Secondary Logo

Journal Logo

Special Report

Development of a Percutaneous Pediatric Ventricular Assist Device

Wang, David H. J.*; Smith, Douglas E.*; Bacha, Emile A.; Hijazi, Ziyad M.; Magovern, James A.§

Author Information
doi: 10.1097/01.mat.0000181635.70527.c1
  • Free


More than 30,000 babies are born each year in the United States with a congenital heart defect.1 Some of these patients need a mechanical circulatory support device during corrective procedure or while waiting for transplant.2 The need for a pediatric ventricular assist device (VAD) is increasing as more complex corrective operations are performed and pediatric heart transplantation becomes more common.3,4 Corrective operations are now done for many previously inoperable forms of congenital heart disease. The patients are younger, smaller, and sicker, and often require preoperative and postoperative cardiac support. The success of pediatric heart transplantation has led to an increasing number of transplant candidates. In 2003, 181 heart transplants were performed in the nation on children 10 years and younger. Almost 700 of the 87,000 people on the nation’s organ-transplant list are 5 years and younger.5 However, the number of donor hearts available has not changed, and therefore an increasing number of pediatric patients need circulatory support while waiting for donor hearts.4

Currently, the most common method of mechanical circulatory support for pediatric patients is extracorporal membrane oxygenation (ECMO).6 Despite reasonable survival rates, the use of ECMO for mechanical support presents many challenges to its users. A typical ECMO circuit consists of a centrifugal or roller pump, oxygenator, drainage and return cannulae, and a heat exchanger.6 The integration of such a large number of devices introduces complications and requires continuous monitoring. When a patient is placed on ECMO, hemodilution can be severe. The large contact area of blood with this circuit creates an intense systemic inflammatory response and increases the likelihood of thrombosis and hemolysis. Because patients have to be heparinized, blood product transfusions are very common.6 Presence of large quantities of anticoagulants exacerbates any bleeding that may occur caused by the immune system’s response to foreign surfaces in contact with the native tissue and blood.6

Compared with ECMO, the use of a VAD in the pediatric population has many advantages. The implementation of a VAD will require less anticoagulation, especially at higher flow rates, and fewer blood and platelet transfusions. The VAD allows for better hemodynamics and less blood trauma than ECMO.7 The use of VADs also requires little technical attention after installation because of the generally simple design. VADs are also less expensive to implement than ECMO systems. Using VADs instead of ECMO can easily reduce intensive care unit and hospital expenses.3 More importantly, many pediatric patients on ECMO do not have pulmonary failure and do not need pulmonary assistance via an oxygenator.6 Despite this increasing demand for a pediatric VAD, a specifically designed pediatric VAD is not currently available in the United States. It is based upon this growing clinical need that the percutaneous pediatric VAD is proposed.

Materials and Methods

System Design

The design of the proposed system is based on the current TandemHeart PTVA (percutaneous transseptal ventricular assist) system (Figure 1). It consists of a centrifugal blood pump outside of the body (Figure 1), a transseptal venous cannula that takes blood from the left atrium (LA) to the pump, an arterial cannula that returns the blood to the arterial system, and a controller to adjust pump speed/flow. Oxygenated blood is withdrawn from the atrium through a venous transseptal cannula. A centrifugal blood pump pumps the blood back to the body through an arterial cannula. Heparinized saline is delivered to the pump at the rate of 10 ml/h and heparin concentration of 900 IU/h. This infusate also provides the pump hydraulic bearing around the rotor.

Figure 1.
Figure 1.

For smaller patients (2–35 kg), the transseptal cannula was designed to be placed in the internal jugular vein and the arterial cannula in the internal carotid artery using the neck cannulation approach. For larger patients (> 35 kg), the femoral cannulation technique through femoral vein and artery would be used as currently used on adults. A spectrum of cannulae was proposed and designed for pediatric patients of different sizes (Table 1).

Table 1
Table 1:
Pediatric Patient Population Segments, Estimated Cannula Sizes, and Flow Requirements for the Proposed Pediatric VAD

Cannula Development and Testing

All the required arterial cannulae sizes (from 8 to 17 Fr) listed in Table 1 are already commercially available. Only venous cannulae were developed. For each patient segment, the size of venous cannula was chosen as the same as that of the venous cannula used in ECMO support. Length was calculated based on the measured distance from internal jugular vein to LA on the chest radiographs of 28 patients. Three sets of venous cannulae for each size listed in Table 1 were made according to the proposed dimensional requirements, a sample of which can be seen in Figure 2. H-Q performance of each cannula was evaluated using Glycerin/Saline solution with a viscosity of 4.0 cP at a temperature of 37ºC. The results of the three duplicates of the each design were averaged to derive the H-Q curve.

Figure 2.
Figure 2.

A controlled study was conducted to assess the degree of hemolysis produced by the developed cannulae prototypes. The test loop consisted of a pediatric pump with an 8-Fr arterial cannula connected to its outlet and an 8-Fr venous cannula to its inlet. The control loop had only a pediatric pump and connecting tubing. The tests were conducted following the guidelines of ASTM standard (ASTM F1841-97). Pump speeds were maintained at 4,050 rpm and pump flow at 0.3 l/min. The tests were run for 6 hours with an initial activated clotting time (ACT) of 200 seconds. Blood samples were taken at each hour and plasma free hemoglobin levels were measured.

Pump Development and Testing

The current adult pump was redesigned ad hoc for the pediatric application. The upper housing was reduced in size to accommodate ¼ connectors and to reduce prime volume. This design change also reduced both the radial and the axial gaps between impeller and pump housing. A computer solid model of the new upper housing design was constructed. The model was sent to Prototech Engineering, Inc. (Willowbrook, IL) for rapid prototyping. A total of 10 upper housings were received. Three pump prototypes were made and tested using a glycerin/saline solution with a viscosity of 4.0 cP at a temperature of 37ºC.

Hemolysis level produced by the developed pediatric pump was assessed by following ASTM standard (ASTM F1841-97: Standard Practice for Assessment of Hemolysis in Continuous Flow Blood Pumps). The current adult pump was used as the control. Anticoagulated fresh bovine blood was used. The temperature of the blood was maintained at 37ºC ± 2ºC. The tests were run for six hours. Blood was sampled before pumping as a baseline and at 1-hour intervals. Plasma free hemoglobin levels were also measured. NIH values for each loop were calculated upon completion of the tests.

Controller Development and Testing

The prototype pediatric controller has been designed and fabricated based on the current adult controller. To broaden the pump speed range from 3,000–7,500 rpm to 3,000–10,000 rpm, the power supply circuit was modified to increase the internal 15-V medical-grade power supply to 25 V.

An infusion pump was incorporated into the prototype pediatric controller to provide lubrication for the blood pump. The infusion pump’s control module was modified to allow the infusion rate to be adjusted. The control module is a potted assembly with an embedded calibration potentiometer. The potentiometer was dissected out to give access to its three wire leads. They were then connected to a like-value potentiometer on the enclosure’s side panel, allowing the operator to adjust the infusion rate. The adjustment range is from 5 to 20 ml/h.

System Testing

For each configuration of the system, a system performance test was conducted to validate that the design could deliver the required flow listed in Table 1. The test was conducted using a 4.0 cp Glycerin/saline mixture as perfusate. Temperature of the loop was maintained at 37ºC ± 2ºC. System inlet and outlet pressure were set to simulate physiological conditions. Pump flow, pump speeds required to deliver the required flow rates, and priming volume of the circuit were measured and recorded for each system configuration.

Hemolysis study was also conducted on the system level, under the simulated use conditions with the current adult system as control. The test loop consisted of a pediatric pump with an 8-Fr arterial cannula connected to its outlet and an 8-Fr venous cannula to its inlet. The control loop had an adult pump with a 15-Fr arterial cannula connected to its outlet and a 21-Fr venous cannula to its inlet. The tests were conducted following the guidelines of ASTM standard (ASTM F1841-97). Anticoagulated fresh bovine blood was used. The temperature of the blood was maintained at 37ºC ± 2ºC. The test was run for 6 hours. Blood was sampled before pumping as a baseline and at 1-hour intervals. Plasma free hemoglobin was measured. NIH value for each loop was calculated upon completion of the test.


The H-Q performance of all the pediatric cannulae prototypes is presented in Figure 3. Table 2 summarizes the pressure drops across each sized cannula at the required maximum flow rates. It is clear that none of the pressure drops exceeds 400 mm Hg. The proposed cannulae can deliver the required flow without causing cavitation and blood cell damage.

Figure 3.
Figure 3.
Table 2
Table 2:
Maximum Pressure Drop at Maximum Required Flow of Venous Cannulae

Calculation of hemolysis level revealed a NIH of 0.08725 for the study group with pediatric cannula and a NIH of 0.08859 for the control group without pediatric cannula. Considering the measurement error, the use of the pediatric cannula induced insignificant levels of hemolysis. Hemolysis study was only conducted on the smallest sized cannulae because of its critical dimension. The experiment also revealed no thrombus formation.

Averaged H-Q curve for pediatric pump is shown in Figure 4. It is clear that the pediatric design adequately achieves the maximum flow requirements for the intended patient population listed in Table 1. Overall pump efficiency was calculated by determining the ratio of electric power input (pump voltage × pump current) to mechanical power output (pressure rise × pump flow). Calculated pump efficiency for pediatric prototypes is presented in Figure 5.

Figure 4.
Figure 4.
Figure 5.
Figure 5.

Testing conditions and calculated NIH values for the pediatric pump as well as the current adult pump are listed in Table 3. The results showed that the pediatric pump produced a higher level of hemolysis than the adult pump at flow rate of 4.5 l/min and pressure rise of 150 mm Hg. This can be explained by the narrower gap and the higher speed of the ad hoc designed pediatric pump. One can argue that the pediatric pump will never run at such a high flow rate, but this high hemolysis level points to the need for further pump design research. A hemolysis study under the simulated use flow condition (flow rate = 0.3 l/min) for the smallest patient segment (2–3 kg) was conducted at the system level and again showed elevated hemolysis (NIH = 0.0873 g/100 l). There was no observable thrombus formation in either pump after the 6-hour test.

Table 3
Table 3:
Hemolysis Study of a Pediatric Pump versus an Adult Pump

A pediatric VAD controller has been successfully produced. It has the capability to drive the pediatric VAD blood pump in the range of 3,000 to 10,000 rpm. It includes an integral infusion pump with a flow rate adjustable between 5 and 20 ml/h. It has the ability to measure pump speed, pump current, pump drive voltage, and infusion pressure. The ability to measure both pump current and pump drive voltage allows the operator to estimate the input power to the pump and therefore compute an estimate of efficiency.

For each configuration of the system, a system performance test revealed that all the required flow rates were successfully achieved. Pump speeds required to deliver the required flow rates for each patient segment, priming volume of the circuit, simulated LA and mean arterial pressures, arterial and venous cannula sizes, and the required flow are listed in Table 4. The priming volume for the largest sized circuit is < 70 cc, which is a great advantage compared with the pediatric ECMO circuit.

Table 4
Table 4:
Priming Volume and Required Pump Speeds to Deliver the Required Flows for System Configurations and Patient Segments

The system hemolysis testing conditions and calculated NIH values are shown in Table 5. The result once again shows that the pediatric pump was more hemolytic than the adult pump under the simulated use conditions. Because the degree of hemolysis produced by pediatric cannulae is negligible, a primary task in future design is to improve the blood pump design to reduce the level of hemolysis. Again, there was no observable thrombus formation in either pump.

Table 5
Table 5:
Hemolysis Study of Pediatric System versus the Current TandemHeart PTVA System


Compared with the conventional ECMO support, the proposed design has many advantages:

Low priming volume. Compared with a standard ECMO circuit with priming volume of 300–600 cc, the proposed design has a much smaller priming volume, from 18 to 65 cc as shown in Table 4, thus eliminating the problem of hemodilution.

Low risk of hemolysis. A centrifugal pump causes less blood damage than the roller pump used in ECMO circuit.9

Low risk of emboli. Lack of an oxygenator makes the circuit simpler. The blood-contacting surface area is significantly reduced. The risks of protein deposition and platelet activation are therefore greatly reduced.

Low cost. The cost of disposables is similar to an ECMO unit but constant attendance by specially trained staff is not needed, thereby significantly reducing operating costs.

Fast placement. The circuit is simple and can be primed and assembled quickly. There’s no need to maintain a specialized on-call staff. Only a few minutes are required to set up and prime the circuit, providing an advantage in the cardiac arrest situation.9 The entire procedure is currently done in < 30 minutes.

Less trauma. No sternotomy, thoracotomy, or other major surgical procedure is involved for device placement and removal. Patients will experience faster recovery with less trauma and infection.

Compared with the surgically implanted pediatric VADs, such as the Berlin Heart, the Medos/HIA assist, and the pediatric VADs under development under recently awarded 5-year NIH contracts, the proposed system has the following advantages:

Less expensive. The current adult system is less expensive than long-term implantable VADs.

More blood friendly because of its small blood-contacting surface and lower pump speeds.

Lower risk of thrombosis due to the unique heparin infusate lubrication system.

Fast response. The system can be assembled quickly to respond to emergency situations.

Lower operating cost. Specially trained staff is not needed to attend to the device 24 hours a day.9

Modular design consisting of a single pump and a spectrum of cannulae to serve pediatric patients of all sizes. The clinicians can choose the proper sized cannulae to build the circuit suited to each patient.

Less invasive and better patient outcome. There is no open chest procedure involved for device placement. Quicker patient recovery, less trauma, and less infection is expected.

Given the broad spectrum of pediatric heart diseases and the great complexity of pediatric cardiac surgeries, it is understood that not all pediatric heart patients can be supported by this device, especially when an extracorporeal oxygenator is needed. Nevertheless, the unique features of this device, i.e., low priming volume, less blood and surgical trauma, simplicity, and low cost, make it a desirable means of supporting many pediatric patients before, during, and after cardiac surgery.

The major challenges of the development of this device lie in cannula development and cannulation techniques, especially for smaller-sized patients. If the proposed approaches fail, a surgical procedure must be performed to place the device. This device will be used as a surgical VAD instead of a percutaneous VAD, but the other benefits of the device would remain.

In summary, a pediatric VAD design is proposed. It is intended to provide short-term (up to 30 days) circulatory support for pediatric patients of all ages, from neonates (> 2 kg) to young adults. A spectrum of different sized cannulae will be developed to facilitate the percutaneous placement of the device. For smaller patients (from 2 to 35 kg), a neck cannulation technique through the internal jugular vein and internal carotid artery will be used. For larger pediatric patients (> 35 kg), the femoral cannulation technique through the femoral vein and the femoral artery will be used. State-of-the-art computational fluid dynamic and flow-visualization techniques will be used to design a safer and more effective blood pump specifically for pediatric application based on the current adult pump design. Safety and functionality of the developed system will be demonstrated in animal models.8 Successful development of the proposed design will lead to clinical trial and commercialization. The proposed system is unique in design. It is simple and low cost. Because it is inserted percutaneously, cardiac assistance can be provided quickly and less invasively, minimizing potential complications and promoting quicker patient recovery. Successful commercialization of this device will bring a revolution in pediatric cardiology and will save many children’s lives.


This work was supported by the National Heart, Lung and Blood Institute of the National Institute of Health (#1 R43 HL078077-01 to Dr. Wang)


1. National Institutes of Health, United Network for Organ Sharing: Chapter V: pediatric transplant. Available at: Accessed May 8, 2005.
2. Konertz W, Hotz H, Schneider M, et al: Mechanical circulatory support in children. Int J Artif Organs 20: 657–658, 1997.
3. Duncan BW, Jonas RA, et al: Mechanical circulatory support in children with cardiac disease. J Thorac Cardiovasc Surg 117: 529–541, 1999.
4. Warnecke H, Berdijs F, Lange P, et al: Mechanical left ventricular support as a bridge to cardiac transplantation in childhood. Eur J Cardiothorac Surg 5: 330–333, 1991.
5. Lipshultz SE, Sleeper LA, Colan SD, et al: The incidence of pediatric cardiomyopathy: The prospective pediatric cardiomyopathy registry. J Am Coll Cardiol 37: 1180–1196, 2001.
6. Levi D, Marelli D, Laks H, et al: Use of assist device and ECMO to bridge pediatric patients with cardiomyopathy to transplantation. J Heart Lung Transplant 21: 760–770, 2002.
7. Thuys CA, Mullaly RJ, Horton SB, et al: Centrifugal ventricular assist in children under 6 kg. Eur J Cardiothorac Surg 13: 130–134, 1998.
8. Burgreen GW, Antaki JF, Griffith BP: A design improvement strategy for axial blood pumps via computational fluid dynamics. ASAIO J 42: M354–M360, 1996.
9. Karl TR, Horton SB: Centrifugal pump ventricular assist device in pediatric cardiac surgery, in Duncan BW (ed), Mechanical Support for Cardiac and Respiratory Failure in Pediatric Patients, New York, Marcel Dekker, 2001, pp. 21–47.
Copyright © 2005 by the American Society for Artificial Internal Organs