in patients with aortic stenosis

19 Pages·2017·7.83 MB·English

RESEARCHARTICLE
Lipoprotein(a) in patients with aortic stenosis:
Insights from cardiovascular magnetic
resonance
VassiliosS.Vassiliou1,2*,PaulD.Flynn3,ClaireE.Raphael1,SimonNewsome1,4,
TinaKhan1,AamirAli1,BrianHalliday1*,AnninaStuderBruengger1,5,TamirMalley1,
PranevSharma1,SubothiniSelvendran1,NikhilAggarwal1,AnitaSri1,HelenBerry6,
JackieDonovan6,WillisLam1,DominiqueAuger1,StuartA.Cook1,7,8,DudleyJ.Pennell1,
SanjayK.Prasad1
a1111111111 1 CMRUnit,RoyalBromptonHospitalandNIHRBiomedicalResearchUnit,RoyalBromptonandHarefield
HospitalsandImperialCollegeLondon,London,UnitedKingdom,2 NorwichMedicalSchool,Universityof
a1111111111
EastAnglia,BobChampionResearch&EducationBuilding,Norwich,UnitedKingdom,3 TheLipidClinic,
a1111111111
Addenbrooke’sHospital,CambridgeUniversityFoundationNHSTrust,UKandUniversityofCambridge,
a1111111111 Cambridge,UnitedKingdom,4 DepartmentofStatistics,LondonSchoolofHygieneandTropicalMedicine,
a1111111111 London,UnitedKingdom,5 ClinicofCardiology,StadtspitalTriemli,Zurich,Switzerland,6 Departmentof
Biochemistry,RoyalBromptonandHarefieldNHSTrust,London,UnitedKingdom,7 DukeNational
UniversityHospital,Singapore,Singapore,8 CardiovascularMagneticResonanceImagingandGenetics,
MRCLondonInstituteofMedicalSciences,ImperialCollegeLondon,HammersmithHospitalCampus,
London,UnitedKingdom
OPENACCESS *[email protected](BH);[email protected](VSV)
Citation:VassiliouVS,FlynnPD,RaphaelCE,
NewsomeS,KhanT,AliA,etal.(2017)Lipoprotein
(a)inpatientswithaorticstenosis:Insightsfrom Abstract
cardiovascularmagneticresonance.PLoSONE12
(7):e0181077.https://doi.org/10.1371/journal.
pone.0181077 Background
Editor:XianwuCheng,NagoyaUniversity,JAPAN Aorticstenosisisthemostcommonage-relatedvalvularpathology.Patientswithaorticste-
Received:November21,2016 nosisandmyocardialfibrosishaveworseoutcomebuttheunderlyingmechanismisunclear.
Lipoprotein(a)isassociatedwithadversecardiovascularriskandiselevatedinpatientswith
Accepted:June26,2017
aorticstenosis.AlthoughmechanisticpathwayscouldlinkLipoprotein(a)withmyocardial
Published:July13,2017
fibrosis,whetherthetwoarerelatedhasnotbeenpreviouslyexplored.Inthisstudy,we
Copyright:©2017Vassiliouetal.Thisisanopen investigatedwhetherelevatedLipoprotein(a)wasassociatedwiththepresenceofmyocar-
accessarticledistributedunderthetermsofthe
dialreplacementfibrosis.
CreativeCommonsAttributionLicense,which
permitsunrestricteduse,distribution,and
reproductioninanymedium,providedtheoriginal Methods
authorandsourcearecredited.
Atotalof110patientswithmild,moderateandsevereaorticstenosiswereassessedbylate
DataAvailabilityStatement:Dataareavailable gadoliniumenhancement(LGE)cardiovascularmagneticresonancetoidentifyfibrosis.
frompublicrepository:https://figshare.com/s/
MannWhitneyUtestswereusedtoassessforevidenceofanassociationbetweenLp(a)
7fdaf66233596705c187.
andthepresenceorabsenceofmyocardialfibrosisandaorticstenosisseverityandcom-
Funding:Theauthorswouldliketoacknowledge
paredtocontrols.Univariableandmultivariablelinearregressionanalysiswereundertaken
financialsupportfromRosetreesCharityTrust
toidentifypossiblepredictorsofLp(a).
(VSV,SKP),London,UK,andtheNIHRBiomedical
ResearchUnit,RoyalBromptonandHarefieldNHS
FoundationTrustandImperialCollegeLondon,UK Results
(TK,DJP,SKP).CRandBHweresupportedbya
Thirty-sixpatients(32.7%)hadnoLGEenhancement,38(34.6%)hadmidwallenhance-
BritishHeartFoundationClinicalResearchTraining
fellowship(FS/14/13/30619andFS/15/29/31492). mentsuggestiveofmidwallfibrosisand36(32.7%)patientshadsubendocardialmyocardial
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 1/19
Lipoprotein(a)inaorticstenosis
SACwassupportedbytheMedicalResearch fibrosis,typicalofinfarction.TheaorticstenosispatientshadhigherLp(a)valuesthancon-
Council,UK. trols,however,therewasnosignificantdifferencebetweentheLp(a)levelinmild,moderate
Competinginterests:ProfessorPennellisa orsevereaorticstenosis.Noassociationwasobservedbetweenmidwallorinfarctionpat-
consultanttoApoPharma,directorof ternfibrosisandLipoprotein(a),inthemild/moderatestenosis(p=0.91)orseverestenosis
CardiovascularImagingSolutions.,London,United
patients(p=0.42).
KingdomandreportspersonalfeesfromSiemens
outsidethesubmittedwork.DrPrasadreports
personalfeesfromBayeroutsidethesubmitted Conclusion
work.ThisdoesnotalterouradherencetoPLOS
ThereisnoevidencetosuggestthathigherLipoprotein(a)leadstoincreasedmyocardial
ONEpoliciesonsharingdataandmaterials.The
otherauthorshavedeclaredthatnocompeting midwallorinfarctionpatternfibrosisinpatientswithaorticstenosis.
interestsexist.
Introduction
Aorticstenosisisthemostcommonage-relatedvalvularpathology.Symptomaticpatientswith
severeaorticstenosishaveapoorprognosisandthereisaneedforidentificationofmarkers
thataremechanisticallyassociatedwithdiseaseprogression.Recently,therehasbeenmuch
interestintheroleofLipoprotein(a)[Lp(a)],alipoproteinsubclassfirstdetectedbyBergin
1963,[1]whosephysiologicalfunctionstillremainselusive.[2]Lp(a)consistsofacholesterol-
richLDLparticlewithonemoleculeofapolipoproteinB100andanadditionalprotein,apoli-
poprotein(a),attachedviaadisulphidebond.[3–5]IncreasedlevelsofLp(a)havebeenassoci-
atedwithincreasedriskofcalcificationoftheaorticvalve,leadingtoaorticstenosis.[6,7]Lp(a)
hasfurtherbeenassociatedwithanincreaseintherateofprogressionofaorticstenosis,and
needforinterventiontorelievethepressureoverload.[8]
VariousmechanismshavebeenproposedasanexplanationfortheassociationbetweenLp
(a)andaorticvalvecalcificationandstenosis.Onepossiblemechanismsuggeststhatafter
transferfromthebloodstreamintothewalloftheaorticvalvecusps,Lp(a)leadstocholesterol
depositioninamannersimilartoLDLcholesterol.Thisissupportedbythesimilarityofthe
structureofLp(a)toLDL,particularlyasLp(a)consistsofalow-densityLDLcholesterol-rich
particleboundcovalentlytoapolipoprotein(a),leadingtothickeningoftheaorticvalvecusps.
[3]AnotherpossiblemechanismrelatestoLp(a)promotingthrombosisbycompetingwith
plasminogenandpreventingplasminfromdissolvingfibrousclots.Thiscouldleadtofibrin
depositionandaorticvalvecalcification.[9]AfurthermechanismsuggeststhatLp(a)maybind
tofibrinanddelivercholesteroltositesoftissueinjury,thuspromotingcalcificationinpatients
withmildaorticstenosis.[10,11]Inaddition,ithasrecentlybeenproposedthatautotaxin
derivedfromLp(a)couldpromoteinflammationandmineralisationpromotingvalvestenosis.
[12]
Aorticstenosisisnotmerelyapathologyofthevalve,butaffectstheleftventricularmyocar-
diumaswell.[13–16]Inarecentstudyonly35%ofpatientswithmoderateorsevereaorticste-
nosishadnormalmyocardiumwhenassessedbycardiovascularmagneticresonance(CMR),
whilst38%hadevidenceofmidwallmyocardialfibrosisand28%hadevidenceofsubendocar-
dialinfarctionpatternfibrosis.Myocardialfibrosis,bothmidwallandinfarctionpattern,isa
strongpredictorofadverseoutcomeinAS.[17]Althoughitisuncertainbywhichmechanism
Lp(a)promotesaorticcalcificationandstenosis,[10]ifanassociationofLp(a)withmyocardial
fibrosisweretobeshownthiscouldhaveclinicalimplicationsaspatientswithfibrosishave
worseoutcome.[17,18]Furthermorethiscouldprovideanexplanationwhysomepatients
developfibrosiswhilstotherswiththesamedegreeofvalvestenosisdonot,andallowusto
betterrisk-stratifypatientsfromtheoutpatientsetting.
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 2/19
Lipoprotein(a)inaorticstenosis
AsLp(a)canaffectmultiplepathwaysatacellularlevelitisuncertainwhatcontribution,if
any,itmighthaveinthedevelopmentofmyocardialfibrosis.OnonehandLp(a)cancompete
withplasminogenforbindingtolysineresiduesonthesurfaceoffibrin,leadingtoareduction
ofplasmingeneration[19]andassociatedfibrinolysis.Thisimpairmentofclotlysiscanthen
leadtoincreasedaccumulationofcholesterol[20]and(micro)thrombosisthusincreasingthe
riskofmyocardialfibrosis.Ontheotherhand,Lp(a)hasbeenshowntodecreasethelevelof
transforminggrowthfactorbeta(TGF-β)[21];afactorpromotingmyocardialfibrosisinaortic
stenosis[15]andotherconditions[22]thereforeleadingtoareducedriskoffibrosis.
ThepotentialassociationofLp(a)withmyocardialfibrosisinpatientswithaorticstenosis
hasnotbeenpreviouslystudied.Inthisstudyweinvestigatedwhethermyocardialfibrosiswas
associatedwithhigherlevelsofLp(a)andcomparedtheLp(a)valuesinthemild/moderateand
severeaorticstenosisgroups.
Methods
Between2011–2013,consecutivepatientswithaorticstenosiswhounderwentCMRwithlate
gadoliniumenhancement(LGE)wereprospectivelyincludedinthissub-studyofCMRusein
cardiomyopathy(ClinicalTrials.govIdentifier:NCT00930735).Thedegreeofseverityofaortic
stenosiswasdefinedaccordingtoAmericanCollegeofCardiology/AmericanHeartAssocia-
tioncriteria.[23]Patientswithclinicalsuspicionorevidenceofcurrentinfectionoracutecoro-
narysyndromewereexcluded.Volunteercontrolswererecruitedfollowinglocaladvertising
andalsounderwentCMR.ThestudywasapprovedbytheRoyalBromptonHospitalInstitu-
tionalReviewBoardandNHSEnglandResearchCommittee,andundertakeninaccordance
withtheethicalstandardsoftheDeclarationofHelsinki.Allpatientsandvolunteersprovided
asignedconsentform.BloodtestswerecollectedonthesamedayastheCMRandanalysedas
onebatchinabiochemistryapprovedlaboratory.
Inourinstitution,CMRisrecommendedroutinelyforallpatientswithsevereaorticsteno-
sisandwheretheclinicalteamrequiresfurtherinformationregardingtheseverityofaorticste-
nosisorleftventricularfunctionoraorticdimensions.Weexcludedpatientswith
disseminatedmalignancy,severeaorticregurgitation,moderateorseveremitralregurgitation/
stenosis,patientswithpreviousvalvereplacementoperations,patientswithcontraindications
toCMR(includingpacemakeranddefibrillatorimplantation)andestimatedglomerularfiltra-
tionrate(Cockcroft-Gaultequation)of<30ml/min.
Datacollection
Demographiccharacteristicsandmedicalhistorywerecollectedfromthepatientaswellas
theirhospitalrecordsorcommunityrecordsonthedayoftheCMR.Allmedicalconditions
andprescribedmedicationwererecorded.Thepresenceofcoronaryarterydiseasewasdefined
aspriorcoronaryrevascularizationorthepresenceofsignificantcoronaryarterystenosisas
assessedbyinvasiveorcomputedtomographycoronaryangiographyby>50%lumendiame-
ternarrowingofavesselof2mmdiameterorgreater.
Cardiovascularmagneticresonance
CMRwasperformedusinga1.5Tscanner(MagnetomSonataorAvanto,Siemens,Erlangen,
Germany)andastandardizedprotocol.Thepatientswerescannedinasupinepositionwith
ananteriorcoilplacedovertheheartandadvancedintothemagnet.Initiallocaliserimages
wereacquiredinthetransaxialplanewithhalf-Fourieracquisitionsingleshortturbospinecho
(HASTE)andfreebreathing.Theseimageswerethenutilisedtoguideacquisitionofavertical
longaxis(VLA)cinewithbalancedsteadystatefreeprecession(SSFP)withbreathholding
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 3/19
Lipoprotein(a)inaorticstenosis
preferablyatendexpiration-asthisismorereproducible.BreathholdSSFPcinesinthe2,3
and4chamberviewswerethentakenusingtheshortaxisscoutandVLAimages.Four-cham-
berand2-chambercineimagesatenddiastolewerethenusedtoplanastackofshort-axis
SSFPcineimages,fromtheleveloftheAVgrooveandperpendiculartotheleftventricular
longaxis.Subsequently,10mmcontiguousshortaxissliceswereacquired(7mmthickness,
3mmgap)frombasetoapex.RetrospectiveECGgatingwaspredominantlyutilisedforthe
cineacquisition.However,prospectivetriggeringwasusedinpatientswitharrhythmia,e.g.
atrialfibrillation.ThesequenceparametersfortheSSFPcineswereTE1.6ms,TR3.2ms,in
planepixelsize2.1x1.3mmandflipangle60˚.AorticvalveplanimetryandLVvolumeand
masswerecalculatedfromSSFPsequencesaspreviouslydescribedbyourgroup.[17]Inthe
aorticstenosispatientstenminutesafterinjectionof0.1mmol/kgofgadoliniumcontrast
agent(Gadovist,ScheringAG,Berlin,Germany)followedby10mlsalineflushtoensurecom-
pletedelivery,inversionrecovery–preparedspoiledgradientechoimageswereacquiredin
standardlong-andshort-axisviewstodetectareasofLGEasdescribedforaorticstenosis
patientspreviously[17][24].Inversiontimeswereoptimizedtonullnormalmyocardiumwith
imagesrepeatedintwoseparatephase-encodingdirectionstoexcludeartifact.
Imageanalysis
ForquantificationofLVfunction,volumes,massandaorticvalveseverityassessmentadedi-
catedsoftwarewasused(CMRTools,www.cmrtools.com,CardiovascularImagingSolutions.,
London,UnitedKingdom)andforquantificationofmyocardialfibrosisaseparatededicated
softwarewasused(CVI42,www.circlecvi.com,CircleCardiovascularImaging,Calgary,
Canada).
InCMRTools theendocardialandepicardialcontoursweresemi-automaticallyappliedin
end-diastoleandend-systoleandthediastolicLVmasswascalculatedfromthetotalend-dia-
stolicmyocardialvolumemultipliedbythespecificdensityofthemyocardium,aspreviously
described[17].TheseverityofaorticstenosiswasassessedusingCMR-derivedplanimetryof
theaorticvalvearea.Thistechniquehasbeenvalidatedagainstechocardiographicmeasuresof
aorticstenosisseverity.[24]TheaorticstenosiswasgradedusingtheCMRaorticvalvearea
(AVA)asfollows:mild,>1.5to2.5cm2;moderate,1.5to1.0cm2;andsevere,<1.0cm2in
accordancewiththeAmericanCollegeofCardiology/AmericanHeartAssociationguidelines.
[23]Forthefinalclassificationofstenosisseverityforourcohortthismethodwasused.
ThepresenceandpatternofLGEwereassessedbytwoindependentexpertobservers
(SCMR/EuroCMRLevelIII)tocategoriseeachpatientaccordingtothevisualpresenceor
absenceofmyocardialfibrosis,andifpresentwhetherthiswasmidwallfibrosisorinfarction
patternfibrosiswithexamplesshowninFig1.Bothobserverswereblindedtoclinicaldata.A
thirdblindedobserveradjudicatedwhentherewasadisparitybetweentheinitialtwoobservers.
PatientswithamixedpatternofLGEwerecategorizedaccordingtothepredominantpattern
offibrosis.Theanonymisedimagesofthepatientswhohadfibrosiswerethenquantifiedusing
CVI42withtheestablished“fullwithhalfmaximum”[17]techniqueandpresentedastheper-
centageofenhancedmassinthelatephasefollowinggadoliniumadministration(LGEmass)
dividedbythetotalLVmassgiving%LGEmass(LGEmass/totalmass)asshownin(Fig2).
Lipoprotein(a)
Lp(a)wasmeasuredusingSentinelDiagnosticsLp(a)Ultra,anisoformindependentlatex
immunoassaydevelopedforLp(a)levels.Whenanantigen-antibodyreactionoccurred
betweenLp(a)inasampleandanti-Lp(a)antibody,thisresultedinagglutinationdetectedas
anabsorbancechange,withthemagnitudeofthechangebeingproportionaltothequantityof
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 4/19
Lipoprotein(a)inaorticstenosis
Fig1. Thetoppanels(A,B,C)representgraphicalsketchesofamid-ventricularshortaxisslicethroughthe
myocardiumusinganinversionrecoverysequence.Thebottompanels(D,E,F)showthecorresponding
imagesobtainedwithCMR.PanelsAandDshownormalmyocardiumwithnoevidenceoffibrosis
(homogeneouslyblackfollowinggadoliniumadministration),panelsBandEshowinfarctionpatternfibrosis
(subendocardialwhiteenhancementfollowinggadoliniumadministration)andpanelsCandFshowmidwall
fibrosis(midwallenhancementfollowinggadoliniumadministrationwithnormal(black)myocardiumboth
towardstheepicardiumandendocardium).
https://doi.org/10.1371/journal.pone.0181077.g001
Lp(a)containedinthesample.Thisanalysiswasundertakenonserumfromourpatientstaken
onthedayofCMRandstoredinadedicatedspaceinabiobankfreezerat-80˚Cuntiltheday
ofanalysis.
Statisticalanalysis
Baselinepatientcharacteristicsarepresentedasmeanandstandarddeviationforcontinuous
variablesandnumber(percentage)forcategoricalvariables.Themildandmoderategroups
weremergedintoonegrouptoincreasegroupnumbersanddirectlycomparedwiththesevere
group.MannWhitneyUtestswereusedtoassesswhethertherewasevidenceofanassociation
betweenLp(a)andaorticstenosisseverity(mild/moderateorsevere),andalsobetweenLp(a)
andpresenceorabsenceofmyocardialfibrosis.Finally,univariableandmultivariablelinear
regressionanalysiswereundertakentoidentifypossiblepredictorsofLp(a).Apvalueof
<0.05wastakenassignificant.AllanalyseswereundertakenusingStata14.0(CollegeStation,
Texas,USA).
Results
Intotal,110patientswithmild/moderateorsevereaorticstenosiswererecruitedandcom-
pletedCMRexaminationand55controlvolunteers.Patientbaselinecharacteristicsareshown
inTable1.ThebaselinepharmacotherapyisshowninTable2.
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 5/19
Lipoprotein(a)inaorticstenosis
Fig2.Exampledemonstratingthequantificationoftheleftventricularmyocardium.PanelAshowsthevisuallategadolinium
enhancementwhilstpanelBshowsthequantifiedenhancedmass.Oncecompletedforallthemyocardialslicesthentheoverallabsolute
enhancedmassor%masscanbecalculated.
https://doi.org/10.1371/journal.pone.0181077.g002
CMRassessmentofmyocardialfibrosis
Ofthecohort,36patients(32.7%)didnotshowanyLGEindicatingthattherewasnomacro-
scopicmyocardialreplacementfibrosis.Atotalof38(34.6%)patientshadmidwallenhance-
mentsuggestiveofmidwallfibrosisand36(32.7%)patientsshowedevidenceof
subendocardialmyocardialfibrosis,apatterntypicalformyocardialinfarction.CMRand
importantbiochemicaldataareshowninTable3.
Lipoprotein(a)level
ThecontrolshadalowermedianLp(a)valuedcomparedtothewholecohortofaorticstenosis
patients(100mg/L(41–266)vs309mg/L(75–688),p<0.001asshowninFig3).
Evenwhencomparedtothemild/moderateandsevereaorticstenosisgroupseparately
therewasstillasignificantdifferenceasshowninFig4.Linearregressionadjustedforageand
sexalsoconfirmedthatcontrolshadlowerLp(a)valuescomparedtothepatientswithmild/
moderate(p=0.013)orsevereaorticstenosis(p=0.019).
Furthermore,therewasnosignificantdifferencebetweentheLp(a)levelseeninthemild,
moderateandsevereaorticstenosisgroup,values541(91–1043),368(94–619)and242(72–
700)respectively(Fig5).
TheconcentrationofLp(a)seeninmild/moderateaorticstenosis(AVA=1.0–2.5cm2)and
severeaorticstenosis(AVA<1.0cm2)wascomparedusingtheMann-WhitneyUtest.The
medianvalueforthemild/moderateaorticstenosisgroupwas384mg/L(91–656)andforthe
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 6/19
Lipoprotein(a)inaorticstenosis
Table1. Patientandcontroldemographiccharacteristics.
Demographics Mild/Moderate(N=35) Severe(N=75) P-Value
Age,years 71±10 78±9 <0.001
Male,n(%) 26(74.3) 51(68.0) 0.66
Hypertension,n(%) 13(40.6) 44(58.7) 0.096
Diabetesmellitus,n(%) 1(3.8) 4(6.1) 1.000
Anycoronaryarterydisease,n(%) 13(37.1) 27(36.0) 1.00
Previousstroke,n(%) 1(2.9) 2(2.7) 1.00
AtrialFibrillation,n(%) 5(14.3) 6(8.0) 0.32
Hypercholesterolaemia,n(%) 18(58.1) 50(67.6) 0.38
NYHA(cid:21)II 19(59.4) 60(81.1) 0.028
Demographics AllAorticStenosis(N=110) Controls(N=55) P-Value
Age,years 76±10 74±7 0.052
Male,n(%) 77(70.0) 39(70.9) 1.00
Hypertension,n(%) 57(53.3) 20(36.4) 0.047
Diabetesmellitus,n(%) 5(5.4) 10(18.2) 0.022
Anycoronaryarterydisease,n(%) 40(36.4) 26(47.3) 0.18
Previousstroke,n(%) 3(2.7) 1(1.8) 1.00
AtrialFibrillation,n(%) 11(10.9) 2(3.6) 0.14
Hypercholesterolaemia,n(%) 68(64.8) 27(49.1) 0.064
NYHA(cid:21)II 79(71.8) 6(11.5) <0.0001
Toppanelcomparisonbetweenmild/moderateandseverepatientswithaorticstenosis.Bottompanelcomparisonbetweenallpatientswithaorticstenosis
andcontrols.NYHA=NewYorkHeartAssociationclassification.
https://doi.org/10.1371/journal.pone.0181077.t001
Table2. Baselinepharmacotherapyofpatientsandcontrolsatthetimeofinclusioninthestudy.
Medicaltherapy Mild/Moderate Severe P-Value
Aspirin,n(%) 19(61.3) 44(59.5) 1.00
Clopidogrel,n(%) 4(13.8) 12(16.7) 1.00
ACEI/ARB,n(%) 13(43.3) 38(51.4) 0.52
BetaBlocker,n(%) 14(46.7) 32(44.4) 1.00
Calciumchannelblocker,n(%) 8(26.7) 6(8.7) 0.028
Diuretic,n(%) 14(43.8) 43(58.1) 0.21
Warfarin,n(%) 4(14.3) 6(8.3) 0.46
Amiodarone,n(%) 0(0.0) 4(5.6) 0.32
Statin,n(%) 20(66.7) 54(72.0) 0.64
Medicaltherapy AllAorticStenosis Controls P-Value
Aspirin,n(%) 63(60.0) 26(47.3) 0.14
Clopidogrel,n(%) 16(15.8) 25(45.5) <0.001
ACEI/ARB,n(%) 51(49.0) 24(43.6) 0.62
BetaBlocker,n(%) 46(45.1) 20(36.4) 0.31
Calciumchannelblocker,n(%) 14(14.1) 7(12.7) 1.00
Diuretic,n(%) 57(53.8) 3(5.5) <0.0001
Warfarin,n(%) 10(10.0) 0(0.0) 0.015
Amiodarone,n(%) 4(4.0) 0(0.0) 0.30
Statin,n(%) 74(70.5) 27(49.1) 0.010
Toppanelcomparisonbetweenpatientswithmild/moderateandsevereaorticstenosis.Bottompanelcomparisonbetweenallpatientswithaorticstenosis
andcontrols.ACEI=Angiotensinconvertingenzymeinhibitor,ARB=AngiotensinIIblocker
https://doi.org/10.1371/journal.pone.0181077.t002
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 7/19
Lipoprotein(a)inaorticstenosis
Table3. PatientbiochemicalandCMRcharacteristicsperaorticstenosisseveritygroup.
BiochemicalandCMRdata Mild/Moderate Severe P-Value
Lp(a),mg/L 420±344 404±390 0.64
Creatinine,μmol/L 93±30 102±36 0.20
CMRaorticvalvearea,cm2 1.2±0.3 0.7±0.1 <0.00001
LVEF,% 62±14 57±17 0.11
LVMass,g 166±47 168±57 0.87
CMRMyocardialTissueCharacterisation
NoMyocardialFibrosis,n(%) 13(37.1) 23(30.7) 0.82
MidwallFibrosis,n(%) 11(31.4) 27(36.0)
InfarctionPatternFibrosis,n(%) 11(31.4) 25(33.3)
Lp(a)byCMRFibrosisGroup
NoMyocardialFibrosis,mg/L 377±416 418±406 0.93
MidwallFibrosis,mg/L 421±333 360±389 0.46
InfarctionPatternfibrosis,mg/L 469±278 438±389 0.74
https://doi.org/10.1371/journal.pone.0181077.t003
severegroupwas242mg/L(72–700).Therewasnosignificantdifferencebetweenaorticsteno-
sisseverity(mild/moderatevssevere)andlevelofLp(a),p=0.64(Fig6).Evenwhenthemean
Lp(a)valueswerecomparedthisdidnotshowanystatisticaldifferencebetweenthemild/
moderategroupandsevereaorticstenosisgroups(420±344vs.404±390,p=0.84).
Althoughwedidnotexpectstatinusetobeaconfounder,asstatinsdonotappeartoaffect
Lp(a)especiallyinnon-FamilialHypercholesterolemiapopulations[25]thiswasfurtherinves-
tigated.MedianLp(a)forthepatientsnottakingstatinvs.patientsonstatinswasnotdifferent
(321mg/L(63–582)vs324mg/L(97–732),p=0.25.Moreover,theeffectofseverityofaortic
stenosisonLp(a)wasassessedusingmultivariableregressionincludingstatinuse,age,sex,
coronaryarterydiseaseandpresenceoffibrosiswhichfailedtoshowanyassociation
(p=0.78).
WefurtherevaluatedwhetherLp(a)wasassociatedwithmidwallorinfarctionpattern
fibrosis.Astherewasnodifferencebetweenthemild/moderateandsevereaorticstenosis
groupsandLp(a)leveltheseweremergedforsubsequentanalysis.Noassociationbetweenthe
presenceandabsenceoffibrosisandLp(a)wasidentified(Fig7).Similarly,therewasnoasso-
ciationbetweenanincreaseintheenhancedabsolutemassor%enhancedmass(definedby
enhancedmass/overallmass)asshowninFig8.
WealsoinvestigatedtheprognosticroleofLp(a)inpatientsdevelopingpost-operative
LBBBorrequiringapacemakerfollowingsurgicalaorticvalvereplacement(AVR)orpercuta-
neoustranscatheteraorticvalvereplacement(TAVI).AsshowninTable4wefoundnosuch
association.
Furthermore,weevaluatedassociationsbetweenotherpotentialadversepredictorsinAS
withLp(a)value.UnivariablelinearanalysisperLp(a)100mg/L,wasundertakenbetween
patientsinthemidwallfibrosisvs.nofibrosis,infarctionpatternfibrosisvs.nofibrosisand
anyfibrosis(midwallorinfarction)vs.nofibrosis.NoassociationwasfoundbetweenLp(a)
andeitherfibrosispattern(midwallp=0.77;infractionpatternp=0.62,anyfibrosisp=0.91).
TherewasnocorrelationbetweenLp(a)andanyotherparametersincludingleftventricular
ejectionfraction(Spearmancorrelation0.14,p=0.14),leftventricularhypertrophy(Mann-
WhitneyUTestp=0.22),leftventricularmass(correlation0.04,p=0.68),gender(female
median=577,(IQR111–741),men172,(72–558),Mann-WhitneyUtestp=0.10);age(Spear-
mancorrelation0.03,p=0.72);aorticvalvearea(Spearmancorrelation0.09,p=0.35),
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 8/19
Lipoprotein(a)inaorticstenosis
Fig3.Boxplotscomparingcontrolsvs.thewholecohortofaorticstenosispatientsindicatingthat
thecontrolshadsignificantlylowerLp(a),median100mg/Lvs309mg/L,p<0.001.
https://doi.org/10.1371/journal.pone.0181077.g003
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 9/19
Lipoprotein(a)inaorticstenosis
Fig4.Boxplotscomparingthecontrolsvsthemild/moderateaorticandsevereaorticstenosis
patientsconfirmingasignificantdifferencebetweenthecontrolsandeitherofthegroups.
https://doi.org/10.1371/journal.pone.0181077.g004
PLOSONE|https://doi.org/10.1371/journal.pone.0181077 July13,2017 10/19