithoutelevationofeithercardiovascularbiomarker during the index hospitalization (Suppl. Tables 7 and 8). The associations between cardiovascular biomarkers and CMR measurements did not change when ad- ditionally adjusting for time from hospitalization to CMR.

Admissionconcentrationsofinflammatorybiomarkers (ie,C-reactiveprotein,IL-6,PCTandferritin)werecom- parableinpatientswithandwithoutpathologyonCMR.

SARS-CoV-2 viremia was also present in an equivalent proportionofpatientswithnormalCMR(n=15,37%) andpatientswithCMRpathology(n=4,33%,P=.84) (TableIII,Figure2).

Therewerenodifferencesindeltavaluesofinflamma- toryandcardiovascularbiomarkersfromhospitaladmis- siontoday3inpatientswithandwithoutpathologyon CMR(TableIII,Figure2).

Discussion

Among58unselectedpatientshospitalizedforCOVID- 19, cardiac pathology on CMR after 6 months was present in 12 (21%) patients. There were no associa-

(7)

TableIII. PathologyonCMRafter6monthsandconcentrationsofcardiovascularandinﬂammatorybiomarkersduringtheindex hospitalizationforCOVID-19

NormalCMRn=46 AbnormalCMRn=12

cTnTadmission(ng/L) 7.0[4.011.0] 12.5[10.5,25.0] 0.003

cTnTdeltatoday3(ng/L) 0.0[-1.0,1.0] 0.0[-2.0,2.0] 0.79

cTnT≥14ng/Lduringhosp. 10(22.2%) 6(50.0%) 0.06

NT-proBNPadmission(ng/L) 97.0[26.0,156.0] 357.0[88.0,615.5] 0.013

NT-proBNPdeltatoday3(ng/L) 15.5[-9.0,91.0] -39.0[-93.0,56.0] 0.18

NT-proBNP≥250ng/Lduringhosp. 13(28.9%) 7(58.3%) 0.06

CRPadmission(mg/L) 70[28,160] 72[47,100] 0.83

CRPdeltatoday3(mg/L) 20[-10,40] 10[-9,30] 0.75

IL-6admission(pg/mL) 37.0[20.8,55.9] 42.5[26.1,69.8] 0.50

IL-6deltatoday3(pg/mL) -8.9[-36.2,17.1] -10.5[-28.8,2.8] 0.95

PCTadmission(g/L) 0.12[0.06,0.21] 0.14[0.09,0.21] 0.43

PCTdeltatoday3(g/L) -0.01[-0.03,0.06] -0.01[-0.04,0.07] 0.98

Ferritinadmission(g/L) 513[265,919] 602[351,1042] 0.65

Ferritindeltatoday3(g/L) 7[-126,138] 13[-98.0,103] 0.89

SARS-CoV-2viremia 15(36.6%) 4(33.3%) 0.84

Concentrations were measured at admission and changes in concentrations to day 3. Also presented by presence of SARS-CoV-2 viremia, elevated cTnT and NT-proBNP at any time during the index hospitalization for COVID-19 are reported.

CRP, C-reactive protein; cTnT, cardiac troponin T; IL-6, interleukin 6; NT-proBNP, N-terminal pro-B-type natriuretic peptide; PCT, procalcitonin; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

tionsbetween CMRfindings andtheneedformechan- ical ventilation,length-of-stay, severity scorings, inflammatory biomarkers or SARS-CoV-2 viremia during the acute COVID-19 hospitalization. Higher concentrations ofcTnTandNT-proBNPduringtheindexhospitalization wereassociatedwithahigherprevalenceofCMRpathol- ogyafter6months,butthiswaslargelyattenuatedafter adjustingfordemographicsandestablishedCVD.

Persistentcardiacpathologyafterrecoveryfrom COVID-19andseverityoftheacuteinfection

Among patients with moderate-to-severe COVID-19, we found a substantially lower prevalence of pathology on CMR than in a Germany study by Puntmann et al.¹⁰They reported abnormal findings 10 weeks after theacute infection in78 of 100 patients who had predominantly mild-to-moderate COVID-19. That prior studyidentifiedfocalscarringin32%andpericardialen- hancementin22%.They alsoreportedelevatedT1and T2values,sensitivemeasuresofmyocardialfibrosisand edema,comparedtohealthyandrisk-matchedcontrols.

AChinesestudyretrospectivelyassessed26patientsre- ferredtoCMRforcardiacsymptomsafterhospitalization forCOVID-19,andfoundpathologicalconventionalCMR findingsin58%ofpatients,myocardialedemabeingthe predominantfinding.¹⁹Inastudyofunselectedpatients 2-3 monthsafter COVID-19cardiac abnormalities were foundin26%,while60%,29%and10%hadabnormalities inthelungs,kidneysandliveronMRI,andthesefindings correlatedwithacutediseaseseverity.²⁰Inastudyofpa- tients who allhad troponinelevation duringtheacute COVID-19,CMRafter median68days demonstratedre- ducedLVEFin11%andmyocardialscarin54%,ofwhich

approximatelyhalfwasmyocarditis-likescarandhalfis- chemicscar.²¹Thisstudyfound noevidence ofdiffuse fibrosis (T1)oredema(T2) intheremotemyocardium comparedtomatchedcontrols.Inourstudy,wefounda lowerproportionoffocalmyocardialscarringthanthese studies,andnopatientswithfocalmyocardialedemaor pericardialenhancement.Also,wefoundnosignificant differencesin T1 orT2 values, or measures of LV and RVstructureandfunction,betweenpatientswithsevere COVID-19(requiring mechanicalventilation) and mod- erateCOVID-19 (requiring hospitalization, but not mechanicalventilation).Therewerealsonoclinicallymean- ingfuldifferencesinCMRmeasurementsbetweenpost- COVID-19patientsandhealthycontrols.NativeT1was slightlyhigher(mean15ms)comparedto healthycontrols, andthis association reachedborderline statistical significance.ElevatedT1hasbeendemonstratedinthe subacutephaseofCOVID-19,andthismodestdifference mayreflectresidualchanges,butmayalsoberelatedto ahigherprevalenceofCVDandriskfactorsamongthe COVID-19.

PossibleexplanationsforthelowerprevalenceofCMR pathology in our COVID-19 patients mayrelate to differences in time since the acute infection. Myocardial edema decreases in the weeks and months after myocardialinjury.²²In our study,median time fromdiag- nosisto CMR was 175 days, which is longer than the aforementionedstudies.Thismaysuggest regressionof post-COVID-19 cardiac pathology with time. Secondly, differencesin baselinerisk factors mayplay a role. Al- thoughthepatientsintheGermanstudywere younger andfree ofheart failureandcardiomyopathy,otherrisk factorsforsubclinicalcardiacremodelingsuchassmok-

(8)

ing, chronic obstructive pulmonary disease and coro- narydiseasewere morecommon.Also,theproportion ofpathologicalCMRfindingsintherisk-matchedcontrol group was ofthe German study was high.The cohort intheChinesestudywasalsoyoung,butwithmoderate to severe COVID-19infection, and allpatients hadcar- diacsymptoms.Finally, technicalCMRdifferencessuch asMRIfieldstrengthandmappingsequencesmaypoten- tiallyexplainsomeofthedifferences.

Elevatedcardiovascularbiomarkersduring

COVID-19andmyocardialpathologyafterrecovery DuringacuteCOVID-19,patientswithelevatedcardiac troponin have been reported to have elevated T1,extracellular volumeand LGEmeasurements, in addition to enhancedmacrophage numbersin myocardialbiop- sies.⁶In ourstudy, higher concentrations ofcTnT and NT-proBNPonadmissionwereassociatedwiththepres- enceofscarorreducedLVEFonCMRafter6months,and thehighest concentrationswereseeninpatientsclassi- fiedwithmajorCMRpathology.However,patientswith scar or reduced LVEF did not experience a greater in- creaseinconcentrationsofthesecardiovascularbiomark- ersduringtheinfection.Moreover,theassociationswere attenuatedwhenadjustingfordemographicsandestab- lishedCVD.Accordingly,webelievethismostlikelyre- flects pre-existing subclinical CVD rather than persis- tentCOVID-19-relatedacutemyocardialinjuryandstress.

ThisissupportedbytheestablishedlinkbetweencTnT andNT-proBNPelevationsandsubclinicalmyocardialfi- brosis and scar in the general population.²³^,²⁴There were limitedassociations between CMR measurements of edema(T2) and elevated cTnTand NT-proBNPdur- ingtheindexhospitalization.Importantly,T2valueshas beenshowntoincreasewithage,²⁵andindeedadjusting forthisattenuatedtheassociationtothesecardiovascular biomarkers.However,theassociationbetweenelevated cTnT concentrations duringhospitalization and higher T2 remainedsignificant after adjusting for demograph- icsandestablishedCVD.Thisfindingmayimplythatpa- tientswithmyocardialinjuryduringtheacuteinfection maybeat riskof persistentmyocardialedemaafter re- coveryfrom theacuteCOVID-19.Still, therewere limited correlationsbetween cTnTand othermeasures of pathologyofCMR.Accordingly,theclinicalsignificance oftheassociationbetweenmyocardialinjuryandpersis- tentmyocardialedemaisuncertainandrequiresmorere- searchwithlongerfollow-upandclinicaloutcomessuch asincidentheartfailure.

InﬂammatoryresponsefromCOVID-19and myocardialpathologyafterrecovery

Greater concentrations of inflammatory biomarkers andthepresenceofSARS-CoV-2RNAinplasma(viremia) areassociatedwithincreaseddiseaseseverityinCOVID- 19.^26-28In ourstudy, we could not identifyany associ-

ationbetweeninflammatory biomarkersorpresenceof SARS-CoV-2viremiaduringacute COVID-19andpatho- logical findings on CMR after recovery. This supports thenotionoflimitedassociationbetweenseverityofthe acuteinfection,reflectedbytheinflammatoryresponse, andpersistentcardiacpathology.COVID-19isknownto causeanoveractiveanddysfunctionalimmuneresponse contributingtodiseaseprogression,andourobservation suggeststhat thedegreeofimmuneactivationdoesnot correlate with persistent cardiac pathology. This is in agreementwithanautopsystudythatfoundpresenceof SARS-CoV-2inthemyocardium tobefrequent, butnot associatedwithinfluxofinflammatorycellsintothemy- ocardiumorlymphocyticmyocarditis.²⁹

Limitations

The study was limited to 49% of survivors from the prospectiveCOVIDMECHstudy.However,patientswho agreedtoparticipateintheCOVIDCMRfollow-upstudy hadcomparable characteristicsand biomarkertrajecto- riescomparedtopatientsnotwillingtoparticipate.Still, wecannotruleoutbiasthatparticipantswerehealthier thannon-participants.Althoughtheclassificationofmi- norandmajorpathology onCMR is clinicallyrelevant, theapplicationof arbitrary cut-offs has obvious limitations. Importantly, our findings were consistent when theCMRvariableswereanalyzedcontinuously.Threepa- tientswerenotassessableforfocalmyocardialscarring.

The multivariable regressionmodels maybe overfitted duetothenumber ofcovariates relativetothenumber ofoutcomes.Moreover,themodestsamplesizeincreases theriskofType2errors,particularlyintheadjustedanal- ysis. Biomarker measurements from the biobank were notavailableinall patients(n= 5), howeverwewere ableto usemeasurementsofcardiovascular biomarkers obtainedinclinicalroutineforthesepatients.Wedidnot have cardiac imaging from thepatients pre-COVID-19, andcanthereforenotwithcertaintydeterminewhether theCMRfindingswerecausedbyCOVID-19orwerepre- existing.Thehealthycontrolgroupwasnotage-matched aspatientswereincludedfromapopulationstudyofpar- ticipantsallbornin1950,andwedidnotincludearisk factormatchedcontrolgroup.

Conclusion

OurfindingsfromCMR6monthsafterCOVID-19con- trastwiththepreviouslyreportedhighprevalenceofmy- ocardialpathologyassessedshorteraftertheacuteinfec- tion.Although we do not have serial CMR to confirm this,itmaysuggestregressionofcardiacpathologyover time.Moreover,wefoundnoassociationbetweenmark- ersof disease severityduringthe indexhospitalization andpathology onCMRafter 6months,suggesting that pre-existingsubclinicalmyocardialdiseasemaybemore

(9)

importantthanCOVID-19fortheobservedCMRpathol- ogy.

Categoryof submission ClinicalInvestigation.

Funding

Dr Myhre is supported by grant number: 2017051 fromtheSouth-EasternNorwayRegionalHealthAuthor- ity.TheCOVIDMECHstudyreceivedassaysformeasur- inginterleukin-6,procalcitonin,ferritin,cardiactroponin T and N-terminal pro-B-type natriuretic free of charge fromRocheDiagnostics.Noextramuralfundingwasdi- rectlyusedtosupportthiswork.

Conﬂict ofinterest

DrMyhre hasservedonadvisoryboards forNovartis and Novo Nordisk, and has received consulting hono- rariafromNovartis,AmGenandNovoNordisk.DrRøsjø hasreceivedpersonalfeesfromNovartisandThermoFis- cher BRAHMS, CardiNor and SpinChipDiagnostics. Dr EinvikhasreceivedresearchsupportfromAstraZeneca andBoehringerIngelheim.DrOmlandhasservedonad- visoryboardsforAbbottDiagnostics,RocheDiagnostics and Bayer andhas received researchsupport fromAb- bottDiagnostics, Novartis,RocheDiagnostics, Singulex and SomaLogic via Akershus University Hospital, and speaker’s orconsulting honorariafrom RocheDiagnos- tics, Siemens Healthineers and CardiNor. All other au- thorsreportnorelevantdisclosures.

Acknowledgments

WearegratefulfortheinvaluablecontributionsbyMy SvenssonMDPhD,RagnhildRøyslandMDPhD,Subaitha Navaruban BSc, Ahmed Meklif MSc, Jannicke Dokken RN,AmylaAbuegRNandLinnBjørnstadHagenRN.We also thank HaldorHusby andtheUnitof Data Analysis atAkershusUniversityHospital,Lørenskog,Norway,for help withclinical data acquisition fromthe dataware- houseatAkershusUniversityHospital.

Supplementary materials

Supplementarymaterialassociatedwiththisarticlecan be found, in the online version, at doi:10.1016/j.ahj.

2021.08.001.

References

1. GuoT,FanY,ChenM,etal.Cardiovascularimplicationsoffatal outcomesofpatientswithCoronavirusDisease2019

(COVID-19).JAMACardiol2020;5:811–18.

2. ZhouF,YuT,DuR,etal.Clinicalcourseandriskfactorsfor mortalityofadultinpatientswithCOVID-19inWuhan,China:a retrospectivecohortstudy.LancetNorthAmEd

2020;395:1054–62.

3.NishigaM,WangDW,HanY,etal.COVID-19and cardiovasculardisease:frombasicmechanismstoclinical perspectives.NatRevCardiol2020;17:543–58.

4.KawakamiR,SakamotoA,KawaiK,etal.Pathologicalevidence forSARS-CoV-2asacauseofmyocarditis.JAmCollCardiol 2021;77:314–25.

5.SandovalY,JanuzziJrJL,JaffeAS.Cardiactroponinfor assessmentofmyocardialinjuryinCOVID-19:JACCreviewtopic oftheweek.JAmCollCardiol2020;76:1244–58.

6.WeckbachLT,CurtaA,BieberS,etal.Myocardialinﬂammation anddysfunctioninCOVID-19-associatedmyocardialinjury.Circ CardiovascImaging2021;14.

7.ShiS,QinM,ShenB,etal.Associationofcardiacinjurywith mortalityinhospitalizedpatientswithCOVID-19inWuhan, China.JAMACardiol2020;5:802–10.

8.OmlandT,PrebensenC,RoyslandR,etal.Established

cardiovascularbiomarkersprovidelimitedprognosticinformation inunselectedpatientshospitalizedWithCOVID-19.Circulation 2020;142:1878–80.

9.Carfì A,BernabeiR,LandiF,ftGAC-P-ACSGroup.Persistent symptomsinpatientsafteracuteCOVID-19.JAMA 2020;324:603–5.

10PuntmannVO,CarerjML,WietersI,etal.Outcomesof cardiovascularmagneticresonanceimaginginpatientsrecently recoveredfromCoronavirusDisease2019(COVID-19).JAMA Cardiol2020;5(11):1265–73.

11RajpalS,TongMS,BorchersJ,etal.Cardiovascularmagnetic resonanceﬁndingsincompetitiveathletesrecoveringfrom COVID-19infection.JAMACardiol2021;6(1):116–18. 12.ClarkDE,ParikhA,DendyJM,etal.COVID-19Myocardial

PathologyEvaluationinathleteswithCardiacMagnetic Resonance(COMPETECMR).Circulation2021;143:609–12. 13StarekovaJ,BluemkeDA,BradhamWS,etal.Evaluationfor

myocarditisincompetitivestudentathletesrecoveringfrom coronavirusdisease2019withcardiacmagneticresonance imaging.JAMACardiol2021;6(8):945–50.

14.MyhrePL,PrebensenC,StrandH,etal.Growthdifferentiation factor-15providesprognosticinformationsuperiortoestablished cardiovascularandinﬂammatorybiomarkersinunselected patientshospitalizedwithCOVID-19.Circulation;0(0).

15RoyalCollegeofPhysicians.Nationalearlywarningscore (NEWS)2:Standardisingtheassessmentofacute-illnessseverity intheNHS.Retrievedfromhttps://www.rcplondon.ac.uk/

projects/outputs/national-early-warning-score-news-2.6August 2020

16.BergeT,VigenT,PervezMO,etal.Heartandbrain

interactions–theAkershusCardiacExamination(ACE)1950study design.ScandCardiovascJ2015;49:308–15.

17.Schulz-MengerJ,BluemkeDA,BremerichJ,etal.Standardized imageinterpretationandpost-processingincardiovascular magneticresonance-2020update:SocietyforCardiovascular MagneticResonance(SCMR):Boardoftrusteestaskforceon standardizedpost-processing.JCardiovascMagnReson 2020;22:19.

18.HeckSL,GulatiG,HoffmannP,etal.Effectofcandesartanand metoprololonmyocardialtissuecompositionduringanthracycline treatment:thePRADAtrial.EurHeartJCardiovascImaging 2018;19:544–52.

19.HuangL,ZhaoP,TangD,etal.Cardiacinvolvementinpatients recoveredfromCOVID-2019identiﬁedusingmagneticresonance imaging.JACCCardiovascImaging2020;13:2330–9.

(10)

20.RamanB,CassarMP,TunnicliffeEM,etal.Medium-termeffects ofSARS-CoV-2infectiononmultiplevitalorgans,exercise capacity,cognition,qualityoflifeandmentalhealth,post-hospital discharge.EClinicalMedicine2021;31.

21KotechaT,KnightDS,RazviY,etal.Patternsofmyocardial injuryinrecoveredtroponin-positiveCOVID-19patientsassessed bycardiovascularmagneticresonance.EuropeanHeartJournal 2021;42(19):1866–78.

22.MonneyPA,SekhriN,BurchellT,etal.Acutemyocarditis presentingasacutecoronarysyndrome:roleofearlycardiac magneticresonanceinitsdiagnosis.Heart2011;97:1312–18. 23.SeligerSL,HongSN,ChristensonRH,etal.High-sensitive

cardiactroponinTasanearlybiochemicalsignatureforclinical andsubclinicalheartfailure:MESA(Multi-EthnicStudyof Atherosclerosis).Circulation2017;135:1494–505. 24.LiuC-Y,HeckbertSR,LaiS,etal.Associationofelevated

NT-proBNPwithmyocardialﬁbrosisintheMulti-EthnicStudyof Atherosclerosis(MESA).JAmCollCardiol2017;70:3102–9.

25.BonnerF,JanzarikN,JacobyC,etal.MyocardialT2mapping revealsage-andsex-relateddifferencesinvolunteers.J CardiovascMagnReson2015;17:9.

26.ZengF,HuangY,GuoY,etal.Associationofinﬂammatory markerswiththeseverityofCOVID-19:ameta-analysis.IntJ InfectDis2020;96:467–74.

27.FajnzylberJ,ReganJ,CoxenK,etal.SARS-CoV-2viralloadis associatedwithincreaseddiseaseseverityandmortality.Nat Commun2020;11:5493.

28PrebensenC,MyhrePL,JonassenC,etal.SARS-CoV-2RNAin plasmaisassociatedwithICUadmissionandmortalityinpatients hospitalizedwithCOVID-19.ClinInfectDis

2021;73(3):e799–802.

29.LindnerD,FitzekA,BräuningerH,etal.Associationofcardiac infectionwithSARS-CoV-2inconﬁrmedCOVID-19autopsycases.

JAMACardiol2020.