The ability to provide mechanical circulatory support for children of all sizes, while meeting the needs that arise from the variety of indications encountered in pediatrics, remains an unrealized goal. In most centers, extracorporeal membrane oxygenation remains the only option for pediatric circulatory support. Because it is suitable only for short-term use, small children with life-threatening heart failure are limited to support periods of days to weeks.1–3 The development of ventricular assist devices (VADs) for children is one of the major unaddressed needs in pediatric cardiology and cardiac surgery. Historically, centrifugal pump systems have been the most commonly used pediatric VAD technology,4,5 while recently, increasing pediatric experience has accumulated with paracorporeal pneumatic systems.6–9 However, existing centrifugal VADs are designed only for short-term use, while available pneumatic VADs require a variety of pump sizes to provide support for the entire range of patient sizes encountered in pediatrics; neither system is designed to be totally implantable.
Ideally, systems designed for children should possess a number of characteristics: 1) the ability to provide support for the entire range of patient sizes encountered in pediatrics with a single pump; 2) small pump size with retention of excellent hemodynamic performance; 3) design that minimizes host impact; and, 4) the ability to be implanted in even the smallest patients. This article describes the PediPump, which is a VAD under development at The Cleveland Clinic designed specifically for children that attempts to address these issues.
Device Characteristics of the PediPump
The enabling technology for the PediPump is derived from an adult “catheter” pump in development by The Cleveland Clinic's Department of Biomedical Engineering and Foster-Miller Technologies, Inc., (Albany, NY) (CCF/FMT team) under National Heart, Lung and Blood Institute SBIR grant HL 67487. This technology has been described before10; briefly, the design is based on a magnetic-bearing–supported, rotary dynamic pump. Because of its capability of being packaged into a small size, a single basic pump design provides support for the entire range of patient sizes encountered in pediatrics as opposed to pneumatic pump–based systems that require a family of pumps to provide similar support.
The pump rotating assembly consists of an impeller in the front, front and rear radial magnetic bearings, and a motor rotor magnet in its center (Figure 1). Blood enters axially at the inlet and is turned to exit the pump at an intermediate angle through the pump outside diameter. Some arterial blood flows through windows at the rear of the pump under the influence of arterial pressure, washing and cooling the motor gap, before returning to the impeller. The rotor is supported on passive, radial magnetic bearings; the absence of a seal with suspension of the rotor on magnetic bearings should result in high durability. The electric motor combines a small diameter with good efficiency and a large, low-blood-shear rotor/stator gap. Titanium shells seal all potentially corrodible components from blood and tissue.
The same basic pump is anticipated for use as a right VAD, a left VAD, or a biventricular assist device (BVAD). Two configurations are currently envisioned for deployment of the PediPump based on patient size: for larger children (>15 kg), the small size of the PediPump may allow completely intravascular implantation (Figure 2A). For smaller children (<15 kg), extravascular, intracorporeal implantation may be performed using standard cannulation strategies used for existing axial flow pumps with inflow and outlet cannulas configured as needed (Figure 2B). The initial PediPump intravascular prototype measures approximately 7 mm × 75 mm with a priming volume of 0.6 ml, which imparts less than 10% of the physical displacement of currently available axial-flow pumps.
Current Status of PediPump Development
Prototype testing was performed during the now completed phase I SBIR program for the adult catheter-based pump.10 Two pump prototypes were tested in a mock loop: data were recorded for motor current, pump speed, pressure and flow. At flows of 3 l/min and 90 mm Hg pressure rise, the PediPump operated at approximately 25,000 rpm and 4.2 W motor output. Despite its small size, the PediPump provided pressure and flow capable of supporting adults, far exceeding the requirements for support of children in the 2–25 kg weight range (Figure 3). Prototypes of the adult catheter-based pump have successfully completed 60-day endurance testing.
In the spring of 2004, The Cleveland Clinic Foundation was awarded a contract under the Pediatric Circulatory Support initiative from the National Heart, Lung and Blood Institute of the National Institutes of Health to support the design and development of the PediPump. There were three overall aims of The Cleveland Clinic proposal:
- To determine the basic engineering requirements for hardware and control logic, including design analysis for system sizing, establishment of the analytical model to evaluate control concepts, and bench testing of prototypes.
- To perform preclinical anatomical fitting studies using computed tomography (CT)-based three-dimensional modeling.
- To perform animal implantations for characterization and reliability testing of the device with assessment of the physiologic impact of chronic support on the host.
Initial hydraulic designs for the first intravascular and extravascular prototypes have been generated and are being used to optimize impeller design using solid three-dimensional CAD modeling and computational fluid dynamic (CFD) studies. CFD simulations of the stator geometry and the impeller hub contour have been performed to evaluate stator pressure loss, wash flow at the contact point, and the resulting flow field for the impeller vane design in both pumps. Initial design stator vanes have been evaluated for their load bearing capacity and several candidate impeller designs have been evaluated for axial and circumferential velocity data as well as for static pressure and total pressure. A matching stator has been designed for the impeller with the best performance for both configurations. The best-performing impeller and stator have been manufactured in stereolithography and their combined performance has been evaluated in a hydraulic test stage with an externalized motor. Data from the hydraulic test stage and CFD simulations have been compared, allowing the CFD models to be refined for more appropriate turbulence modeling and meshing. Design of test loops for pump prototypes is also proceeding.
The magnetic bearing design of the proposal has been reviewed and detailed for the prototype. A test fixture has been designed and built to evaluate different surfaces and geometries for the axial touch points in this design. Vibration testing can also be performed with the wear test apparatus. As part of the design process, vibration will be reduced to an absolute minimum; resonances seem to be outside the operating speed range of the pump. Pump component configuration and assembly procedures have been evaluated to achieve crevice-free geometries, ease of assembly, and reasonable tolerance stack-up. Tooling has been designed to support assembly and adjustment of key features of the pump. Device assembly will begin shortly, after the durability testing described above finalizes the details of finishing in the thrust bearing area of the hardware.
Anatomical Fitting Studies
Archived CT scans from children with congenital heart disease have been converted into three-dimensional digital models using Mimics software (Materialise, Leuven, Belgium). The original CT scans were obtained for clinical purposes and are representative of the anatomical variation and range of patient sizes encountered in pediatric cardiology and cardiac surgery. Datasets were generated which included cardiac structures, the great vessels, and the chest wall. This CT data, consisting of images made up of pixels with differing gray values, were then stacked using Mimics software to allow visualization of the data in three orthogonal views. On-screen models were then generated from the data by selecting and highlighting the pixels that made up the desired structures. The software was used to calculate the three-dimensional versions of the highlighted pixels and generate model images on the computer screen. An iterative process was then used to refine the image further by adding and erasing pixels until the most accurate reproduction of the anatomy was achieved; these on-screen images were stored as .stl files.
On-screen digital renderings have subsequently been used to generate to-scale physical models, using a variety of rapid prototyping techniques. Standard stereolithography methods have been used to create rigid physical models. Flexible physical models have also been generated using a three-dimensional printer (ZPrinter 310 System; Z Corporation, Burlington, MA). The ZPrinter builds models in layers using an ink-jet printer head to apply a liquid binder to thin layers of fine, starch-based powder to create rigid, thin-walled physical models derived from the on-screen renderings. These thin-walled models are then coated with a liquid polymer that, when cured, produces flexible models. Although standard stereolithography reconstructions with hard resins allow some determination of cannulation and placement strategies during VAD development, it is anticipated that these flexible models will be particularly useful. The accurate modeling provided by these tools will hopefully allow preoperative planning on a case-by-case basis, which will be of particular importance for patients with abnormalities of the great arteries and veins.
Animal studies will be undertaken to provide both characterization and reliability testing of the PediPump. Careful assessment of the physiologic impact of the PediPump and the ensuing host response will be determined from these studies. Animal implantations will begin after further basic engineering and anatomical studies are completed.
At the completion of studies in this proposal, the following will have been achieved: 1) development and testing of a small pediatric VAD, including the acquisition of initial multiyear durability data; 2) determination of deployment methods for the wide range of sizes and anatomical variation encountered in children with heart disease using three-dimensional modeling techniques based on CT imaging; and 3) documentation of the responsiveness and reliability of the pump in the implant environment and determination of the host response to the presence of the pump from animal implantations.
Reduction of pump size will be combined with other, ongoing CCF/FMT programs to produce fully implantable pediatric VAD systems with a downsized internal battery, controller, and transcutaneous energy transmission. To produce a complete, ambulatory ventricular assist system, this pump technology will ultimately be combined with supporting technology based on the experience of CCF/FMT for an implantable total artificial heart, which should be easily adaptable to this new application. Circuit changes and new software for the implantable total artificial heart controller will be made to generate a small, wearable pediatric controller. In the future, use of custom integrated circuitry is envisioned, which will make the CCF/FMT controller much more compact and implantable for some size ranges of children.
The PediPump is much smaller than currently available VADs, allowing use for even newborn circulatory support, and yet seems to retain excellent hemodynamic performance. Current and future development goals include: 1) determination of the basic engineering requirements for hardware and control logic, including design analysis for system sizing, evaluation of control concepts, and bench testing of prototypes; 2) performance of preclinical anatomical fitting studies using CT based three-dimensional modeling and intraoperative fitting of device mock-ups; and 3) evaluation with animal studies to provide characterization and reliability testing of the device.
This project was funded in whole or in part with federal funds from the National Heart, Lung and Blood Institute, National Institutes of Health, under contracts HHSN268200448188C and HL67478.
1.Black MD, Coles JG, Williams WG, et al: Determinants of success in pediatric cardiac patients undergoing extracorporeal membrane oxygenation. Ann Thorac Surg
60: 133–138, 1995.
2.Duncan BW: Mechanical circulatory support for infants and children with cardiac disease. Ann Thorac Surg
73: 1670–1677, 2002.
3.Duncan BW, Hraska V, Jonas RA, et al: Mechanical circulatory support in children with cardiac disease. J Thorac Cardiovasc Surg
117: 529–542, 1999.
4.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.
5.Karl TR, Sano S, Horton S, Mee RB: Centrifugal pump left heart assist in pediatric cardiac operations: Indication, technique, and results. J Thorac Cardiovasc Surg
102: 624–630, 1991.
6.Shum-Tim D, Duncan BW, Hraska V, et al: Evaluation of a pulsatile pediatric ventricular assist device in an acute right heart failure model. Ann Thorac Surg
64: 1374–1380, 1997.
7.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.
8.Konertz W, Hotz H, Schneider M, et al: Clinical experience with the MEDOS HIA-VAD system in infants and children. Ann Thorac Surg
63: 1138–1144, 1997.
9.Stiller B, Dahnert I, Weng Y, et al: Children may survive severe myocarditis with prolonged use of biventricular assist devices. Heart
82: 237–240, 1999.
10.Duncan BW, Lorenz M, Kopcak MW, et al: The PediPump: A new ventricular assist device for children. Artif Organs
29: 527–530, 2005.