Left ventricular assist devices (LVADs) have become a routine part of the armamentarium of treatments available for patients with advanced heart failure. Over the last 40 years, changes in the technology of the pumps and perioperative care of the patient have allowed LVAD therapy to progress to its wide acceptance, use, and success. The Jarvik 2000 is an axial flow LVAD that was first implanted in humans in the year 2000, after 13 years of laboratory research and development. Since 2000, more than 900 implants have been performed, resulting in more than 900 cumulative patient years of support. Worldwide, the Jarvik 2000 has been utilized for bridge to recovery, bridge to transplant, and destination therapy.1,2 The first bridge to transplant patient survived 13 years post-transplant, and the longest support on the Jarvik 2000 was more than 9.5 years before eventual successful transplant (direct communication from Jarvik Heart). LVADs and patients both experience a wide range of operating conditions while using mechanical support that cannot always be anticipated or appreciated during the experimental design phase of pump development. Since its implementation, many stepwise, systematic improvements of the Jarvik 2000 led to the current iteration of the heart pump (Table 1). Herein, we describe some of the key observations and interventions as it relates to three main areas of evolutionary improvement during the last decade.
Adverse events related to pump thrombus and embolic events remain a significant issue for all mechanical circulatory devices and recently have become more discussed.3 Three main design changes have been implemented to address these issues: intermittent low speed (ILS) controller, microsphere coating, and cone-bearing pump.
The Jarvik 2000 was one of the first pumps to use alternative outflow connections. In particular, the configuration to the descending aorta left concerns for stasis in the aortic root that could potentially lead to both coronary and neurologic events.4,5 The Jarvik 2000 is a five speed pump that is manually set to selected speeds ranging from 8,000 to 12,000 revolutions per minute (RPM) in increments of 1,000 RPM from speed 1 (8,000 RPM) up to speed 5 (12,000 RPM). To wash out the aortic root, the controller was redesigned to automatically reduce pump speed to 7,000 RPM for 8 seconds each minute, regardless of the speed setting of the device.6 This causes increased preload, and the natural heart responds with increased contractility for a few beats. By acutely loading the left ventricle (provided the patient has some intrinsic contractile function), the heart will eject, thereby, opening the aortic valve and washing the aortic root. In addition to clearing the aortic valve sinuses, this ILS controller serves to provide some pulsatility to the circulation. This feature could prove valuable for potentially enhancing vasomotor tone and manipulation of the peripheral vasculature.7 Indeed, this ILS configuration is conceptually used in some of the designs of newer pump controllers (i.e., HeartMate 3) that wish to highlight pulsatile flow.
Compared with other continuous flow pumps, the Jarvik 2000 has a wide range of speeds. Although few patients routinely use the higher speeds (except perhaps during periods of exercise or larger patients), one wonders whether the large and high range of operational speeds subject blood to enhanced shear. Although higher speeds might increase biochemical parameters of hemolysis (lactate dehydronase (LDH), albeit not statistically different), platelet number and function remained similar over the range of speeds.8,9 In extensive bench and animal testing of the infant/pediatric pump, little hemolysis was found at speeds from 12K to 22K. However, at higher speeds (25K–30K), hemolysis was severe (personal communication, Robert Jarvik, MD). The highest shear occurred in the tiny gap of the cone bearings where the stationary bearing posts contact the rotating cone. To extrapolate from these data, the adult Jarvik 2000 would likely have unacceptably high level of hemolysis at speeds >15K.
In addition to modifications of the controller, two specific changes in the external and internal design of the pump have focused on promoting thromboembolic resistance. The angle between the intraventricular pump and the surrounding left ventricular endocardium is an area of relative stasis that is prone to develop thrombus. To prevent thrombus that may form in the crevice between the pump and the apical endocardium from creeping near the inflow, the outer pump surface was modified by adding a titanium microsphere coating (Figure 1).
However, the biggest change related to pump design was the conversion of the rotor support from pin bearings to cone bearings (Figure 2). With the original pin-bearing configuration, a circumferential ring of thrombus was often observed originating in a small crevice between the rotating bearing pin and the stationary bearing sleeve. Although often quite small (2–3 mm), this was observed in four of six of the initial qualification calf experiments and, furthermore, observed on explanted human pumps as well (direct communication, Jarvik Heart; Figure 2). Although one could not correlate these findings with neurologic or thromboembolic events, this observation raised the hypothesis that these pathological findings could serve as a nidus for clot formation. The cone-bearing modification eliminates any circumferential crevice by supporting the rotating bearing cone on three short blades so that the blood flows between the blades and greatly improves the washing of the bearings. The diameter where the cone bearing support blades contact the tapered shaft of the cone is very small, less than 2 mm. This keeps the surface friction and bearing wear to a minimum. With the geometry of the cone bearings, there is no longer any circumferential crevice that can contribute to thrombus generation and retention. Optimal fluid film lubrication at the points of contact between the rotating ceramic bearing cones and the tips of the support blades contribute to the very low wear measured during durability testing (less than 1 μm per year). In vitro, the cone configuration demonstrated an enhanced hydrodynamic profile compared with the pin design.10 In 2009, six 60 day implants in 80–95 kg calves demonstrated that the cone bearings had no thrombus around the bearings and no hemolysis (personal communication, Jarvik Heart). In 2010, the Food and Drug Administration approved an investigational device exemption (IDE) supplement that allowed the cone-bearing pumps to replace the pin-bearing pumps in the pivotal trail. The only published data comparing the two designs is a retrospective review of the Italian Mechanical Support Registry; 99 patients supported with the Jarvik 2000 were retrospectively compared—39 with the pin and 60 with the cone. In their analysis, the pin-bearing configuration was an independent risk factor for cardiovascular death, right ventricular failure, and both ischemic and hemorrhagic stroke.11
While much attention is currently given to alternative (left thoracotomy) approaches for the HeartWare HVAD device, many of these techniques were pioneered in the Jarvik population because of several innovative features of the Jarvik system. In particular, as an intraventricular, intrapericardial pump with a relatively low profile, this pump has long been very amenable to alternative surgical approaches.12–14 The design of the sewing ring and its association with its coring knife makes off-pump insertion relatively straightforward. Along the way, several changes in technique have been incorporated in to the surgical repertoire.
Avoidance of sternotomy in patients with a hostile mediastinum is a particular advantage to the left thoracotomy approach. As the pump does not come with an attached bend relief, utilizing a radially supported polytetrafluoroethylene (PTFE) graft sutured just to the flange of the outflow graft allows a smooth transition of blood flow through the outflow, especially as it traverses toward the diaphragm and then up to the descending aorta (usual attachment is near the inferior pulmonary vein). This addition is also important with sternotomy approaches because of an even sharper angle coming off the intraventricular position of the pump.
Although off-pump approaches can be advantageous, concerns for incomplete coring, left ventricular thrombus, and cannula position often make it wise to have direct visualization of the left ventricular cavity. While femoral–femoral bypass is traditionally used to assist, dealing with femoral vessels and passing wires in the decubitus position can be challenging. We have most recently adopted an approach of direct descending aortic cannulation and a small cannula in the inferior pulmonary vein, thereby, using left heart bypass (with a reservoir in place for air and blood return) to provide circulatory support. This technique keeps the operation in the chest and provides the visualization necessary for successful implants.
Other surgical innovations have predated many of the approaches currently used with other pumps. Although not routinely performed, the Jarvik 2000 can be successfully used for total ventricular assist with biventricular devices.15,16 This technique has also incorporated bilateral skull pedestal implants for the driveline providing for total heart assist as destination therapy.
In addition to the ILS program, the Jarvik 2000 controller is unique in that it allows patients to set the speed of the pump. LVAD systems are not intelligent; they spin at the set speed regardless of the demands that the patient requires. This notably influences exercise capacity. The Jarvik device allows a patient (when informed and educated by the care team) to flip the speed up during periods of exercise and down during periods of rest. We have also used this novel feature to help recondition ventricles to promote myocardial recovery and LVAD explantation.17 In our current practice, we prescribe the “dose” or speed of the pump to meet the needs of the patient, just like one would do with heart failure medications. For example, if a patient was in a compensated state at speed 3 and felt fatigued with exercise, we would empower them to turn the controller to speed 4 during periods of strenuous activity and then back to speed 3 when they return to normal activity.
The longest patient supported with his original Jarvik 2000 pump lived for 9.5 years. This is only possible if both internal and external components are durable. Early on, the Jarvik 2000 moved to a Y cable to allow for battery changes without pump stoppage. In the first 269 patients who received the original Jarvik 2000 abdominal driveline, which had straight wires, there were no fractures of the implanted portion of the drivelines, but there were 12 (4.5%) driveline fractures outside the body where it is exposed to more flexing stress. Although all fractured drivelines were repaired or the pump was replaced, the abdominal driveline has now been modified to the current spiraled wire configuration that provides for a highly flexible cable.
Although driveline infections anecdotally occur less with the Jarvik pump, all abdominal drivelines provide a continuous source of potential infection. In addition, the abdominal driveline can be uncomfortable for patients and limits their ability to bathe. The Jarvik 2000 system with the postauricular driveline has been long used in Europe and is just now being adopted in the United States as part of the Jarvik 2000 RELIVE trial for Destination Therapy. The RELIVE trial randomizes patients to the postauricular device or HeartMate II system. To date, nearly 20% of the prespecified target population has been enrolled. With the skull pedestal implant, patients can fully immerse themselves in water and, because of its solid fixation in well-vascularized tissue, is very resistant to infection.
This report summarizes the evolutionary changes of the Jarvik 2000 over the past decade. No perfect LVAD exists. The ability to nimbly modify both controller and physical design, thereby, striving to continually improve the LVAD system, are important lessons for the mechanical circulatory community. Finally, thorough vetting of clinical outcome data of the Jarvik 2000 must be contextually weighed by the wide spectrum of changes that occurred within the pump over the last decade.
We fully recognize that this article is a reflection of years of hard work put in by the team at Jarvik Heart. In particular, we acknowledge the efforts of Dr. Jarvik and his team of dedicated engineers and technicians including Kamal Gandhi, Robert Scharp, Vern Fulmer, Mark Baker, Latha Kavala, and Doug Longo.
1. Sorensen EN, Pierson RN 3rd, Feller ED, Griffith BP. University of Maryland surgical experience with the Jarvik 2000 axial flow ventricular assist device. Ann Thorac Surg. 2012.93: 133140.
2. Yoshioka D, Matsumiya G, Toda K, et al. Clinical results with Jarvik 2000 axial flow left ventricular assist device: Osaka University Experience. J Artif Organs 2014.17: 308314.
3. Starling RC, Moazami N, Silvestry SC, et al. Unexpected abrupt increase in left ventricular assist device thrombosis. N Engl J Med 2014.370: 3340.
4. Frazier OH, Myers TJ, Westaby S, Gregoric ID. Use of the Jarvik 2000 left ventricular assist system as a bridge to heart transplantation or as destination therapy for patients with chronic heart failure
. Ann Surgery. 2003.237: 631636; discussion 636–637.
5. Gazzoli F, Panzavolta M, Grande AM, Viganò M. Right coronary thrombosis in patient supported by Jarvik 2000 left ventricular assist device. Eur J Cardiothorac Surg 2011.39: 1076.
6. Tuzun E, Gregoric ID, Conger JL, et al. The effect of intermittent low speed mode upon aortic valve opening in calves supported with a Jarvik 2000 axial flow device. ASAIO J 2005.51: 139143.
7. Witman MA, Garten RS, Gifford JR, et al. Further peripheral vascular dysfunction in heart failure
patients with a continuous-flow left ventricular assist device: the role of pulsatility. JACC Heart Fail 2015.3: 703711.
8. Löffler C, Straub A, Bassler N, et al. Evaluation of platelet activation in patients supported by the Jarvik 2000* high-rotational speed impeller ventricular assist device. J Thorac Cardiovasc Surg 2009.137: 736741.
9. Mondal NK, Sorensen EN, Feller ED, Pham SM, Griffith BP, Wu ZJ. Comparison of intraplatelet reactive oxygen species, mitochondrial damage, and platelet apoptosis after implantation of three continuous flow left ventricular assist devices: HeartMate II, Jarvik 2000, and HeartWare. ASAIO J. 2015;61:244252
10. Stanfield JR, Selzman CH. In vitro
hydrodynamic analysis of pin and cone bearing designs of the Jarvik 2000 adult ventricular assist device. Artif Organs 2013.37: 825833.
11. Tarzia V, Di Giammarco G, Di Mauro M, et al. From bench to bedside: Can the improvements in left ventricular assist device design mitigate adverse events and increase survival? J Thorac Cardiovasc Surg 2016.151: 213217.
12. Frazier OH, Gregoric ID, Cohn WE. Initial experience with non-thoracic, extraperitoneal, off-pump insertion of the Jarvik 2000 Heart in patients with previous median sternotomy. J Heart Lung Transplant 2006.25: 499503.
13. Selzman CH, Sheridan BC. Off-pump insertion of continuous flow left ventricular assist devices. J Card Surg 2007.22: 320322.
14. Zucchetta F, Tarzia V, Bottio T, Gerosa G. The Jarvik-2000 ventricular assist device implantation: how we do it. Ann Cardiothorac Surg 2014.3: 525531.
15. Frazier OH, Myers TJ, Gregoric I. Biventricular assistance with the Jarvik FlowMaker: a case report. J Thorac Cardiovasc Surg 2004.128: 625626.
16. Yoshioka D, Toda K, Yoshikawa Y, Sawa Y. Over 1200-day support with dual Jarvik 2000 biventricular assist device. Interact Cardiovasc Thorac Surg 2014.19: 10831084.
17. Healy AH, Koliopoulou A, Drakos SG, McKellar SH, Stehlik J, Selzman CH. Patient-controlled conditioning for left ventricular assist device-induced myocardial recovery. Ann Thorac Surg 2015.99: 17941796.