Disclosure
Dr. Russell is a consultant for Medtronic and serves on the Data and Safety Monitoring Board for the HeartMate 3 trial sponsored by Abbott. Dr. Teuteberg has industry relationships with Medtronic (advisory board, speaking), Abiomed (advisory board), CareDx (advisory board, speaking), Abbott (HeartMate 3 Clinical Events Committee), and EcoR1 (consulting).
Overview of clinical trials of mechanical circulatory support devices
Clinical trials of mechanical circulatory support devices evolved from simple studies evaluating the efficacy of the device in a few patients to large-scale clinical trials with composite endpoints including quality of life, improvements in functional capacity, and freedom from major complications in addition to mortality. Mechanical circulatory support devices were first used clinically in 1963 when Dr. Michael DeBakey placed a device in the left thorax, connected to the left atrium and the descending thoracic aorta, and used for a patient who had a cardiac arrest 18 hours after an aortic valve replacement. The patient had clinical improvement in pulmonary edema and neurologic function, but unfortunately, he remained anuric and died 4 days later despite the continued function of the ventricular system device. In 1975, based on case reports of similar patients, two short-term trials were initiated for patients who could not be weaned from cardiopulmonary bypass. The trials were performed with a pneumatically powered single-chamber pump that could be placed either intraabdominally or extracorporeally. One of these case series included a 21-year-old man who successfully underwent heart transplantation after 6 days of support, which encouraged the field to move forward.
In 1976, the US Congress passed the Medical Device Amendments Act, which gave the US Food and Drug Administration (FDA) the authority to preapprove the use of medical devices in the United States. The FDA’s influence on clinical testing of these devices began to shape the early patient populations studied in clinical trials and the endpoints being evaluated. In 1977, the National Heart, Lung, and Blood Institute requested grant applications for the development of mechanical, electrically powered blood pumps. A second application request was issued in 1980 for long-term (2 years) left ventricular assist devices (LVADs). This support resulted in the development of the Novacor LVAD and the Thoratec LVAD, which eventually became the HeartMate XVE. Because of FDA concerns about the durability of these devices, initial clinical trials could enroll only patients who were eligible for cardiac transplantation. As devices demonstrated the ability to provide long-term support, the patient populations studied expanded to include non–transplant eligible patients. These early clinical trial populations were the source of the current indications for mechanical support: bridge to recovery, bridge to transplantation (BTT), and destination therapy (DT).
Since the mid-1980s, there have been changes in pump design from paracorporeal pumps to intracorporeal configurations and from pulsatile to continuous-flow designs. In this chapter, we review the major clinical trials that have been performed that have propelled the field forward for both BTT and DT.
Pulsatile flow
The first successful use of an LVAD as a BTT was reported by Hill in 1986 describing the use of the pneumatically actuated Thoratec Pierce-Donachy paracorporeal ventricular assist device (PVAD) ( Fig. 15.1 ) in a patient after a large myocardial infarction with cardiogenic shock. This was soon followed by a large series of patients successfully bridged to cardiac transplant in 1988. Twenty-nine patients underwent placement of either left or left and right (14 patients) ventricular assist devices in a paracorporeal position. Twenty-one of the patients survived to transplantation after a median of 3 days of support. For those who did not survive to transplantation, the two most common causes of death were multiorgan failure and bleeding. It should be emphasized that all 29 patients had severe cardiogenic shock preoperatively, requiring multiple vasopressors, an intraaortic balloon pump ( n = 25), and mechanical ventilation ( n = 21). Based on these early reports, LVADs were approved for BTT in 1995.
Some of the drawbacks of the PVAD were its large size, need for the patient to remain hospitalized while supported, two large percutaneous cannulas, external pumping chamber, and patient discomfort. Given these limitations, an implantable version of the device was developed (Thoratec Implantable Ventricular Assist Device; Fig. 15.2 ). Between 2001 and 2004, 39 patients underwent implantation of this device as either BTT or for postcardiotomy failure. Twenty-four of the patients had the device as left ventricular support and 15 as biventricular support, for a mean duration of 101 days. Eighteen patients were discharged from the hospital with ongoing support for a mean duration of 96 days. No survival difference was observed between the BTT and the postcardiotomy groups, and 69% of the patients were either transplanted or weaned from the device. The adverse effects of the device were quite similar to the PVAD and included infection in 22 patients, bleeding in 18 patients, and neurologic events in 14 patients. The incidence of ischemic or hemorrhagic stroke was 7.7%, which was somewhat lower than the 12.0% in the PVAD population.
With these early successes, pump design moved to the development of long-term, left ventricular only support with intracorporeal devices. The pumps that came to clinical trial included the HeartMate IP1000 (Thoratec Corporation), HeartMate VE/XVE (Thoratec Corporation), and Novacor LVAD (World Heart Corporation).
HeartMate IP1000
In 1992, Frazier reported on 34 patients who received the HeartMate IP1000 ( Fig. 15.3 ) as BTT and compared them to six control patients who met the enrollment criteria but did not receive the device. Sixty-five percent of the patients receiving the device went on to transplantation, with a posttransplantation survival of 80%. This compared to a 50% survival to transplantation in the control arm, with no patients surviving posttransplantation. Complications of the device included bleeding in 39%, infection in 25%, and right heart failure in 21%. No thromboembolic events were reported, with the maximum time of support being 324 days. A larger study of 75 patients compared to 33 control patients was reported in 1994. The survival rate to transplantation with the device was 71% compared to 36% in the controls. The adverse event profile was similar to the prior trial, with the exception of three patients having a thromboembolic complication: two with strokes and one with an embolus to the subclavian artery, which was successfully removed. This trial required a preimplantation pulmonary capillary wedge pressure > 20 mm Hg with either a cardiac index < 2.0 L/min/m 2 or a systolic blood pressure < 80 mm Hg. In spite of these promising early results, this pump was short lived due to the development of the HeartMate VE and XVE, which had a similar pump housing but were driven by a mechanical rotor rather than pneumatically. This development allowed patients to be discharged from the hospital and have greater mobility because the pump could be powered by two smaller wearable external batteries rather than a large pneumatic console.
HeartMate VE/XVE
The HeartMate VE/XVE ( Fig. 15.4 ) was studied in a prospective, multicenter trial of 280 transplant candidates who had poor hemodynamics despite being inotrope dependent. Forty-nine percent of the patients also required a balloon pump and had a mean pulmonary capillary wedge pressure of 27 mm Hg with a systolic blood pressure of 76 mm Hg and a cardiac index of 1.67 L/min/m 2 . The outcomes were compared to a control group of 48 patients who had higher blood pressure and cardiac index. The average pump support was 112 days, with 54 patients supported for over 180 days. Adverse events included bleeding in 11%, infection in 40%, neurologic dysfunction in 5%, and thromboembolic events in 6%. Only 29% of the device-supported patients died before transplantation, compared to 67% of the controls. The other highlight of this trial was that 115 of the patients were discharged from the hospital after a median of 82 days, and 86% of those patients received a transplant. This was the first major study that demonstrated that patients could be supported and successfully discharged home to await cardiac transplantation and introduced the possibility of long-term mechanical support for patients who were not eligible for cardiac transplantation, so-called destination therapy (DT).
The first trial of DT was the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) study, which used the Thoratec VE pump. This study was supported by the National Institutes of Health and was performed at 20 experienced transplant centers. Enrollment criteria included the presence of New York Heart Association (NYHA) functional class IV symptoms for at least 90 days despite optimal medical therapy, an ejection fraction (EF) < 25%, and a peak oxygen consumption of < 12 mL/kg/min or the need for continuous intravenous inotropic therapy due to symptomatic hypotension, decreasing renal function, or worsening pulmonary congestion. Due to slow enrollment, the criteria were later eased to include patients with 60 days of NYHA class IV symptoms and a peak oxygen consumption < 14 mL/kg/min or patients who were class III/IV for 28 days plus either 14 days of either intraaortic balloon or inotropic support with at least 2 failed weaning attempts. A total of 129 patients were enrolled and randomized to either placement of an LVAD or continued medical therapy. A statistically significant improvement in survival was noted with LVAD therapy at both 1 year (52% (vs.) 25%) and 2 years (23% (vs.) 8%) ( Fig. 15.5 ), and median survival was longer in the LVAD group, 408 days (vs.) 150 days with medical therapy. The LVAD group also had significant improvements in NYHA class and scored better on the physical-function and emotional-role subscales of the 36-item Medical Outcomes Study Short-Form General Health Survey (SF-36) and the Beck Depression Inventory. The Minnesota Living With Heart Failure score improved in the LVAD arm but not significantly ( P = 0.11). As expected, patients in the LVAD arm had more complications, including sepsis, LVAD failure, and neurologic complications. However, the tradeoff was less mortality due to heart failure and an increase in days spent out of the hospital. This trial resulted in FDA approval of the HeartMate XVE for DT.
Improved surgical technique and enhancement of the pump design resulted in improved outcomes. Long et al. reported on 42 patients who were supported for a mean of 232 days, with a 1-year survival rate of 61% and with lower rates of adverse events and death due to sepsis. Despite this, there was not widespread clinical adoption of DT for a variety of reasons, including a lack of long-term durability, requiring device replacement in over 50% of the patients by 2 years ( Fig. 15.6 ).
Novacor
The other durable pulsatile pump that was developed and studied was the Novacor left ventricular assist system ( Fig. 15.7 ). Similar to the Thoratec device, it was initially used as bridge to transplant (BTT) in 1984 and was optimized over the years for improved outcomes. In contrast to the Thoratec device, it did not have a textured blood-contacting surface and instead used a smooth-surfaced polyurethane sac that required anticoagulation with warfarin. Registry data were published describing the outcomes in over 100 patients in a variety of US and European studies. For BTT patients, 77% of the patients were transplanted, with a mean of 80 days of support compared to 37% of the control patients. However, the stroke rate with the Novacor was higher than seen in the HeartMate XVE experience, although the device was more durable. During the 76-year cumulative support period, there were no pump failures, which set the scene for a DT trial.
The Investigation of Nontransplant-Eligible Patients who are Inotrope Dependent (INTrEPID) trial enrolled 55 patients to either the Novacor LVAD or medical therapy. All patients had to meet similar eligibility criteria to the HeartMate for enrollment, and the control group consisted of patients who chose not to undergo LVAD implantation, had a mechanical aortic valve, or did not have the financial resources to cover the cost of the device or follow-up. Enrollment criteria included inotrope-dependent patients with an EF < 25% and NYHA functional class IV symptoms for at least the last 3 months who were not candidates for transplant. Additionally, all patients had to have two unsuccessful weaning attempts from inotropes. The LVAD patients had improved survival at both 6 (46% (vs.) 22%) and 12 (27% (vs.) 11%) months. Unfortunately, 62% of the LVAD patients experienced a stroke or transient ischemic attack (TIA) compared to 11% of the medically managed patients.
There has been one randomized trial of the Novacor LVAD (vs.) HeartMate XVE. Forty BTT patients were randomized, and there were no significant differences in survival to transplantation. The mean duration of support was 235 ± 210 days for Novacor and 175 ± 175 days for the XVE. More neurologic complications occurred with the Novacor pump, and more infections and pump durability issues with the XVE. Survival to transplantation was 65% for Novacor and 60% for the XVE. Based on the high stroke rate with the Novacor device, further development was halted in 2007.
Continuous flow—axial
With the demonstration of safety and improved survival compared to medical therapy, pulsatile pumps became the mainstay of VAD therapy worldwide through the 1990s into the early 2000s. However, major limitations related to pump design, including poor durability, pump size which did not allow implantation into shorter and smaller patients, and a large percutaneous lead required to vent air resulted in the use of these pumps being mostly restricted to short-term BTT use. The introduction of continuous-flow designs which were smaller, more durable (due to fewer moving parts), and noiseless was a significant improvement in pump design and has resulted in an increase in the use of LVADs for both BTT and DT indications.
HeartMate II
The first continuous-flow pump approved by the FDA was the HeartMate II (Thoratec) ( Fig. 15.8 ). This axial flow rotary pump produces continuous flow to the patient at a pump speed of 8000–12,000 rpms. The first trial included 133 patients awaiting heart transplantation with no control group. The primary outcome was patient survival at 180 days with either continued support, cardiac recovery, or transplantation. At 24 hours after implantation, the cardiac index increased from 2.0 ± 0.6 to 2.8 ± 0.7 L/min/m 2 , with a decrease in pulmonary capillary wedge pressure from 36 ± 10 to 26 ± 7 mm Hg, showing that the pump was hemodynamically effective. Of the 133 patients, 100 (75%) reached the endpoint at 180 days, with 56 being transplanted, 43 on continued support, and 1 with recovery of function and device explantation. Thirty-three patients had an unsuccessful outcome, with 25 deaths, 5 patients removed from the transplant list due to irreversible complications, and 3 who underwent change to a different LVAD due to complications and were removed from the study. One hundred patients were discharged from the hospital, with a median of 60 days out of the hospital. Similar to the pulsatile pumps, the most common adverse events included bleeding in 31%, infection in 28%, and stroke in 8%. Causes of death included sepsis, ischemic stroke, multisystem organ failure, hemorrhagic stroke, anoxic brain injury, and right heart failure. Importantly, there was improvement in 6-minute walk distance and quality of life measured by both the Minnesota Living With Heart Failure and Kansas City Cardiomyopathy questionnaires. This study resulted in FDA approval of the device for BTT. A follow-up study was performed to evaluate longer-term outcomes with this device. A total of 469 patients were enrolled and 281 patients who completed study endpoints or had 18 months of follow-up were evaluated. At 18 months, 79% of the patients were either transplanted (55.8%), had ongoing LVAD support (20.6%), died (19.9%), or had device removal (3.5%) for either recovery or exchange to another pump. Survival for patients who remained on pump support was 73% at 1 year, which was a significant improvement in survival compared to pulsatile pumps. This study demonstrated the improved durability of continuous-flow pumps, resulting in improved survival with no obvious signals of untoward effects from the lack of pulsatility.
Similar to the pulsatile clinical trial experience, after approval for BTT, the HeartMate II was studied in a DT population. In contrast to a previous trial where the control arm was medical therapy, the landmark DT study randomized patients to either HeartMate II or the Thoratec XVE device in a 2:1 design. Patients ineligible for transplant, usually because of advanced age, were randomized with a primary endpoint of survival free from disabling stroke and reoperation to repair or replaced the device at 2 years. Two hundred patients were randomized with a mean age of 64 and a mean EF of 17% and almost 80% were on inotropic therapy. At 2 years, the survival with HeartMate II was 58% compared to 24% with the pulsatile device. Additionally, both groups showed an improvement in NYHA class, 6-minute walk distance, and quality of life questionnaires from baseline to 24 months. Compared to the pulsatile pump, there were significant reductions in the adverse event rate, with the exception of stroke, which was similar between the two pump groups (17% (vs.) 13%). The pump replacement rate was also much lower at 9% (vs.) 34% with the HeartMate XVE. Unfortunately, despite the improvement in most side effect profiles and quality of life, the HeartMate II did not result in a reduction in hospitalizations, with 94% of the patients compared to 96% of the patients rehospitalized during the study.
The impact of clinical experience being associated with improved outcomes was also demonstrated with the HeartMate II. Park and colleagues compared 281 patients who received the HeartMate II for DT to the initial 133 patients. There was a nonsignificant improvement in survival (73% (vs.) 68%, P = 0.21) at 1 year and a reduction in bleeding, sepsis, device-related infections, and hemorrhagic stroke.
To evaluate if the results of this study could be replicated in a nonclinical trial setting, the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) registry was used to assess outcomes in DT patients receiving the HeartMate II after FDA approval for DT. The first 247 patients who were enrolled in the registry were followed for 2 years and the outcomes were compared to those of the patients enrolled in the pivotal DT trial. This study demonstrated the benefits of improvement in clinical care with experience. Survival at 2 years increased to 62% from 58%, with a decrease in device related infections, postoperative bleeding, and length of stay. They also demonstrated that survival varied based on the status of the patient preoperatively, with non–inotrope-dependent patients having an 82% survival at 1 year and 69% survival at year 2.
Jarvik 2000
Other axial flow LVADs have been evaluated that are not currently approved in the United States. The Jarvik 2000 (Jarvik Heart) ( Fig. 15.9 ) pump has a pumping chamber located inside of the left ventricle, as opposed to a preperitoneal space with the HeartMate II, but the outflow graft is anastomosed to the descending thoracic aorta, in contrast to most LVADs. Frazier et al. reported the results of their first 10 implants in patients awaiting heart transplant. The device was implanted through a left thoracotomy during partial cardiopulmonary bypass. The mean support period was 84 days, and seven of the patients survived to transplantation. There was a significant improvement in hemodynamics, with an improvement in cardiac index from 1.77 to 3.45 L/min/m 2 and a reduction in pulmonary capillary wedge pressure from 21.7 to 10.3 mm Hg by 48 hours. The authors did not report on adverse events in this small trial.
A report on 83 BTT patients from the Japanese Registry for Mechanically Assisted Circulatory Support was published in 2018. The median support duration was 191 days, with a maximum of 758 days. Survival at 1 and 2 years was 85.0% and 79.3%, respectively. Most side effects were similar to other pumps, including a 27.7% incidence of major bleeding and 31.3% new infection. The stroke rate was high at 20.5%, with 18 hemorrhagic and 6 ischemic events. One patient developed pump thrombosis and required device exchange, and in another patient, thrombus was found at autopsy following a fatal stroke.
A clinical trial evaluating the efficacy of this pump in 150 patients has been completed, but no results have been published yet (NCT00591799). This company has also developed a driveline that is placed in a postauricular position, rather than in the upper abdomen, which they believe may reduce the incidence of serious driveline infections.
MicroMed DeBakey/HeartAssist 5
Another axial flow pump that has not yet been approved is the MicroMed DeBakey ventricular assist device, which currently is called the HeartAssist 5 (Reliant Heart) ( Fig. 15.10 ). This device requires a small preperitoneal pocket, and the outflow graft is attached to the ascending aorta. Some unique aspects of this pump are that it is small, weighing 84 g, and it has an ultrasound flow probe attached to the outflow graft to directly measure pump flow. Goldstein reported the outcomes of the first 150 patients who underwent placement of this pump as BTT between 1998 and 2002. The mean support time was 75 ± 81 days, and 55% of the patients were either bridged to transplantation, had successful explantation, or were on continued support at the end of the study. Unfortunately, 45% of the patients died, although the causes of death were not described. Major adverse events included reoperation for bleeding in 32%, hemolysis in 12%, thromboembolic events in 10.7%, and pump thrombus in 11.3%. This pump is currently approved for both BTT and DT in Europe. A clinical trial in the United States (NCT02205411) was terminated after the enrollment of 17 patients.
