Secondary Logo

Journal Logo

Special Reports

Artificial Heart Research and Present Status of Clinical Application in Japan

Imachi, Kou

Author Information
doi: 10.1097/01.mat.0000196711.98042.12
  • Free


Japan’s history of artificial heart (AH) development began with Atsumi in 1959 at the University of Tokyo. I joined the project in 1970. Since that time, many AH types, as well as materials, blood pumps, valves, driving mechanisms, and control methods, have been investigated at the University of Tokyo. Figure 1 provides a timeline of AH development.1

Figure 1.
Figure 1.:
Timeline of total artificial heart development at the University of Tokyo

Total Artificial Heart Development

With the goal of developing an implantable total artificial heart (TAH) (Figure 2), four types of TAHs have been investigated: (1) TAH with fibrillated natural heart (FTAH), placed paracorporeally, 1970–1980; (2) a hybrid TAH (HTAH), placed paracorporeally and used with the beating natural heart in which the roots of the ascending aorta and pulmonary artery were completely occluded, 1978–1983; (3) the total replacement artificial heart (TRAH), placed paracorporeally with the natural heart resected, 1981–1995; and (4) the implantable TAH system (ITAH), completely implanted into the body with electrical energy sent through a transcutaneous energy transmission system (TETS), 1996 to present. Goat survival was 174 days with FTAH, 288 days with HTAH, and 532 days with TRAH.2 Presently, we are developing a fully implantable TAH using an undulation pump that will be mentioned later.

Figure 2.
Figure 2.:
Various total artificial heart methods developed at the University of Tokyo

Pump Materials

As shown in the timeline, our first blood pump was made of natural rubber. Silicone rubber was used between 1964 and 19713; since 1972, polyvinyl chloride (PVC) paste resin has been used as the base material of the blood pump because of its easy fabrication. A coating of Avcothane (name changed to Cardiothane in 1983) block copolymer of segmented polyurethane with polydimethyl siloxane on the blood contact surface of the PVC blood pump was added in1978, which greatly improved antithrombogenicity.4 This technique was transferred to Nippon Zeon Co. Ltd, and they developed a sac-type blood pump for clinical use in 1983. Many kinds of segmented polyurethane were evaluated for long-term blood compatibility during this time (Figure 3). KIII block copolymer of polyurethane with polydimethyl siloxane, made by Nippon Zeon Co. Ltd., was selected and has been used for pump material since 1990.6

Figure 3.
Figure 3.:
Multimaterials test method of in vivo antithrombogenicity using artificial heart blood pump (left) and its scanning electron microscopy (SEM) result at the interface between a test material and base material (right).

Blood Pumps

Various types of blood pumps (tube, bellows, diaphragm, sac) have been developed (Figure 4). After the pneumatically driven method was adopted, sac-type pumps came into use, which were superior to the diaphragm pump with regard to blood compatibility, durability, and performance.

Figure 4.
Figure 4.:
Various kinds of blood pumps and valves developed at the University of Tokyo.


Numerous types of valves were incorporated into the blood pumps, such as the ball and disc valve, pneumatically driven valve, oblique elliptical seat door (OED) valve, OED valve with heparin gland, Bjork-Shiley (B-S) valve, and jellyfish valve. The OED valve (Figure 4) was a homemade door-type valve in which an elliptical valve plate was supported by two polyester sutures that worked as hinges. Although it had very good performance of the short opening and closing time, thrombi were often formed around the polyester sutures. Heparin sacs, from which heparin was exuded to the blood through hinges, were attached. These worked like heparin glands and were very effective in protecting against thrombus formation. However, the OED could not be used after we started coating the PVC pump surface with Avcothane, because the Avcothane layer came off at the valve seat as a result of the repetitive hitting by the valve plate at its closure. In 1978, the OED was replaced with the B-S valve, which worked well up to 344 days in goat experiments and was used also in a clinical pneumatic VAD pump made by Nippon Zeon Co. Ltd. (Figure 4). However, ring thrombus was often formed at the small gap between the valve ring and pump housing.

To eliminate the gap and fabricate seamlessly between valve ring and the housing, the jellyfish valve was designed.5Figure 5 shows the mechanism and simple structure of the jellyfish valve, which is composed of a valve membrane and a valve seat. The valve membrane, made of segmented polyurethane (40 to 50 μm thick), is adhered at its center to the center of a valve seat. The valve seat, with 12 spokes (20 mm diameter) to prevent prolapse of the membrane when it closes, is made of urethane resin and coated with segmented polyurethane. The performance of the jellyfish valve is superior to that of the B-S valve, with low flow resistance, 1/4 of regurgitant flow during the valve closure, no leakage flow after the valve closure, and high-frequency response of more than 300 beats/min. The blood pump with the jellyfish valves seamlessly incorporated showed very good performance and antithrombogenicity in long-term TRAH goat experiments without anticoagulant. However, creep fatigue of the membranes was found after 312 days in the left side and 414 days in the right side. The membrane was stretched repetitively between and on the spokes when it closed, and calcification was found only on the stretched side of the membrane.6 Modification of the valve seat design to avoid strain concentration of the membrane has been done by Iwasaki et al.7 using the finite element analysis (FEM) method. According to FEM analysis, an additional thin rim was added as shown in the lower right panel of Figure 5. It works very effectively to prevent creep fatigue as well as calcification of the membrane.

Figure 5.
Figure 5.:
The jellyfish valve and its principle.

Drive Mechanisms

The drive mechanism used in the first animal experiment of AH with a dog in 1960 was a hydraulically driven unit. The microroller pump, DC micromotor, and cam-driven bellows pump were developed as shown in Figure 4 (1960 to 1962). A digital computer-controlled drive unit, composed of a pulse motor and cardan gear that changed piston speed, was developed to control blood pressure waveform between 1964 and 1966. A pneumatic driven unit was developed in 1963. A compact drive unit controlled by fluid amplifier was developed for left ventricular assist between 1966 and 1970. After several improvements in the 1970s, a clinical pneumatic drive for VADs named the Corart system was developed in 1983 in cooperation with Aisin Seiki Co. Ltd. Recently, brushless DC motor-driven undulation pump TAHs and VADs are under development and will be mentioned later.

Control Methods

The control method of the TAH was the most prominent factor in prolonging the survival period of TAH goats at the University of Tokyo. Table 1 shows the relationship between control method and pathophysiological state of our TAH goats. TAH control began with Starling’s law control method in 1970, in which the entire venous return was sent to systemic circulation through the right and left pumps. This method was initiated by Kolff’s group in 1963 and has been investigated by many institutes since that time. However, we found that cardiac output (CO) increased to 150 to 200 ml·kg–1·min–1 within a few days under this control method, and accordingly, TAH goats could not stand up, had no appetite, and developed serious pathophysiological states such as high central venous pressure (CVP), strong anemia, low total protein, and anuria. Severe congestion and central necrosis of liver, cortical necrosis of kidney, constriction of arteriole were common pathological finding. TAH goats couldn’t survive for more than 10 days. In these experiments, we placed the TAH paracorporeally and used large blood pumps (compared with the goat’s size), so CO was easily increased by the limit of pump stroke volume. We thought the cause of these pathophysiological abnormalities was hyper-CO and named this phenomenon hyper-CO syndrome.8 Then, we began to restrict CO between 80 and 100 ml·kg–1·min–1, which was considered the physiologically normal level (fixed CO control method). The TAH goats’ pathophysiological states were dramatically improved as shown in Table 1. The mean and the longest surviving period of our TAH goats increased by 10 times, and their general condition, including kidney function, became almost normal. However, several pathophysiological states remained, such as mild anemia (hematocrit was 20–25%), low thyroid hormones, and high CVP, followed by large amounts of ascites and liver congestion. We found that in HTAH goats, whose natural hearts kept beating, these pathophysiological states almost disappeared. We speculated that the cardiac nerve impulses to the cardiovascular center of the brain must be very important to maintain homeostasis of the living body (cardiac receptor hypothesis). We then tried to add electrical stimulation to a fixed CO controlled TRAH goat (whose natural heart was resected) through two electrodes put on the left atrium and the skin with a cardiac pacemaker instead of a natural heart.9 This effectively improved all pathophysiological states except high CVP, followed by large amounts of ascites and liver congestion. The TRAH goat survived for 344 days in 1984, the world’s longest survival record of a TAH animal at that time. Finally, we tried to control CO according to an inverse function of total peripheral resistance (TPR). We named this the 1/R control method and tried to have the TAH goat herself control CO using biofeedback. The basic principle of the 1/R control method is that, because there is no nervous connection between the TAH and the cardiovascular center of the TAH animal, the animal cannot control CO even if it is inadequate. However, the animal still has the ability to change its TPR with biofeedback. So, if CO can be changed according to a function of TPR, the TAH animal can choose adequate CO by changing his TPR by itself. We developed the following equation through acute and chronic animal experiments:

Table 1
Table 1:
Influence of the Control Method on Pathophysiology of TAH Animals

Co(n) = (AoP.set(n)–Rap.set) / Tpr + Cp(AoP.set(n)–Aop)

Where CO(n) is next target CO, AoP.set(n) is setting point of aortic pressure calculated as the running mean value during 12 hours, RAP.set is setting point of right atrial pressure, CP is gain of correcting term, and AoP is mean aortic pressure. The equation was installed into a computer. The left pump output, left and right atrial pressure, and aortic pressure were measured, and TPR was calculated automatically. The next target of CO was calculated according to the equation, and the driving parameters of TAH were set as the target CO was attained. The result under 1/R control method was amazing. CO increased automatically in accordance with the animal’s states such as eating, drinking, or exercising on treadmill. All the pathophysiological abnormalities disappeared almost completely as shown in Table 1, and the longest surviving period of the TRAH goat was prolonged to 532 days with no abnormality until she died due to human error.10

Implantable TAH

Because of the successful progression from FTAH through TRAH, we began to study implantable TAHs and VADs. We developed a new blood pump, the undulation pump (UP), composed of a housing, disc, driving shaft, and brushless DC motor. The UP is a special kind of rotary blood pump in which the rotational motion of the motor is converted into undulation motion of the disc by a special mechanism of the driving shaft. The small clearance between the disc and housing moves from the pump inlet to the outlet according to the undulation motion, which squeezes out the blood to the outlet and sucks the blood from the inlet in the same moment. The UP has several advantages compared with other rotary blood pumps and the pulsatile AH: (1) a large amount of output can be produced with comparatively low rotational speed; (2) arbitrary flow patterns from continuous flow, pulsatile flow on continuous flow, to complete pulsatile flow can be easily produced by controlling the motor drive current wave form; (3) the compact size of the pulsatile TAH and VAD makes them suitable for implantation (UPTAH is 76 mm in diameter, 79 mm long; UPVAD is 69 mm in diameter, 35 mm thick, Figure 6); and (4) no compliance chamber is required when it is implanted inside the body.11

Figure 6.
Figure 6.:
Undulation pump total artificial heart (UPTAH) and undulation pump ventricular assist device (UPVAD)

Goat survival with the UPTAH was 63 days and with the UPVAD, 4 months. Figure 7 is a comparison of our UPTAH with the AbioCor TAH (AbioMed Co. Ltd). The UPTAH system can be implanted into a goat with less than 40 kg body weight. Since 2002, we have been developing an implantable UPVAD system under a multiinstitutional research project supported by The Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation. Six universities (Hokkaido University, Hokkaido Tokai University, Tohoku University, Waseda University, University of Tokyo, and Kyushu University) and two companies (Tonokura Ika Kogyo Co. Ltd, and Nemoto Project Industry Co. Ltd.) are participating in this project. Blood pump, internal battery unit including motor drive and control circuit and transcutaneous information transmission system, and TETS are being developed and implantations into goats have been performed.

Figure 7.
Figure 7.:
Size comparison between the UPTAH and AbioCor.

Miniaturized AH Technique

Several miniaturized techniques have been developed by our group. A small pneumatic blood pump with 2 ml of stroke volume was developed for a rabbit model around 1972, by which we tried to observe microcirculation during AH pumping through the rabbit’s ear chamber. A small jellyfish valve (4 mm diameter) was developed for installation into a catheter type AH in 1990.12 A miniaturized undulation pump (38 mm diameter, 33 mm length) was developed for a distributed AH that is my future dream for AHs in which they will be distributed for each organ and tissue (Figure 8). As the first step, we tried to connect it to the renal artery to perfuse the kidney in situ.13 These techniques offer strong support for the development of a pediatric AH in the near future.

Figure 8.
Figure 8.:
Concept of distributed artificial heart.

Other AH Research in Japanese Institutes

In Japan, 13 Institutes are participating in AH development (Table 2). AH research at Tohoku University was started by Professor Nitta in 1970. A polyurethane blood pump with endothelial cell cultured surface was developed as a first step. A pneumatically driven VAD for clinical use, a vibration blood pump, and other types have been developed. The TETS, which provides 88% of DC/DC efficiency, was also developed at Tohoku University by Professor Matsuki’s group in the Department of Electrical Engineering. The National Cardiovascular Center (NCVC) in Osaka, beginning in 1976, has developed a pneumatically driven VAD, a TAH, an electrohydraulic driven TAH, and a centrifugal pump (CFP). Hokkaido University is developing an electromechanically driven VAD and axial flow blood pump (VALVO pump), which will be implanted at the aortic valve position. They also established the computational fluid dynamic technique to investigate blood flow pattern in a pump and to estimate blood compatibility. The Tokyo Medical & Dental University is developing an electromechanical VAD and TAH, and a centrifugal flow pump (CFP) VAD. The Ibaragi University group is working on a magnetically levitated CFP and axial flow pump with a direct drive motor. Recently, Tokyo Denki University started to develop a VAD driven with a linear oscillatory actuator. A pneumatically driven spiral vortex with a mock circulatory loop that can generate a pressure waveform similar to that of the living body was developed at Waseda University, as well as test systems for accelerated durability and in vitro antithrombogenicity. Hokkaido Tokai University is developing a transcutaneous information transmission system and an implantable early diagnostic probe for pump failure.

Table 2
Table 2:
Japanese Institutions Conducting Artificial Heart Development and Research

Circulatory Support Devices for Clinical Use in Japan

Two pneumatically driven ventricular assist devices, approved by our government in 1990, were the first commercial devices in the world at that time. One was the Aisin-Nippon Zeon system originally developed at the University of Tokyo and the other was the Toyobo system originally developed at NCVC (Figure 9). They are still being used clinically. Four pulsatile VADs—Novacor, HeartMate IP, HeartMate VE, and BVS5000—are imported from the United States. Novacor and BVS5000 were approved by our government and were applied to the health insurance by the Japanese Government. HeartMate VE is undergoing clinical testing. In addition to these devices, six centrifugal pumps (three domestic and three imported) are available for short-term cardiac assist, percutaneous cardiopulmonary support (PCPS), and extracorporeal membrane oxygenation (ECMO). Recently, two domestic semi-implantable centrifugal pump VADs for long-term use were developed and are undergoing clinical testing. One is the DuraHeart made by TERUMO Corporation, and the other is the EVAHEART made by Sun Medical Technology Research Corporation.

Figure 9.
Figure 9.:
Clinical pneumatic VADs used in Japan (left: Aisin-Nippon Zeon system; right: Toyobo system).

Present Clinical Status of Adult Cardiac Support in Japan

The first clinical application of VAD in Japan occurred in 1980 at Mitsui Memorial Hospital. Pneumatically driven sac-type blood pumps, developed by the University of Tokyo group in cooperation with Toray Co. Ltd., were used as the biventricular bypass for 3 days. Clinical trial of the Aisin-Zeon and Toyobo systems was initiated in 1982. Their official clinical testing began in 1986 and the systems were approved by our government as commercially available devices in 1990. They were applied to health insurance by the Japanese Government in 1994. In 1997, the “organ transplantation law from the brain death patient” was passed through the Diet, and heart transplantation was reopened in 1999.

According to the registry of the Japanese Society for Clinical Ventricular Assist Systems, 697 cases of VADs were used between 1980 and August of 2004. Toyobo systems were used for 411 cases (59%), Aisin-Zeon systems for 158 cases (22.7%), BVS5000 for 61cases (8.8%), Novacor for 22cases, HeartMate IP for 18 cases, HeartMate VE for 7 cases, and others for 20 cases. Excluding BVS5000 cases, 417 cases were noncardiomyopathy (non-CM) patients and 219 cases were CM patients. As for the VAD application method for non-CM patients, 347 cases (83.8%) were left VADs, 24 cases (5.8%) were right VADs, and 43 cases (10.4%) were bi-VADs. The weaning rate of left VADs, right VADs, and bi-VADs was 41.8%, 58.3%, and 18.6%, respectively; the survival rate was 25.1%, 45.8%, and 11.6%, respectively. The total weaning rate and survival rate was 40.3% and 24.9%, respectively. The mean supporting duration for non-CM patients was 19.6 days for total cases, 18.6 days for nonweaned cases, and 21.0 days for weaned cases. The main causes of death for non-CM patients were: multiorgan failure (MOF), 77 cases (59 nonweaned, 18 weaned); heart failure, 63 cases (54 nonweaned, 9 weaned); bleeding, 29 cases (24 nonweaned, 5 weaned); brain damage, 28 cases (21 nonweaned, 7 weaned); infection, 22 cases (15 nonweaned, 7 weaned); respiratory failure, 15 (13 non-weaned, 2 weaned); and renal failure, 15 cases (6 nonweaned, 9 weaned).

Among 219 CM patients, 185 cases were dilated cardiomiopathy (DCM), 13 hypertrophic cardiomiopathy (HCM), 14 ischemic, 4 post partum CM, and 3 others. Age was between 8 and 78 (mean ± SD, 38 ± 14) years. Blood pumps and pumping duration used for CM patients were: Toyobo 161 cases, 4 to 1,245 days; Nippon Zeon 15 cases, 1 to 178 days; Novacor 19 cases, 6 to 1,090 days; HeartMate IP 17 cases, 2 to 689 days; and HeartMate VE 7 cases, 25 to 993 days. Twenty-eight VAD patients received heart transplantation (15 cases in Japan and 13 cases overseas), and 34 patients were weaned from VADs resulting in 24 cases of survival, 1 case of heart transplantation, 1 case of re-VAD, and 8 deaths. The causes of death for CM VAD patients were brain damage in 47 patients, MOF in 35 patients, infection in 16 patients, heart failure in 9 patients, hepatic failure in 3 patients, bleeding in 3 patients, renal failure in 1 patient, and device failure in 1 patient.

Pediatric VADs in Japan

There is no VAD system for pediatric patients in Japan. Toyobo made a pneumatically driven pediatric pump that was used for 9 patients around 1986, but its development was stopped because of low performance. According to the registry of Japanese Society for VAS, 38 cases of VAD were performed for patients under 18 years (0.8–17 years) since 1980. The blood pumps used were Toyobo in 30 cases, Nippon Zeon in 4 cases, Novacor in 1 case, HeartMate IP in 1 case, and BVS5000 in 2 cases. The pumping duration was between 0.04 and 689 days (mean 112 ± 170 days). Their clinical results were: 7 heart transplantations, 5 weaned and survived, 4 weaned but died, 17 died, and 5 cases ongoing at that moment.

Percutaneous Cardiopulmonary Support In Japan

According to the registry of Japanese Society of PCPS, clinical use of PCPS In Japan began in 1989, and by 2002 the number of cases had increased to 600 to 800 per year, and included 40 to 60 cases involving patients under 20 years (Figure 10). The application of PCPS between 1997 and 2002 was 3,708 cases: 46% for acute heart and lung insufficiency, 28% for after open heart surgery, 17% for emergency, 6% for supported percutaneous transluminal coronary angioplasty, and 3% for after lung and bronchial surgery. The total survival rate of PCPS patients between 2000 and 2002 was approximately 45%; however, the survival rate after lung and bronchial surgery was 85%. The causes of death in PCPS were cardiogenic in 71% of patients and noncardiogenic in 29% (12% MOF, 7% respiratory failure, 5% bleeding, 3% neurologic complication, and 2% infection). One of the reasons for the low number of thromboembolic complications is that a heparin-coated circuit was used for more than 90% of recent cases.

Figure 10.
Figure 10.:
Transition of annual clinical cases of PCPS in Japan

At the present time, there are no precise registry data for PCPS in Japan. Table 3 shows PCPS data of 43 patients under 18 years from three representative heart institutes, Tokyo Women’s Medical University, Osaka University, and the National Cardiovascular Center. Age of patients, case number, duration of support, and clinical outcome are shown. The longest supporting duration was 25 days. Total survival rate was 28% and 19% of patients moved to VAD. The death rate of children under 6 months was 71%, which was much higher than that of children older than 2 years (42%).

Table 3
Table 3:
Results of PCPS for 43 Patients Under 18 Years


Because of the increased reliability of AHs for adult patients, the requirement for pediatric VADs has become stronger year by year. Two German companies, Medos and Berlin Heart, have developed small-sized pneumatically driven blood pumps for pediatric use and have applied them in cases of small children and infants. In the United States in 2004, the National Institutes of Health contracted with five institutes for the development of a pediatric VAD within 5 years.

Unfortunately, there is neither a blood pump for pediatric use nor a project to develop pediatric VADs in Japan at this moment. As mentioned above, Japan has one of the greatest shortages of heart donors in the world. Moreover, because of the present law regarding organ transplantation from a brain-dead patient, it is prohibited for pediatric patients under 6 years to donate their organs, even if there is agreement by patients and their families.

Although the number of candidates for pediatric VAD in Japan is small compared with that in the United States and Europe, PCPS has been performed for more than 50 pediatric patients a year, resulting in 30% to 40% survival rate. It is very difficult to keep patients alive for more than 2 weeks under PCPS; therefore, developing a pediatric VAD in Japan is an urgent requirement.


I express my sincere appreciation to Dr. Kawai (Tokyo Women’s Medical University), Professor Kyo (Saitama Medical College), Dr. Matsumiya (Osaka University), and Dr. Nakatani (National Cardiovascular Center) for providing information concerning pediatric mechanical circulatory support at their institutions. The clinical data for VAD and PCPS was provided by the registry of Japanese Society for VAS and Japanese Society for PCPS, respectively.


1. Imachi K, Chinzei T, Abe Y, et al: Implantable total artificial heart: History and present status at the University of Tokyo. J Artif Organs 2: 13–23, 1999.
2. Atsumi K, Fujimasa I, Imachi K, et al: Three goats survived for 288 days, 243 days and 232 days with hybrid total artificial heart. Trans ASAIO 27: 77–82, 1981.
3. Atsumi K, Hori M, Ikeda S, et al: Artificial heart incorporated in the chest. Trans ASAIO 9: 292–297, 1963.
4. Imachi K, Fujimasa I, Miyake H, et al: Evaluation of antithrombogenicity, durability and biocompatibility of an artificial heart system for more than 100 days. Artif Organs 5: 423–429, 1981.
5. Imachi K, Fujimasa I, Mabuchi K, et al: A newly designed jellyfish valve for an artificial heart blood pump. Trans ASAIO 34: 726–728, 1988.
6. Imachi K, Chinzei T, Abe Y, et al: A new hypothesis on the mechanism of calcification formed on a blood contacted polymer surface. J Artif Organs 4: 74–82, 2001.
7. Iwasaki K, Umezu M, Imachi K, Fujimoto T: Design improvement of the jellyfish valve for long-term use in artificial heart. IJAO 24: 463–469, 2001.
8. Atsumi K, Sakurai Y, Fujimasa I, et al: Hemodynamic analysis on prolonged survival cases (30 days and 20 days) of artificial total heart replacement. Trans ASAIO 21: 545–554, 1975.
9. Imachi K, Fujimasa I, Nakajima M, et al: Overall analysis of the causes of pathophysiological problems in total artificial heart in animals by cardiac receptor hypothesis. Trans ASAIO 30: 591–596, 1984.
10. Abe Y, Chinzei T, Mabuchi K, et al: Physiological control of a total artificial heart: Conductance and arterial pressure based control. J Appl Physiol 84: 868–876, 1998.
11. Abe Y, Chinzei T, Isoyama T, et al: Third model of the undulation pump total artificial heart. ASAIO J 49: 123–127, 2003.
12. Imachi K, Mabuchi K, Chinzei T, et al: Fabrication of a jellyfish valve for use in an artificial heart. ASAIO J 38: 237–242, 1992.
13. Saito I, Abe Y, Chinzei T, et al: Miniature undulation pump for the study of renal circulation. J Congest Heart Fail Circ Sup 1: 321–325, 2001.
Copyright © 2006 by the American Society for Artificial Internal Organs