Durable left ventricular assist device therapy in advanced heart failure: Patient selection and clinical outcomes

Durable

left

ventricular

assist

device

therapy

in

advanced

heart

failure:

Patient

selection

and

clinical

outcomes

Sachin

P.

Shah

,

Mandeep

R.

Mehra

*

a_{CenterforAdvancedHeartDisease,BrighamandWomen'sHospitalHeartandVascularCenter,Boston,MA,} UnitedStates

b_{HarvardMedicalSchool,Boston,MA,UnitedStates} c

DivisionofCardiovascularMedicine,LaheyHospitalandMedicalCenter,Burlington,MA,UnitedStates d_{TuftsUniversitySchoolofMedicine,Boston,MA,UnitedStates}

The increasing adoption of left ventricular assist devices (LVADs)into clinicalpracticeisrelatedtoacombinationof engineeringadvancesinpumptechnologyandimprovements inunderstandingtheappropriateclinicaluseofthesedevices inthemanagementofpatientswithadvancedheartfailure. This review intends to assist the clinician in identifying candidates for LVAD implantation, to examine long-term outcomesandprovideanoverviewofthecommon complica-tionsrelatedtouseofthesedevices.Intheearly1990s,larger pulsatileLVADs(i.e.NovacorLVADandThoratecHeartMate XVE)wereinitiallyusedforleftventricularsupportinpatients awaitingcardiactransplantation.Thisstrategywasnotbased onrandomizeddatabutwasadoptedoutofnecessity,given the long waiting times for cardiac transplantation.1 _As confidencegrew,theRandomized Evaluationof Mechanical

Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial was launched, which randomized 129 nontransplant candidates with end-stage heart failure to ongoing medical therapy (72% of patients supported with continuousintravenous inotropes)versusapulsatile Heart-MateXVEanddemonstratedadramaticsurvivaladvantageat oneyear(53%survivalintheLVADgroupand25%survivalin themedicaltherapygroup).2_{Thisaffirmedtheproofofconcept} andviabilityoflifetimeordestinationtherapyusing mechan-icalcirculatorysupport(MCS)systems.Theinitialadoptionof pulsatiledeviceswaslowduetolesserdevicedurabilityand frequent morbidity,but as LVAD technology advanced, the adventofsmallerandmorereliablecontinuousflowdevices (HeartMateIILVADandHeartWareHVAD)ledtoadramatic rise inutilizationof LVADsinthepastdecade.In2013,the

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received11January2016 Accepted25January2016 Availableonline8February2016 Keywords: HeartMateXVE REMATCH NovacorLVAD INTERMACS

a

b

s

t

r

a

c

t

Theincreasingadoptionofleftventricularassistdevices(LVADs)intoclinicalpracticeis relatedtoacombinationofengineeringadvancesinpumptechnologyandimprovementsin understandingtheappropriateclinicaluseofthesedevicesinthemanagementofpatients with advanced heart failure.This review intends to assist the clinicianin identifying candidatesforLVADimplantation,toexaminelong-termoutcomesandprovideanoverview ofthecommoncomplicationsrelatedtouseofthesedevices.

#2016CardiologicalSocietyofIndia.PublishedbyElsevier,adivisionofReedElsevier India,Pvt.Ltd.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http:// creativecommons.org/licenses/by-nc-nd/4.0/).

*Correspondingauthor.

E-mailaddress:[email protected](M.R.Mehra).

Available

online

at

www.sciencedirect.com

ScienceDirect

journalhomepage:www.elsevier.com/locate/ihj

http://dx.doi.org/10.1016/j.ihj.2016.01.017

0019-4832/#2016CardiologicalSocietyofIndia.PublishedbyElsevier,adivisionofReedElsevierIndia,Pvt.Ltd.Thisisanopenaccess articleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

numberofdurableMCSdevicesintheUnitedStates(n=2642, 97% continuous flow LVADs, 44% for destination therapy) exceededthenumberofcardiactransplantsperformed.3,4_For thepastdecade,thenumberofcardiactransplantsperannum worldwidehasremainedstagnantataround4000relatedto the relatively fixed donor pool, with the vast majority of cardiactransplantsbeingperformedintheUnitedStatesand Europe. Attempts at increasing the donor pool have been outpacedbytheimprovingclinicaloutcomesexperiencewith currentgenerationLVADscontributingtoagrowing popula-tion of transplant ineligible patients supported with these devicesasdestinationtherapy.Balancingtherisk–benefitratio to match the device to the patient's condition will be paramountinprolongingandimprovinglifewhileachieving cost-effectiveness.5

1.

Patient

selection

1.1. Selectingpatientswiththeappropriateseverityof heartfailure

Determining a level of severity of illness in patients with advanced heart failure relies heavily on the degree of symptoms, refractoriness to traditional disease modifying therapy, and the worsening hemodynamic profile. As ad-vanced-stageheartfailuresetsin,thetraditionalNewYork HeartAssociationclassificationsystemisnolongersufficient tocharacterize patients.Thus, the InteragencyRegistry for Mechanically Assisted Circulatory Support (INTERMACS) profilesystemassignsalevel(INTERMACSlevel1through7) topatientsbasedontheseverityofillness(Table1).Currently, themajorityof patientsundergoingLVAD implantationare eithercategorizedasINTERMACSlevel1(criticalcardiogenic shock),level2(progressivedeclineoninotropictherapy),or level3(stablebutinotropictherapydependent).3_INTERMACS level 1patients (thosein criticalcardiogenic shock)pose a challenge to LVAD implantation. Specifically, there is an increasedriskofperioperativemortality(relativerisk1.55).6In responsetotherealizationoftheperioperativeriskinpatients withcardiogenicshock,theproportionofpatientsundergoing durable MCS implantation at INTERMACS level 1 has

decreasedoverthepastdecade.3,6,7_{Thisincreasedriskisat} leastinpartrelatedtotheend-organdysfunctionassociated withcardiogenicshockandtheinflammatorystateofsevere shock, leadingtoincreasedpostimplant bleeding,infection, multisystemorganfailure,andneedforrightheartsupport. Avoidance of support is not an adequate strategy and management algorithms in INTERMACS I patients have evolved tousetemporaryMCSdevices,suchasintra-aortic ballooncounter-pulsation,percutaneousorsurgical centrifu-gal devices (TandemHeart, Centrimag), percutaneous axial flow devices (Impella),or venoarterialextracorporeal mem-braneoxygenation(VAECMO)inanefforttorestoreend-organ perfusion,stabilizethepatient,andpotentiallyreducetherisk ofsubsequentdurableLVADimplantation.8

In general, patientswith cardiogenic shock who can be stabilizedwithpercutaneoussupportmaybecandidatesfora durableLVAD,whereaspatientswhosufferfromirreversible end-organ dysfunction (renal, hepatic, neurologic, etc.) or refractory shockdespite temporaryMCSare likelyathigher riskandaresuboptimalcandidatesforLVADimplantation.

Patientswhoareinotropictherapydependent(INTERMACS levels 2and 3) currently representnearly two-thirdsof all LVAD implantations and likely also represent the most appropriate use of thecurrent technology. Patients treated withinotropictherapyduetorefractoryend-organ hypoperfu-sionorrefractorysymptomsrelatedtoadvancedheartfailure have an overall very poor prognosis withmedical therapy alone. Ofthe61patientsinthemedicaltherapyarmofthe REMATCHtrial,72%ofpatientswereoncontinuousinotropic infusion, and by one year, only 25% of medically treated patientswerealive,whichdecreasedto8%bytwoyears.2_The Investigation of Non-Transplant Eligible Patients who are Inotrope Dependent (INTrEPID) and Continuous Outpatient Support with Inotropes (COSI) trials are two other small prospectiveanalyses,whichdemonstrateda1-yearsurvivalof 11% and 6%, respectively in patients bound to continuous inotropictherapysupport.9,10 _{Incontrast, theexpected} one-yearsurvivalofpatientsfollowingimplantationofa continu-ous flow LVADnow approaches 80%(Fig.1).6Although the patient populationsinthesetrialsdiffer,thesedatainfera dramaticsurvivaladvantagefordurableMCSinINTERMACS levels2and3patientswithadvancedheartfailure.

Patients in INTERMACS levels 4through 7 suffer from advancedheartfailurebutarenotinotropedependent.Not onlyarepatientswiththisseverityofillnessmoredifficultto define, itispossibleforanindividual patienttotransition between levelsovertime.Currently,only18.5%ofpatients who undergo durable MCS arelevels 4 through 7(mostly INTERMACS level4).6 _{Of these patients,the INTERMACS 4} profileisincreasinglygainingacceptanceasanappropriate candidategroup.Suchpatientsexhibitsymptomsofdyspnea andfatigueonminimalactivity,aretypicallyhousebound due tothe severity of symptoms, and sufferfrom a poor quality of life and excess 1-year mortality. The recently concludedROADMAPtrialprovidesinsightintotheselection ofpatientsforLVADtherapyfromthisgroupofindividuals.11 Further estimation ofprognosisin this‘‘lesssick’’patient population is warranted with the use ofother prognostic indicatorsinchronicheartfailure,manyofwhicharelisted inTable2.12–19

Table1–Currentdistributionofdurablemechanical

circulatorysupportdevicesacrossINTERMACSlevels.

INTERMACS Level

Definition %Ofdurable MCS 1 Criticalcardiogenicshock 14.3%

2 Progressivedecline 36.0%

3 Stablebutinotropedependent 29.6%

4 Restingsymptoms 14.5%

5 Exertion-intolerant 3.0%

6 Exertion-limited 1.2%

7 AdvancedNYHAClass3 0.7% INTERMACS,InteragencyRegistryfor MechanicallyAssisted Cir-culatory Support; MCS, mechanical circulatory support; NYHA, NewYorkHeartAssociation.

MEDAMACSisanongoingregistrycoupledtothe INTER-MACS registry that aims to further characterize medically managedpatientsinINTERMACSlevels4–7.20_Thepilotdata fromthefirst166patientsenrolledinMEDAMACSdemonstrate

a1-yearsurvivalof78%inINTERMACSlevel6/7,a67%1-year survivalinINTERMACSlevel5,anda39%1-year survivalin INTERMACSlevel4.Basedonthislimiteddata,thereisunlikely tobeasignificantsurvivaladvantagetoLVADimplantationin INTERMACSlevel6or7.INTERMACSlevels4–6patientswere analyzed prospectively in the recently reported prospective observational ROADMAP trial, which offered all enrolled patients implantationofacontinuousflowLVAD,and com-paredthosewhounderwentimplantationtothosewhochose medicaltherapy.11_{Ofthe196patientsenrolled,95underwent} implantationofanLVAD.Byintentiontotreat,therewasno difference in mortality (since advancing symptoms in the medicaltherapyarmresultedintransitiontoLVADimplant), butshowedimprovementsinfunctionalcapacity,qualityoflife, anddepressionintheLVADgroup.

PatientsinINTERMACSlevels6or7arelikelytoowellto benefitsubstantially fromdurableMCSandimplantationin thiscohortofpatientswithcurrentgenerationdevicesislikely tobetooearly.AdecisiontoimplantadurableMCSdevicein INTERMACS4patientsshouldbebasedonthepotentialbenefit in functional capacity and quality of life versus the risks (namely stroke, hemorrhage, thrombosis, and infection) associated with durable MCS therapy. An appropriate in-formedconsentprocessthatsimplifiestheeducation regard-ing LVAD therapy and defines these risk-benefit ratios objectivelyisneededtoinvolvethepatientastheprincipal decisionmakerinsuchscenarios.

1.2. Understandingtheroleofheartfailurephenotypein decisionmaking

PriortoconsiderationofLVADimplantation,thelikelihoodof spontaneousrecoveryshouldbeassessed.Patientswho are likelytorecoverandabletobestabilizedonmedicaltherapy should not have a durable mechanical support device implanted, and those who are likely to benefit from valve surgery or revascularization should have these attempted prior to LVAD implantation. There is often significant uncertaintyastowhetherornotrevascularizationorvalvular intervention prior to LVAD will preclude the need for mechanicalsupport;inthosecasesitisprudenttoevaluate a patient fully for candidacy of LVAD therapy prior to the proposed intervention incaseof the needfor a ‘‘bail out’’ strategy.Cardiacresynchronizationtherapy(CRT)in inotrope-dependentpatientshasbeenassociatedwithpooroutcomes, andthereforeimplantingaCRTdeviceinpatientsoninotropic therapymayonlycausedelaysinLVADimplantationrisking furtherend-organdysfunctionordeath.21_{However,evenif1in} 3patientsweretobenefitsubstantially fromCRT,it makes prudentsensetoemploythisstrategyasapreemptivemove, butitisimperativethattheresponsetoCRTbeassessedearly withplannedescalationoftherapyoncenonresponderstatus isconfirmed.

Typically,patientsundergoingMCSarethosewithadilated leftventriclesincecannulapositioningisdifficultinthosewith smaller ventricles. Thosewitharestrictive orhypertrophic cardiomyopathyposeatechnicalchallengerelatedtoinflow cannulaplacementinasmallventricularcavity;however,in thosepatientsinwhomimplantationofanLVADistechnically feasible,perioperativeandone-yearoutcomesdonotappear

Fig.1–One-yearsurvivalofpatientsoncontinuous

inotropicsupportcomparedtothosesupportedwitha

durablecontinuousflowleftventricularassistdevice. One-yearsurvivalwas25%inthemedicallytreatedarmofthe

REMATCHtrial,72%ofwhomweredependentoninotropic

therapy.Survivalwas11%and6%intheprospective

analysesINTrEPIDandCOSI,respectivelyatoneyear.In

comparison,basedonthe7thINTERMACSannualreport,

theone-yearsurvivalofpatientssupportedwithadurable continuousflowleftventricularassistdeviceis

approximately80%.REMATCH,RandomizedEvaluationof

MechanicalAssistancefortheTreatmentofCongestive

HeartFailure;INTrEPID,InvestigationofNon-Transplant EligiblePatientswhoareInotropeDependent;COSI,

ContinuousOutpatientSupportwithInotropes.

Table2–Poorprognosticmarkersinheartfailure.

Symptomsatrest(NYHAclassIV)orwithminimalexertion

Recurrentheartfailurehospitaladmissions

Tachycardia

Sustainedventriculararrhythmia

Hypotension Cardiaccachexia Hyponatremia Renaldysfunction Hepaticdysfunction Anemia

Priororcurrentneedforinotropictherapy

Highdiureticrequirement(furosemideequivalent>160mg/day)

Circulatoryorrenallimitationstoneurohormonalantagonists

Nonrespondertocardiacresynchronizationtherapy

LowerEFandlargeleftventricularvolume

Mitralregurgitation

Pulmonaryhypertensionandrightventriculardysfunction

Six-minutewalkdistance<300m

PeakVO2<14mL/kg/min(notonbetablocker)

or<12mL/kg/min(onbetablocker),or<50%ofpredicted

Ve/VCO2slope>35

NYHA,New York Heart Association; EF,ejection fraction; VO2,

oxygenconsumption;Ve/VCO2,ventilatoryefficiency(Ve,minute

to differ from those without a restrictive or hypertrophic cardiomyopathy.22Patientswitharecentmyocardial infarc-tioninvolvingtheleftventricularapextheoreticallyhavean increasedsurgicalrisk.Certainvalvularpathologiesmayneed tobeaddressedatthetimeofMCSimplantationandshouldbe fully evaluated prior tosurgery. Typically, while supported with a continuous flow LVAD, the aortic valve opens infrequently and may be at increased risk for thrombosis; therefore, mechanical aortic prostheses are often replaced withabioprostheticvalve.Aorticregurgitationoftenworsens following continuous flow LVAD implantation, which decreases pump efficiency by recirculation, imposing an increasedvolumeloadontheleftventricle.Therefore,patients withmoderateor severeaortic regurgitationoftenundergo concurrentclosure,oversewing,orreplacementoftheaortic valve.23_{Mitralregurgitationimprovesinmostofthepatients} followingLVADimplantationwhenitisfunctionallyrelatedto left ventricular dilation and volume, and does not often require surgical intervention; conversely, significant mitral stenosislimitsLVADfillingandrequiresreplacementofthe stenotic mitral valve with a bioprosthesis.24 _{Atrial septal} defectsandpatentforamenovaleshouldbeidentifiedpriorto surgeryandshouldbeclosedatthetimeofLVADimplantation given the decrease in left atrial pressure following LVAD implantationandthepossibilityofpostoperativerighttoleft shunt,facilitatedbyresidualrightventricularfailure. 1.3. Assessmentoftherightventricle

Ongoingrightventriculardysfunctionfollowingisolatedleft ventricularmechanicalsupportmayposeapersistent limita-tion to functional capacityand impair end-organ function. Right ventricular failure following LVAD implantation increasestheriskofdeathandmaynecessitatethe implanta-tionofarightventricularmechanicalsupportdeviceoruseof prolongedcontinuousintravenous inotropictherapy.Inthe current era, right ventricular mechanical support is only available to those patients who are listed for cardiac transplantation (sincedischarge to homeisoftendifficult), therebyplacingprofoundimportanceinidentifyingthoseat riskfor postoperativeright ventricularfailure, especiallyin patientsnoteligibleforcardiactransplantation.

LVADimplantationmaynotonlyunmaskpreexistingright ventriculardysfunction, but severalfactorsassociated with left ventricular MCS may further impact right ventricular performance.Ischemic injuryassociated with cardiopulmo-narybypassmayatleasttemporarilyworsenrightventricular function.Increasedsystemicvenousreturnwithleft ventric-ularmechanicalsupportcausesanincreaseinpreloadtothe right ventricle, which in concert with a reduction of left ventricular pressure due todirect unloading may shiftthe septum toward the left ventricle causing worsening right ventriculardilationandtricuspidregurgitationanddecreased septalcontributiontorightventricularforwardoutput.25

Assessmentofrightventricularfunctionshould incorpo-rate a combination of imaging findings (to assess RV contractility and diastolic compliance) with hemodynamic (toassesssystemicelevationoffillingpressuresand contrac-tileinsufficiency)andbiochemicaldata(toassesshepatorenal end-organ perfusion). Preoperative severe right ventricular

dysfunctionbyqualitativeandquantitativeechocardiographic evaluation, inabilitytodecrease the rightatrialpressureto below 15mmHgwith optimal medical therapy, alow right ventricularstrokeworkindex,lowrightatrialtopulmonary artery pulse pressure ratio, high right atrial to pulmonary artery wedge pressure ratio, and persistent elevations in creatinine,bilirubin,andinternationalnormalizedratio(INR) allindicateanincreasedriskforpostoperativeright ventricu-larfailure.25_{Nosinglefactorhasbeenshowntobeadequatein} thepredictionofrightventricularfailure;therefore,duringthe evaluationofapatientreferredforleftventricularmechanical support, all of these factors should be comprehensively examined.

Inadditiontoassessingrightventricularfunction,special attentionshouldbedirectedtowardthearrhythmiaburden. Significantventriculararrhythmiasmayresultina hemody-namicembarrassmentinpatientswithisolatedleft ventricu-larmechanicalsupport;therefore,strongconsiderationshould begiventobiventricularMCSinpatientswithveryfrequentor refractoryventriculararrhythmia.

1.4. Estimatingtheimpactofcomorbidconditions Data regarding the impact of advanced age on LVAD outcomes are variable.A single center reportedoutcomes of30patientsovertheageof70witha97%30-dayand70% two-yearsurvivalfollowingLVADimplantation,whichwas notdifferentfrompatientsatthesameinstitution undergo-ingLVADimplantationatanagelessthan70yearsold.26This iscontrarytotheseventhINTERMACSregistryreport,which revealed an increased perioperative and late mortality in olderpatients.6

Given the hemorrhagic and thrombotic complications, whichareunfortunatelycommonfollowingLVAD implanta-tion,patientswithhypercoagulabledisorders or coagulopa-thiesshouldbeconsideredatincreasedriskforpostoperative complications.AllpatientswhoundergoLVADimplantation mustbeabletotoleratesystemicanticoagulation,whichatthe currenttimeconsistsofwarfarinwithanINRof2.0–3.0and aspirin(oranalternateantiplateletagent).25

Renal and hepatic dysfunction are associated with an increaseinperioperativemortality,theriskofwhichrelatesto the degreeof impairment. These two end organs typically predict underlying chronicity of the disease, a worse right heart function, and predispose to a greater propensity for bleeding and infection-related complications. Despite this association, on average, both renal and hepatic function improve by six months following LVAD implantation.27 Patients on hemodialysis have an overall poor prognosis andarelikelyatincreasedriskofdeviceinfectionandmost centers consider LVAD implantation in this population unattractive. In patients with marginal renal function, improvement with a trial of inotropic therapy and lack of significant proteinuria may suggest a greater likelihood of renal recoveryfollowingLVADimplantation.Chronic eleva-tioninrightatrialpressureresultinginlongstandinghepatic congestionmayultimatelyleadtohepaticcirrhosis.Patients withchemistryorimagingdataindicatingthepossibilityfor cirrhosis(orevenextensivehepaticfibrosis)shouldundergo liver biopsy, as patients with cirrhosis are suboptimal

candidates for LVAD implantation given an increase in operative bleeding and perioperative mortality, especially thosewhoareChilds-PughclassBorCorhaveanelevated ModelforEnd-StageLiverDisease(MELD)score.28

Thepresenceofperipheralarterialdiseaseatthetimeof LVADimplantationmayincreasetheriskofstroke,andlimb andmesentericischemia;thisdependslargelyontheextentof vasculardisease,andinsomecases,thismaybeconsidereda relative contraindication to durable LVAD implantation. In addition, significant disease of the ascending aorta may contraindicatetheanastomosisoftheoutflowgraft, necessi-tating alternative connections, such as in the descending aorta,insomecases.

Psychosocialbarriersincludinginadequatesocialsupport, substanceabuse,andmedical noncomplianceall shouldbe assessed prior to consideration of LVAD therapy. Poor nutritional status and frailty increase the risk of death followingLVADimplantation,althoughnomethodof improv-ingtheseparameterstomitigatethisriskhasbeen demon-strated.29_{Frailtyisincreasinglyrecognizedasariskmarkerfor} pooroutcomeand shouldbeevaluated independentof the patient'sage.Asimple6-second‘‘shuttle’’testtoseeifthe patient can walk 5meters and assessment of hand grip strength(<20thpercentileforageandgender)canbeusedto establishfrailty.29,30_{Anotherimportantconsiderationisinthe} neuropsychological evaluation of cognitive function; recent studieshave shownthat cognitivefunction,which isoften poorinadvanced heartfailure, improvessignificantlyafter LVADimplantation.However,advancedcognitivedeclinemay beaharbingerforpooradherencetothemedicalregimenand complicateLVADmanagement.31

2.

Outcomes

following

LVAD

implantation

2.1. Survival

BasedondatafromtheINTERMACSregistry,perioperative (30-day)survivalis95%followingLVADimplantation.32Survival followingLVADimplantationhasimprovedasdevice technol-ogyandpatientselectionhaveadvanced.One-year survival followingLVADimplantationfortransplantineligible candi-dates was 53% inthe pulsatile HeartMate XVE arm of the REMATCHtrial;mostrecently,theINTERMACSregistrydata fromtheyears2008to2014demonstrateda1-yearsurvivalof 80% and median survivalof nearly four years for patients supportedwithacontinuousflowLVAD.Thegreatestriskof deathfollowingLVADimplantationisintheearly postopera-tiveperiodandreachesanadirby3monthspostoperatively. Factors that have the greatest impact on perioperative mortality include age, femalesex, prior stroke, mechanical ventilation, INTERMACS level 1 or 2, LVAD for destination therapy, hepatic or renal dysfunction, right ventricular dysfunctionorneedforrightventricularmechanicalsupport, andpriororotherconcurrentcardiacsurgery.3,6

2.2. QualityofLife

TheimpactofLVADsonsymptomburdenandqualityoflifein patientswithsevereheartfailurehasbeenfavorable.6_Inthe

HeartMateIIdestinationtherapytrial,allpatientsexperienced NYHAclassIIIorIVheartfailuresymptomsatthestartofthe trial,andbytheend,80%ofthoseundergoingsupportwitha continuousflowLVADwereNYHAclassIorII.33_Inaddition, patientsinthistrialdemonstratedmeaningfulimprovements intheMinnesotaLivingwithHeartFailurequestionnaireand KansasCityCardiomyopathyquestionnaire,withasignificant increaseina6-minwalkdistanceby12months.Thisshouldbe temperedbytheunderstandingthatcomplicationsfollowing LVADimplantationareoftenunrelatedtoheartfailure,and therefore,heartfailurespecificqualityoflifeassessmentsmay overestimatethebenefitofLVADtherapy.

2.3. ComplicationsfollowingLVAD

Themostcommoncomplicationsinpatientssupportedwith anLVADare bleeding,LVADthrombosis,strokeorsystemic thromboembolism,andinfection.

Followingthe30-dayperioperativeperiod,bleeding(mostly gastrointestinal) occurs at a rate of 8–23% by one year.34 SeveralfactorsplacepatientswithcontinuousflowLVADsat riskforgastrointestinalbleeding.Backgroundantithrombotic therapy with warfarin and an antiplatelet agent (usually aspirin)increasetheoverallbleedingrisk.Shearstressfrom the pump itself may lead to platelet dysfunction and degradationoflargevonWillebrandfactormultimersinthe setting of low pulsatility leads to further coagulopathy.35 Gastrointestinal arteriovenous malformations are common anddifficulttotreatlesionsinpatientswithLVADsandare likely related to the reduced arterial pulsatility related to continuous flow support; this phenomenon is similar to Heyde's syndrome,which was initiallydescribed incalcific aortic stenosis.35 _{The effects of low pulsatility on the} microcirculation, coupled with increased oxidative stress, arethemostlikelymechanisticcandidatesthatpredisposeto developmentofthisuniquegastrointestinalcomplication,in concert with the hematological abnormality encountered universallywithcontinuousflowLVADs.

Pumpthrombosisoccurswithanannualincidenceof6– 12%andisassociatedwithanincreaseinneurologicevents andahigherrateofmortality.36–38_{LVADthrombosisisoften} initially suspected when there is biochemical evidence of hemolysiscausedbyturbulentVADfloworelevationinthe devicepower;later,signsincludeaninabilitytounloadtheleft ventricleasdeterminedbynoninvasiveimaging orinvasive hemodynamic study, decompensated heart failure, and possiblyhemodynamiccompromise.39_{Lactatedehydrogenase} (LDH)hasproventobeanexcellentbiomarkerofhemolysis and hence impending or established pump thrombosis. ElevationofLDHoftenprecedespumpthrombosisbyseveral weeks,and isthereforeoftenmonitored asroutine surveil-lanceforLVADthrombosisinanefforttoinitiateintensified therapy to prevent surgical pumpexchange, although this strategy may be fraught with peril.36 _{From the time of} confirmedpumpthrombosis,thereisatwo-foldincreasein mortalityat30days,90days,and6months.37Patientswho havesuspectedLVADthrombosisanddonotundergoLVAD exchangeorcardiactransplantationhavea6-monthmortality of 48.2%inferringthatmedicaltherapyfor VADthrombosis may be inadequate or cause harm (as in the case of

thrombolytics).Reoperation(pumpexchange)carriesa mod-est6.5%perioperativemortalityriskanda65%2-yearsurvival followingexchange.40

Infectioniscommonandoccursinabout20%ofpatients followingLVADimplantation,whichmaymanifestassepsisor adrivelineinfection.34_{InfectionassociatedwithLVADtherapy} should be treated aggressively and may require long-term suppressiveantibioticsunlessthedeviceisexchangedorthe patientundergoescardiactransplantation.Infection andits inflammatorysequelaepredisposetoaprothromboticmilieu, aswellasheightentheriskofneurologicalcomplications.

Cerebrovascularcomplications,especiallystrokes,remain anAchillesheelofLVADimplants,withanannualincidence exceeding6%.Thiscomplicationiscertainlymorecommon withcertaindevicessuchastheHVAD(HeartWare, Framing-ham,MA)whichexhibitisanexcessivelyhighrateofstrokeat 2years (29%), in women, and inthose with a pre-implant history of stroke or atrialfibrillation. A complex weave of stroke-relatedeventsinthesettingofinfectionorwithpump thrombosis hasbeen observed. Unless this complication is reduced with newer devices or enhanced management strategies, adoption of LVADs to less sicker patients will remainlimitedsince theoccurrenceofadisablingstrokeis tantamounttoalessmeaningfulsurvivalbenefit.41

Inaggregate,byoneyear,80%ofpatientsundergoingLVAD implantation will be alive and on average will have an improvedqualityoflife.Fifty-fivepercentwillbe rehospita-lizedforanycause,30%willhavemajorbleedingwithinthe first month, and 20% will have major bleeding over the following 11 months. Ten percent of patients will have a stroke,5%willhaveadevicemalfunctionrelatedtoclotting, 20%willhaveaseriousinfection,and18%willhaveongoing heartfailure.34_{Thesesummarystatisticsareareasonableway} of developing a well-crafted informed consent process for patientsand their caregivers to helpthem understand the expectationspostimplantation.

3.

Summary

Theapplication of LVAD therapy continues toevolve with technologicaladvancesonthebackgroundofalimiteddonor supplyandagrowingpopulationofpatientswithadvanced heartfailure.TheidealpatientsfordurableLVADimplantation eitherasabridgetotransplantationorasdestination(lifetime) therapy are those requiring continuous inotropic support (INTERMACS 2, 3) or nearing the need for such therapy (INTERMACS 4). This cohort of patients should be stable enoughtoundergoanoperationsafelyandcontinuedmedical therapy is associated with a poor prognosis (<50% 1-year survival). Patients in cardiogenic shock who ultimately undergodurableLVADimplantationhaveamediansurvival ofthreeandahalfyears,andthereforeshouldbeconsidered candidatesfordurablemechanicalsupportideallyfollowinga periodoftemporarymechanicalsupporttoimproveend-organ functionandpotentiallyreduceperioperativerisk.

Despite the success of the contemporary devices in providingcirculatorysupport,thebaneofcomplicationslimits theexpansionofthistherapyandwillneedtobeovercome with future device iterations. The HeartMate III is a new

promising centrifugal pumpcurrently intrial (MOMENTUM III),whichaimstoreducecomplicationsbyaddressingsomeof thepathophysiologicmechanismsthoughttoberesponsible forbleedingandthrombosisassociatedwithcurrent genera-tiondevices.42_{TheHeartMateIIILVADfeaturesanautomated} speedvariabilityat30cyclesperminute,whichproducesflow pulsatility, large blood flow gaps to reduce hemolysis, and texturedsurfaces,whicharemeanttoimprove hemocompat-ibility.Asdevicetechnologycontinuestoadvance,LVADswill inevitablymovebeyondanoptiononlyforselectpatientsand willmigrate towarda broaderpopulationof lesssickheart failurepatients.Pushingtechnologyforwardwilllikelyneedto be coupled witha reduction incomplications and a better understandingofthecost-effectivenessandfinancialimpact ofthistransformativetherapy.

Con

fl

icts

of

interest

Theauthorshavenonetodeclare.

r

e

f

e

r

e

n

c

e

s

1. StewartGC,StevensonLW.Keepingleftventricularassist deviceaccelerationontrack.Circulation.2011;123:1559–1568. 2. RoseEA,GelijnsAC,MoskowitzAJ,etal.Long-termuseofa leftventricularassistdeviceforend-stageheartfailure.N EnglJMed.2001;345:1435–1443.

3. KirklinJK,NaftelDC,PaganiFD,etal.SixthINTERMACS annualreport:a10,000patientdatabase.JHeartLung Transplant.2014;33:555–564.

4. LundLH,EdwardsLB,KucheryavayaAY,etal.Theregistryof theinternationalsocietyforheartandlungtransplantation: thirtiethofficialadulthearttransplantreport-2013;Focus theme:Age.JHeartLungTransplant.2013;32:951–964. 5. RogersJG,BosticRR,TongKB,etal.Cost-effectiveness

analysisofcontinuous-flowleftventricularassistdevicesas destinationtherapy.CircHeartFail.2012;5:10–16.

6. KirklinJK,NaftelDC,PaganiFD,etal.SeventhINTERMACS annualreport:15,000patientsandcounting.JHeartLung Transplant.2015;34:1495–1504.

7. KirklinJK,NaftelDC,StevensonLW,etal.INTERMACS databasefordurabledevicesincirculatorysupport:First annualreport.JHeartLungTransplant.2008;27:1065–1072. 8. RihalCS,NaiduSS,GivertzMM,etal.2015SCAI/ACC/HFSA/

STSClinicalexpertconsensusstatementontheuseof percutaneousmechanicalcirculatorysupportdevicesin cardiovascularcare.JAmCollCardiol.2015;65:e8–e26. 9. RogersJG,ButlerJ,LansmanSL,etal.Chronicmechanical

circulatorysupportforinotrope-dependentheartfailure patientswhoarenottransplantcandidates:resultsofthe INTrEPIDtrial.JAmCollCardiol.2007;50:741–747.

10.HershbergerRE,NaumanD,WalkerTL,DuttonD,Burgess D.Careprocessesandclinicaloutcomesofcontinuous outpatientsupportwithinotropes(COSI)inpatients withrefractoryendstageheartfailure.JCardiacFail. 2003;9:180–187.

11. ClinicalTrials.gov.ROADMAPclinicalstudyprotocol.2015. Availableat:https://clinicaltrisl.gov/ct2/show/NCT01452802. 12. FrancisDP,ShamimW,DaviesLC,etal.Cardiopulmonary

exercisetestingforprognosisinchronicheartfailure: continuousandindependentprognosticvaluefromVe/ VCO2slopeandpeakVO2.EurHeartJ.2000;21:154–161.

13. GodaA,WilliamsP,ManciniD,LundLH.Selectingpatients forhearttransplantation:comparisonoftheHeartFailure SurvivalScore(HFSS)andtheSeattleHeartFailureModel (SHFM).JHeartLungTransplant.2011;30:1236–1243. 14. SetoguchiS,StevensonLW,SchneeweissS.Repeated

hospitalizationspredictmortalityinthecommunity populationwithheartfailure.AmHeartJ.2007;154:260–266. 15. KleinL,O'ConnorCM,LeimbergerJD,etal.Lowerserum

sodiumisassociatedwithincreasedshort-termmortalityin hospitalizedpatientswithworseningheartfailure:results fromtheOutcomesofaProspectiveTrialofIntravenous MilrinoneforExacerbationsofChronicHeartFailure (OPTIME-CHF)study.Circulation.2005;111:2454–2460. 16. KittlesonM,HurwitzS,ShahMR,etal.Developmentof

circulatory-renallimitationstoangiotensin-converting enzymeinhibitorsidentifiespatientswithsevereheart failureandearlymortality.JAmCollCardiol.2003;41: 2029–2035.

17. KetchumES,LevyWC.Establishingprognosisinheart failure:amultimarkerapproach.ProgCardiovascDis. 2011;54:86–96.

18. StewartGC,GivertzMM.Mechanicalcirculatorysupportfor advancedheartfailure.Patientsandtechnologyin evolution.Circulation.2012;125:1304–1315.

19. FangJC,EwaldGA,AllenLA,etal.Advanced(stageD) heartfailure:astatementfromtheHeartFailureSociety ofAmericaGuidelinesCommittee.JCardiacFail. 2015;21:519–534.

20. StewartGC,KittlesonMM,CowgerJA,etal.Whowantsaleft ventricularassistdeviceforambulatoryheartfailure?Early insightsfromtheMEDAMACSscreeningpilot.JHeartLung Transplant.2015;34:1630–1633.

21. ImamuraT,KinugawaK,NittaD,HatanoM,KomuroI. Shouldcardiacresynchronizationtherapybearescue therapyforinotrope-dependentpatientswithadvanced heartfailure?JCardiacFail.2015;21:535–538.

22. TopilskyY,PereiraNL,ShahDK,etal.Leftventricularassist devicetherapyinpatientswithrestrictiveandhypertrophic cardiomyopathy.CircHeartFail.2011;4:266–275.

23. RajagopalK,DaneshmandMA,PatelCB,etal.Natural historyandclinicaleffectofaorticvalveregurgitationafter leftventricularassistdeviceimplantation.JThorac CardiovascSurg.2013;145:1373–1379.

24. KitadaS,KatoTS,ThomasSS,etal.Pre-operative echocardiographicfeaturesassociatedwithpersistent mitralregurgitationafterleftventricularassistdevice implantation.JHeartLungTransplant.2013;23:897–904. 25. SlaughterMS,PaganiFD,RogersJG,etal.Clinical

managementofcontinuous-flowleftventricularassist devicesinadvancedheartfailure.JHeartLungTransplant. 2010;29:S1–S39.

26. AdamsonRM,StahovichM,ChillcottS,etal.Clinical strategiesandoutcomesinadvancedheartfailure patientsolderthan70yearsofagereceivingthe HeartMateIIleftventricularassistdevice.JAmColl Cardiol.2011;57:2487–2495.

27. RussellSD,RogersJG,MilanoCA,etal.Renalandhepatic functionimproveinadvancedheartfailurepatientsduring

continuous-flowsupportwiththeHeartMateIIleft ventricularassistdevice.Circulation.2009;120:2352–2357. 28. MatthewsJC,PaganiFD,HaftJW,KoellingTM,NaftelDC,

AaronsonKD.Modelforend-stageliverdiseasescore predictsleftventricularassistdeviceoperativetransfusion requirements,morbidityandmortality.Circulation. 2010;121:214–220.

29. DunlaySM,ParkSJ,JoyceLD,etal.Frailtyandoutcomes afterimplantationofleftventricularassistdeviceas destinationtherapy.JHeartLungTransplant.2014;33:359–365. 30. DobblesF,MauthnerO,MilisenK.Frailtyinleftventricular assistdevicedestinationtherapy:puttinganewmotorina ricketyoldcarrunningoutofgas?JHeartLungTransplant. 2014;33:347–349.

31. BhatG,YostG,MahoneyE.Cognitivefunctionandleft ventricularassistdeviceimplantation.JHeartLung Transplant.2015;34:1398–1405.

32. KirklinJK,NaftelDC,KormosRL,etal.FifthINTERMACS annualreport:riskfactoranalysisfrommorethan6000 mechanicalcirculatorysupportpatients.JHeartLung Transplant.2013;32:141–156.

33. SlaughterMS,RogersJG,MilanoCA,etal.Advancedheart failuretreatedwithcontinuous-flowleftventricularassist device.NEnglJMed.2009;361:2241–2251.

34. McIlvennanCK,MagidKH,AmbardekarAV,etal.Clinical outcomesaftercontinuous-flowleftventricularassist device.CircHeartFail.2014;7:1003–1013.

35. SuarezJ,PatelCG,FelkerM,etal.Mechanismsofbleeding andapproachtopatientswithaxial-flowleftventricular assistdevices.CircHeartFail.2011;4:779–784.

36.KirklinJK,NaftelDC,KormosRL,etal.Interagency RegistryforMechanicallyAssistedCirculatorySupport (INTERMACS)analysisofpumpthrombosisinthe HeartMateIIleftventricularassistdevice.JHeartLung Transplant.2014;33:12–22.

37. StarlingRC,MoazamiN,SilvestrySC,etal.Unexpected abruptincreaseinleftventricularassistdevicethrombosis. NEnglJMed.2014;370:33–40.

38. NajjarSS,SlaughterMS,PaganiFD,etal.Ananalysisof pumpthrombosiseventsinpatientsintheHeartWare ADVANCEbridgetotransplantandcontinuedaccess protocol.JHeartLungTransplant.2014;33:22–34.

39. MehraMR,StewartGC,UberPA.Thevexingproblemof thrombosisinlong-termmechanicalcirculatorysupport.J HeartLungTransplant.2014;33:1–11.

40.MoazamiN,MilanoCA,JohnR,etal.Pumpreplacementfor leftventricularassistdevicefailurecanbedonesafelyand isassociatedwithlowmortality.AnnThoracSurg.

2013;95:500–505.

41. WileyJZ,DemmerRT,TakayamaH,ColomboPC,LazarRM. Cerebrovasculardiseaseintheeraofleftventricularassist deviceswithcontinuousflow:riskfactors,diagnosisand treatment.JHeartLungTransplant.2014;33:878–887. 42. HeatleyG,SoodP,GoldsteinD,etal.Clinicaltrialdesignand

rationaleofMOMENTUM3:Multi-centerstudyofMagLev technologyinpatientsundergoingMCStherapywith HeartMate3IDEclinicalstudyprotocol.HeartLung Transplant.2016[publishedonlineJanuary29,2016].