Intrapericardial Left Ventricular Assist Device for Advanced Heart Failure

Transcription

1 The new england journal of medicine Original Article Intrapericardial Left Ventricular Assist Device for Advanced Heart Failure Joseph G. Rogers, M.D., Francis D. Pagani, M.D., Ph.D., Antone J. Tatooles, M.D., Geetha Bhat, M.D., Mark S. Slaughter, M.D., Emma J. Birks, M.B., B.S., Ph.D., Steven W. Boyce, M.D., Samer S. Najjar, M.D., Valluvan Jeevanandam, M.D., Allen S. Anderson, M.D., Igor D. Gregoric, M.D., Hari Mallidi, M.D., Katrin Leadley, M.D., Keith D. Aaronson, M.D., O.H. Frazier, M.D., and Carmelo A. Milano, M.D. ABSTRACT BACKGROUND Mechanical circulatory support with a left ventricular assist device (LVAD) is an established treatment for patients with advanced heart failure. We compared a newer LVAD design (a small intrapericardial centrifugal-flow device) against existing technology (a commercially available axial-flow device) in patients with advanced heart failure who were ineligible for heart transplantation. METHODS We conducted a multicenter randomized trial involving 446 patients who were assigned, in a 2:1 ratio, to the study (centrifugal-flow) device or the control (axial-flow) device. Adults who met contemporary criteria for LVAD implantation for permanent use were eligible to participate in the trial. The primary end point was survival at 2 years free from disabling stroke or device removal for malfunction or failure. The trial was powered to show noninferiority with a margin of 15 percentage points. RESULTS The intention-to treat-population included 297 participants assigned to the study device and 148 participants assigned to the control device. The primary end point was achieved in 164 patients in the study group and 85 patients in the control group. The analysis of the primary end point showed noninferiority of the study device relative to the control device (estimated success rates, 55.4% and 59.1%, respectively, calculated by the Weibull model; absolute difference, 3.7 percentage points; 95% upper confidence limit, percentage points; P = 0.01 for noninferiority). More patients in the control group than in the study group had device malfunction or device failure requiring replacement (16.2% vs. 8.8%), and more patients in the study group had strokes (29.7% vs. 12.1%). Quality of life and functional capacity improved to a similar degree in the two groups. From Duke University School of Medicine, Durham, NC (J.G.R., C.A.M.); University of Michigan, Ann Arbor (F.D.P., K.D.A.); Advocate Christ Medical Center, Oak Lawn (A.J.T., G.B.), and University of Chicago Medicine (V.J.) and Northwestern Memorial Hospital (A.S.A.), Chicago all in Illinois; University of Louisville, Louisville, KY (M.S.S., E.J.B.); MedStar Heart Institute, Washington, DC (S.W.B., S.S.N.); University of Texas Health Science Center (I.D.G.) and Texas Heart Institute (H.M., O.H.F.), Houston; and HeartWare, Framingham, MA (K.L.). Address reprint requests to Dr. Rogers at the Division of Cardiovascular Medicine, Duke Clinical Research Institute, Duke University Medical Center, Box 3034 DUMC, Durham, NC 27710, or at joseph. rogers@ duke. edu. Drs. Rogers and Pagani contributed equally to this article. N Engl J Med 2017;376: DOI: /NEJMoa Copyright 2017 Massachusetts Medical Society. CONCLUSIONS In this trial involving patients with advanced heart failure who were ineligible for heart transplantation, a small, intrapericardial, centrifugal-flow LVAD was found to be noninferior to an axial-flow LVAD with respect to survival free from disabling stroke or device removal for malfunction or failure. (Funded by HeartWare; ENDURANCE ClinicalTrials.gov number, NCT ) n engl j med 376;5 nejm.org February 2,

2 The new england journal of medicine A dvanced heart failure is characterized by profound limitations in survival, functional status, and quality of life despite treatment with evidence-based therapies. 1 Left ventricular assist devices (LVADs) offer patients with this condition a therapeutic option that provides circulatory support while they are awaiting heart transplantation or that can serve as a permanent alternative to transplantation (i.e., destination therapy ). The Food and Drug Administration (FDA) approved one continuous-flow LVAD with an axialflow design for destination therapy after outcomes with the device were shown to be superior to those with older, pulsatile LVAD technology. 2 Despite the efficacy of the approved axial-flow LVAD, adverse events still lead to substantial morbidity and mortality and represent an unmet need for the development of long-term support devices that have improved efficacy and safety. 3,4 We report the results of a randomized clinical trial evaluating outcomes in patients who were ineligible for heart transplantation and were treated with either the FDA-approved axial-flow device or a centrifugal-flow LVAD that is smaller (50-cc displacement, vs. 65-cc displacement for the axialflow device), lies entirely within the pericardial space, and incorporates a bearingless design with magnetic and hydrodynamic levitation of the internal rotor. 5,6 Methods Trial Objective and Organization The purpose of the ENDURANCE trial was to assess the safety and effectiveness of the centrifugal-flow LVAD relative to those of a control, axial-flow LVAD in patients with advanced heart failure. The trial was designed by the sponsor (HeartWare) in consultation with FDA and clinical investigators. The trial was conducted at 48 U.S. sites (a list is provided in the Supplementary Appendix, available with the full text of this article at NEJM.org) and was supervised by the sponsor and an independent clinical research organization (ICON) that was paid by the sponsor. ICON managed the electronic data-capture system, conducted all statistical analyses of the primary end point, and provided the final data listings. Coordinators at each site captured all trial data electronically. The sponsor had access to the safety data that were required for regulatory monitoring. The authors had unrestricted access to the data and attest to the completeness and accuracy of the data and to the fidelity of the trial to the protocol, which is available at NEJM.org. A data and safety monitoring board independently monitored and reviewed compliance, adverse events, and outcomes. The trial was approved by the institutional review board at each clinical site. Trial Participants Patients were eligible for enrollment if they had chronic, advanced left ventricular failure with New York Heart Association (NYHA) functional class IIIB or IV limitations despite receiving currently recommended medical therapy, had a left ventricular ejection fraction of 25% or less, and were ineligible for transplantation at the time of enrollment. A complete list of inclusion and exclusion criteria is provided in Tables S1 and S2 in the Supplementary Appendix. All participants or their authorized representatives provided written informed consent. Trial Procedures Participants were randomly assigned, in a 2:1 ratio, to receive either the centrifugal-flow (study) device or the axial-flow (control) device (Fig. S1 in the Supplementary Appendix). Randomization was performed with the use of a permuted-block, central randomization scheme and was implemented with a Web-based interactive response system. Baseline data were obtained after patient consent and enrollment. We collected data on demographic characteristics, health history, and medications, as well as hematologic, biochemical, and hemodynamic data. The severity of illness was classified with the use of the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) patient profile 7 (determined by an independent assessor; see Table S3 in the Supplementary Appendix). Additional assessments included the Kansas City Cardiomyopathy Questionnaire (KCCQ) (scores range from 0 to 100, with higher scores indicating better health status), the European Quality of Life 5 Dimensions visualanalogue scale (EQ-5D VAS) (scores range from 0 to 100, with higher scores indicating better health status), 6-minute walk distance, NYHA functional class, and neurologic status (as measured on the modified Rankin scale, on which scores range from 0 to 6, with 0 indicating no symptoms and 6 indicating death). 452 n engl j med 376;5 nejm.org February 2, 2017

3 Intrapericardial Left Ventricular Assist Device After implantation, device performance, laboratory data, and medications were recorded until hospital discharge, at follow-up visits scheduled at 3 and 6 months, and then every 6 months until the assessment of the primary end point at 24 months. Functional capacity and quality-of-life measurements were performed at 3 and 6 months and then at 12 and 24 months. Study Devices Enrolled participants received either the study device (HeartWare Ventricular Assist Device System, HeartWare) 5,6,8-10 or the control device (Heart- Mate II Left Ventricular Assist System, St. Jude Corporation) (Fig. 1). 2,11-15 The study device is a continuous-flow LVAD that has magnetic and hydrodynamic levitation of the internal rotor and is implanted within the pericardial space. 6,10 The study device was changed during the course of the trial by applying titanium microspheres to the outer surface (i.e., sintering) of the inflow cannula to promote tissue ingrowth, and the coring knife was enlarged to excise more muscle at the left ventricular apex for insertion of the inflow cannula (Fig. S2 in the Supplementary Appendix). 16,17 Two thirds of the participants who were assigned to the study device in the trial received the currently available sintered device that is approved for the bridge-to-transplantation indication in the United States. The control device is a continuous-flow LVAD with an axial design that is approved for both bridge-to-transplantation and destination indications in the United States. 2,11-15 Anticoagulation recommendations for the trial are provided in Table S4 in the Supplementary Appendix but were individualized by the clinical investigators. End Points The primary end point was a composite of 2-year survival free from disabling stroke (with disabling defined as a modified Rankin scale score of 4 assessed 24 weeks after the stroke), with the patient alive with the originally implanted device, having undergone elective transplantation, or with the device explanted because of left ventricular recovery. Therapy was considered to have failed if, within the first 2 years after enrollment, the patient died, had a stroke with a modified Rankin scale score of 4 or higher assessed 24 weeks after the stroke, or underwent either urgent transplantation or surgery for LVAD removal or replacement because of a failure of the original device. Prespecified secondary end points included the incidence of major adverse events (classified according to the INTERMACS definitions), 18 overall survival, changes in quality of life and health status as assessed by the KCCQ and the EQ-5D VAS, and functional status, as measured by NYHA functional class and 6-minute walk distance. Table S5 in the Supplementary Appendix includes the full list of prespecified end points. All outcomes were adjudicated by an independent clinical events committee. Statistical Analysis The study device was tested for noninferiority to the control device with regard to success in meeting the primary end point. The noninferiority margin was set at 15 percentage points, under the assumption that the observed success rate of the control device would be 55% or greater. Estimates of event-free survival with the originally implanted device were calculated for each device with the use of a Weibull model in which the shape parameter was the same for both devices and for all analysis sites (participants from study sites that had numbers of participants below a certain threshold were pooled into analysis sites; see the protocol). The scale parameter for the model depended on device as a fixed effect and on the combinations of device and analysis site as random interaction effects in comparisons between the study group and the control group, with a one-sided alpha level of We determined that noninferiority would be established if the one-sided upper confidence limit for the difference in proportions was lower than the noninferiority margin. The principal analysis of the primary end point was conducted in the intention-to-treat population; additional analyses involved the per-protocol and as-treated populations (Fig. S1 and Table S5 in the Supplementary Appendix). Because the design of the device was modified during the trial by sintering of the inflow cannula as discussed above, a separate analysis of the primary end point was conducted in the subgroup of patients who had received a sintered device. Fisher s exact test was used to compare the prevalence of events between the treatment groups. Survival analysis was performed with the use of the Kaplan Meier method. P values of less than n engl j med 376;5 nejm.org February 2,

4 The n e w e ng l a n d j o u r na l of m e dic i n e A Study Device Centrifugal-Flow Pump Blood flow from left ventricle Aorta Short inflow cannula Left ventricle Outflow graft Heart Motor Pericardial sac Diaphragm Pump housing Percutaneous drive line Centrifugal-flow pump designed for intrapericardial placement Percutaneous drive line connects to external battery pack and controller Blood flow to aorta Magnetic hydrodynamically levitated impeller 40mm B Control Device Axial-Flow Pump Blood flow to aorta Outlet stator and diffuser Rotor bearing Inlet stator and blood-flow straightener Motor Aorta Rotor bearing Left ventricle Outflow graft Pump housing Heart Pericardial sac Percutaneous drive line Inflow cannula Diaphragm Percutaneous drive line connects to external battery pack and controller 454 Rotor Axial-flow pump designed for intraabdominal placement n engl j med 376;5 nejm.org February 2, 2017 Blood flow from left ventricle

5 Intrapericardial Left Ventricular Assist Device Figure 1 (facing page). Study Device and Control Device. Panel A shows the study device, which is a compact centrifugal-flow left ventricular assist device (LVAD) with an integrated inflow cannula designed for intrapericardial placement. The device incorporates a bearingless design with magnetic and hydrodynamic levitation of the internal impeller. Panel B shows the control device, which is an axial-flow LVAD that requires bearing support of the internal impeller and is implanted outside the pericardium in a preperitoneal (pump) pocket were considered to indicate statistical significance. All safety assessments were conducted in the as-treated population (Fig. S1 and Table S5 in the Supplementary Appendix). P values were not adjusted for multiple testing. All P values are two-sided, with the exception of the noninferiority P value for the primary end point, which is one-sided. Additional information regarding the statistical analysis is provided in the Additional Statistical Analyses section in the Supplementary Appendix. Results Randomization and Patient Characteristics From August 2010 through May 2012, we screened 559 patients. Of these patients, 113 did not meet entry criteria, most commonly because of large body habitus, an unwillingness to adhere to study requirements, or a left ventricular ejection fraction greater than 25%. The remaining 446 patients were randomly assigned to receive the study device or the control device (Fig. S1 in the Supplementary Appendix). The intention-to-treat population consisted of 297 patients randomly assigned to the study device and 148 patients randomly assigned to the control device; 1 patient who had been assigned to the control device did not receive any device. Four patients who had been assigned to the study device received the control device, and 3 patients who had been assigned to the control device received the study device. Follow-up at 2 years was completed for 99.6% of the patients. The patients in the trial had advanced heart failure associated with a substantial risk of death, as evidenced by a high prevalence of ischemic heart disease, abnormal renal function with hyponatremia, ejection fractions lower than 20%, and requirement for intravenous inotropic support (Table 1, and Table S6 in the Supplementary Appendix). The reasons for the patients ineligibility for transplantation are provided in Table S7 in the Supplementary Appendix. On average, the control group was older, and the study group had a higher prevalence of severe tricuspid insufficiency. The groups did not differ significantly with respect to severity of illness, baseline hemodynamic characteristics (Table S6 in the Supplementary Appendix), or treatment with evidence-based therapy for heart failure at the time of enrollment. Similar rates of tricuspid valve repair or replacement occurred among participants who received the study device and those who received the control device (11.1% and 12.1%, respectively) (Table S8 in the Supplementary Appendix). The mean (±SD) length of the initial hospital stay was 26.0±20.5 days in the study group and 25.0±15.6 days in the control group. The device settings at implantation, discharge, and follow-up are provided in Table S9 in the Supplementary Appendix. Primary End Point The analysis of the primary end point in the intention-to-treat population showed that the study device was noninferior to the control device (estimated success rates, 55.4% and 59.1%, respectively, on the basis of the Weibull model; absolute difference, 3.7 percentage points; upper one-sided 95% confidence limit, percentage points; P = 0.01 for noninferiority). The results of additional analyses of the primary end point in the per-protocol and as-treated populations were consistent with those in the intention-to-treat population (Tables S10 and S11 in the Supplementary Appendix). The Kaplan Meier estimates of survival free from disabling stroke or need for device replacement at 2 years were 55.0% with the study device and 57.4% with the control device (P = 0.67 by the log-rank test) (Fig. 2A). A binary analysis of the primary end point in the as-treated population showed that the rate of overall success and failure was similar in the study group and the control group (Table 2). With regard to the individual components of the primary end point, there were no significant between-group differences in the number of deaths, participants with disabling stroke, or imputed failures. There were significantly more participants with device malfunction or failure requir- n engl j med 376;5 nejm.org February 2,

6 The new england journal of medicine Table 1. Baseline Characteristics of Patients in the Intention-to-Treat Population.* Characteristic Study Group (N = 297) Control Group (N = 148) Age yr 63.9± ±10.2 Male sex no. (%) 227 (76.4) 122 (82.4) Race no. (%) White 228 (76.8) 115 (77.7) Black 66 (22.2) 32 (21.6) Other 3 (1.0) 1 (0.7) Height cm 173.8± ±9.1 Weight kg 82.1± ±18.4 Body-surface area m 2 2.0± ±0.3 INTERMACS profile no. (%) 1 10 (3.4) 5 (3.4) 2 86 (29.0) 46 (31.1) (40.4) 60 (40.5) 4 59 (19.9) 27 (18.2) (7.4) 10 (6.8) Ischemic cause of heart failure no. (%) 172 (57.9) 89 (60.1) Previous stroke or transient ischemic attack no. (%) 57 (19.2) 24 (16.2) Diabetes no. (%) 132 (44.4) 65 (43.9) Hypertension requiring medication no. (%) 194 (65.3) 105 (70.9) Serum creatinine mg/dl 1.5± ±0.5 Serum sodium mmol/liter 135.5± ±4.8 Hematocrit % 35.1± ±5.1 Platelet count per mm 3 200,400±71, ,400±75,700 Albumin g/dl 3.5± ±0.53 Severe tricuspid valve insufficiency no. (%) 35 (11.8) 8 (5.4) Left ventricular ejection fraction % 17.1± ±4.8 Concurrent medication or intervention no. (%) Intravenous inotrope 209 (70.4) 108 (73.0) Diuretic 239 (80.5) 121 (81.8) Angiotensin-converting enzyme inhibitor 93 (31.3) 41 (27.7) Beta-blocker 160 (53.9) 85 (57.4) Internal cardiac defibrillator 254 (85.5) 135 (91.2) Cardiac resynchronization therapy 83 (27.9) 37 (25.0) * Plus minus values are means ±SD. There were no significant differences between groups in any characteristic with the exception of age (P = 0.04) and the prevalence of severe tricuspid valve insufficiency (P = 0.04). To convert the values for serum creatinine to micromoles per liter, multiply by Race was self-reported. Interagency Registry for Mechanical Circulatory Support (INTERMACS) profiles range from 1 to 7; a profile of 1 represents the most severe illness, and a profile of 7 the least severe illness. Additional details are provided in Table S3 in the Supplementary Appendix. 456 n engl j med 376;5 nejm.org February 2, 2017

7 Intrapericardial Left Ventricular Assist Device ing exchange, explantation, or urgent transplantation in the control group. The primary end point was also assessed for participants who received the sintered study device in the as-treated population (Fig. S3 in the Supplementary Appendix). The sintered device was found to be noninferior to the control device in this analysis. A comparison of outcomes of death and device exchange among participants who received either the sintered or the nonsintered study device is provided in Table S12 in the Supplementary Appendix. Adverse Events and Overall Survival The rates of major bleeding, cardiac arrhythmias, renal dysfunction, and infections, including percutaneous drive-line infections, were similar in the study group and the control group (Table 3, and Table S13 in the Supplementary Appendix). Significantly more participants in the study group than in the control group had ischemic or hemorrhagic stroke (29.7% vs. 12.1%, P<0.001). Modified Rankin scale data for strokes, including shifts in scores from 0 weeks to 12 weeks to 24 weeks after the stroke, are provided in Table S14 in the Supplementary Appendix. Most strokes in the study group occurred in the first 6 months (Fig. S4 in the Supplementary Appendix). A post hoc analysis of participants who received the study device revealed that mean arterial blood pressure measurements of 90 mm Hg or lower after implantation were associated with a lower frequency of strokes, particularly hemorrhagic strokes (Fig. S5 in the Supplementary Appendix). Participants in the study group had a higher rate of right heart failure; however, the need for support with a right ventricular assist device (RVAD) after LVAD implantation was similar in the two groups. Participants in the study group had more sepsis events than did patients in the control group. Overall mortality rates did not differ significantly between the two groups; the Kaplan Meier rate of overall survival at 2 years was 60.2% in the study group and 67.6% in the control group (P = 0.17 by the log-rank test) (Fig. 2B). The causes of death are provided in Table S15 in the Supplementary Appendix. Functional Status and Quality of Life After LVAD implantation, the functional status of approximately 80% of the participants in both groups improved to NYHA class I or II. The mean A B Event-free Survival Rate (%) No. at Risk Study group Control group Overall Survival Rate (%) No. at Risk Study group Control group P=0.67 by log-rank test P=0.01 for noninferiority by Weibull model Days P=0.17 by log-rank test Study group Control group Figure 2. Kaplan Meier Survival Curves. Panel A shows event-free survival (the primary end point) in the intentionto-treat population. The P value for noninferiority is significant, indicating the noninferiority of the study device to the control device. The dashed line indicates the time of assessment of the primary end point. Panel B shows overall survival in the as-treated population. The log-rank P value is nonsignificant, indicating that there was no significant difference in survival between the study and control groups baseline 6-minute walk distance was 100.2± m in the study group and 91.9±125.3 m in the control group. At 3 months after LVAD implantation, the mean 6-minute walk distance increased to 199.4±183.4 m in the study group and to 190.1± m in the control group. These improvements were sustained through 2 years (Fig. S6 in the Supplementary Appendix). At baseline, participants had poor quality of life and health status, as assessed by the KCCQ and the EQ-5D VAS (Fig. S7 in the Supplementary Ap Days Control group Study group n engl j med 376;5 nejm.org February 2,

8 The new england journal of medicine Table 2. Binary Analysis of the Primary End Point and Its Components.* End Point or Component Study Group (N = 297) no. of patients (%) pendix). At 3 months, the mean KCCQ score had improved by 25.8 points in the study group and by 25.3 points in the control group. At 3 months, the mean EQ-5D VAS improved by 22.5 points in the study group and by 25.5 points in the control group. Improvements in the KCCQ and EQ-5D VAS were sustained during the follow-up period. Discussion Control Group (N = 148) Success 164 (55.2) 85 (57.4) Failure 133 (44.8) 63 (42.6) Death 103 (34.7) 39 (26.4) Device malfunction or failure 26 (8.8) 24 (16.2) requiring exchange, explantation, or urgent transplantation Exchange 23 (7.7) 20 (13.5) Explantation 0 1 (0.7) Urgent transplantation 3 (1.0) 3 (2.0) Disabling stroke 3 (1.0) 0 Imputed failure 1 (0.3) 0 * A post hoc statistical analysis of success versus failure with the study device as compared with the control device revealed a P value of The P values for comparisons between the study device and the control device with regard to the reasons for failure were as follows: death, P = 0.08; device malfunction or failure requiring exchange, explantation, or urgent transplantation, P = 0.03; disabling stroke, P = 0.55; and imputed failure, P>0.99. Of the device exchanges in the study group, 8 of 97 (8.2%) were from patients with nonsintered pumps and 15 of 200 (7.5%) were from patients with sintered pumps. A disabling stroke was defined as a score of 4 or higher on the modified Rankin scale at 24 weeks (scores range from 0 to 6, with 0 indicating no symptoms and 6 indicating death). The patient had a stroke before the 2-year end point and died after the 2-year end point but before the 24-week post-stroke assessment on the modified Rankin scale. In the ENDURANCE trial, participants with advanced heart failure who were ineligible for transplantation were randomly assigned to either a centrifugal-flow LVAD or an axial-flow LVAD. The centrifugal-flow device was found to be noninferior to the axial-flow device with respect to survival free from disabling stroke or need for device replacement. An analysis of the components of the primary end point showed no differences between the study group and the control group with regard to death, disabling strokes, or imputed study failures. Participants who received the control device were more likely to require device replacement, explantation, or urgent transplantation. The study device was associated with more strokes, right heart failure, and sepsis. Treatment with either device was associated with sustained improvements in functional and quality-of-life measures. The rate of stroke was higher in the study group than in the control group. Most strokes occurred within the first 6 months, after which the rate of stroke decreased. In a post hoc analysis, we found that a recorded mean arterial blood pressure of higher than 90 mm Hg was associated with a risk of stroke predominantly hemorrhagic strokes that was 34% higher than that among patients whose blood pressure was 90 mm Hg or lower. Our data are consistent with previous assessments of the centrifugal-flow device, in which an association was found between elevated blood pressure and stroke. 16 These findings suggest that management strategies, including early postoperative management, may influence the stroke risk associated with the study device. Patients who received the control device had a higher rate of device replacement, explantation, or need for urgent transplantation for device malfunction than did patients who received the study device. An increase in the risk of LVAD thrombosis after implantation of the control device has previously been reported in retrospective observational studies. 3,4,19-21 There was an expectation that the bearingless design of the study device would lower the rate of pump thrombosis by removing the bearing as a potential nidus for thrombus formation. In addition, the modifications in device design that were made during the trial were found to reduce the rate of pump thrombosis in another study of the study device, the ADVANCE trial. 9,17 However, no significant difference between the two devices in the rate of pump exchange because of device thrombosis was found in ENDURANCE. This trial had several limitations. First, guidelines for anticoagulation and antiplatelet management were provided to investigators, but the investigators individualized these recommendations. The antithrombotic regimen may have affected the incidence of adverse events. Second, 458 n engl j med 376;5 nejm.org February 2, 2017

9 Intrapericardial Left Ventricular Assist Device Table 3. Adverse Events in the As-Treated Population.* Event Study Group (N = 296) Control Group (N = 149) P Value no. of patients (%) no. of events events/ patient-yr no. of patients (%) no. of events events/ patient-yr Bleeding events 178 (60.1) (60.4) >0.99 Requiring reoperation 45 (15.2) (18.1) Requiring transfusion of >4 units of 45 (15.2) (22.1) packed red cells within 7 days Gastrointestinal bleeding 104 (35.1) (34.2) Cardiac arrhythmia 112 (37.8) (40.9) Hepatic dysfunction 14 (4.7) (8.1) Hypertension 47 (15.9) (16.8) Sepsis 70 (23.6) (15.4) Drive-line exit-site infection 58 (19.6) (15.4) Stroke 88 (29.7) (12.1) <0.001 Ischemic cerebrovascular event 52 (17.6) (8.1) Hemorrhagic cerebrovascular event 44 (14.9) (4.0) <0.001 Transient ischemic attack 25 (8.4) (4.7) Renal dysfunction 44 (14.9) (12.1) Respiratory dysfunction 86 (29.1) (25.5) Right heart failure 114 (38.5) (26.8) Need for RVAD 8 (2.7) (3.4) Pump replacement 23 (7.8) NA NA 20 (13.4) NA NA 0.06 Exchange owing to pump thrombosis 19 (6.4) NA NA 16 (10.7) NA NA 0.12 Device malfunction or failure 93 (31.4) (25.5) Rehospitalization 249 (84.1) (79.2) Death 116 (39.2) NA NA 48 (32.2) NA NA 0.18 * For the study device, patient-years were evaluated; for the control device, patient-years were evaluated. Death and device thrombosis are not INTERMACS-defined events. The following adverse-event subcategories are reported as additional information for purposes of clarification: bleeding requiring reoperation, bleeding requiring transfusion, and gastrointestinal bleeding under the bleeding category; ischemic cerebrovascular event, hemorrhagic cerebrovascular event, and transient ischemic attack under stroke; and need for a right ventricular assist device (RVAD) under right heart failure. Major bleeding, infection, and death were considered to be both adverse events and secondary outcomes in the trial (details are provided in the Additional Statistical Analyses section in the Supplementary Appendix). P values are for comparisons of the percentage of patients with each event, as determined by the Fisher s exact test proportion method; they are post hoc and are included for information purposes only. Events in this category were site-reported rather than adjudicated by the clinical events committee. This category includes transient ischemic attacks that were resolved in less than 24 hours. Events and events per patient-year are not applicable (NA) because only the first event is counted. a change in study-device design that included sintering of the inflow cannula and a modified coring tool for the ventricle, along with a change in anticoagulation and antiplatelet therapy recommendations, occurred concurrently after the first third of the patients had been enrolled in the trial. This may have affected the outcomes for the latter two thirds of the patients in the study group. Finally, this trial did not address the durability of the devices beyond 2 years, and longer-term data would be useful given that, for patients who are not eligible for transplantation, lifelong LVAD support is required. In summary, in this trial, patients with advanced heart failure who were ineligible for transplantation were randomly assigned to receive either the HeartMate II axial-flow LVAD or the HeartWare centrifugal-flow device, which is n engl j med 376;5 nejm.org February 2,

10 Intrapericardial Left Ventricular Assist Device a smaller intrapericardial LVAD. The centrifugalflow device was found to be noninferior to the axial-flow device with respect to survival free from disabling stroke or need for device replacement. Use of the study device was associated with higher risks of stroke, right heart failure, and sepsis, whereas use of the control device was associated with more frequent device malfunction or failure requiring surgical intervention. Supported by HeartWare. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. References 1. Go AS, Mozaffarian D, Roger VL, et al. Executive summary: heart disease and stroke statistics 2014 update: a report from the American Heart Association. Circulation 2014; 129: Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009; 361: Starling RC, Moazami N, Silvestry SC, et al. Unexpected abrupt increase in left ventricular assist device thrombosis. N Engl J Med 2014; 370: Kirklin JK, Naftel DC, Kormos RL, et al. Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) analysis of pump thrombosis in the HeartMate II left ventricular assist device. J Heart Lung Transplant 2014; 33: Aaronson KD, Slaughter MS, Miller LW, et al. Use of an intrapericardial, continuous-flow, centrifugal pump in patients awaiting heart transplantation. Circulation 2012; 125: Larose JA, Tamez D, Ashenuga M, Reyes C. Design concepts and principle of operation of the HeartWare ventricular assist system. ASAIO J 2010; 56: Interagency Registry for Mechanically Assisted Circulatory Support. Appendix O profiles of advanced heart failure: patient profile at time of implant. Birmingham: University of Alabama School of Medicine, 2016 ( / medicine/ intermacs/ appendices/ app-o-5-0). 8. Wieselthaler GM, O Driscoll G, Jansz P, Khaghani A, Strueber M. Initial clinical experience with a novel left ventricular assist device with a magnetically levitated rotor in a multi-institutional trial. J Heart Lung Transplant 2010; 29: Slaughter MS, Pagani FD, McGee EC, et al. HeartWare ventricular assist system for bridge to transplant: combined results of the bridge to transplant and continued access protocol trial. J Heart Lung Transplant 2013; 32: Slaughter MS. Implantation of the HeartWare left ventricular assist device. Semin Thorac Cardiovasc Surg 2011; 23: Park SJ, Milano CA, Tatooles AJ, et al. Outcomes in advanced heart failure patients with left ventricular assist devices for destination therapy. Circ Heart Fail 2012; 5: Jorde UP, Kushwaha SS, Tatooles AJ, et al. Results of the destination therapy post-food and Drug Administration approval study with a continuous flow left ventricular assist device: a prospective study using the INTERMACS registry (Interagency Registry for Mechanically Assisted Circulatory Support). J Am Coll Cardiol 2014; 63: Miller LW, Pagani FD, Russell SD, et al. Use of a continuous-flow device in patients awaiting heart transplantation. N Engl J Med 2007; 357: Pagani FD, Miller LW, Russell SD, et al. Extended mechanical circulatory support with a continuous-flow rotary left ventricular assist device. J Am Coll Cardiol 2009; 54: Starling RC, Naka Y, Boyle AJ, et al. Results of the post-u.s. Food and Drug Administration-approval study with a continuous flow left ventricular assist device as a bridge to heart transplantation: a prospective study using the INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support). J Am Coll Cardiol 2011; 57: Teuteberg JJ, Slaughter MS, Rogers JG, et al. The HVAD left ventricular assist device: risk factors for neurological events and risk mitigation strategies. JACC Heart Fail 2015; 3: Najjar SS, Slaughter MS, Pagani FD, et al. An analysis of pump thrombus events in patients in the HeartWare ADVANCE bridge to transplant and continued access protocol trial. J Heart Lung Transplant 2014; 33: Interagency Registry for Mechanically Assisted Circulatory Support. Appendix A adverse event definitions. Birmingham: University of Alabama School of Medicine, 2013 ( / medicine/ intermacs/ protocol-3-0 -archive-pedimacs1/ appendices -pedimacs/ appendix-a-pedimacs). 19. Smedira NG, Blackstone EH, Ehrlinger J, et al. Current risks of HeartMate II pump thrombosis: non-parametric analysis of Interagency Registry for Mechanically Assisted Circulatory Support data. J Heart Lung Transplant 2015; 34: Kirklin JK, Naftel DC, Pagani FD, et al. Pump thrombosis in the Thoratec HeartMate II device: an update analysis of the INTERMACS Registry. J Heart Lung Transplant 2015; 34: Jeffries N, Miller MA, Taddei-Peters WC, Burke C, Baldwin JT, Young JB. What is the truth behind pump thrombosis in the HeartMate II device? A National Heart, Lung, and Blood Institute perspective based on data from the Interagency Registry for Mechanically Assisted Circulatory Support. J Heart Lung Transplant 2015; 34: Copyright 2017 Massachusetts Medical Society. 460 n engl j med 376;5 nejm.org February 2, 2017