Initial Clinical Experience with the HeartMate® II Axial-Flow Left Ventricular Assist Device

Abstract

The redesigned HeartMate® II, an axial-flow left ventricular assist device, is simpler, smaller, and easier to operate than are pulsatile pumps. These design characteristics should make the HeartMate II more reliable and durable and broaden the eligible population base.

We implanted the HeartMate II in 43 patients (average age, 42 yr). The indication for use was bridge-to-heart transplantation in 26 patients and destination therapy in 17.

The average duration of device support was 258 days (range, 1–761 days), and cumulative duration, more than 31 patient-years. Hemodynamic function improved in all patients during support. By 48 hours after implantation, the mean cardiac index had increased from 1.9±0.27 L/(min·m²) (baseline) to 3.5±0.8 L/(min·m²), and the pulmonary capillary wedge pressure had decreased from 24.8±11 mmHg to 18.5±5.3 mmHg. Of the 43 patients, 35 were discharged from the hospital. Support is ongoing in 27 patients (longest duration, >700 days). Nine patients died during support. Four patients had sufficient heart recovery to undergo pump explantation. Three patients underwent transplantation. One patient underwent device replacement after the pump driveline was fractured in a skateboarding accident; the device was removed in another patient because of a pump-pocket infection after 749 days of support. Of the 10 patients in whom the HeartMate II replaced a failed HeartMate I, 8 were discharged from the hospital.

We have seen excellent results with use of the HeartMate II. Functional status and quality of life have greatly improved in patients who survived the perioperative period.

Axial-flow left ventricular assist devices (LVADs) have been under development for the past decade and are now being tested in clinical trials. These devices are smaller and simpler than pulsatile pumps.^1–4 Having only 1 moving component, no valves, and no vent or compliance chamber greatly reduces the complexity of these pumps and should lead to better reliability than that shown by the widely used implantable pulsatile pumps. Because of their small size, axial-flow LVADs can be used in smaller patients, including children. Furthermore, the procedure for implantation is less invasive, which reduces operative complications such as bleeding, infection, and thromboembolism.

The 1st design of the HeartMate® II (Thoratec Corp.; Pleasanton, Calif) was used in a European study beginning in February 2001. The study was halted early, in 2002, when it became evident that thrombosis related to the texturing of the internal blood-contacting surfaces was resulting in poor outcomes. Extensive experience with the HeartMate I (IP and VE LVAS [left ventricular assist system]) had shown that textured surfaces promote the development of a biocompatible tissue lining, thus eliminating the blood–biomaterial interface.⁵ The proliferation of tissue on this surface within an axial-flow pump, however, narrows the blood-flow path already compromised by the design of the pump. This narrowing, combined with the low pressure generated by the failing heart, may have contributed to primary pump thrombosis in the European study. The textured surface in the HeartMate II pump housing was eliminated, and clinical studies resumed in 2003 with the new design.

In November 2003, we were the first to implant the newly designed HeartMate II pump.⁴ Since then, we have implanted the device in 42 additional patients. We have completed the feasibility phase of the clinical trial and are enrolling patients in the phase II pivotal trial. In this report, we describe our experience from both the pilot and pivotal trials with the first 43 patients supported by the HeartMate II at the Texas Heart Institute (THI).

Methods

The HeartMate II is an axial-flow pump that is implanted through a median sternotomy with cardiopulmonary bypass (CPB) support. The surgical approach is similar to that of the HeartMate I, except that the subdiaphragmatic abdominal space required for pump placement is considerably smaller than that for the HeartMate I. Detailed descriptions of the HeartMate II system and the implantation procedures have been published.^2–4 Once implanted, the pump sits just below the left hemidiaphragm; the inflow cannula is within the left ventricle and the outflow graft is anastomosed to the ascending aorta (Fig. 1). The pump is powered by a pair of lead-acid batteries or by a bedside power base unit. The microprocessor-based system controller, worn on a belt, maintains pump speed, monitors device function, and provides audible and visual alerts if system problems occur. Because of the wearable design of this system, patients are mobile and can live outside the hospital and resume their normal, daily activities.

Fig. 1 The HeartMate II® consists of an implanted blood pump, a microprocessor-based system controller, and a 12-volt power supply. The pump lies below the left hemidiaphragm. The inflow conduit is placed within the left ventricle, and the outflow graft is anastomosed to the ascending aorta. The electrical lead (driveline) is tunneled subcutaneously and exits the right lateral abdominal wall.

We enrolled 43 patients with severe heart failure in the study. There were 32 men and 11 women with a mean age of 42 years (range, 14–73 yr), and a mean body surface area (BSA) of 1.94 m² (range, 1.50–2.47 m²). The indication for use was bridge-to-heart transplantation in 26 patients and destination therapy in 17. The diagnosis was idiopathic cardiomyopathy in 30 patients and ischemic cardiomyopathy in 13. All patients gave informed written consent to participate in the study. At the time of preoperative evaluation, all 43 patients were classified in New York Heart Association (NYHA) functional class IV and were receiving optimal medical management in the hospital. Fourteen patients were being supported by an intra-aortic balloon pump, and 6 patients had been supported by a TandemHeart percutaneous ventricular assist device (pVAD) (CardiacAssist, Inc.; Pittsburgh, Pa) before receiving the HeartMate II. The HeartMate II was placed in 10 patients after their HeartMate I devices failed (Table I).

TABLE I. Patients' Characteristics and Outcomes

Patients received routine, postimplantation medical support after their operations; however, a more aggressive anticoagulation protocol was recommended after the experience with pump thrombosis in the initial HeartMate II implants. A combination of heparin, warfarin, aspirin, and dipyridamole is used for anticoagulation and suppression of platelet function during HeartMate II support (Table II). Once patients are stabilized and become ambulatory, proper nutrition, rehabilitation, and education become a focus of care. After discharge from the hospital, patients return to our heart failure clinic for routine follow-up at a decreasing frequency, depending on their needs. Serial echocardiographic studies are performed at regular intervals for inpatients and outpatients to evaluate the adequacy of ventricular unloading.

TABLE II. Anticoagulation Guidelines

Results

Patient Support. The average duration of support was 258 days (range, 1–761 days). Excluding perioperative and intraoperative deaths, the average duration of support was longer than 300 days. The total duration of support was more than 31 patient-years. Of the 43 implanted patients, 35 (81%) survived the operation without significant complications and were discharged from the hospital in NYHA functional class I. Protocol restrictions prevented our 1st patient from being discharged until 167 days after pump implantation. All other HeartMate II recipients were discharged a mean of 36 days (range, 14–106 days) after implantation. Three patients underwent heart transplantation after 175, 641, and 187 days of support. After 749 days of support, our 1st patient, whose device was implanted in November 2003, had the device explanted emergently because of a pump-pocket infection; he underwent heart transplantation 4 months later. Support is ongoing in 27 patients (longest duration, >700 days).

Nine patients died while receiving device support. Five patients died during the early postoperative period because of a combination of right-sided heart failure, multisystem organ failure, and bleeding complications. In a 6th patient, bleeding associated with small-bowel arteriovenous malformations necessitated the discontinuation of all anticoagulation and antiplatelet therapy. Although the pump flow was adequate for circulatory support, the patient's left ventricular dimensions were not reduced by the pump. Poor myocardial contractility resulted in a lack of aortic valve opening and thrombus generation in the non-coronary sinus (Fig. 2). The patient developed ventricular arrhythmias and, eventually, ventricular fibrillation. The thrombus embolized during cardiac massage for resuscitation and caused neurologic death. The 7th patient who died had severe, generalized atherosclerotic vascular disease, coronary artery occlusive disease, and diabetes. She had been discharged but was readmitted 125 days after implantation because of a transient ischemic attack with minimal neurologic residual effects. To avoid further arterial blockage, systemic anticoagulation with heparin was administered. The next day, she became obtunded, with progressive neurologic deterioration, and studies showed a massive hemorrhagic stroke. Support was discontinued. Two patients experienced sudden death at home at 326 and 81 days. Although autopsies were performed, no obvious cause of death was determined in either patient. Both pumps appeared to be functioning normally at the time of death.

Fig. 2 Echocardiogram shows thrombus in the non-coronary sinus of a patient with persistent gastrointestinal bleeding, in whom all anticoagulant therapy was discontinued. This patient never developed aortic valve opening and is the only patient to have experienced thrombotic complications in this series.

Real-time motion images are available at texasheart.org/journal

The postimplant course has been uncomplicated in the 34 remaining patients, all of whom are in NYHA functional class I. In 1 patient, the HeartMate II device was exchanged on an emergency basis after the pump driveline was fractured in a skateboarding accident⁶; the patient was discharged 1 week later. Eight of the 10 HeartMate I to II pump-exchange patients have been discharged and are doing well. The other 2 pump-exchange patients died during the perioperative period, one of bleeding and the other of diffuse intravascular clotting that was thought to be secondary to factor VII administration. Our current method of pump exchange is via subcostal incision alone; with this approach, patients have experienced rapid postoperative recovery and none have died. Superficial driveline infections occurred in 3 patients; 2 of these were treated successfully with local wound care. However, the superficial infection in our 1st implant patient from 2003 was never satisfactorily controlled and worsened with the patient's weight gain (>13 kg). Eventually a pump-pocket infection developed, and the device had to be removed more than 2 years after the initial implantation.

Hemodynamic and Echocardiographic Evaluation. Hemodynamic function improved from preoperative levels in all patients during support with the HeartMate II. By 48 hours after device implantation, the average cardiac index had increased significantly, from 1.9±0.27 to 3.5±0.8 L/(min·m²), and the pulmonary capillary wedge pressure had decreased significantly, from 24.8±11 to 18.5±5.3 mmHg. Inotropic support was reduced after implantation, and all patients were weaned within the 1st week. Serial echocardiographic studies showed improvement in left ventricular dimensions. The average left ventricular diastolic dimension decreased by 16% (from 6.7 cm to 5.6 cm) 1 month after implantation.

Hematology. All 43 patients received a combination of heparin, warfarin, aspirin, and dipyridamole as anticoagulant therapy to achieve an international normalized ratio between 2.5 and 3.5 during support. Anticoagulation was temporarily discontinued in 1 patient because of gastrointestinal bleeding from arteriovenous malformations. Eighteen devices were explanted: 3 after successful bridging to transplantation, 4 after recovery of native heart function sufficient to warrant pump explantation, 1 after damage in a skateboarding accident, 1 for pump-pocket infection, and 9 post mortem. None of the removed devices showed evidence of thrombus. No bearing wear was detected in any of the explanted pumps. Two pumps had been operational longer than 2 years. In the 35 patients who were discharged, hemoglobin, hematocrit, and end-organ function had reached normal levels by the time of discharge and remain normal in all current outpatients. Hemolysis was seen in the patient who had gastrointestinal bleeding and aortic thrombosis but has not been observed in the other surviving patients.

HeartMate II System Operation. The technical performance of the HeartMate II has been excellent, and patients have been extremely satisfied with the small, quiet pump. An overall Kaplan-Meier survival rate of 80% has been gratifying in this very ill patient cohort (Fig. 3).

Fig. 3 Kaplan-Meier survival curve for the HeartMate II patients at 1 year (n=43).

Before discharge, all patients were trained in the care and use of their equipment and batteries. The patients participated in a physical rehabilitation program that comprised graduated ambulation within the hospital and treadmill exercise under observation. The speed setting of the treadmill was predetermined by use of echocardiographic studies to determine the degree of unloading and exercise tolerance. Life-threatening arrhythmias, possibly generated by contact of the intraventricular cannula with the endocardium, developed in 3 patients.

No device malfunction or system problem has been seen in the outpatient setting. In our 1st patient, the device was removed after 749 days of use, and physical examination of the pump showed no wear of the bearing. Moreover, measurements of function showed no difference in technical operation of the pump before implantation and after explantation (Fig. 4).

Fig. 4 Photographs of the inlet bearing's ball (A) and cup (B) of a HeartMate II removed 749 days after implantation in the 1st patient of our series. Radial score marks are visible on the ruby ball but are considered normal for all pumps after initial bearing wear-in. No physical wear of the bearing was detected. Test measurements of operation for this pump were identical before implantation and after explantation.

Discussion

Our initial experience with the redesigned HeartMate II at THI has been favorable and has shown this model to be safe, durable, and effective. This implantable axial-flow pump evolved from a prototype, originally developed by the Nimbus Corporation in the early 1990s. Richard Wampler, then an employee of Nimbus, demonstrated that effective, short-term circulatory support was possible with a small axial-flow pump operating at 25,000 rpm (the Hemopump). The Hemopump was first used clinically at THI in 1988.³ Enlargement and modification of the original Hemopump, combined with the addition of blood-lubricated bearings, culminated in the development of the long-term, implantable pump described in this report—the HeartMate II.

After modifications to the initial clinical pump, the current HeartMate II design was finalized and pump implantations began at our institution in 2003.⁴ The technical function of this design has been excellent. We have seen no pump failures or operational problems with this technology to date. In our experience, the Jarvik 2000, another axial-flow pump, has shown reliability and durability for more than 6 years, and studies are ongoing. The HeartMate II has a similar pump design and the potential for comparable long-term efficiency.

In this series, 35 of 43 patients were discharged from the hospital, and problems with the HeartMate II system have been minimal. Arrhythmias have been noted, particularly if the patient becomes volume depleted or if the pump's unloading of the ventricle is excessive (Fig. 5). Lowering the rpm and ensuring adequate volume status, especially during the postoperative period, can resolve this issue. Our patient group differs slightly from the typical bridge-to-transplant population, because our patients were younger (mean age, 42 years vs 50–62 years in other studies) and were small (with a mean BSA, 1.94 m²; range, 1.42–2.57 m²).^7–9 The small size and reduced weight of this technology have enabled us to implant this device in more women who have terminal heart failure. Women, because of their comparatively small size, have heretofore received limited benefit from the larger pulsatile pumps. Younger, smaller patients can also benefit. Our 3 youngest patients to receive the HeartMate II were 14, 15, and 17 years old at pump implantation.

Fig. 5 Increasing pump speed associated with ventricular arrhythmias may be caused by contact of the cannula with the endocardium.

The introduction of continuous-flow left ventricular assist devices has created a new physiology of cardiac support. This technology has definitively shown that ambulatory, effective circulatory support can be achieved long-term in patients with advanced heart failure who require cardiac support, even when they do not have a detectable pulse. However, by changing the internal physiology, this therapy introduces physiologic phenomena and accompanying clinical problems, such as arteriovenous malformations leading to gastrointestinal bleeding, septal shift with resultant right-sided heart failure, thrombosis of the aortic valve noncoronary sinus, aortic valve fusion, and aortic valve insufficiency. These problems are medically manageable, but physicians must be aware of them when applying this life-saving technology.

We pay particular attention to the proper level of anticoagulation and to the measurement and control of blood pressure in patients without a detectable pulse. To date, experience with these pumps in a limited patient population has suggested an increased incidence of hemorrhagic stroke. This may be related to the difficulty of properly monitoring blood pressure due to the absent or dampened pulse imparted by this technology. If hypertension is present but not easily measurable, inadequate pharmacologic control of blood pressure may result, in turn subjecting these patients to an increased risk of hypertensive hemorrhagic stroke. An additional concern is that the altered stress on the aortic valve may result in valve leaflet fusion and aortic insufficiency and stenosis. This has been reported in patients with pulsatile assist pumps, but we have also seen this problem in patients with continuous-flow pumps.¹⁰

Our initial experience with the 43 patients described here has been favorable. We are encouraged by our results with the use of the redesigned HeartMate II in the treatment of patients with advanced heart failure. We believe that this small, axial-flow blood pump is durable and safe and will offer an extended, improved quality of life for critically ill heart failure patients of varying sizes and clinical statuses.