Durable Mechanical Circulatory Assist Devices for Heart Failure

A Golden Era of Mechanical Circulatory Support

Mechanical circulatory support (MCS) for the failing heart is now the mainstay of the modern management of patients with both acute and chronic heart failure (HF) refractory to pharmacologic and other usual interventions. In fact, the successes realized to date have been so significant that the main focus of this arena has now shifted away from simple survival to mitigation of risk and minimization of adverse events. Undeniably, continued advances in device technology have made this possible, but when coupled with analyses of the ever-growing patient management experience, better understandings exist regarding: optimal patient selection and timing of intervention, significant improvement in multiorgan function during the time spent on ventricular assist device (VAD) support, and preexisting and demographic risk factors resulting in complications.

Though some of the data taken from the experience with the first generation of pulsatile devices may no longer be applicable in the modern era of nonpulsatile support, the valuable lessons learned still strongly help shape management and clinical decision-making. All of these factors have now resulted in more widespread acceptance of VADs by physicians and patients as a management strategy. There is also earlier utilization of VADs in the course of a patient’s cardiac deterioration. Previously used “indications” for MCS (e.g., “bridge to recovery,” “bridge to transplantation” [BTT], etc.) are now simply termed “short-term use” and “long-term use.” Regardless of the terminology, MCS is currently employed for a variety of both short- and long-term modern indications including acute rescue of patients from acute low cardiac output situations (“bridge to immediate survival”); prevention of further myocardial damage following an ischemic event (“bridge to recovery”); prevention of deterioration in multisystem organ function; as a temporizing measure to buy time for recovery; as a bridge to the next step of management (“bridge to next decision,” “bridge to a bridge”); as a bridge to “improved candidacy” (for transplantation); and/or, in many cases, as a final management strategy for end-stage HF (“Destination Therapy” [DT]).

Specifically, for “long-term” MCS with durable, implantable left ventricular assist devices (LVADs), it was recently suggested that such terminology as “bridge to transplantation” and “destination therapy” be abandoned and replaced with a “single preimplant strategy ... to extend the survival and improve the quality of life of patients with medically refractory heart failure.” This recommendation came from a comprehensive analysis of data from the MOMENTUM 3 trial (discussed in detail later), which revealed no significant differences in outcomes when patients were stratified by previously used granular “indications” for LVAD implantation.

Equally important to the advances in MCS technology has been the formal sharing of outcomes data from centers nationwide through INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support), a North American registry database sponsored by the National Heart, Lung and Blood Institute, the US Food and Drug Administration (FDA), and the centers for Medicare and Medicaid Services. INTERMACS was established in 2005 for adult patients receiving long-term MCS device therapy to treat advanced HF. A similar database called EuroMACS exists in Europe, JMACS is the Japanese database, the United Kingdom maintains its own registry, and IMACS is an international database maintained by the International Society for Heart and Lung Transplantation (ISHLT). PEDIMACS collects pediatric MCS data, and MEDAMACS collects data about adult HF patients still being medically managed because they are not yet “sick enough” to receive MCS (see discussion of INTERMACS profiles later).

INTERMACS collects clinical data about patients implanted with durable MCS devices at 1 week, 1 month, 3 months, 6 months, and every 6 months thereafter. Major outcomes after implant (eg, death, explant, rehospitalization, and adverse events) are updated frequently and also as part of the defined follow-up intervals. Additional endpoints focus on the patient’s level of function and quality of life; the reported improvements in these areas have been compelling. These data have proven invaluable to appropriate risk stratification and patient selection, and, as new devices are introduced, documentation of complex outcomes beyond simple survival will clearly help to suggest “superiority” of one device over another.

The 2020 STS-INTERMACS annual report, released in 2021, summarizes the enrollment and outcomes of over 25,551 continuous flow (CF) LVAD patients implanted between January 1, 2010 and December 31, 2019 stratified according to era (2010–2014 and 2015–2019). An examination of this latest INTERMACS report reveals the dynamic and expanding landscape of modern MCS:

• Patient accrual includes a record number of 3198 CF LVAD implants in 2019 in the United States alone.

• Overall all-comer survival at 2 years with a durable CF LVAD device was approximately 71%. Three-year survival approached 61%, and 4-year survival approached 52%. DT survival was >80% at 1 year and 61% at 3 years.

Continued improvement in survival has shifted the focus to reduction of adverse events and repeated hospitalizations. Survival has improved dramatically in recent years, but it is affected by the timing of implantation in the course of a patient’s cardiac deterioration. Classification in this regard is denoted by INTERMACS “level,” a scale of clinical condition ranging from 1 to 7. An INTERMACS 7 patient is simply in the advanced stages of HF, with the clinical condition of the patient worsening as the INTERMACS profile number gets lower. An INTERMACS 4 patient has symptoms at rest, an INTERMACS 3 patient is hemodynamically stable but inotrope dependent, an INTERMACS 2 patient is clinically deteriorating with signs of end-organ dysfunction even with the use of inotropes, and an INTERMACS 1 patient is essentially in cardiogenic shock ( Box 22.1 ).

BOX 22.1

Intermacs Profiles

Level	Hemodynamic Status
1 “Crash and burn”	Persistent hypotension despite rapidly escalating inotropic support and eventually IABP, and critical organ hypoperfusion.
2 “Sliding on inotropes”	Intravenous inotropic support with acceptable values of blood pressure and continuing deterioration in nutrition, renal function, or fluid retention.
3 “Dependent stability”	Stability reached with mild-to-moderate doses of inotropes but demonstrating failure to wean from them due to hypotension, worsening symptoms, or progressive renal dysfunction.
4 “Frequent flyer”	Possible weaning of inotropes but experiencing recurrent relapses with symptoms at rest, usually fluid retention.
5 “Housebound”	Severe limited tolerance for activity: comfortable at rest with some volume overload and often with some renal dysfunction.
6 “Walking wounded”	Less severe limited tolerance for activity and lack of volume overload. Fatigue easily.
7 “Placeholder”	Patient without current or recent unstable fluid balance. NYHA class II or III.

IABP, Intraaortic balloon pump; NYHA, New York Heart Association.

Based on the collective outcome data contained within the INTERMACS registry, guidelines have been developed for device implantation. For early elective implantation of a durable LVAD at numerically higher INTERMACS levels (5–7), the risks of adverse events may outweigh the benefits. Conversely, waiting until the patient is at too low an INTERMACS level (1–2) with multisystem organ failure is associated with a low probability of rescue and poor survival. Consequently, at least in the United States, elective LVAD patients are being implanted with durable LVADs when at an INTERMACS level 3 (and in some cases 4), because this seems to be the best timing to balance the risks and benefits, and results in the best outcome.

Until 2009, MCS was used most often as a BTT, but DT has grown exponentially since 2010 when the HeartMate II (HM II) (Abbott, Abbott Park, IL) received approval as a DT device. INTERMACS data show that DT is now the most common utilization of MCS in the United States, accounting for 56.1% of all LVAD implants in the time period 2015 to 2019 (compared to 14.7% in 2006–2007). More patients were treated with temporary MCS before durable LVAD implantation in the later era than in the earlier one (36.8% as opposed to 26%). “Bridge to candidacy” (BTC) (for transplantation) continues to be the second most common modern indication for LVADs. In fact, the percentage of patients listed for transplantation at the time of LVAD implantation has decreased to 20.3% in 2015 to 2019 (compared to 24.2% in 2010–2014). “Bridge to recovery” and “bridge to next decision” with short-term VADs and/or extracorporeal membrane oxygenation (ECMO) currently constitute only a very small percentage of the overall usage of this technology by comparison to the long-term indications.

A revised US heart allocation system went into effect in October 2018 that has ushered in a significant change in the durable LVAD implantation strategy (see Chapter 19 ). Before 2018, approximately 50% of patients received an LVAD as DT, with an additional 25% used for BTT and bridge to therapy each. After 2018, 70% of LVADs are being implanted as DT and only a minority are being used as BTT or BTC.

Mechanical Circulatory Support: Theory and Practice

Cardiogenic shock can be defined as the inability of the heart to deliver sufficient blood flow to meet the metabolic requirements of the body despite the presence of adequate intravascular volume. Generally, cardiogenic shock entails sustained hypotension (systolic blood pressure [SBP] < 90 or 30 mm Hg below baseline), low cardiac output with high central filling pressures (e.g., cardiac index < 2.2 L/min/m ² with pulmonary capillary wedge pressure [PCWP] > 12 mm Hg), and signs of diminished tissue perfusion.

What specifically distinguishes cardiogenic shock from the other forms of shock is the mechanical impairment of pump function. Once a patient develops mechanical pump failure, and the intracardiac volumes and pressures begin to rise, the well-known “vicious cycles of deterioration” ( Fig. 22.1 ) can result in an imbalance of myocardial oxygen supply and demand, worsening (or creating) ischemia, and resulting in further decreases in ventricular function. Cardiogenic shock may ultimately result if the cycle is not broken.

Manipulations and optimization of preload, afterload, heart rate, and contractility are generally the first-line treatments for acute HF (see Chapter 26, Chapter 34, Chapter 8 ). If the heart is to recover, it is necessary to maintain myocardial oxygen supply while decreasing myocardial oxygen demand. One common pharmacologic strategy to accomplish this is to optimally increase “afterload,” decrease “preload,” and slow the heart rate, the combined effect of which will be to maximize coronary perfusion pressure and perfusion time.

Pharmacologic therapies can potentially improve hemodynamics and stabilize the patient with mild or moderate cardiac failure. In severe HF, management with inotropes and vasopressors comes at a very high price from the perspective of the myocardium and the peripheral and splanchnic circulations in the attempts to attain acceptable central hemodynamics. Moreover, fixing numbers is not always the same thing as fixing patients , and there are choices to be made because the pharmacologic management of HF is sometimes fraught with seemingly conflicting goals, especially when ischemia plays a role. From the perspective of the myocardium, β-adrenergic stimulation may improve contractility of areas that are well perfused, but it will greatly increase myocardial oxygen demand, feeding into and fueling the vicious cycle.

Vasoconstriction may improve coronary and systemic perfusion pressures, but depending on which vasoconstrictor is used, alpha-adrenergic stimulation will increase both the systemic and pulmonary vascular resistances (PVRs), making it harder for failing ventricles to eject. This is especially problematic when there is right ventricular (RV) failure, as this will increase the workload of the already struggling RV. Further, intentional vasoconstriction generally leaves the peripheral and splanchnic beds underperfused.

Afterload reduction with vasodilators is a common strategy to assist the failing heart because the physiologic principle of ventriculo-arterial coupling holds that regardless of the poor intrinsic systolic mechanics of the failing ventricle, its overall ability to function as a pump can be improved by decreasing the afterload against which it must pump. Nonspecifically decreasing “afterload” in the setting of developing cardiogenic shock leads to predictable hypotension that will be problematic because a prolonged period of poor tissue perfusion will predispose the patient to multisystem organ failure and a poor outcome.

This is where mechanical circulatory assistance can play an important role, effectively breaking the cycle and improving the balance between myocardial supply and demand, as well as systemic perfusion. By decompressing the failing ventricle, the increased wall tension that is adversely affecting the supply/demand ratio is addressed, which potentially sets the stage for myocardial recovery. Concurrently, effective perfusion is resumed to the heart and the rest of the body, which can stave off multisystem organ failure.

Thus by using a mechanical device to take over the pumping function of the failing ventricle, the ravages of cardiogenic shock can often be addressed with the one intervention, albeit an extremely invasive one that has potential advantages and disadvantages. Thus the implementation of mechanical assistance is often approached in a stepwise fashion. “Short-term support” of the failing heart with the intraaortic balloon pump (IABP), Impella, Tandem Heart, and ECMO are discussed in Chapter 2, Chapter 26, Chapter 35 . The remainder of this chapter focuses on MCS with implanted, durable LVADs.

Long-Term Support

As revealed by INTERMACS data, long-term, durable devices are being implanted in thousands of people each year who otherwise would succumb to HF. The technologic advancement in the design of implantable LVADs in the last 20 years has led to significant improvements in the outcomes and reductions in certain complication rates. The first-generation LVADs (c. 1990s) were bulky, large devices that produced pulsatile output by mechanical compression of a filling chamber. The early 2000s saw second-generation devices, which were more compact, miniaturized, and utilized an axial-flow impeller mounted on bearings to produce a CF output. The third-generation devices, introduced in the mid-to-late 2000s also produced continuous output, but instead of mechanical compression of the blood chamber, magnetic and/or hydrodynamic suspension was used for the impeller. While all of the third-generation devices utilize magnetic drive of the impeller, those utilizing hydrodynamic levitation of the impeller were termed “hybrid devices,” while those that used magnetism to both levitate the impeller and drive its rotation are termed “true maglev” devices. The absence of bearings (or other points of physical contact) to mechanically drive impeller rotation was demonstrated to substantially improve the durability of the devices. Improved hemocompatibility of the new-generation LVADs with reduced thrombotic complications of the impeller and blood contacting surfaces is characteristic of some third-generation devices. The majority of the devices implanted after 2018 utilize magnetic levitation and have consistently proved to be superior to the older generation devices with respect to complications.

At present, the HeartMate 3 (HM3; Abbott, Fig. 22.2A –E) is the most frequently implanted FDA-approved durable LVAD in the United States (and many other countries). The HM3 was FDA approved for short-term use in 2017 and for long-term use in 2018. Until recently, the HeartWare HVAD (Medtronic, Framingham, MA) was still being implanted in some US centers and countries around the world, but it was taken off the market in June 2021 after a class 1 recall was issued for electrical problems. However, it is possible to still encounter patients supported by the HVAD and some of the older devices (eg, the HM II) in the United States and elsewhere. A variety of other devices are used worldwide; not all devices are available and/or yet approved for use in all countries; and some countries manufacture their own devices. For example, in Japan, the EVAHEART (Sun Medical, Nagano, Japan) is a commonly implanted hydraulically levitated maglev centrifugal flow LVAD that has recently expanded into North America (the EVAHEART 2 is currently the subject of clinical investigation in conjunction with US centers). Meanwhile, China is developing the full maglev centrifugal CH-VAD (CH Biomedical Inc., Suzhou, China), and Russia has the axial-flow Sputnik (JSC ZITC, Moscow, Russia).

HeartMate 3 Left Ventricular Assist Device

The HM3 LVAD is a third-generation CF centrifugal pump based on Full MagLev Flow technology ( Fig. 22.2A ). Blood enters the inflow cannula implanted in the LV apex along the central axis of the device and is expelled from the device at a right angle by the impeller blades of a magnetically levitated magnetically driven rotor through an outflow graft connected to the ascending aorta ( Fig. 22.2B ). Fig. 22.2C shows how the device is configured internally and the only visible external component is a tunneled driveline that exits the skin of the abdomen, usually on the right somewhere convenient between the upper and lower quadrants. The driveline is connected by a modular cable to the system controller ( Fig. 22.2D ) that can be connected to mains A/C power (for sedentary periods or while sleeping) or external batteries (for active periods). When connected to mains A/C power, it is possible to use the bedside clinical control screen that displays various parameters of LVAD function, allows configuration of the pump settings and alarm limits, and allows review of device logs of system function and events ( Fig. 22.2E , clinical screen). When connected to battery power, several of the key parameters of LVAD function are visible, and alerts can be annunciated from the system controller ( Fig. 22.2D , system controller). When fully charged, two 14-volt HeartMate batteries can power the system for 10 to 17 hours depending on power utilization and the number of previous charging cycles. In addition, the HM3 system controller has a backup battery that can power the device for approximately 15 minutes in case of electrical interruption and incorporates a communication link for transfer of event/period logs and alarm information ( Box 22.2 ).

BOX 22.2

HM3 Key Features

HM3, HeartMate 3; PI, pulse index; rpm, rotations per minute.

• Third-generation miniaturized continuous flow centrifugal pump

• Full MagLev Flow Technology

  • Impeller is magnetically suspended in all three axes

• Enhanced biocompatibility

  • Textured interior surfaces

  • Wide fluid paths

• Artificial pulse

  • Impeller rotational speed automatically ramps down by 2000 rpm from the set speed and back up to 2000 rpm more than the set speed every 2 seconds to create an artificial pulse

• PI event detection

  • If the PI decreases by >45% from the previous averaged value for >10 seconds the low speed limit kicks in automatically until the PI event is resolved and then gradually ramps back up to the set speed

During clinical operation, the usual rotational speed of the HM3 is in the order of 5000 rotations per minute (rpm) (5400 rpm is often ultimately used, and the maximum speed is 9000 rpm). If the volume in the LV and set pump speed are optimal, the LVAD flow will be unidirectional toward the aorta. If the LV volume is suboptimal and/or the LVAD speed is set too high, it is possible that continuous LVAD action will collapse the LV, resulting in decreased LVAD flow and hypotension (a so-called “suction event”). In rare cases, if the aortic pressure is significantly higher than the LV pressure, it is possible to see retrograde flow through the LVAD.

Flow through the LVAD is in parallel to the natural flow through the LV and out the aortic valve, with the amount of blood flowing through the pump directly proportional to the pressure difference between the aortic and LV pressures, and inversely related to the systemic vascular resistance (SVR). The proportional amount of blood flowing out the aortic valve will depend on the contractility of the LV, as will the overall pulsatility of the circulation of the CF LVAD-supported patient.

Among the parameters displayed on the HM3 system controller ( Table 22.1 ) is a “pulsatility index” (PI) indicating the magnitude of sensed increases in the velocity of flow through the LVAD during LV systole (averaged over 15 s). Possible values of this unitless index range from 1 to 10, with higher values reflecting a higher level of LV contractility and less LVAD assistance. Lower PI values reflect lower levels of LV contractility and more LVAD assistance. When the PI falls below a preset value, the system automatically reduces the rotor speed to a preset low speed limit (to allow increased LV filling and aversion of a suction event) and then gradually returns to the previous set speed. In clinical practice, a drop in the PI reflects an inadequate LV volume, a decrease in LV contractility, or both.

The HM3 also incorporates a feature called “artificial pulse,” in which a brief intentional reduction and then ramping back up of rotor speed (30 times per minute) creates a small artificial “pulse.” Taken all together, the artificial pulse, the textured interior surfaces, and the wide fluid paths within the pump improve hemocompatibility and are intended to minimize shear forces (e.g., to decrease platelet activation and destruction of von Willebrand factor [vWF]) to reduce thrombus formation, and to potentially decrease certain other complications of CF (e.g., renal insufficiency from absence of pulsatility, and gastrointestinal [GI] and/or intracerebral bleeding from arteriovenous malformations [AVMs] induced by continuous, nonpulsatile flow).

Complications of Left Ventricular Assist Device Support

As outcomes have improved, simple survival of mechanically supported patients has become less of an issue, and the primary focus of MCS research has shifted toward optimizing outcomes through limiting adverse events. Unfortunately, no single risk stratification method or scoring system has yet been discovered that can adequately predict the various adverse events inherent to the MCS population. For example, though the preimplant Sequential Organ Failure Assessment (SOFA) score was reported to reliably predict survival after 6, 9, 12, 24, and 36 months of support, perhaps not surprisingly, no correlation was able to be demonstrated between SOFA scores and other long-term adverse events (e.g., stroke, bleeding, infection, need for pump replacement). It is also important to understand that due to the distinct mechanical underpinnings, materials, and functional specifications of modern devices, all data, predictive indices, and risk-stratification scores generated during the era of first-generation pulsatile devices cannot be extrapolated to the current generation of nonpulsatile devices.

Overall, while there has been a major decrease in the rates of specific adverse events with the CF devices compared to the first generation of pulsatile devices, according to INTERMACS and other data, there has been only a minor decrease reported in the total burden of adverse events. While the rates of some classic problems have decreased significantly (e.g., mediastinal bleeding, RV failure, stroke), the rates of some important ones (e.g., infection, renal failure, respiratory failure) have not changed, and new complications have appeared that essentially did not exist with the first generation of pulsatile devices (AVMs in the GI tract, von Willebrand Syndrome [vWS] resulting in GI and intracerebral bleeding, and pump thrombosis).

New information is rapidly becoming available, and certain complications of modern VAD support have now been linked with certain preexisting factors, management decisions, and/or aspects of modern MCS technology. INTERMACS data demonstrate that certain complications tend to occur most often within the first 90 days of implantation, while the rates of others (notably infection, GI bleeding, stroke, pump malfunction) increase with time spent on support. Table 22.7 shows the rates of freedom from GI bleeding, infection, and stroke as a function of the time spent on support (INTERMACS data).