Secondary Logo

Journal Logo

Special Report

Feasibility of a Miniature Centrifugal Rotary Blood Pump for Low-Flow Circulation in Children and Infants

Takatani, Setsuo*; Hoshi, Hideo*; Tajima, Kennichi*; Ohuchi, Katsuhiro*; Nakamura, Makoto*; Asama, Junichio; Shimshi, Tadahiko; Yoshikawa, Masaharu

Author Information
doi: 10.1097/01.mat.0000176139.01152.5b
  • Free


Thousands of pediatric patients around the world with cardiomyopathy or single-ventricular physiologies secondary to debilitating heart defects may benefit from long-term (more than 2 weeks) mechanical circulatory support. Because there is a greater possibility of heart recovery in infants and children, ventricular assist devices (VADs) may contribute to the postoperative management of intractable heart failure after congenital heart surgery, and may also be beneficial in the long-term support of end-stage cardiac patients, successfully bridging to heart transplantation and/or to a viable solution for recovery of heart function.

In Europe, both Medos-HIA and Berlin Heart pulsatile pediatric devices are available and have been used to bridge patients to heart transplantation as well as for recovery of heart function.1,2 In the United States, however, clinically approved pediatric blood pumps are not available, although development efforts were expended in the 1990s.3–5 In the absence of approval by the US Food and Drug Administration (FDA) of pediatric-size blood pumps, clinicians are forced to use BioMedicus centrifugal blood pumps for the treatment of combined heart and lung failure in extracorporeal membrane oxygenation or ventricular support during postoperative ventricular management.6–10 Although the adult-size pulsatile VADs, such as HeartMate-I, Thoratec, and BVS-5000, are available,11 their size limits usage in pediatric patients. The MicroMed, Jarvik-2000, HeartMate-II, and Gyro continuous-flow pumps would be promising because of their small size, but may be difficult in infants with a body surface area of less than 0.8 m2.12–14 The pediatric VAD projects started in 2004 by the National Institutes of Health are expected to deliver tiny mechanical circulatory support devices intended for children and infants.15–19

Until 1997, when the Japanese government passed a law allowing organ transplantation from brain-dead subjects, VAD use in Japan was strictly limited to postcardiotomy support. Along with the development of adult-size pneumatic VADs in the late 1980s, postcardiotomy applications of 20 cc pneumatic pulsatile VADs for 3 and 7 days have been reported in two patients.20,21 With the approval of adult-size Toyobo and Zeon VADs for clinical use in 1990,22 adult-size extracorporeal VADs have been used in the support of pediatric patients. Since 1997, the Toyobo VAD has been used for long-term support of cardiomyopathy patients as a bridge to transplantation.23 A severe donor heart shortage in Japan, however, has hindered the prevalence of heart transplantation not only in adults, but also in pediatric patients. Out of 26 transplants performed to date, only two involved patients who were younger than 15 years. Because the law does not allow heart transplantation from children younger than 15 years, heart transplantation to children younger than 10 years is difficult because of the shortage of donor hearts. A safe and reliable pediatric mechanical circulatory support device can help bridge patients not only to transplantation (possibly abroad), but also to an appropriate therapeutic means for heart recovery.

The implantable continuous-flow pumps such as EVAHEART that are now coming into the clinical arena in Japan may have long-term applications for pediatric patients in the future.23,24 Adult-size VADs, however, may invite some problems because of differences in flow capability, priming volume, and inflow and outflow cannula size. The size mismatch may create thromboembolic complications because the pumps will be forced to operate at the lower operating point. There is an increasing need for the development of downsized VADs for pediatric applications in Japan.

Continuous-flow rotary pumps offer advantages over the pulsatile devices because 1) no heart valves are required, 2) they can be made compact-size with a single moving element, and 3) most of all, they require an extremely low priming volume. These favorable features lead to easier implantation and removal, simpler flow control to cover a wide range of patients, and estimation of pump flow through noninvasive monitoring of motor power consumption and rotational speed. However, there are some drawbacks. Reverse flow through the pump at low pump speeds can be detrimental to the heart. Mechanical durability and biocompatibility of the shaft seal may be questionable for long-term durability. Downsizing the pump size boosts the rotational speed and consequently the shear rate, which in turn increases the chance for mechanical damage to the blood cell elements. Finally, the physiologic effects of reduced pulsatility due to continuous-flow devices are not yet fully known for adults; therefore, for children, serious consideration of possible effects is necessary before extended clinical application.

In this study, we designed and evaluated a tiny, centrifugal rotary blood pump (CRBP) to find solutions for the development of a successful implantable device intended for children and/or infants. The feasibility of the pump design, in vitro performance, hemolysis test results, and reverse-flow characteristics were evaluated.

Materials and Methods

Pump Design

The design aimed at a compact, seal-less CRBP with the following features: 1) seal-less structure with mechanical noncontact bearing, 2) external pump-head diameter of less than 50 mm, 3) pump flow of 0.5–4.0 l/min against a head pressure of 40–100 mm Hg at a rotational speed of 1,500–4,000 rpm, 4) durability of 3–6 months, and 5) hemolysis index of less than 0.0060 g/100 l (BioPump Bp-80).

Figure 1 shows the design of a CRBP. Radial magnetic coupling with the driver magnets located outside the impeller follower magnets was used. The impeller was supported by a hydrodynamic needle bearing at the center of the impeller. The 30-mm-diameter impeller with six straight vanes and an inlet angle of 40 degrees was proposed to meet the flow requirement.25 The inflow and outflow port diameter was adjusted to accommodate a 1/4” internal diameter cannula. The impeller and top and bottom housing were fabricated with acrylic resin to yield a see-through view of the pump inside, whereas the hydrodynamic needle bearing’s male and female parts were made with titanium and polyethylene material, respectively.

Figure 1.
Figure 1.:
(a) Schematic of a seal-less, radial magnetic-driven miniature centrifugal rotary blood pump. Top and side view showing its construction. (b) Schematic of the needle hydrodynamic bearing at the center of the impeller.

Hydrodynamic Performance Testing

Hydrodynamic performance was studied in a simulated mock circulatory loop. The flow rate and head pressure were measured with an electromagnetic flow meter and a disposable pressure transducer (Nihon Kohden Co. Ltd., Tokyo, Japan). The power requirement of the DC brushless motor was measured with a digital power analyzer (AZ-4000, Yokogawa Co. Ltd., Tokyo, Japan). Fresh porcine blood was used as a working fluid at a room temperature of 24°C. Pump efficiency was calculated from the following equation: Efficiency = (Hydraulic power output /Motor electrical power requirement) × 100 [%]

The reverse flow through the pump at the reduced pump speed was evaluated in the setup shown in Figure 2. The pump afterload was varied from 50 to 100 mm Hg using an overflow tank. For the selected afterload, the rotational speed of the pump was reduced from the high level where forward flow was produced to the lower level, whereas the bypass flow rate was monitored using an electromagnetic flow meter.

Figure 2.
Figure 2.:
The mock circulatory loop used to evaluate reverse-flow characteristics of the miniature centrifugal rotary blood pump.

Mechanical Damage to Blood Cells

Hemolysis testing was conducted at a flow rate of 1.7 l/min against a head pressure of 70 mm Hg. The Citrate Phosphate Dextran (CPD) anticoagulated fresh porcine blood, which was obtained from a local slaughterhouse, was used as a working fluid at 24°C. The experiments were conducted with the impeller in the top and bottom contact mode with the two different sizes of the bearing clearance. The normalized index of hemolysis (NIH) was calculated using the following equation:

where Hct is the hematocrit value (in %), fHb is the free plasma hemoglobin (in mg/100 ml), V is the priming volume (in ml), Q is the pump flow (in ml/min), and t is the test duration (in min).


Prototype Pump

Figure 3 shows the first prototype model, a tiny centrifugal blood pump intended for pediatric application. The pump dimensions are summarized in Table 1. The pump head height was 20 mm, with an external diameter of 49 mm, weight of 50 g, and priming volume of 5 ml. The impeller had six straight vanes, an inlet angle of 40 degrees, and a diameter of 30 mm. The impeller was driven with a six-pole radial magnetic coupler made of a Neodium-iron magnet, with the driver magnet attached to the motor shaft and rotated outside the impeller to create a radial coupling force.

Figure 3.
Figure 3.:
(a) Prototype miniature centrifugal rotary blood pump, (b) its components, and (c) secondary flow through a needle-bearing area.
Table 1
Table 1

The impeller was supported by a needle-type bearing attached to the bottom pump housing (Figure 3b). The polyethylene female bearing through which a titanium needle-type male bearing was inserted was fixed at the center of the impeller. The top and bottom journal-type bearings created a hydrodynamic effect around the support area (Figure 3c). A 3-mm-diameter ball was fixed at the top of the male needle bearing to prevent movement of the impeller in an upward direction due to lift-off force at the higher pump rpm (Figure 1).

Figure 4 shows the prototype pediatric rotary blood pump in comparison to the BioPump BP-80 and Medos-HIA 9cc pediatric VAD. When the priming volume of these devices was examined, our rotary pump VAD showed the smallest value of 5 cc in comparison with 85 ml of the BP-80 and 24 ml of the Medos-HIA. From the point of priming volume, it is apparent that the prototype rotary VAD is less invasive to pediatric patients, particularly infants.

Figure 4.
Figure 4.:
Prototype miniature centrifugal rotary blood pump (center) in comparison with the Medos-HIA pediatric-size pulsatile pump (right) and BioPump BP-80 (left).

Hydrodynamic Performance

Figure 5a shows the head pressure versus pump flow of the prototype rotary blood pump obtained using porcine blood. A target flow rate of 0.5–4.0 l/min against a head pressure of 40–100 mm Hg was generated with a rotational speed of 2,000–4,000 rpm. The electrical-to-hydraulic efficiency ranged from 7.5–20% at an operational flow between 0.5 and 3.0 l/min with power requirements of the DC brush-less motor of 1–6 W (Figure 5b). The maximum efficiency was 22.5% at a flow rate of 3.0 l/min against a head pressure of 170 mm Hg and a rotational speed of 4,400 rpm.

Figure 5.
Figure 5.:
(a) Pressure-flow characteristics of the M-CRBP and (b) the electrical-to-hydraulic efficiency characteristics of the device.

Figure 6 shows the analysis of the magnetic field generated by the magnetic coupler based on the finite element method. The stiffness in the axial direction was 2.4 N/mm and the maximum magnetic coupling torque was 0.043 N-m. No decoupling between the driver magnet and impeller magnet was observed during operation. The impeller lift-off phenomenon was observed at around 3,200 rpm but with no hysteresis between going up-mode and going down-mode due to a strong coupling force between the driver and follower magnets.

Figure 6.
Figure 6.:
Finite element method analysis of the magnetic coupler. (a) Axial restoring force, and (b) magnetic coupling torque.

The forward and reverse flow through the rotary pump versus pump speed is shown in Figure 7. Backflow through the centrifugal pump occurred at low speeds. The rate of change in the flow rate as a function of the pump speed became steeper with each increase in the afterload level. For a flow rate above 0.5 l/min, stable performance can be guaranteed with this pump for any pump afterload.

Figure 7.
Figure 7.:
Forward- and reverse-flow characteristics of the M-CRBP versus pump speed for different pump afterloads.

Hemolysis Test

Figure 8 shows the hemolysis test results. With the impeller in the top contact mode (rpm > 3,200) and with a bearing clearance of 0.2 mm (model 1), the hemolysis level was highest at 0.089 g/100 l. With the impeller in the down-mode (rpm < 3,200) and with a bearing clearance of 0.1 mm (model 3), the hemolysis level was reduced to 0.0076 g/100 l, which was close to the target value of 0.0060 g/100 l. In Figure 8, model 2 has the impeller in the down position with a bearing clearance of 0.2 mm.

Figure 8.
Figure 8.:
Hemolytic performance of the M-CRBP for top contact mode with 0.2 mm bearing clearance (model 1), bottom contact mode with 0.2 mm bearing clearance (model 2), and bottom contact mode with 0.1 mm bearing clearance (model 3)


Controversy continues regarding the use of pulsatile versus nonpulsatile or reduced-pulsatile perfusion during support of the failing heart. From the device point of view, the continuous-flow devices offer advantages because 1) no inflow and outflow valves are required, reducing the chance of thromboembolic complications related to pump components; 2) devices can be made small and compact, allowing easier implantation and removal; and 3) devices are less invasive, requiring smaller priming volume than the pulsatile devices. Where children and infants are concerned, issues of priming volume and inflow and outflow conduit size become very critical. Our rotary VAD has a priming volume of 5 ml compared with 85 ml for the BioPump BP-80 and 24 ml for the Medos-HIA pediatric VAD. Although the Medos-HIA is intended for use in infants, the volume-displacement–type VAD requires a higher priming volume. Non–volume-displacement pumps, such as centrifugal- or axial-flow devices, are better suited for smaller subjects. As for the issue of pulsatility, there is not yet conclusive evidence. Although continuous-flow devices have been used for the maintenance of normal circulation in adult subjects, their applicability to pediatric subjects must await the results of clinical applications.

Because of the various advantages of the continuous-flow devices, a tiny CRBP was designed, and its pressure-flow and hemolytic performance were studied. The external diameter of the CRBP is 49 mm, and the impeller diameter is 30 mm. The pump head weighs only 50 g (150 g when the motor driver is included). It met our design criteria by providing flows of 0.5–4.0 l/min against a head pressure of 40–100 mm Hg at rotational speeds of 2,000–4,000 rpm. Because of the wide flow range from 0.5 to 4.0 l/min, the device may be suited to support the circulation of pediatric patients, from infants to small-size children. Assuming a normal cardiac output requirement of 80 ml·kg–1·min–1and 4.0 l/min, pump output is sufficient to support circulation of 50-kg patients. A small priming volume of 5 ml was also achieved, which is favorable for application to infants. The inflow and outflow conduit size of 1/4” will be suitable to cannulate pediatric patients. The electrical-to-hydraulic efficiency ranged from 5% to 22.5%, requiring extremely low power from 1 to 6.0 W.

To attain stable operation, which may be important to obtain good biocompatibility, the 30-mm-diameter impeller with six straight vanes was magnetically coupled to the external drive magnets. A six-pole magnet coupler made of Neodium-iron magnets was specially designed to attain high magnetic torque and stable operation. The driver magnets were attached to the shaft of a small DC brushless motor and placed radially external to the impeller magnets (Figure 1). Finite element method analysis revealed an axial restoring force of 2.4 N/mm, and 0.043 N-m magnetic torque was sufficient to maintain stable operation without decoupling. Although the lift-off of the impeller occurred above 3,200 rpm, there was no hysteresis in the pump speed between the up-mode and down-mode directions, possibly because of the high coupling force.

Hemolytic performance of the pump revealed that, with the impeller in the up-mode (pump speed > 3,200 rpm) and a bearing clearance of 0.2 mm, the NIH was highest at 0.087 g/100 l, whereas it was reduced to 0.061 in the down-mode with the bearing clearance of 0.2 mm. In the up-mode, it is probable that the rubbing of the impeller against the stopper ball damaged red blood cells, increasing the NIH value. The pump speed reduction had little effect on reducing the NIH level. In the down-mode, however, when the bearing clearance was reduced to 0.1 mm, the NIH was significantly reduced to 0.0076 g/100 l. The bearing design and magnetic coupling force must be further optimized to stabilize the impeller rotation, which seems to strongly affect the hemolytic performance of the pump. Computational fluid dynamic studies are currently ongoing to quantify the secondary flow through the hydrodynamic bearing used in this pump and to possibly predict the flow and shear stress field that seems to affect the biocompatibility of the device.26–28

One disadvantage of the continuous-flow device in comparison to the pulsatile device is its reverse flow through the pump that occurs at low flow and in case the pump stops. In the low-flow condition, reverse flow through the pump from the aorta to ventricle may develop, reducing the net efficiency of the device and negatively affecting the ventricle. This feature requires careful attention when the device is used and when weaning is performed with the recovery of heart function. In addition, an emergency outflow occlusive system must be considered to cope with a sudden stop of the pump.29

As for the biocompatibility of the device, we need to evaluate its performance in vivo using an animal model. Particularly, the secondary flow through the bearing area is important to maintain stable flow dynamics and antithrombogenic performance of the pump. Computational fluid dynamics may be a useful tool in predicting the problem areas of the pump along with the in vivo evaluation in animals.


A prototype, seal-less, tiny CRBP with an external pump-head diameter of 49 mm and an overall height of 42 mm, including the motor driver, was designed, and its flow dynamics and hemolytic performance were evaluated. The prototype pump met the design criteria of delivering 0.5–4.0 l/min against a head pressure of 40–100 mm Hg between pump speeds of 2,000–4,000 rpm. The hemolytic performance was 0.0076 g/100 l, which was close to the target value of 0.006 g/100 l of the BioPump. Further refinement in the hydrodynamic bearing at the center of the impeller and magnetic coupler should improve its overall performance to meet clinical CRBP requirements.


This research was partially supported by grants-in-aid from the Japan Society for the Promotion of Science under Project #14208103 and #16500290 (principal investigator: Prof. Setsuo Takatani).


1.Hetzer R, Loebe M, Potapov EV, et al: Circulatory support with pneumatic paracorporeal ventricular assist device in infants and children. Ann Thorac Surg 66: 1498–1506, 1998.
2.Stiller B, Hetzer R, Weng Y, et al: Heart transplantation in children after mechanical circulatory support with pulsatile pneumatic assist device. J Heart Lung Transplant 22: 1201–1208, 2003.
3.Daily BB, Pettitt TW, Sutera SP, Pierce WS: Pierce-Donachy pediatric VAD: Progress in development. Ann Thorac Surg 61: 437–443, 1996.
4.Litwak P, Butler KC, Thomas DC, et al: Development and initial testing of a pediatric centrifugal blood pump Ann Thorac Surg 61: 448–451, 1996.
5.Yoshikawa M, Nakata K, Ohtsuka G, et al: Feasibility of a tiny Gyro centrifugal pump as an implantable ventricular assist device. Artif Organs 23: 774–779, 1999.
6.Reinhartz O, Stiller B, Eilers R, Farrar DJ: Current clinical status of pulsatile pediatric circulatory support. ASAIO J 48: 455–459, 2002.
7.Duncan BW, Hraska Y, Jonas RA, et al: Mechanical circulatory support in children with cardiac disease. J Thorac Cardiovasc Surg 117: 529–542, 1999.
8.Ibrahim AE, Duncan BW, Blume ED, Jonas RA: Long-term follow-up of pediatric cardiac patients requiring mechanical circulatory support. Ann Thorac Surg 69: 186–192, 2000.
9.Duncan BW: Mechanical circulatory support for infants and children with cardiac disease. Ann Thorac Surg 73: 1670–1677, 2002.
10.Undar A, McKenzie ED, McGarry MC, et al: Outcomes of congenital heart surgery patients after extracorporeal life support at Texas Children’s Hospital. Artif Organs 28: 963–966, 2004.
11.Takatani S: Beyond implantable first generation cardiac prostheses for treatment of end-stage cardiac patients with clinical results in a multicenter. Ann Thorac Cardiovasc Surg 8: 253–263, 2002.
12.Morales DLS, DiBardino DJ, McKenzie ED, et al: Lessons learned from the first application of the DeBakey VAD child: An intracorporeal ventricular assist device for children. J Heart Lung Transplant 24: 331–337, 2005.
13.Jarvik R, Griffith B, Burgreen G, et al: Initial results with the pediatric Jarvik 2000 heart. ASAIO J 51: 28A, 2005.
14.Butler KC, Farrar DF: Long-term reliability testing of the HeartMate II LVAD. ASAIO J 51: 32A, 2005.
15.Duncan B: The “PediPump”: A versatile, implantable, pediatric ventricular assist device. 1st International Conference on Pediatric Mechanical Support Systems and Pediatric Cardiopulmonary Perfusion. Hershey, PA, May 19-2, 2005 [invited lecture].
16.Gartner M: Integrated pediatric cardiopulmonary assist system (pCAS). 1st International Conference on Pediatric Mechanical Support Systems and Pediatric Cardiopulmonary Perfusion. Hershey, PA, May 19–22, 2005 [invited lecture].
17.Griffith BP: Infant and child-size Jarvik 2000 hearts. 1st International Conference on Pediatric Mechanical Support Systems and Pediatric Cardiopulmonary Perfusion. Hershey, PA, May 19–22, 2005 [invited lecture].
18.Weiss B: Pulsatile pediatric ventricular assist devices (pulsatile PVADs). 1st International Conference on Pediatric Mechanical Support Systems and Pediatric Cardiopulmonary Perfusion. Hershey, PA, May 19–22, 2005 [invited lecture].
19.Borovetz H: The pediatric ventricular assist device (Pedia Flow). 1st International Conference on Pediatric Mechanical Support Systems and Pediatric Cardiopulmonary Perfusion. Hershey, PA, May 19–22, 2005 [invited lecture].
20.Taenaka Y, Takano H, Noda H, et al: Experimental evaluation and clinical application of a pediatric ventricular assist device. ASAIO Trans 35: 606–608, 1989.
21.Taenaka Y, Takano H, Noda H, Kinoshita M: A pediatric ventricular assist device: Its development and experimental evaluation of hemodynamic effects on postoperative heart failure of congenital heart diseases. Artif Organs 14: 49–56, 1990.
22.Takano H, Nakatani T: Ventricular assist systems: Experience in Japan with Toyobo pump and Zeon pump. Ann Thorac Surg 61: 317–322, 1996.
23.Takatani S, Matsuda H, Hanatani A et al: Mechanical circulatory support device in Japan: Current status and future direction. J Artif Organs 8: 13–27, 2005.
24.Yamazaki K, Litwak P, Tagusari O, et al: An implantable centrifugal blood pump with a recirculating purge system (cool-seal system). Artif Organs 22: 466–473, 1998.
25.Yoshino M, Uemura M, Takahashi K, et al: Design and evaluation of a single-pivot supported centrifugal blood pump. Artif Organs 25: 683–687, 2001.
26.Watanabe N, Masuda T, Iida T, et al: Quantification of the secondary flow in a radial coupled centrifugal blood pump based on particle tracking velocimetry. Artif Organs 29: 26–35, 2005.
27.Throckmorton AL, Untaroiu A, Allaire PE, et al: Computational analysis of an axial flow pediatric ventricular assist device. Artif Organs 28: 881–891, 2004.
28.Bachmann C, Hugo G, Rosenberg G, et al: Fluid dynamics of a pediatric ventricular assist device. Artif Organs 24: 362–372, 2000.
29.Ohtsuka G, Yoshikawa M, Sueoka A, et al: An emergency balloon occlusion system for a rotary blood pump left ventricular assist system. Artif Organs 23: 704–707, 1999.
Copyright © 2005 by the American Society for Artificial Internal Organs