Baylor College of Medicine, Center for Artificial Organ Development, headed by Professor Yukihiko Nosé, started in 1990 a series of heart assist device development programs to expand centrifugal blood pump technologies with the aid of several medical industries.1,2 During these pump research and development (R&D) programs, a 2-day pump for cardiopulmonary bypass1 and a 1-month pump for extracorporeal membrane oxygenation (ECMO) or postcardiotomy short-term ventricular assist device (VAD) applications1 were commercialized. This 1-month pump, the Kyocera Gyro C1E3 centrifugal pump, was developed by a joint project initiated in 1993 between Kyocera Corporation (Kyoto, Japan) and Baylor College of Medicine (Houston, Texas). It successfully demonstrated superior bearing durability, antithrombogenicity, and minimal hemolysis properties.3–5 At the present time, this extracorporeal pump has shown to be a promising circulatory support device in various short-term and mid-term applications; thus, over 1,200 pump heads are clinically used in Japan every year. However, the original concept of the Gyro pump was to pursue longer-term VAD applications such as bridge-to-heart transplantation (BTT). Based upon the design concept of the Kyocera Gyro C1E3, a totally implantable titanium pump project was initiated in 1995 to develop longer-term VADs for possible permanent applications. NEDO (New Energy and Industry Development Organization) is a sub-agency of the Japanese Ministry of Economy, Trade and Industry, and has supported this project during the past 10 years, ending in March, 2005.6,7 The NEDO implantable centrifugal pump system is a totally implantable biventricular assist device (BVAD) and a longer-term device applicable for longer than 5 years. This paper aims to summarize this 10-year NEDO progress and also to address further strategies for moving forward into the clinical arena.
Pulsatile implantable VADs are universally used for BTT, primarily supporting the left ventricular function. However, 20–30 % of left ventricular assist device (LVAD) patients develop multiorgan failure, which is a major cause of mortality at subacute or chronic clinical phases after LVAD implantation.8 This phenomenon is generally associated with an impairment of the portal vein circulation, high bilirubinemia, consequently caused by the right heart dysfunction. Therefore, a BVAD is becoming a more essential therapeutic tool to improve the clinical outcomes of mechanical circulatory support.8–11 Maintenance of the required total blood flow is occasionally difficult with only LVAD support in patients with severely deteriorated heart function of dilated cardiomyopathy, yet right ventricular assist device (RVAD) circulatory support in conjunction with an LVAD would be highly effective in achieving a total pump flow with appropriate pulmonary and systemic circulation. Contrary to this clinical demand, the current bulky pulsatile VADs are only capable of assisting left ventricular function. In addition, only large patients, who have enough volume capacity for device implantation in the body, will receive these LVADs. To be able to implant a BVAD inside of the body, the two pump systems should be well miniaturized. Particularly, a compact BVAD system, implantable in small patients (i.e., Asian female patients, adolescents), would be ideal for expanding the number of VAD patients.
The project aims to achieve: (1) dual centrifugal pumps with the ability for full biventricular support, (2) a compact system implantable in small patients, (3) a totally implantable system with transcutaneous energy transmission system (TETS), (4) a durable system with a lifetime of over 5 years, and (5) a system free of thrombus and with minimal hemolysis. Within the project framework, the final goal is to complete preclinical readiness studies and commence the clinical trials.
Project Transition from Phase I (1995–1999) to Phase II (2000–2005)
The initial 5 years (Phase I) were focused on developing a miniaturized implantable pump system based upon the Kyocera Gyro C1E3 together with the actuator system developed by the University of Vienna group.12 To assemble the totally implantable system, all of the necessary sub-components, including the TETS, were obtained internationally from available manufacturing companies or institutions. The titanium pump head was fabricated at the machine shop of Baylor College of Medicine, Center for Artificial Organ Development.
After confirming the basic mechanical feasibility and biocompatibilities, this Phase I program progressed into the Phase II program (2000–2005) in conjunction with the Japanese medical device company Miwatec (Tokyo, Japan.) During Phase II, the Phase I prototype design was improved to be a preclinical model. The system component technologies, such as the pump head, actuator, cannulae, TETS, and controller, were successfully transferred to Japanese companies (Table 1). Furthermore, flow visualization/simulation analyses were achieved by Dr. Takashi Yamane’s group, the National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan. The NEDO Phase II preclinical BVADs were tested under an extensive number of in vitro studies, endurance studies, and in vivo animal implantations to assess the device safety, effectiveness, durability, and reproducibility. Figure 1 shows the general system configuration of the NEDO totally implantable system in a human.
Kyocera Gyro C1E3 is a 1-month centrifugal pump made of polycarbonate with unique pump designs such as an eccentric inlet port and double pivot suspending impeller in conjunction with a seal-less magnetic coupling pump structure. This mother pump design was introduced and modified to meet the requirements for implantable pump applications. The diameter of the pump impeller was miniaturized from 63 mm to 50 mm, and the material of the pump housing was changed to titanium alloy (4Al6V Ti). However, the rest of the basic structure was unchanged, including the Kyocera fine ceramic shaft (Figure 2).
In addition to the material conversion and miniaturization of the impeller diameter, the straight inlet port was bent to improve anatomical fitting and surgical feasibility during device implantation. This curved inlet design was introduced in the Phase II program. The total height of the pump-actuator system was minimized by removing dead space in the pump head and actuator head, reducing the total height from 62 mm to 54 mm, further improving the anatomical fitting property (Figure 3). Specifications of the preclinical pump-actuator system are shown in Table 2. The size benefit of the preclinical pump-actuator system is shown in Figure 4 when it is compared with commercially available pulsatile LVADs.
The NEDO pump inherited the Kyocera gyro pump double-pivot impeller suspension system. Furthermore, the impeller magnet and actuator magnet balance were properly adjusted to generate a levitating impeller position, or “top-contacting mode,” by optimizing the magnet-to-magnet distance with the appropriate thickness of the spacer. During the pump hydraulic performance testing, standard PQ curves were achieved, and the impeller levitation point was plotted at several RPM points. The impeller position, both top-contacting mode and bottom contacting mode, were identified visually through direct endoscope examination from the blood outlet port. The NEDO actuator employed a fixed RPM control, instead of a voltage control. Whenever the operational RPM is set within a levitation range, there is no need to have a cumbersome control program for maintaining the levitated impeller position (Figure 5). Because the pump system will be implanted under a beating heart, which creates pulsatile pressure head, the spinning impeller will move up-and-down, and right-and-left. This hydrodynamic impeller suspension mode is analogous to mechanical heart valve prosthesis; thus, it is important to demonstrate antithrombogenicity for longer-term pump usage.
Flow Visualization Analysis
A centrifugal pump generates a vortex blood stream that produces a high-and-low pressure distribution in the pump head, which pushes out the blood through the outlet port. There are various rheological changes inside the pump head containing the impeller spinning at high RPM. Particularly, blood stagnation and high shear stress play a key role in terms of blood clot formation and red blood cell trauma, “hemolysis.” The NEDO pump was evaluated with particle tracing flow visualization analysis, with the aid of Dr. Takashi Yamane, director of medical technology, National Institute of Advanced Industrial Science and Technologies, Tsukuba, Japan.13 In his series of flow visualization studies, it was observed that the stagnant point was off center. Because the impeller is usually operated in the top-contacting mode, the bottom pivot space behind the impeller is continuously washed out (Figure 6). To confirm antithrombogenicity of the top-contacting mode of the impeller, a platelet adhesion study was conducted in a closed mock circuit primed with heparinized blood to compare the top-contacting and bottom-contacting mode. The top-contacting mode having continuous washout flow could significantly inhibit platelet adhesion at the bottom pivot area compared to the bottom-contacting mode of the impeller (p < 0.05) (unpublished data, T. Motomura, 2003). In contrast, the hemolysis value did not change regardless of the impeller operational mode. The Normalized Index of Hemolysis (NIH) was 0.0039 g/100 l and 0.0043 g/100 l in the bottom- and top-contacting mode, respectively, and the differences were insignificant. Therefore, the fixed RPM and top-contacting mode of the impeller were routinely applied for Phase II animal experiments. The preclinical in vitro hemolysis study simulating various pump positions is described in a later section.
For a clinical pump system, biocompatibility is an important issue. Particularly, long-term rotary blood pumps need to prove a low hemolysis property and antithrombogenicity. A series of biocompatibility studies were performed: antiplatelet effects of the blood contacting surface in different materials (normal titanium and chemically treated titanium surface) and different surface roughness of titanium plates, and also hemolysis studies simulating potential pump positions, which vary depending upon patient postures (Table 3).
In the initial material antithrombogenicity studies, polished titanium (0.2-micron surface roughness) demonstrated a 60% reduction in platelet adhesion compared to commercial grade polycarbonate surfaces. Material conversion from polycarbonate into titanium alloy (medical grade 4Al6V titanium alloy) was ideal from the standpoint of long-term endurance and also antithrombogenicity. From the platelet adhesion studies of different surface roughness of titanium, it was determined that a surface roughness of <0.2 microns, which showed a mirror-like appearance, is required for a long-term blood pump system. During Phase I and the first 2 years of Phase II, the Baylor group attempted to develop a chemically treated titanium surface to enhance antithrombogenicity of the original material. The “Titanium gel layer” was formulated by H2O2 chemical oxidization. It effectively inhibited platelet adhesion and thrombin generation.14 However, only a fine-polished titanium surface of <0.2-micron surface roughness was reproducible and feasible enough to demonstrate antithrombogenicity during in vivo pump implantation for over 3 months. Therefore, a simple smooth surface finish was selected rather than a cumbersome and costly chemical treatment, or any other coating procedures.
ASTM standard procedures were performed to evaluate the hemolysis characteristics of the NEDO preclinical pump system, which demonstrated the clinically acceptable NIH. value at 0.0039 g/100 l.15–17 In a clinical setting, the pump system is practically operated under various positions of the patient, such as a standing position, a reclined position, and others. Therefore, in addition to the standard hemolysis study at the horizontal position, 90- and 180-degree-angled (upside-down) positions of the pump head were tested to simulate different patient postures. In contrast to the horizontal position, the NIH. value was slightly increased to 0.007 g/100 l and 0.009 g/100 l in 90- and 180-degree-angled conditions, respectively. However, these results also indicated that the NEDO pump would be secure to operate within the range of a clinically acceptable NIH value of <0.02 g/100 l, regardless of the possible patient postures.
Actuator and Pump Controller Development
The Phase I voltage control actuator was developed by Dr. H. Schima, University of Vienna. The actuator and motor driver were independently fabricated. During the Phase II program, a new brushless DC motor system with built-in controller driver was developed by a Japanese motor company (Softronics, Co., Ltd, Saitama, Japan). The operational RPM had a control voltage ranging from 0 to 5 V, with an RPM ranging from 0 to 2400. This actuator system was equipped with a feedback control system to maintain the steady setup RPM. This all-in-one (built-in controller unit) RPM control actuator was tested using an in vitro mock loop, presurgical burn-in testing, 2-year endurance studies, and long-term animal experiments. The preclinical pump-actuator system demonstrated superior pump efficiency. The power consumption was 6–6.4 W under LVAD operational condition in the in vitro studies (room temperature, 37% glycerin water, closed mock loop with 400 ml venous reservoir, 5 l/min pump flow against 100 mm Hg pressure head). A multistranded double helical lead configured at six concentric positions was incorporated into the actuator cable to improve high-frequency flex durability. The hermetic property of the preclinical actuator was tested under positive and negative pressures, and no water invasion was identified. Currently, eight actuator systems have been installed in the total endurance stations to confirm long-term mechanical liability.
Even though the NEDO pump-actuator system uses a fixed RPM system, the pump flow can be autoregulated depending upon the change of inflow pressure, which usually represents the left ventricular end-diastolic pressure. Figure 7 indicates that the NEDO pump flow varies from 3.5 to 6 l/min at 2000 fixed RPM pursuant to the increment of inflow pressure. This fixed RPM pump controlling method is clinically used in several axial flow and centrifugal implantable VADs, and is widely accepted.18,19 It is believed that a fixed RPM control will be also clinically feasible for the NEDO pump system, yet another feedback control algorithm may be necessary to avoid the inflow sucking in case of recovered heart. Pump flow autoregulation has been show to be effective during long-term animal BVAD experiments. In fact, the LVAD flow, RVAD flow, and total cardiac output were increased, simultaneously, during treadmill exercise testing, and no abnormal physiologic findings were observed (unpublished data, T. Motomura, D. Ogawa, A. Hata et al., 2004). Independent LVAD and RVAD flow control is needed to meet the systemic and pulmonary circulatory flow requirements; however, an automatic feedback control program will not be incorporated into the system. LVAD and RVAD flow balance is a controversial issue. Conventional extracorporeal BVAD flow rates are manually adjusted by cardiac surgeons or cardiologists in the operating room and intensive care unit. Thus, during initial clinical trials of our external BVAD system will also employ manual adjustment and control of this flow balance.
Transcutaneous energy transmission is an essential component needed to make a totally implantable VAD possible and beneficial by avoiding a drive-line infection. In Phase I, the WorldHeart prototype TETS (WorldHeart Corporation, Ottawa, Canada), was incorporated into the total system (Table 1). In Phase II, Miwatec started the development of a more efficient TETS in transmitting power with the technical support from Hokkaido Tokai University (Sapporo, Japan).20 For improvement of power transmission and voltage adjustment for the NEDO system, the TETS internal and external coil were optimized in size and other coil materials including potting materials. The TETS internal circuit was equipped with a position sensor of the external-internal coil to prevent dislodgement of the two transmission coils. When dislodgement of a coil occurs, a warning signal is sent to the internal TETS circuit immediately, and simultaneously, the pump operation is switched from the TETS power provision to the internal battery operational mode without interruption of the motor rotation. The NEDO-Miwatec TETSs were specially designed to be compatible with the NEDO pump system. Furthermore, Miwatec developed a more widely compatible TETS unit, which could be used for other VAD systems with a maximum power capacity of 60 W. In comparison with the Phase I prototype, the Miwatec TETS improved the DC-DC power transmission efficiency. The Miwatec version 3.0 TETS demonstrated over 85% efficiency of the power transmission between 10–20 W power consumption (Figure 8). The design improvement of the TETS coil is one important step in Phase II before moving forward towards the clinical stage. A rigid shell was used so that the external coil has a self-fitting feature with minimum device fixation. No skin problem, such as infection and other necrosis injuries, was observed in the in vivo animal studies. The external coil fitting was very secure. For further improvements, a ring-type external coil fully potted with silicone rubber was developed for minimizing the skin irritation for a longer period of time and helped miniaturize the coil in size and weight. Approximately 30% miniaturization was achieved, maintaining the same power transmission efficiency. A ring-type external coil was believed to release heat and humidity between the contacted skin and the external coil surface. The preclinical TETS is shown in Figure 9. The latest model was used in a bench test circuit at Hokkaido Tokai University and has been operating for over 400 days (as of June 2005). This TETS was incorporated into the total NEDO BVAD endurance station in 2004 to evaluate the totally system endurance at Baylor College of Medicine, Center for Artificial Organ Development.
A preliminary BVAD endurance study was performed using a Donovan-type biventricular mock circuit for 1 year to independently evaluate wear to the left and right pump polyethylene pivot bearing. Since LVAD pumps need to operate at higher RPM to generate pump flow against the higher after load of the RVAD, the bearing wear increases in LVAD pumps, indicating the need for a bearing life expectancy of over 2 years.21 There was no significant wear of the ceramic shaft or the bottom polyethylene pivot bearing. Considering the information from previous preliminary bearing studies, we selected the LVAD condition, 5 l/min pump flow against 100 mm Hg (peak pressure) of afterload, as a more severe operating condition for longer-term total system endurance evaluation. A pusher plate-type pulsatile master pump was fabricated and used in the endurance station, maintaining approximately 2l/min pulsatile pump flow with 78 bpm pulse rate and 28 ml of the pump stroke volume. The mock loop was primed with 0.9 % saline solution, and the total system was immersed into 0.9 % saline to simulate physiological condition (Figure 10). So far, eight endurance stations have been testing the LVAD condition for over 1 year, and even up to 450 days. Accumulated continuous running dates of these endurance stations are 4,126 days (as of June 1st, 2005), and there were no mechanical or electrical failures throughout the series of the experiments. Since the TETS and information transmission units were employed in the endurance station on the final year of Phase II (year 2004), long-term durability has only been tested for < 1 year in the total endurance station. This 2-year endurance study will be continued for at least 2 years, and will continue to 3 years to confirm total system durability.
The final goal of the animal study is to achieve eight chronic implantations for longer than 3 months. Six to eight-month-old miniaturized Dexter strain calves weighing 70–100 kg were used to rule out potential residual cardiac shunt anatomy (i.e., atrial septal defect, patent ductus arteriosus). The pump-actuator system was implanted under beating condition without cardiopulmonary bypass. The LVAD was implanted at the bypass position between the left ventricular apex and the descending aorta, and the RVAD between the right ventricle and the main pulmonary artery to establish biventricular bypass. Heparin was used for acute postoperative anticoagulation, and warfarin was administered as chronic anticoagulant therapy in order to maintain the international normalized ratio between 2.5 and 3.5.
During the entire 10-year program, a total of 108 animal experiments were used for acute experiments and long-term chronic implantation studies. Out of 108 experiments, 91 cases were acute studies to confirm the basic system function, concept feasibility, and control algorithm of the pump. Chronic studies longer than 1 month were performed in 17 cases.
Out of the 17 cases, which achieved chronic implantation for longer than 1 month, 12 cases did not reach the 3 month-study period due to several reasons: (1) actuator cable breakage, (2) outflow graft occlusion due to towing of flow probe cable, (3) inflow obstruction due to pannus over growth around the inflow titanium tip ostium, (4) bleeding from gastric ulcer, and (5) power outage due to a natural disaster.
Studies longer than 3 months were completed in two LVAD cases (208 days and 280 days) and three BVAD cases (scheduled termination at 3 months) evaluating the eight pump-actuator systems. Within these eight pump systems that operated for longer than 3 months, there was no evidence of pump thrombus, neurologic complication, or mechanical failure. Plasma free hemoglobin during the entire experiment course was maintained within normal range (average value was 4.5 mg/dl). LVAD and RVAD pump flow were well maintained at 4.6 ± 0.8 l/min and 4.3 ± 1.1 l/min, requiring an RPM of 1823 ± 206 and 1732 ± 254, respectively (Table 4). The average pump power consumption was 5.0 W in LVAD and 4.5 W in RVAD.
Moving Forward to the Clinical Arena
NEDO preclinical studies were completed with an extensive number of in vitro studies, chronic animal studies, and ongoing bench tests for device safety, effectiveness, reliability, and durability. In preparing for clinical application, final anatomical fitting studies using human cadavers were performed, particularly to determine the appropriate cannula design and length for human application. The NEDO total BVAD system includes two pump-actuator systems, TETS, and an internal lithium ion battery, which can be fully implanted into a small human cadaver, whose body surface area is <2.0 m2. However, further system miniaturization such as integration of TETS and information transmission unit should be accomplished for better anatomical fitting in smaller patients. Because the NEDO cannula system is composed of a combination of wire-reinforced silicone tubing and vascular graft, it is possible to use it in either implantable or extracorporeal (wearable) applications. In the wearable application, a silicone tubing portion will penetrate the patient abdominal skin area to provide pump flow from the pump head, which will be placed outside of the body.
As mentioned in the endurance testing paragraph, the TETS and information transmission units should be tested for additional years. Meanwhile, the NEDO pump system will be applied as wearable VAD (Step 1). The system can be used either as a single VAD or BVAD depending upon each clinical demand. The system will be implanted intracorporeally in Step 2, but it will have a drive line to connect the pump-actuator and external battery unit. Step 3 will be the project goal: a totally implantable system with TETS. This stepwise strategy is being considered to propel the clinical program (Figure 11).
For several biologic safety studies, and other laboratory-based studies, study protocols were established to implement Good Laboratory Practice with proper documentation including standard operating procedures and a wide spectrum of data assurances. These extensive protocol and documentation procedures include in vitro biocompatibility studies, endurance bench tests, and in vivo animal experiments, and were designed based upon the STS-ASAIO Long-Term Mechanical Circulatory Support System Reliability Recommendation published in 1998,22 to confirm long-term device reliability. In addition to laboratory-based preclinical studies, the system manufacturers are currently compiling data to comply with the US Food and Drug Administration Quality Regulation System, ISO 13485, and other ISO documents associated with the system risk analyses, which are mandatory for future clinical trials.
The NEDO centrifugal VAD is a highly versatile system that can be used for either a single VAD or BVAD and is also applicable for either wearable or implantable use (Figure 12). A versatile NEDO VAD/BVAD system will be an ideal mechanical circulatory assistance for various clinical demands, including postcardiotomy heart failures, BTT, and destination therapy.
1. Orime Y, Takatani S, Sasaki T, et al: Cardiopulmonary bypass with Nikkiso and BioMedicus centrifugal pumps. Artif Organs
18: 11–16, 1994.
2. Takano T, Nakata K, Yoshikawa M, et al: Development of an antithrombogenic and antitraumatic blood pump: The Gyro C1E3. ASAIO J
46: 123–127, 2000.
3. Takami Y, Nakazawa T, Makinouchi K, et al: Biocompatibility of alumina ceramic and polyethylene as materials for pivot bearings of a centrifugal blood pump. J Biomed Mater Res
36: 381–386, 1997.
4. Tayama E, Niimi Y, Takami Y, et al: Hemolysis test of a centrifugal pump in a pulsatile mode: the effect of pulse rate and RPM variance. Artif Organs
21: 1284–1287, 1997.
5. Nakata K, Yoshikawa M, Takano T, et al: Antithrombogenicity evaluation of a centrifugal blood pump. Artif Organs
24: 667–670, 2000.
6. Yoshikawa M, Nonaka K, Linneweber J, et al: Development of the NEDO implantable ventricular assist device with Gyro centrifugal pump. Artif Organs
24: 459–467, 2000.
7. Ichikawa S, Linneweber J, Motomura T, et al: In vivo
evaluation of the NEDO biventricular assist device with an RPM dynamic impeller suspension system. ASAIO J
49: 578–582, 2003.
8. Nosé Y, Ohtsubo S, Tayama E: Therapeutic and physiological artificial heart: future prospects. Artif Organs
21: 592–596, 1997.
9. Schiessler A, Friedel N, Weng Y, et al: Mechanical circulatory support and heart transplantation: Pre-operative status and outcome. ASAIO J
40: M476–481, 1994.
10. Potapov EV, Merkle F, Guttel A, et al: Transcontinental transport of a patient with an AbioMed BVS 5000 BVAD. Ann Thorac Surg
77: 1428–1430, 2004.
11. Drews T, Loebe M, Jurmann M, et al: Outpatients on biventricular assist devices. Thorac Cardiovasc Surg
49: 296–299, 2001.
12. Ohtsuka G, Nakata K, Yoshikawa M, et al: Long-term in vivo
left ventricular assist device study with a titanium centrifugal pump. ASAIO J
44: M619–623, 1998.
13. Yamane T, Kodama T, Yamamoto Y, et al: Flow visualization in a centrifugal blood pump with an eccentric inlet port. Artif Organs
28: 564–570, 2004.
14. Ishitoya H, Kawamura M, Linneweber J, et al: Titania gel reduces thrombin generation. Artif Organs
26: 959–963, 2002.
15. Standard Practice for Selection of Blood for In Vitro Evaluation of Blood Pumps, 2005. American Standard for Testing and Materials
: F1830–1897, 2005.
16. Standard Practice for Assessment of Hemolysis in Continuous Flow Blood Pumps, 2005. American Standard for Testing and Materials
: F1841–1897, 2005.
17. Nosé Y: Design and development strategy for the rotary blood pump. Artif Organs
22: 438–446, 1998.
18. Nojiri C, Kijima T, Maekawa J, et al: Terumo implantable left ventricular assist system: Results of long-term animal study. ASAIO J
46: 117–122, 2000.
19. Frazier OH, Myers TJ, Westaby S, Gregoric ID: Clinical experience with an implantable, intracardiac, continuous flow circulatory support device: Physiologic implications and their relationship to patient selection. Ann Thorac Surg
77: 133–142, 2004.
20. Okamoto E, Inoue T, Watanabe K, et al: Development of an implantable high-energy and compact battery system for artificial heart. Artif Organs
27: 184–188, 2003.
21. Asai T, Watanabe K, Ito S, et al: Real-time studies of the pivot bearings in the NEDO Gyro PI-710 centrifugal blood pump. Artif Organs
28: 899–903, 2004.
22. Altieri F, Berson A, Borovetz H, et al: Long-term mechanical circulatory support system reliability recommendation. American Society for Artificial Internal Organs and Society of Thoracic Surgeons: Long-term mechanical circulatory support system reliability recommendation. ASAIO J
44: 108–114, 1998.