Circulatory Support for Long-Term Treatment of Heart Failure
Experience With an Intraventricular Continuous Flow Pump
Background— A lifetime mechanical solution for advanced heart failure must be reliable, with a low risk of life-threatening complications. After extensive laboratory testing, we began clinical trials with an axial flow pump for long-term treatment of New York Heart Association class IV, transplant-ineligible patients.
Methods and Results— The Jarvik 2000 is a continuous flow device that is implanted in the apex of the left ventricle with offloading to the descending thoracic aorta. Skull-based percutaneous power delivery was derived from cochlear implant technology. We used this system in 4 patients with end-stage dilated cardiomyopathy. Exercise capacity, quality of life, device parameters, and native heart function were monitored serially. One patient died from right heart failure at 3 months. The other patients were discharged from hospital between 3 and 8 weeks postoperatively and are currently New York Heart Association I or II. Follow-up lasted between 9 and 20 months. There has been no device failure or hemolysis. Native heart function and quality of life were markedly improved.
Conclusions— The Jarvik 2000 is a true assist (rather than replacement) device that functions synergistically with the native left ventricle and provides excellent quality of life. Adverse events are infrequent. This blood pump may provide a mechanical solution for end-stage heart failure in the community.
Received March 22, 2002; accepted April 2, 2002.
End-stage heart failure patients are leading consumers of healthcare resources. Medical management may slow the process of deterioration, but in New York Heart Association class IV patients, the quality of life remains poor. Availability of cardiac transplants is now severely limited. Lifetime mechanical circulatory support is a potential alternative which was recently tested in the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial.1 Although the REMATCH trial demonstrated survival benefit in favor of left ventricular assist device (LVAD) use, both longevity and quality of life were limited by device-related complications. The success of future long-term LVAD use will depend on implantable, user-friendly blood pumps with less risk of infection, thromboembolism, or device failure. After a comprehensive laboratory research program, we began a clinical trial of a continuous-flow pump in transplant-ineligible patients with terminal dilated cardiomyopathy.2 This report describes our early experience.
The Jarvik 2000 is a titanium axial flow impeller pump (5.5 cm × 2.5 cm, weight 90 g) that is implanted into the apex of the failing left ventricle (LV). In vitro pressure flow relationships have been described previously.2 Pump output is delivered to the descending thoracic aorta, thereby perfusing the brachiocephalic and coronary arteries in retrograde fashion. In patients, the device provides continuous non-pulsatile flow of 2.5 to 6 L/min at between 8000 to 12 000 rpm, depending on afterload. We used an angiotensin-converting enzyme (ACE) inhibitor with or without β-blockers to maintain mean blood pressure between 70 and 90 mm Hg. Total systemic blood flow is the combination of pump output and LV ejection through the aortic valve during partial offloading. LV ejection causes pulse pressure in the systemic circulation. The higher the pump speed, the greater the flow through the apex is. Further hemodynamic information is provided in the study by Frazier et al.3 Power consumption is 3 to 7 watts. The external controller and Lithium ion battery together weigh 2400 g and are easily portable. A skull-mounted pedestal transmits the power cable through the skin (Figure 1), which resists infection by rigid bone fixation and the healing capacity of highly vascular scalp skin.4
Criteria for entry to the trial were the presence of “end stage” dilated cardiomyopathy (New York Heart Association [NYHA] class IV on maximum medical treatment), ineligibility for transplantation, cardiac index <2.0 L/min per m2, left ventricular ejection fraction (LVEF) <25%, peak oxygen consumption rate (peak V̇o2) <16 mL · kg−1 · min−1, and creatinine clearance >25 mL/min. The clinical protocol and informed consent were endorsed by the Central Ethics Committee of the Oxford Radcliffe Hospital. Patients with insulin-dependent diabetes, previous cardiac surgery, active malignancy, or a contra-indication to anticoagulants were excluded.
Four implants took place in a 12-month period beginning June 2000. All patients were males aged 61 to 72 years (median 62 years). Preoperative patient characteristics are shown in Table 1.
Postoperative medication was tailored to individual requirements based on repeated examination, body weight, and echo-cardiographic assessment. Warfarin was given to achieve an International Normalized Ratio (INR) between 2.5 and 3.0.
Patients were assessed at 2 to 4 week intervals using 2-dimensional echocardiography to monitor LV and right ventricular (RV) cavity size and ejection fraction (LVEF and RVEF, respectively). Doppler echocardiography was performed through the range of speeds and with the pump switched off for up to 5 minutes. A pump speed setting was then recommended based on effectiveness of LV offloading (LV cavity size 5 to 6 cm). This setting was either 10 000 or 11 000 rpm.
Respiratory exercise testing was undertaken using a couch with bicycle ergometer (Redbuld Medical System Inc). Device-related parameters and adverse events were recorded continuously. NYHA functional status was defined at each follow-up, and the Minnesota “Living with Heart Failure” questionnaire was completed at 3 monthly intervals.5
Patients 1, 3, and 4 left hospital between 3 and 8 weeks postoperatively. The duration of survival was 20, 12, and 9 months, respectively. Patient 2 weighed 120 kg. He initially awoke after surgery and moved all limbs on command. On discontinuation of sedation, right side weakness was noted, and a CT scan showed a left subdural hematoma beneath the skull pedestal. Evacuation was undertaken, but repeated attempts to wean him from the ventilator proved unsuccessful. He died from right heart failure 3 months later. Autopsy showed the pump and vascular graft to be free from thrombus. Subsequently, each patient’s skull thickness was measured by computerized tomography and the surgical technique was modified to prevent penetration of the internal table.
Postoperative patient parameters are shown in Table 2. Exercise tolerance increased gradually and was accompanied by a decrease in body weight and a disappearance of peripheral edema and ascites (Figure 2). None of the patients required diuretics after 3 months. Patient 1 had return of dyspnea at 4 months. He had elevated blood pressure (mean >100 mm Hg), and echocardiography showed inadequate left ventricular offloading (LVEDD 6.5cm). He became asymptomatic after optimizing angiotensin converting enzyme (ACE) inhibitor dosage and initiation of β-blockade. In subsequent patients we used maximum ACE inhibition.
Echocardiography showed early (within 3 months) and sustained improvement in LV and RV function. Pulsatility between 10 to 30 mm Hg was detectable in the circulation at all but the highest pump speed (12 000 rpm). There was a discrepancy between hemodynamic parameters and symptomatic recovery from heart failure. The pump restored cardiac index and left atrial pressure to within normal levels (operating room measurements). Resolution of breathlessness and fatigue lagged behind hemodynamic improvement as demonstrated by the time-related increase in peak V̇o2 and exercise duration (Figure 2) During exercise, systemic blood pressure increased by 10% to 15%.
At 11 months, patient 3 suffered ventricular tachycardia at 250 beats/min. He arrived in the hospital 4 hours later, still fully conscious with a venous pH of 7.35. He was cardioverted and prescribed amiodarone, and he recovered uneventfully.
Quality of Life
Patients 1 and 3 are NYHA class I. Patient 4 is NYHA class II. They have recorded major improvement in quality of life scores (Table 2). Activity is unrestricted by the device. Patient 1 can walk up to 5 miles daily and has completed a long distance charity walk. He has undertaken 3 international return flights, the last of which was unaccompanied. Skull-based power delivery is user friendly. There has been no power line infection.
There was no clinically important hemolysis (Table 2). Patient 1 suffered 3 prolonged interruptions of power supply. On one occasion, the bag containing the controller and battery was stolen by a thief, and the external cable was avulsed from the pedestal. Power was restored by re-attaching the connector. The “pump off” situations were well tolerated, and there was no thromboembolism on restoration of flow.
Patient 4 had dietary deficiency anemia and required a blood transfusion at 16 weeks postoperatively. After intravenous cannula infection, he suffered a Staphylococcus aureus bacteremia infection and a transient neurological event involving the left hand. He recovered in 48 hours, but was treated with intravenous antibiotics for 6 weeks. There was no device infection shown on repeated echocardiography or blood cultures.
Survival to 20 months is a landmark in continuous flow pump technology. Demonstration that stroke volume and pulse pressure are not required from the LVAD itself now facilitates the introduction of miniaturized, less intrusive devices as a practical alternative to cardiac transplantation.
With axial flow pumps, pulse is transmitted to the systemic circulation not only by limited antegrade stroke volume ejection, but also by increased preload to the device during LV contraction. If the force of contraction improves, a greater proportion of pump flow occurs during systole. Increasing pump speed raises the flow via the pump during diastole and reduces stroke volume ejection through the aortic valve. At higher speeds (11 000 to 12 000 rpm), all transmitral flow is captured by the device, the LV is fully offloaded, and the aortic valve remains closed. Our preferred approach is to preserve about one-third of total flow by ejection through the valve. By partially offloading, the device improves the position of the LV on the Frank-Starling curve and minimizes distortion of RV geometry and function. This is a different strategy from pusher plate LVADs, which are programmed to empty and completely replace the native LV. Survival of patient 3 after prolonged tachycardia indicates that the pump can sustain both systemic and pulmonary circulations in the event of dysrhythmia.
Exercise increases cardiac output through intrinsic mechanisms (elevated heart rate, contractility, and venous return). Flow increases both through the device and via the LV outflow tract. Consequently, activity responsive electronic mechanisms are unnecessary. The intraventricular position has important advantages. There is no inflowcannula, which may cause flow restriction, hemolysis, thrombosis, and thromboembolism in other LVADs. An inflow cannula can also tether the apex and adversely affect native LV function.
All patients showed improvement in native heart function. As the molecular mechanisms of myocardial recovery are defined, the combination of mechanical LV offloading with agents to promote myocardial recovery is developing into an exciting new strategy for heart failure management.6 We conclude that long-term mechanical circulatory support with smaller continuous flow LVADs may evolve into realistic therapy for advanced heart failure in the community.
We are grateful for financial support from the National Heart Research Fund (UK) and an anonymous benefactor.
Dr Jarvik is president of Jarvik Heart Inc.
Guest editor for this article is Myron L. Weisfeldt, MD, Johns Hopkins School of Hygiene and Public Health, Baltimore, Md.
- ↵Frazier OH, Myers TJ, Gregoric ID, et al. Initial clinical experience with the Jarvik 2000 implantable axial-flow left ventricular assist system. Circulation. In press.
- ↵Rector TS, Cohn JN. Assessment of patient outcome with the Minnesota Living with Heart Failure questionnaire: reliability and validity during a randomized, double-blind, placebo-controlled trial of pimobendan. Pimobendan Multicenter Research Group. Am Heart J. 1992; 124: 1017–1025.
- ↵Muller J, Wallukat G, Weng YG, et al. Weaning from mechanical cardiac support in patients with idiopathic dilated cardiomyopathy. Circulation. 1997; 96: 542–553.