Author + information
- Received November 5, 2013
- Revision received December 28, 2013
- Accepted January 28, 2014
- Published online May 6, 2014.
- Ulrich P. Jorde, MD∗∗ (, )
- Sudhir S. Kushwaha, MD†,
- Antone J. Tatooles, MD‡,
- Yoshifumi Naka, MD, PhD∗,
- Geetha Bhat, MD‡,
- James W. Long, MD, PhD§,
- Douglas A. Horstmanshof, MD§,
- Robert L. Kormos, MD‖,
- Jeffrey J. Teuteberg, MD‖,
- Mark S. Slaughter, MD¶,
- Emma J. Birks, MD¶,
- David J. Farrar, PhD#,
- Soon J. Park, MD†,
- HeartMate II Clinical Investigators
- ∗Columbia University Medical Center–New York Presbyterian Hospital, New York, New York
- †Mayo Clinic, Rochester, Minnesota
- ‡Advocate Christ Medical Center, Oak Lawn, Illinois
- §INTEGRIS Baptist Medical Center, Oklahoma City, Oklahoma
- ‖University of Pittsburgh, Pittsburgh, Pennsylvania
- ¶University of Louisville, Louisville, Kentucky
- #Thoratec Corporation, Pleasanton, California
- ↵∗Reprint requests and correspondence
: Dr. Ulrich P. Jorde, Division of Cardiology, Columbia University Medical Center, College of Physicians & Surgeons, 622 West 168th Street, PH 12 – Room 134, New York, New York 10032.
Objectives A post-approval (PA) study for destination therapy (DT) was required by the Food and Drug Administration (FDA) to determine whether results with the HeartMate (HM) II (Thoratec, Pleasanton, California) left ventricular assist device (LVAD) in a commercial setting were comparable to results during the DT multicenter pivotal clinical trial.
Background New device technology developed in the clinical research setting requires validation in a real-world setting.
Methods The PA study was a prospective evaluation of the first 247 HM II patients identified pre-operatively as eligible for DT in the national INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) registry. Patients were enrolled from January to September 2010 at 61 U.S. centers and followed for 2 years. A historical comparison group included patients (n = 133 at 34 centers) enrolled in the primary data cohort in the DT pivotal trial (TR). Survival rates and adverse events for the PA group were obtained from the INTERMACS registry.
Results Baseline characteristics were similar for PA versus TR. Forty-five percent of PA patients were in INTERMACS profiles 1 to 2 and 28% were in profile 3. Adverse events in the PA group were similar or lower than those in the TR group, including improvements in device-related infection (0.22 vs. 0.47) and post-operative bleeding requiring surgery (0.09 vs. 0.23) events per patient-year. Kaplan-Meier survival at 2 years was 62% (PA group) versus 58% (TR group). PA group survival at 1 and 2 years was 82 ± 5% and 69 ± 6% for INTERMACS profiles 4 to 7 (n = 63) versus 72 ± 3% and 60 ± 4% for profiles 1 to 3 (n = 184). The median length of stay after surgery was reduced by 6 days in the PA group versus the TR group.
Conclusions Results in a commercial patient care setting for the DT population supported the original pivotal clinical trial findings regarding the efficacy and risk profile of the HM II LVAD. Survival was best in patients who were not inotrope-dependent (INTERMACS profiles 4 to 7).
Continuous-flow (CF) left ventricular assist devices (CF-LVADs) have successfully replaced the previous generation of pulsatile flow devices for patients with advanced heart failure (1). Moreover, observational studies following initial approval for the bridge-to-transplant (BTT) indication by the Food and Drug Administration (FDA) have demonstrated excellent outcomes with the HeartMate (HM) II (Thoratec, Pleasanton, California) when used by clinicians in the commercial (i.e., outside of the clinical trial) setting (2). Of note, success for the BTT indication is generally defined as “alive on device or transplanted by 180 days,” and is by definition accomplished in patients who are candidates for cardiac transplantation (i.e., relatively young and without major comorbidities).
More recently, following a pivotal multicenter trial conducted from 2005 to 2010 (3), which had a primary endpoint of survival at 2 years on the original device and without disabling stroke, the HM II was also approved for destination therapy (DT). DT candidates are not eligible for cardiac transplantation, are typically older, and frequently have a higher burden of comorbidities. Management of this patient cohort is more complex, and results obtained in a strictly regulated clinical trial setting may not be achievable in clinical commercial use. Accordingly, and to fulfill requirements by the FDA for a post-approval (PA) study, we compared outcomes of the first HeartMate II DT patients in commercial use to DT patients in the pivotal trial (TR). We hypothesized that LVAD patients who underwent implantation for DT in commercial use would have comparable outcomes to the clinical trial.
Study design and patient population
The study was a prospective evaluation of the first 247 consecutive HM II patients who underwent implantation after FDA approval of the device and who were preoperatively identified for DT in INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support). Patients in the PA study group were enrolled at 61 centers from January to September 2010 and followed for 2 years after implantation. These patients were compared with a historical control group comprised of the primary data cohort of 133 patients implanted with the HM II in the multicenter clinical pivotal trial (TR) from March 2005 to May 2007 at 34 centers, which led to FDA approval of the device for DT (3). Patients with advanced heart failure who were ineligible for heart transplantation and who were refractory to optimal medical management were considered for enrollment. Details of the inclusion and exclusion criteria for the DT clinical trial have been published elsewhere (see the Online Appendix in Slaughter et al. ). All patients had a minimal clinical follow-up since implantation for at least 2 years or until they met the primary endpoint of the study.
For the TR group, data were retrieved from the original clinical trial database. For the PA group, data were obtained from the INTERMACS registry. Baseline demographics included age (in ranges as used by INTERMACS), sex, heart failure etiology, New York Heart Association class, history of stroke, body surface area, and weight. Other baseline data included left ventricular ejection fraction, hemodynamics (cardiac index), central venous pressure, pulmonary capillary wedge pressure, pulmonary vascular resistance, pulmonary arterial pressures, systemic blood pressure, laboratory values (creatinine, blood urea nitrogen, alanine aminotransferase, aspartate aminotransferase, total bilirubin, serum sodium), and baseline device or medical therapy (cardiac resynchronization therapy, implantable cardioverter-defibrillator, ventilator support, intra-aortic balloon pump, angiotensin-converting enzyme inhibitors, beta blockers, and inotropes). INTERMACS profiles were only determined for patients in the PA group, because the clinical trial preceded the introduction of INTERMACS profiles.
Follow-up after device implantation
Post-operative medical care (including inotropic, antiarrhythmic, anticoagulant, and heart failure therapy) was managed according to each investigator’s and/or clinician’s preference and usual practice.
The primary endpoint was survival at 2 years without re-operation to repair or replace the device and/or disabling stroke (Rankin scale >3). Secondary endpoints included frequency of adverse events, functional status, and quality-of-life assessments. Adverse events were determined for events that had equivalent definitions in both studies. Functional assessments and quality-of-life questionnaires were obtained at baseline when possible before LVAD implantation and at 3, 6, 12, 18, and 24 months. Functional status measurements included 6-min walk distances. Heart failure–related quality of life was assessed for the PA group using the European Quality of live 5-dimensional utility score, and for the TR group using the Minnesota Living with Heart Failure (MLWHF) and the Kansas City Cardiomyopathy questionnaires (KCCQ). Adverse events were recorded throughout the study until the analysis cutoff.
Data for the PA group were collected through the INTERMACS registry. Statistical analysis was performed by the sponsor, Thoratec Corporation. Differences between groups for independent, normally distributed, and continuous variables were evaluated using the t test. Variables that were not normally distributed were evaluated using the nonparametric Mann-Whitney U test. Differences in categorical variables were evaluated using Fisher exact test or the Pearson chi-square test for more than 2 groups. Survival analysis was performed using the Kaplan-Meier method, with patients censored for transplantation, recovery of native heart function with device removal, or withdrawal from the study. Comparison of survival between the 2 groups was performed using the log-rank test. Adverse events were presented as both percentages of patients and event rates (events per patient-year). Comparison of adverse event rates between the two groups were performed using Cochran-Mantel-Haenszel statistics. Adverse events in the PA group were compared with those for the TR group for events in which the definitions were equivalent. Changes for statistical significance over time for EQ-5D and 6-min walk tests were evaluated using linear mixed-effect modeling. All comparisons were 2-sided with the level of significance set at p < 0.05. Statistical analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, North Carolina).
Baseline parameters were characteristic of extremely ill patients with advanced heart failure and were similar between groups (Table 1). The percent of patients older than age 60 years was 70% and 62% for the PA and TR groups, respectively. INTERMACS profiles were only determined for patients in the PA group, of whom 47% were the most severely ill (profiles 1 and 2).
Median duration of support was 2.0 years (range 0 to 2.6) for the PA group and 1.7 years (longest 6.0) for the TR group, with a cumulative follow-up duration of 386 and 280 patient-years, respectively. Kaplan-Meier survival (Fig. 1) at 12 and 24 months for the PA group was 74 ± 3% and 61 ± 3% compared with the TR group of 68 ± 4% and 58 ± 4% (p = 0.2081; log-rank test). Survival for the PA group at 12 and 24 months for INTERMACS profiles 4 to 7 was 82 ± 5% and 67 ± 6% compared with INTERMACS profiles 1 to 3 of 71 ± 3% and 59 ± 4% (p = 0.179 log-rank test) (Fig. 2). The percent of patients who reached the endpoint of survival free from any stroke and re-operation to replace the device at 2 years in the PA group was 54% (135 of 247 patients) compared with 44% (58 of 133 patients) for the TR group (p = 0.042). Additional subgroup survival analysis showed no difference in patients younger than and older than age 60 years (Fig. 3), by sex (Fig. 4), or by race. The median length of initial hospital stay after LVAD implantation was 21 days in the PA group versus 27 days in the TR group.
Functional assessment and quality of life
Early and sustained improvements in quality of life were seen in the PA group using the EQ-5D visual analog scale and the EQ-5D total score (Fig. 5). By 3 months, the visual analog scale had increased approximately 30 points, from 40 to approximately 70, and remained stable through 24 months. Improvements were noticeable in all 5 components of the EQ-5D: activities, anxiety, mobility, pain, and self care (Fig. 6). Results for the trial group have been previously published using the KCCQ and MLWHF questionnaires, which demonstrated a similar degree of improvement over time (3,4).
Significant improvements in functional status over time as determined by the 6-min walk test were observed in both the PA and TR groups. Many patients were not able to walk before LVAD implantation. The 6-min walk distances for the 19% of the PA patients and 38% of the TR patients who walked at baseline were 183 ± 97 m and 182 ± 140 m for the PA and TR groups, respectively, and increased to 297 ± 118 m and 372 ± 191 m by 24 months.
A comparison of adverse events between the 2 groups is shown in Table 2. There were reductions or favorable trends in adverse events between the TR and the PA patients in most major categories, including bleeding, device-related infections, ischemic stroke, and overall pump replacement for all causes, whereas hemolysis was the only event that trended toward an increase. Figure 7 shows a significant improvement in the Kaplan-Meier estimate of survival free of stroke, device-related infections, or pump replacement.
We examined outcomes of the first 247 US patients who received the HM II LVAD as DT in commercial practice following FDA approval. We then compared these outcomes with those achieved in the initial clinical trial cohort to see whether clinical trial outcomes were successfully translated into commercial practice. Our principal findings were as follows:
First, dissemination of the HM II technology PA was associated with continued excellent results with 1- and 2-year survival of 74% and 61%, respectively. This survival outcome was independent of age, sex, or race, and not only matched, but demonstrated a favorable trend to that observed in the initial trial (1 year 68%; 2 year 58%).
Second, we observed a lower prevalence of serious adverse events, namely, bleeding, device-related infection, and stroke in the PA cohort.
Lastly, and because DT is currently being tested in less sick patients (INTERMACS profiles 4 to 7), it is of particular note that our survival results in this subgroup (1 year 82%; 2 year 67%) were the best in any multicenter cohort described to date.
Our results serve as a reminder that overall improvement in outcomes of advanced heart failure patients over the past decade has been nothing less than dramatic (Fig. 8): In REMATCH (Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure), 2-year survival of advanced heart failure patients randomized to medical therapy (mainly continuation of inotropic support) was 8% (5–7). Although the HM 1 outperformed medical therapy at that time, 2-year survival with this pulsatile device was only 23%. Our results, which showed 2-year survival of 61% in the overall cohort and 67% in less sick patients, clearly indicate that the hurdle of providing survival benefit has been more than cleared by currently available technology. Simultaneously, the bar for future destination therapy devices has risen from providing mainly a survival benefit to also improving quality of life with minimal adverse events. As such, careful patient selection is paramount and can be guided by risk scores specifically developed for patients receiving long-term mechanical circulatory support (8).
The reduction in stroke from 19% to 11.7 % (0.13 to 0.08 events per patient-year) we observed is particularly encouraging, because disabling stroke is arguably the adverse event with the highest negative impact on quality of life. Although it is unlikely that this complication can be entirely eliminated, we are hopeful that attention to anticoagulation therapy and other management practices may lead to a further reduction in the stroke rate. With regard to infectious complications, there are also encouraging trends, but transcutaneous energy transfer and the elimination of drivelines will likely provide the greatest reduction in rate of infection with future devices.
The observed reduction in the device replacement rate is likely a function of improved surgical technique and lower frequency of driveline infections because the reported rate of device thrombosis was unchanged. Although it is somewhat difficult to consolidate the increased rate of hemolysis with a steady rate of device thrombosis because these should go hand in hand (9), this discrepancy may simply underscore the wide variety in the clinical management of hemolysis and diagnosis of device thrombosis between centers (10). In this context, it is of particular note that the present data—including the hemolysis rate of 6.5% in the PA cohort on the basis of the previous INTERMACS definition (plasma free hemoglobin > 40 mg/dl)—was collected on device implantations performed before a reported increase in the prevalence of device thrombosis began in 2010 (11) or 2011 (12). Clearly, a prospective study of hemolysis and device thrombosis using strict definitions is needed, and newer generation devices and improved management practices should aim to eliminate both.
Our results in the real world should further encourage cardiologists who are familiar with the clinical trial results to discuss DT with their patients. Such discussions should entail the distinct survival benefit of CF mechanical circulatory support in advanced heart failure, without omitting the still highly significant adverse event burden, including the need for readmission for device-related complications (13).
The principal limitation of this prospective study is that data for the PA cohort were collected through INTERMACS. Adverse events were reported by treating physicians and not adjudicated by a clinical events committee. We do not believe that this form of data collection had any effect on our major endpoint (survival), but we cannot exclude underreporting of minor adverse events.
In summary, we report excellent survival in the first 247 patients treated in a commercial setting with the HM II LVAD as DT. Although randomization to medical therapy for ethical reasons is no longer possible in this setting, we believe that the results seen here likely exceed those that would have been seen with medical therapy by a survival difference of approximately 50% (absolute). Although it is clear that further improvement in survival should occur, and further reduction in serious adverse events must occur to make LVAD therapy the right choice for heart failure patients who are less sick, we believe that our data reflect the state of the art, and as such, set the bar for all future endeavors in circulatory support as DT.
The authors wish to acknowledge the valuable statistical analyses performed by Jerry Heatley, MS, and Mary McGarigle, MS.
Dr. Jorde has consulted for and received honoraria from the Thoratec Corporation, HeartWare, and Jarvik. Dr. Teuteberg has been consultant for HeartWare, Sunshine Heart; and XDx; and a member of the speaker's bureau for HeartWare. Dr. Horstmanshof has been a consultant and a member of the speaker's bureau for the Thoratec Corporation. Drs. Naka and Tatooles have been consultants for the Thoratec Corporation. Dr. Long is a consultant (educational activities) for Thoratec Corporation. Dr. Farrar is an employee of the Thoratec Corporation. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- continuous flow
- destination therapy
- Food and Drug Administration
- Kansas City Cardiomyopathy Questionnaire
- left ventricular assist device
- Minnesota Living with Heart Failure
- pivotal trial
- Received November 5, 2013.
- Revision received December 28, 2013.
- Accepted January 28, 2014.
- American College of Cardiology Foundation
- Starling R.C.,
- Naka Y.,
- Boyle A.J.,
- et al.
- Rogers J.G.,
- Aaronson K.D.,
- Boyle A.J.,
- et al.
- Park S.J.,
- Milano C.A.,
- Tatooles A.J.,
- et al.,
- HeartMate II Clinical Investigators
- Rogers J.G.,
- Butler J.,
- Lansman S.L.,
- et al.
- Cowger J.,
- Sundareswaran K.,
- Rogers J.G.,
- et al.
- Uriel N.,
- Morrison K.A.,
- Garan A.R.,
- et al.
- Hasin T.,
- Marmor Y.,
- Kremers W.,
- et al.