J A C C : H E A R T F A I L U R E VO L . 6 , N O . 9 , 2 0 1 8

ª 2 0 1 8 T H E A U T H O R S . P U B L I S H E D B Y E L S E V I E R O N B E H A L F O F T H E AM E R I C A N

C O L L E G E O F C A R D I O L O G Y F O U N DA T I O N . T H I S I S A N O P E N A C C E S S A R T I C L E U N D E R

T H E C C B Y - N C - N D L I C E N S E ( h t t p : / / c r e a t i v e c o mm o n s . o r g / l i c e n s e s / b y - n c - n d / 4 . 0 / ) .

HVAD: The ENDURANCESupplemental Trial

Carmelo A. Milano, MD,a Joseph G. Rogers, MD,a Antone J. Tatooles, MD,b Geetha Bhat, MD, PHD,b

Mark S. Slaughter, MD,c Emma J. Birks, MB, BS, PHD,c Nahush A. Mokadam, MD,d Claudius Mahr, MD,d

Jeffrey S. Miller, MD,e David W. Markham, MD, MSC,e Valluvan Jeevanandam, MD,f Nir Uriel, MD, MSC,f

Keith D. Aaronson, MD, MS,g Thomas A. Vassiliades, MD,h Francis D. Pagani, MD, PHD,g

for the ENDURANCE Investigators

ABSTRACT

ISS

Fro

Tra

De

Ca

of

va

Me

He

acc

rel

Ma

OBJECTIVES The aim of this study was to prospectively evaluate the impact of blood pressure management on stroke

rates in patients receiving the HeartWare HVAD System.

BACKGROUND The ENDURANCE trial demonstrated noninferiority of the HeartWare HVAD System versus control

(HeartMate II) in patients with advanced heart failure ineligible for heart transplantation. However, stroke was more

common in HVAD subjects. Post hoc analyses demonstrated increased mean arterial blood pressure as a significant in-

dependent risk factor for stroke.

METHODS The ENDURANCE Supplemental Trial was a prospective, multicenter evaluation of 465 patients with

advanced heart failure ineligible for transplantation, randomized 2:1 to HVAD (n ¼ 308) or control (n ¼ 157). The primary

endpoint was the 12-month incidence of transient ischemic attack or stroke with residual deficit 24 weeks post-event.

Secondary endpoints included the composite of freedom from death, disabling stroke, and need for device replacement

or urgent transplantation, as well as comparisons of stroke or transient ischemic attack rates in HVAD cohorts in

ENDURANCE Supplemental and ENDURANCE.

RESULTS The enhanced blood pressure protocol significantly reduced mean arterial blood pressure. The primary

endpoint was not achieved (14.7% with HVAD vs. 12.1% with control, noninferiority [margin 6%] p ¼ 0.14). However, the

secondary composite endpoint demonstrated superiority of HVAD (76.1%) versus control (66.9%) (p ¼ 0.04). The

incidence of stroke in HVAD subjects was reduced 24.2% in ENDURANCE Supplemental compared with ENDURANCE

(p ¼ 0.10), and hemorrhagic cerebrovascular accident was reduced by 50.5% (p ¼ 0.02).

CONCLUSIONS The ENDURANCE Supplemental Trial failed to demonstrate noninferiority of HVAD versus control

regarding the pre-specified primary endpoint. However, the trial confirmed that BP management is associated

with reduced stroke rates in HVAD subjects. HVAD subjects, relative to control subjects, more commonly achieved

the composite endpoint (freedom from death, disabling stroke, and device replacement or urgent transplantation).

(A Clinical Trial to Evaluate the HeartWare� Ventricular Assist System [ENDURANCE SUPPLEMENTAL TRIAL]

[DT2]; NCT01966458) (J Am Coll Cardiol HF 2018;6:792–802) © 2018 The Authors. Published by Elsevier on behalf

of the American College of Cardiology Foundation. This is an open access article under the CC BY-NC-ND license

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

N 2213-1779 https://doi.org/10.1016/j.jchf.2018.05.012

m the aDepartments of Surgery and Medicine, Duke University School of Medicine, Durham, North Carolina; bCenter for Heart

nsplant and Assist Devices, Advocate Christ Medical Center, Oak Lawn, Illinois; cCardiovascular and Thoracic Surgery and

partment of Cardiovascular Medicine, University of Louisville, Louisville, Kentucky; dDivisions of Cardiothoracic Surgery and

rdiology, University of Washington, Seattle, Washington; eDepartments of Cardiac Surgery and Medicine, St. Joseph’s Hospital

Atlanta, Atlanta, Georgia; fHeart and Vascular Center, University of Chicago Medicine, Chicago, Illinois; gDivisions of Cardio-

scular Medicine and Cardiothoracic Surgery, University of Michigan, Ann Arbor, Michigan; and the hDepartment of Clinical and

dical Affairs, Medtronic (formerly HeartWare), Framingham, Massachusetts. This study was supported by Medtronic (formerly

artWare). HeartWare (now Medtronic) sponsored this clinical trial and partnered with us in this analysis, but the authors had

ess to all the data and provided critical review, writing, and content control. The authors have reported that they have no

ationships relevant to the contents of this paper to disclose. John R. Teerlink, MD, served as Guest Editor for this paper.

nuscript received November 15, 2017; revised manuscript received April 25, 2018, accepted May 3, 2018.

AB BR E V I A T I O N S

AND ACRONYM S

HCVA = hemorrhagic

cerebrovascular accident

HR = hazard ratio

IBPM = improved blood

pressure management

LVAD = left ventricular assist

device

MAP = mean arterial pressure

mRS = modified Rankin scale

NYHA = New York Heart

Association

TIA = transient ischemic attack

J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8 Milano et al.S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2 HVAD: The ENDURANCE Supplemental Trial

793

H eart failure is one of the leading causes ofdeath in the developed world and is char-acterized by limitations in survival despite

treatment with guideline-directed medical therapies(1). Projections indicate that the prevalence of heartfailure will increase by 46% from 2012 to 2030 (2).Patients with heart failure progress to advancedstages and require cardiac transplantation or implan-tation of durable left ventricular assist devices(LVADs) to extend survival and improve quality oflife and functional status (1). As the number of suit-able heart donors and patient comorbidities place re-strictions on the feasibility of cardiac transplantation,implantation of durable LVADs has emerged as themost frequently applied surgical treatment for end-stage heart failure as either a bridge to transplanta-tion or permanent therapy (i.e., destination therapy)(3,4). Considerable data have documented improve-ments in survival, functional status and quality oflife offered by destination therapy (5–7).

We previously reported the outcomes of theENDURANCE trial, a prospective, multicenter, ran-domized controlled trial evaluating the use of theHVAD LVAD (HeartWare, Miami Lakes, Florida)compared with control (HeartMate II, a U.S. Food andDrug Administration–approved LVAD for destinationtherapy, Abbott, Abbott Park, Illinois) for destinationtherapy in 445 patients with advanced heartfailure not eligible for cardiac transplantation (8).ENDURANCE demonstrated noninferiority of theHVAD compared with control in survival at 2 yearsfree from disabling stroke and alive on the originallyimplanted device. Although the rate of deviceexchange was lower with the HVAD, the rate of strokefor the HVAD cohort was significantly greater than therate for the control device, with the greatest differ-ence seen in the rate of hemorrhagic events. Retro-spective multivariable analyses of the data from theHVAD pivotal trial for bridge-to-transplantationindication (9,10) determined that elevated bloodpressure was a highly significant, independent riskfactor for stroke in patients on HVAD support (11).This observation was corroborated in a post hocmultivariable analysis of data from ENDURANCE(completed by the sponsor but not yet published). Itwas also observed in ENDURANCE that most strokesoccurred early, in the first 6 months post-implantation. The ENDURANCE Supplemental Trialwas designed to prospectively determine effective-ness of a blood pressure management strategy toreduce neurological injury in patients receiving theHVAD System. The data from the ENDURANCE trialand the Supplemental Trial data presented here led toFood and Drug Administration approval in September

2017 for the HVAD System as destinationtherapy in patients with advanced heartfailure.

METHODS

The ENDURANCE Supplemental Trial was aprospective, randomized controlled, un-blinded, multicenter trial in patients withchronic American Heart Association stage D/New York Heart Association (NYHA) func-tional class IIIB or IV heart failure who hadoptimal medical management was unsuc-cessful and who were deemed ineligible fortransplantation. A total of 465 subjects

enrolled at 47 centers were randomly assigned in a 2:1ratio to receive either the study device (HeartWareHVAD System) (12) or control (HeartMate II). Afterimplantation, device performance, laboratory data,and medications were recorded until hospitaldischarge and at follow-up visits scheduled at 3, 6,and 12 months. Functional capacity and quality-of-life measurements were performed at 3, 6, and 12months. All HVAD subjects received oral anti-coagulation with a target international normalizedratio of 2.0 to 3.0 and antiplatelet therapy (recom-mended starting dose of aspirin 325 mg/day). Man-agement of patients who received the control devicewas at the discretion of their providers and by thedevice-specific instructions for use.

The primary endpoint was the incidence ofneurological injury (defined as any stroke with amodified Rankin scale [mRS] score >0 at 24 weekspost-stroke or any transient ischemic attack [TIA] orspinal cord infarct) at 12 months, including only timeon the originally implanted LVAD. Strokes with mRSscores of 0 at 24 weeks (n ¼ 13 for HVAD, n ¼ 5 forcontrol) were not included in the primary endpoint ofneurological injury. The composite secondary effi-cacy endpoint was freedom from death, disablingstroke (mRS score $4), and device malfunctions orfailures requiring exchange, explantation, or urgenttransplantation over 1 year from implantation,including only time on the originally implantedLVAD. Additional endpoints also included rates foradverse events, which were classified according tothe Interagency Registry for Mechanically AssistedCirculatory Support definitions and were adjudicatedby an independent clinical events committee (13).Additionally, a pre-specified secondary endpointanalysis called for a comparison of the stroke or TIArate for the HVAD subjects in the ENDURANCE Sup-plemental Trial against a performance goal of 17.7%,which was based on the lower bound of the

Milano et al. J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8

HVAD: The ENDURANCE Supplemental Trial S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2

794

confidence interval in the sintered cohort of theENDURANCE trial.

A post hoc hierarchical analysis of all randomizedand implanted subjects in the study was performed aspart of the secondary composite endpoint analysis.Transplantation, recovery, and device exchange weretreated as terminal events, meaning that any neuro-logical event (stroke or TIA) or death that occurredafter those events was not counted, and death wasalways counted as the failing event in the binary-components analysis (i.e., subjects were only coun-ted once as a failure, and death was considered theworst-case failure event in subjects with multipleevents). For subjects with multiple events other thandeath, only the first event was counted.

Primary and secondary endpoints as well asadverse events were adjudicated by an independentclinical events committee. A data and safety moni-toring board monitored and reviewed study compli-ance, adverse events, and outcomes. The study wasconducted in compliance with Food and DrugAdministration regulations for good clinical practiceand approved by each clinical site’s Institutional Re-view Board. All subjects or their authorized repre-sentatives provided informed consent.

Changes in quality of life were measured using theKansas City Cardiomyopathy Questionnaire, adisease-specific, 23-item, self-administered instru-ment that quantifies physical function, symptoms(frequency, severity, and recent change), socialfunction, self-efficacy, knowledge, and quality of life,and the EuroQoL EQ-5D-5L, an assessment of generalwell-being.

Changes in functional status were measured usingNYHA classification and the 6-min walk test. NYHAclass was assessed by a qualified person at the clinicalsite not directly involved with this clinical trial. The6-min walk test was conducted by the site staffaccording to the American Thoracic Society’s guide-lines for the 6-min walk test (14).

BLOOD PRESSURE MANAGEMENT PROTOCOL.

Following discharge from the index hospitalization,all HVAD subjects were required to adhere to a bloodpressure management protocol in addition toreceiving standard of care. Prior to discharge, subjectsand caregivers were trained to measure blood pres-sures and record values in a diary. Subjects with apalpable pulse used a Terumo Elemano (modelESH5503, Terumo, Somerset, New Jersey) automatedblood pressure cuff. Subjects without a palpable pulsewere trained on the use of a Doppler and manual cuffto obtain an opening pressure. In the latter case, theopening pressure was assumed to be 5 mm Hg higher

(a correction factor) than the mean arterial bloodpressure. Subjects were to record their blood pressuretwice per day for at least the first 3 months followingdischarge and to continue this routine after the first 3months until blood pressure consistently remainedwithin the recommended range. The recommendedblood pressure range was defined as mean arterialpressure (MAP) #85 mm Hg for the automated cuffmethod and an opening pressure #90 mm Hg for theDoppler cuff method.

Subject compliance was assessed weekly duringthe initial 3 months, after which the investigatordetermined if the subject’s blood pressure was stableor unstable. If the blood pressure was determined tobe unstable, the subject continued the blood pressuremonitoring for an additional 3 months, with reas-sessment for stability. Subjects with 4 consecutiveblood pressure measurements outside the targetrange were instructed to contact the study site, withthe expectation that antihypertensive therapy wouldbe adjusted. Compliance was assessed via review ofdiaries by the study site and the sponsor.

Management of blood pressure in subjectsreceiving the control device was not standardized bystudy protocol. The manufacturer’s instructions foruse for the control device allows treatment of post-implantation hypertension at the discretion of thephysician and notes that maintaining MAP at <90mm Hg should be adequate (15). Also, the current In-ternational Society for Heart & Lung Transplantationguidelines for patients receiving mechanical circula-tory support recommend blood pressure control,specifically maintaining blood pressure at <85 mmHg, as a standard of care in these patients (16).

STATISTICAL ANALYSIS. The difference in the per-centage of patients meeting the primary endpointbetween the HVAD and control devices was testedusing the Farrington-Manning asymptotic test ofnoninferiority with a noninferiority margin of 6%.The confidence interval was reported using the pre-specified exact method, using SAS code (SAS Insti-tute, Cary, North Carolina) to perform the exactmethods. However, this method was found to be tooconservative, so an alternative method, the d-pro-jected Z statistic without restricted search for thenuisance parameter, is also presented (17).

The composite secondary endpoint was tested us-ing the Farrington-Manning asymptomatic test ofnoninferiority with a noninferiority margin of 15%.Because noninferiority was statistically significant,superiority testing was performed without alphapenalty. Analyses of the primary and secondaryendpoints were conducted on the modified intention-

FIGURE 1 ENDURANCE Supplemental Trial Design

HMII ¼ HeartMate II; ITT ¼ intention-to-treat; LVAD ¼ left ventricular assist device.

J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8 Milano et al.S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2 HVAD: The ENDURANCE Supplemental Trial

795

to-treat subjects who did not exit prior to an event. Apost hoc comparison of the stroke or TIA rates in theHVAD subjects between the 2 studies was also per-formed using the chi-square test.

For blood pressure analysis, 5 mm Hg was sub-tracted from the Doppler opening pressure values toestimate the MAP measurement. Comparison ofgroups’ mean MAPs was done using 2-sampleStudent’s t tests. Post-transplantation values werenot included in the analysis.

The Fisher exact test was used to compare the ratesof adverse events between the treatment and controlcohorts. Survival analysis was performed usingKaplan-Meier methods. Patients were censored fromanalysis at the time of original device explantation. Ap value < 0.05 was considered to indicate statisticalsignificance. P values were not adjusted for multipletesting.

RESULTS

Between October 2013 and August 2015, 494 subjectswere screened at 48 sites in the United States

(Figure 1). There were 19 screen failures, resulting inrandomization of 475 subjects in a 2:1 fashion, studydevice to control. Ten subjects did not undergo im-plantation, resulting in a modified intention-to-treatpopulation of 465 subjects (n ¼ 308 HVAD, n ¼ 157control) from 47 sites. All subjects underwent im-plantation of the devices to which they wererandomized.

The baseline characteristics of the 2 groups weresimilar (Table 1). Study and control cohorts did notdiffer with respect to severity of illness or treatmentsat the time of enrollment. International normalizedratio was measured monthly post-implantation, andthe percentage of HVAD subjects within the interna-tional normalized ratio target range of 2.0 to 3.0 rangedbetween 49% and 62% during the 12-month follow-upperiod, while the percentage of control subjectsranged from 47% to 61%. Additionally, after adjust-ment for multiple comparison, there were no statisti-cally significant differences in the percentage in thetarget range between HVAD and control at any timepoint during the follow-up period. A comparison of

TABLE 1 Baseline Characteristics of Randomized Subjects

HVAD(n ¼ 308)

Control Device(n ¼ 157)

pValue

Age (yrs) 63.3 � 11.4 64.2 � 11.1 0.39

Female 18.2 20.4 0.62

Race (% white) 71.8 75.2 0.51

Height (cm) 175.0 � 9.4 175.1 � 9.8 0.91

Body mass index (kg/m2) 28.2 � 5.5 27.4 � 5.2 0.13

INTERMACS profile* 0.90

1 3.9 2.5

2 32.8 32.5

3 43.3 43.3

4–7 20.0 21.7

Ischemic etiology of heart failure 55.2 58.0 0.62

History of smoking 68.2 65.6 0.60

Diabetic 49.4 48.4 0.92

Previous stroke 10.4 8.3 0.51

Hypertension requiring medication 75.0 72.0 0.50

Atrial fibrillation 50.6 51.0 >0.99

Mean arterial blood pressure(mm Hg)

78.9 � 11.5 (n ¼ 296) 77.6 � 11.1 (n ¼ 153) 0.23

Tricuspid regurgitation (more thanmoderate)

40.4 (n ¼ 302) 44.2 (n ¼ 154) 0.48

Left ventricular ejection fraction (%) 17.3 � 5.1 18.2 � 4.5 0.07

Previous intervention

ICD 80.8 82.2 0.80

CRT 28.9 28.7 >0.99

IABP 19.2 15.9 0.45

Values are mean � SD or %. The p values are post hoc and are included for information purposes only. The pvalues comparing categorical values are from the Fisher exact test. The p values comparing continuous values arefrom a 2-sample Student’s t test. Three HVAD subjects were missing INTERMACS profile and are not included.*Total number of HVAD subjects in the denominator ¼ 305, as INTERMACS profiles were incomplete for 3subjects.

CRT ¼ cardiac resynchronization therapy; IABP ¼ intra-aortic balloon pump; ICD ¼ implantable cardioverter-defibrillator; INTERMACS ¼ Interagency Registry for Mechanically Assisted Circulatory Support.

Milano et al. J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8

HVAD: The ENDURANCE Supplemental Trial S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2

796

baseline characteristics between the HVAD cohortin the ENDURANCE Supplemental Trial and theENDURANCE trial revealed no statistically significantdifferences among any of the demographics or medicalhistory, with the exception of a significantly largermean body mass index in the ENDURANCE Supple-mental patients receiving the HVAD compared withthose in the ENDURANCE trial (28.2 vs. 27.2 kg/m2;p ¼ 0.03) and a significantly greater history ofhypertension in those patients in the ENDURANCESupplemental Trial (75.0% vs. 65.5%; p ¼ 0.01).Interestingly, these 2 parameters could have biasedagainst success in a trial targeting blood pressuremanagement.

Subjects receiving the HVAD and management withthe blood pressure management protocol had lowerMAP compared with subjects receiving the controldevice without protocol-directed management ofblood pressure (Figure 2A). Additionally, the meanMAPwas significantly reduced at all post-implantation

time points in subjects receiving the HVAD in theENDURANCE Supplemental Trial compared with theENDURANCE trial. However, there was also a reduc-tion in the mean blood pressure of the control devicewhen comparing ENDURANCE with ENDURANCESupplemental (Figure 2A), although the reductionswasnot statistically significant. An analysis of the propor-tion of subjects meeting the target MAPs was alsosimilar over time for HVAD and control, althoughHVAD subjects had nominally better target adherencein the post-discharge periods than at 6 and 12 months(Figure 2B). A comparison of the percentage of subjectswith strokes between 3 and 12 months between HVADand control subjects by those who were outside of thetarget blood pressure range at 3 months showed nosignificant difference between the 2 cohorts (7.8% [6 of77] vs. 6.7% [4 of 60], respectively; p > 0.99), althoughthe number of strokes in each group was small. Addi-tionally, the mean MAP in patients with strokes(averaging MAPs up to the day prior to a stroke) weresimilar in the HVAD and control groups (80.7 mm Hgvs. 78.1 mm Hg).

Neurological injury as defined for the primaryendpoint occurred in 14.7% of the HVAD cohort and12.1% of the control cohort (Figure 3). The primaryendpoint as defined in this study was not achieved,with an upper limit of the confidence interval (pre-specified method) of 10.7%, which was above thepre-specified noninferiority margin of 6% (p ¼ 0.14).A more reliable, but post hoc, confidence limitmethod produced an upper confidence bound of7.9% (17).

An analysis of the composite secondary endpointof freedom from death, disabling stroke, and devicemalfunction or failure requiring exchange, explanta-tion, or urgent transplantation (the primary outcomemeasure used in most clinical trials of LVAD therapyas well as from the ENDURANCE trial but measured at1 year) demonstrated noninferiority of the HVADgroup compared with control (76.1% vs. 66.9%;p < 0.0001). A subsequent test of superiority showedthe HVAD to be significantly better than control(superiority p ¼ 0.04). A Kaplan-Meier representationof the results is presented in Figure 4.

Although the mortality rate was similar betweenHVAD and control (16.9% HVAD, 17.8% control), ahigher percentage of control subjects experienceddisabling (mRS score $4) strokes (1.6% vs. 3.2%) anddevice exchange (3.2% vs. 10.2%) as their worst event.Strokes with mRS scores <4 were more common inthe HVAD group (9.1% vs. 5.7%) (Figure 5A).

The comparison of the rate of strokes betweenHVAD (16.9%) and control (14.6%) was not

FIGURE 2 Analysis of Mean Arterial Blood Pressure Over Time

(A) Comparison of mean arterial pressure (MAP) at each follow-up visit from the ENDURANCE trial and ENDURANCE Supplemental Trial

between the HVAD and control groups and (B) analysis of subjects versus MAP targets over time in the ENDURANCE Supplemental Trial. For

ENDURANCE Supplemental, 5 mm Hg was subtracted from MAP when measured by Doppler. Blood pressure collection methods were not

collected in ENDURANCE, so those data are not adjusted. LVAD ¼ left ventricular assist device.

J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8 Milano et al.S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2 HVAD: The ENDURANCE Supplemental Trial

797

statistically significantly different (p ¼ 0.60). Therewas also no significant difference between ischemiccerebrovascular accident (13.0% vs. 7.6%) and HCVA(5.2% vs. 7.0%) for HVAD versus control, respectively.However, a significantly greater percentage of sub-jects experienced TIAs in the HVAD group versuscontrol (4.2% vs. 0.6%, respectively; p ¼ 0.04). Theoverall level of disability post-stroke was propor-tionately lower in the HVAD group compared withcontrol (Figure 5B).

The secondary endpoint of incidence of stroke orTIA in patients receiving the HVAD was pre-specifiedto be compared with a 17.7% performance goaland did not pass (p ¼ 0.74). However, the incidenceof stroke was reduced by 24.2% in the ENDURANCESupplemental Trial compared with the HVAD cohortof the ENDURANCE trial, although superiority wasnot shown (chi-square superiority p ¼ 0.10). Howev-er, an analysis of the overall strokes in HVAD patientsin ENDURANCE compared with ENDURANCE

FIGURE 3 Primary Endpoint: Freedom From Neurological Injury (Stroke With Modified Rankin Scale Score >0 at 24 Weeks Post-Stroke or

Transient Ischemic Attack) at 12 Months

(A) Histogram represents the incidence of the pre-specified endpoint of neurological injury (any stroke with a modified Rankin scale [MRS]

score >0 at 24 weeks or a transient ischemic attack [TIA]) in each cohort at 1 year. (B) The 95% confidence interval of the primary endpoint,

including the pre-specified zone of noninferiority as defined in the statistical methods.

FIGURE 4 Surviva

Freedom from death

malfunction or failur

presented as a Kapla

test comparing prop

Milano et al. J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8

HVAD: The ENDURANCE Supplemental Trial S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2

798

Supplemental revealed an improvement of theneurological event profile in general, and hemor-rhagic strokes in particular were statistically signifi-cantly reduced by 50.5% (p ¼ 0.02) (Table 2). Theoverall stroke rates for the control device were notsignificantly different between the control arms inthe 2 trials, including the rate of disabling strokes,which occurred in 4.7% of patients (7 of 149) in theENDURANCE trial and 6.4% of those (10 of 157)

l Analysis of Composite Secondary Endpoint

, disabling stroke (modified Rankin scale score >4), and device

e requiring exchange, explantation, or urgent transplantation is

n-Meier survival plot. The p value presented describes the chi-square

ortions.

enrolled in the ENDURANCE Supplemental Trial at12 months.

Adverse event rates for major bleeding, cardiacarrhythmias, renal dysfunction, and infections,including percutaneous driveline infections, weresimilar between HVAD and control subjects (Table 3).TIAs were more frequent in the HVAD cohort, whilehemolysis, pump replacement, and exchange due topump thrombus were all significantly higher in thecontrol cohort. Additionally, freedom from any pumpthrombus event through 1 year was significantlygreater for subjects receiving the HVAD, 91.3%,compared with 81.6% for those receiving the control(log-rank p ¼ 0.0013).

Although the overall rate of device malfunctionswas similar between the 2 cohorts, there were dif-ferences in the types of malfunctions. In the HVADcohort, 19.6% of the device malfunctions were relatedto battery problems or electric faults, whereas 26.2%were due to controller faults and another 26.2% of thetotal malfunctions to suspected or confirmed pumpthrombus events. In the control group, there were nobattery-related malfunctions, 23.4% of the total mal-functions were related to controller faults, and 57.4%of the total device malfunctions consisted of pumpthrombosis.

Overall quality-of-life scores (Table 4) were simi-larly significantly improved at 12 months comparedwith baseline in both HVAD and control subjects.Functional capacity, as measured by NYHA classifi-cation improvements and increase in 6-min walk

FIGURE 5 Comparisons of Outcomes in the ENDURANCE Supplemental Trial

(A) Hierarchical comparison of outcomes in the ENDURANCE Supplemental Trial and (B) proportion of total neurological events, by severity

(or recovery). CVA ¼ cerebrovascular accident; mRS ¼ modified Rankin scale; TIA ¼ transient ischemic attack.

J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8 Milano et al.S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2 HVAD: The ENDURANCE Supplemental Trial

799

distance, were comparably improved in both cohorts.Although data were somewhat incomplete (51.9% ofHVAD and 43.3% of control subjects had paired data),improvements of $100 m in the 6-min walk totaldistance were similarly achieved by 47.5% of thesubjects in the HVAD group and 51.5% of the subjectsin the control group.

DISCUSSION

Previous analyses of data from the ADVANCE andENDURANCE trials suggested that increased MAP is

an independent predictor of stroke among patientssupported with the HVAD (11). In the ENDURANCESupplemental Trial, a blood pressure algorithm,most often using a Doppler/cuff method, wasapplied to subjects receiving the HVAD to measureMAP. Application of this protocol successfully low-ered MAP in the HVAD cohort relative to the con-trol cohort (managed according to the standard ofcare). The absolute difference in MAPs was small,on the order of 4 to 5 mm Hg, but was observed atall time points throughout the trial. This smalldifference may be clinically important compared

TABLE 2 Percentage of HVAD Patients Who Experienced

Neurological Events at 12 Months in the ENDURANCE Trial and the

ENDURANCE Supplemental Trial

HVADENDURANCE(n ¼ 296)

HVADENDURANCESupplemental(n ¼ 308)

pValue

All stroke 22.3 16.9 0.10

Ischemic cerebrovascular event 13.9 13.0 0.81

Hemorrhagic cerebrovascularevent

10.5 5.2 0.02

Disabling stroke 8.1 6.5 0.53

Transient ischemic attack 5.4 4.2 0.57

Values are %. The p values were calculated using the Fisher exact test.

TABLE 3 Summary o

Major bleeding

Cardiac arrhythmia

Hepatic dysfunction

Hypertension

Major infection

Driveline exit site inf

Device malfunction/fail

Hemolysis

Stroke

Ischemic cerebrovasc

Hemorrhagic cerebrov

TIA

Renal dysfunction

Respiratory failure

Right heart failure

Pump replacement

Exchange for pump t

Values are n (%). The p va

NA ¼ not applicable; TIA

Milano et al. J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8

HVAD: The ENDURANCE Supplemental Trial S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2

800

with other studies of blood pressure treatment(18,19).

Although the difference in neurological injury forthe primary endpoint between the HVAD cohort andcontrol was only 2.6%, this failed to achieve the pre-specified test of noninferiority, which called for a6% noninferiority margin. It is important to note thatalthough TIAs, which do not represent persistentinjury, were included in the definition of neurologicalinjury for the primary endpoint, the primary endpointdid not include strokes with mRS scores of 0 at the24-week follow-up point (fully recovered strokes withno residual deficit).

f Adverse Events Occurring Through 1 Year

Study Device(n ¼ 308)

Control Device(n ¼ 157)

pValue

Patients WithEvents

Number ofEvents

PatientsWith Events

Number ofEvents

159 (51.6) 310 89 (56.7) 196 0.33

105 (34.1) 151 49 (31.2) 56 0.60

12 (3.9) 12 6 (3.8) 6 >0.99

40 (13.0) 54 20 (12.7) 21 >0.99

166 (53.9) 300 93 (59.2) 181 0.28

ection 50 (16.2) 59 19 (12.1) 22 0.27

ure 74 (24.0) 107 38 (24.2) 47 >0.99

4 (1.3) 5 9 (5.7) 9 0.01

52 (16.9) 75 23 (14.6) 25 0.60

ular event 40 (13.0) 58 12 (7.6) 14 0.09

ascular event 16 (5.2) 17 11 (7.0) 11 0.53

13 (4.2) 13 1 (0.6) 1 0.04

32 (10.4) 35 23 (14.6) 25 0.22

61 (19.8) 77 31 (19.7) 37 >0.99

109 (35.4) 116 60 (38.2) 65 0.61

16 (5.2) NA 18 (11.5) NA 0.02

hrombosis 14 (4.5) NA 16 (10.2) NA 0.03

lues compare the percentage of patients with events using the Fisher exact test.

¼ transient ischemic attack.

Although the primary endpoint was not met, ex-amination of the totality of data from this trialsuggests that the HVAD cohort managed accordingto this algorithm experienced a significant decreasein the rate of stroke. The total stroke rate ascaptured in the summary of adverse events dem-onstrates no statistical difference in rates for anystroke, ischemic stroke, and hemorrhagic stroke forthe HVAD cohort versus control. Although not sta-tistically significant, the rate of hemorrhagic strokewas less for the HVAD cohort versus control, rep-resenting an improvement from the originalENDURANCE trial, which showed a 2- to 3-foldincreased stroke rate for HVAD patients comparedwith control. Similarly, the 2.6% higher rate ofneurological injury for the HVAD group in the pri-mary endpoint is much improved from the differ-ences seen in the ENDURANCE trial for stroke rateor TIA. Finally, the rates of stroke and TIA for theHVAD cohort in the ENDURANCE SupplementalTrial were all meaningfully lower than those for theHVAD cohort in the original ENDURANCE trial. Theimprovement included a 50.5% reduction for hem-orrhagic events in the ENDURANCE SupplementalTrial versus the ENDURANCE trial, which is consis-tent with the therapy focused on blood pressurereduction. Therefore, the examination of all data inthis trial would suggest a meaningful associationbetween reduced MAP for the HVAD patients andreduced risk for stroke events.

Previous studies in non-LVAD patients haveshown an association between reduced stroke andbetter blood pressure management (20). Therefore,the reduced rate of hemorrhagic stroke for theHVAD patients in this trial may have been antici-pated. There also appears to be a beneficial effectwith regard to ischemic stroke. The exact mecha-nism for this is unclear. One theory is that reducedMAP results in reduced afterload for the LVADpump, which translates into greater pump flow andbetter washing of the entire left heart and LVADpump. This could reduce stasis and result in lessthrombus formation at different sites. A greaterreduction in afterload could also promote greaternative left ventricular ejection and intermittentopening of the aortic valve, which has beendemonstrated to reduce areas of blood flow stag-nation in the left ventricle (21).

We also examined the conventional compositeoutcome of survival free from death, disabling stroke(mRS score $4), or need for device replacement orurgent transplantation. This endpoint is a conven-tion used in trials to assess LVAD outcomes and was

TABLE 4 Quality of Life Measures

Study Device(n ¼ 308)

Control Device(n ¼ 157)

KCCQ overall summary score

Percentage of subjects withimprovements of $10 pointsfrom baseline at 12 months

80.0 (n ¼ 195) 86.5 (n ¼ 89)

EQ-5D visual analog score

Mean change from baseline at 12months (minimum, maximum)

23.6 � 26.5 (�90, 86)(n ¼ 197)

27.2 � 32.4 (�55, 98)(n ¼ 87)

Values are % or mean � SD. Summary of KCCQ overall summary score improvement and mean change frombaseline of the EQ 5D visual analog score at 12 months. There was no statistically significant difference betweenstudy device and control for either measure.

KCCQ ¼ Kansas City Cardiomyopathy Questionnaire.

J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8 Milano et al.S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2 HVAD: The ENDURANCE Supplemental Trial

801

used in the ENDURANCE trial (8) as well as in therecently published MOMENTUM trial (22). At 1 year,the HVAD group was superior to the control group atachieving this endpoint. This difference is duemainly to greater need for pump replacementbecause of thrombosis observed in the control group.One limitation for this composite endpoint is that all3 types of events that constitute it are not equal. Forexample, death is a greater failure than need fordevice replacement. To examine this more carefully,the rates of different events are presented in hier-archical severity (i.e., death first) (Figure 4).

STUDY LIMITATIONS. Improved blood pressuremanagement was not mandated for the controlcohort in this trial. Clinical trial design rules donot allow alterations in the standard of care in acontrol population. Therefore, it is unknownwhether improved blood pressure managementwould have resulted in any impact on the strokerate for the control cohort. However, the MAPs ofpatients with stroke events in both cohorts weresimilar, and conversely the incidence of neurolog-ical events was similar in the HVAD and controlcohorts that did not have adequately controlledblood pressure.

HVAD patients were not randomized to receive theblood pressure management protocol; rather, out-comes were compared with results from the ENDUR-ANCE trial. The blood pressure measures in theENDURANCE trial were not collected with the samefrequency or methodology as in the ENDURANCESupplemental Trial, and the analysis was retrospec-tive. Other changes may have occurred in the deviceor patient management that may account for thereduced rates of stroke.

Additionally, elements of the ENDURANCE Sup-plemental Trial design contributed to limitations ofthe results. First, the primary endpoint excludedpatients who experienced strokes but went on tohave complete recovery (mRS score ¼ 0) at 24weeks, while including patients who experiencedTIAs. In retrospect, it might have been more appro-priate to include or exclude both. Second, sample-size calculations for the primary endpoint usedlower power (80%) and a small noninferiority margin(6%), underestimated the event rate, and assumed asuperior rate in the HVAD group compared withcontrol subjects, which could have led to thenonsignificant result of the primary endpoint.Finally, no adjustments were made to p values re-ported, so multiple comparisons should be takeninto account when interpreting p values.

CONCLUSIONS

The ENDURANCE Supplemental Trial failed todemonstrate noninferiority of the HVAD versus con-trol with respect to the primary endpoint; there was ahigher incidence of neurological injury for the HVADversus control. However, the trial confirmed thatblood pressure management was associated with areduced risk for stroke in HVAD subjects. Addition-ally, with regard to the composite endpoint of sur-vival free from disabling stroke and need for deviceexchange or urgent transplantation or death, theHVAD System was superior to the control. Therefore,the HVAD is a reasonable alternative and may be su-perior to the HeartMate II for some patients, as therewere fewer thrombi requiring device exchange, andthe majority of strokes were not disabling. This studyalso provides support for the existing InternationalSociety for Heart & Lung Transplantation recom-mendations for blood pressure management in pa-tients with mechanical circulatory support (16),perhaps further defining that blood pressure bemonitored twice daily for at least the first 3 monthspost-implantation to ensure appropriate blood pres-sure control. Finally, these results support the use ofthe HVAD as a safe and effective system for patientsineligible for cardiac transplantation.

ACKNOWLEDGMENTS The authors acknowledgeMary V. Jacoski, MS, of Medtronic for assistance inthe analysis and preparation of the manuscript, andJeffrey Cerkvenik, also of Medtronic, for statisticalsupport.

ADDRESS FOR CORRESPONDENCE: Dr. Carmelo A.Milano, Duke University School of Medicine, 4532Hospital South–Blue Zone, Durham, North Carolina27710. E-mail: carmelo.milano@duke.edu.

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: For

clinicians, this study has demonstrated that careful atten-

tion to blood pressure management may lead to lower

rates of strokes in patients supported with the HVAD

System, thus improving overall outcomes for patients with

advanced heart failure ineligible for transplantation.

TRANSLATIONAL OUTLOOK: For investigators, these

data have prospectively validated the preliminary data

found through a post hoc analysis of clinical trial data,

which demonstrated through a multivariable analysis that

elevated blood pressure is significant risk factor for stroke

in patients with advanced heart failure implanted with

LVADs. This study is an example of the utility and value of

data-mining efforts and post hoc analyses of clinical trial

data.

Milano et al. J A C C : H E A R T F A I L U R E V O L . 6 , N O . 9 , 2 0 1 8

HVAD: The ENDURANCE Supplemental Trial S E P T E M B E R 2 0 1 8 : 7 9 2 – 8 0 2

802

RE F E RENCE S

1. Ponikowski P, Voors AA, Anker SD, et al. 2016ESC guidelines for the diagnosis and treatment ofacute and chronic heart failure. The Task Force forthe Diagnosis and Treatment of Acute and ChronicHeart Failure of the European Society of Cardiol-ogy (ESC) developed with the special contributionof the Heart Failure Association (HFA) of the ESC2015. Eur Heart J 2016;37:2129–200.

2. Mozaffarian D, Benjamin EJ, Go AS, et al. Heartdisease and stroke statistics—2016 update.A report from the American Heart Association2016;133:e38–360.

3. Stewart GC, Givertz MM. Mechanical circulatorysupport for advanced heart failure. Circulation2012;125:1304–15.

4. Kirklin JK, Naftel DC, Pagani FD, et al. SeventhINTERMACS annual report: 15,000 patients andcounting. J Heart Lung Transplant 2015;34:1495–504.

5. Slaughter M, Rogers J, Milano C, Russell S,Conte J, Feldman D. Advanced heart failuretreated with continuous-flow left ventricularassist device. N Engl J Med 2009;361:2241–51.

6. Park SJ, Milano CA, Tatooles AJ, et al. Out-comes in advanced heart failure patients with leftventricular assist devices for destination therapy.Circ Heart Fail 2012;5:241–8.

7. Jorde UP, Kushwaha SS, Tatooles AJ, et al.Results of the destination therapy post-food anddrug administration approval study with acontinuous flow left ventricular assist device: aprospective study using the INTERMACS registry(Interagency Registry for Mechanically Assisted

Circulatory Support). J Am Coll Cardiol 2014;63:1751–7.

8. Rogers JG, Pagani FD, Tatooles AJ, et al.Intrapericardial left ventricular assist device foradvanced heart failure. N Engl J Med 2017;376:451–60.

9. Aaronson KD, Slaughter MS, Miller LW, et al.Use of an intrapericardial, continuous flow, cen-trifugal pump in patients awaiting heart trans-plantation. Circulation 2012;125:3191–200.

10. Slaughter MS, Pagani FD, McGee EC, et al. Useof the HeartWare� ventricular assist system forbridge to transplant: combined results of theADVANCE and CAP trial. J Heart Lung Transplant2013;32:675–83.

11. Teuteberg JJ, Slaughter MS, Rogers JG, et al.The HVAD left ventricular assist device: risk fac-tors for neurological events and risk mitigationstrategies. J Am Coll Cardiol HF 2015;3:818–28.

12. LaRose JA, Tamez D, Ashenuga M, Reyes C.Design concepts and principle of operation of theHeartWare ventricular assist system. ASAIO J2010;56:285–9.

13. Kirklin J. Interagency Registry for MechanicallyAssisted Circulatory Support (INTERMACS�).AppendixA—adverseevent definitionsprotocol 4.0.

14. ATS Committee on Proficiency Standards forClinical Pulmonary Function Laboratories. ATSstatement: guidelines for the six-minute walk test.Am J Respir Crit Care Med 2002;166:111–7.

15. Thoratec Corporation. HeartMate II left ven-tricular assist system instructions for use. Pleas-anton, California: Thoratec, 2017.

16. Feldman D, Pamboukian SV, Teuteberg JJ,et al. The 2013 International Society for Heart andLung Transplantation Guidelines for mechanicalcirculatory support: executive summary. J HeartLung Transplant 2013;32:157–87.

17. Chan ISF, Zhang Z. Test-based exact confi-dence intervals for the difference of two binomialproportions. Biometrics 1999;55:1202–9.

18. Akioyamen L, Levine M, Sherifali D, et al.Cardiovascular and cerebrovascular outcomes oflong-term angiotensin receptor blockade: meta-analyses of trials in essential hypertension. J AmSoc Hypertens 2016;10:55–69.e1.

19. Xie X, Atkins E, Lv J, et al. Effects of intensiveblood pressure lowering on cardiovascular andrenal outcomes: updated systematic review andmeta-analysis. Lancet 2016;387:435–43.

20. Gorelick PB. New horizons for stroke preven-tion: PROGRESS and HOPE. Lancet Neurol 2002;1:149–56.

21. Prisco AR, Aliseda A, Beckman JA,Mokadam NA, Mahr C, Garcia GJM. Impact of LVADimplantation site on ventricular blood stagnation.ASAIO J 2017;63:392–400.

22. Mehra MR, Naka Y, Uriel N, et al. A fullymagnetically levitated circulatory pump foradvanced heart failure. N Engl J Med 2017;376:440–50.

KEY WORDS advanced heart failure,destination therapy, HeartWare HVADsystem, LVAD, stroke